UNIVERSITY OF CALIFORNIA 
 
 CTURAL DEPARTMENT LIBRARY 
 
 GIFT OF 
 .'Irs. Lycila Bart)' 
 
CEMENT AND CONCRETE 
 
 BY 
 
 LOUIS CAELTON SABIN, B. S., C. E. 
 
 ASSISTANT ENGINEER, ENGINEER DEPARTMENT, U. S. ARMY; MEMBER OF THE 
 AMERICAN SOCIETY OF CIVIL ENGINEERS 
 
 NEW YORK 
 
 McGRAW PUBLISHING COMPANY 
 
 1905 
 
S'a- 
 
 COPYBIGHT, 1904 
 BY 
 
 L. C. SABIN 
 
 Stanbope 
 
 BOSTON, U.S.A. 
 
PREFACE 
 
 THAT the use of cement has outstripped the literature on 
 the subject is evidenced by the number and character of the 
 inquiries addressed to technical journals concerning it. This 
 volume is not designed to fill the proverbial "long felt want/' 
 for until within a few years the number of engineers using 
 cement in large quantities was quite limited. These American 
 pioneers in cement engineering, under one of whom the author 
 received his first practical training in this line, needed no formal 
 introduction to the use and properties of cement; their knowl- 
 edge was born and nurtured through intimate association and 
 careful observation. 
 
 To-day the young engineer frequently finds a good working 
 knowledge of cement one of the essentials of success, and the 
 gaining of this knowledge by experience alone is likely to be 
 too slow and expensive, judged by twentieth century standards. 
 In fact, the variety and extent of the uses to which cement is 
 applied, and the knowledge concerning its properties, have of 
 late increased so rapidly that even the older engineer, whose 
 practice may have directed his special attention ' along other 
 channels for a few years, finds it difficult to follow its progress. 
 
 One who wishes only a catechetical reply to any question 
 that may arise concerning cement and its use will be somewhat 
 disappointed in these pages; on the other hand, he who would 
 devote special attention to the subject must, of course, go far 
 beyond them. The author has attempted to take a middle 
 course, avoiding on the one hand a dogmatic statement of facts, 
 and on the other too detailed and extended series of tests, but 
 giving, where practicable, sufficient tests to support the state- 
 ments made, and endeavoring to show the connection between 
 theory and practice, the laboratory and the field. 
 
 The original investigations forming the basis of the work 
 were made in connection with the construction of the Poe 
 Lock at St. Marys Falls Canal, Michigan, under the direction 
 
 iii 
 
iv PREFACE 
 
 of the Corps of Engineers, U. S. Army. To the late General 
 O. M. Poe, the Engineer officer in charge of the district at that 
 time, and to Mr. E. S. Wheeler, his chief assistant engineer, 
 m&y be credited a very large share of the value of the results 
 obtained, since the accomplishment of a series of experiments 
 of so comprehensive a character was made possible only through 
 the broad views held by them as to the value of thorough tests 
 of cement. 
 
 The author wishes to express his appreciation of the courtesy 
 of General G. L. Gillespie, Chief of Engineers, U. S. A., in grant- 
 ing permission to use the data collected, and of the kindness 
 of Major W. H. Bixby in presenting a request for this per- 
 mission. 
 
 When not otherwise stated, the tables in the work are con- 
 densed from the results of the above mentioned investigations. 
 In supplementing this original matter, much use has been made 
 of the experiments of others as published in society transac- 
 tions, technical journals, etc., to all of whom credit has been 
 given in the body of the work. 
 
 If this attempt to place in one volume a connected story of 
 the properties and use of cement serves to make the road to 
 this knowledge a little less devious than that followed by the 
 writer, the latter will be rewarded. 
 
 L. C. S. 
 
 SAULT STE. MARIE, MICH. 
 January 3, 1905. 
 
CONTENTS 
 
 PART I. CEMENT: CLASSIFICATION AND 
 MANUFACTURE 
 
 CHAFFER I. DEFINITIONS AND CONSTITUENTS 
 
 PAfiK 
 
 ART. 1. GENERAL CLASSIFICATION OF HYDRAULIC PRODUCTS .... 1 
 
 ART. 2. LIME: COMMON AND HYDRAULIC 3 
 
 ART. 3. PORTLAND CEMENT 4 
 
 ART. 4. SLAG CEMENT 7 
 
 ART. 5. NATURAL CEMENT 8 
 
 CHAPTER II. MANUFACTURE 
 
 ART. 6. MANUFACTURE OF PORTLAND CEMENT 10 
 
 Materials. Wet Process. Dry Process. Semi-dry Process. 
 Details of the Manufacture: Burning, Grinding. Sand-Cement. 
 
 ART. 7. OTHER METHODS OF MANUFACTURE OF PORTLAND .... 22 
 
 ART. 8. MANUFACTURE OF SLAG CEMENT 23 
 
 ART. 9. MANUFACTURE OF NATURAL CEMENT 24 
 
 PART II. PROPERTIES OF CEMENT AND 
 METHODS OF TESTING 
 
 CHAFfER III. INTRODUCTORY 
 Desirable Qualities. Uniform Methods of Testing 28 
 
 CHAPTER IV. CHEMICAL TESTS 
 ART. 10. COMPOSITION AJMD CHEMICAL ANALYSIS 31 
 
 CHAPTER V. THE SIMPLER PHYSICAL TESTS 
 
 ART. 11. MICROSCOPICAL TESTS. COLOR 36 
 
 ART. 12. WEIGHT PER CUBIC FOOT, OR APPARENT DENSITY .... 37 
 ART. 13. SPECIFIC GRAVITY, OR TRUE DENSITY. . 39 
 
 CHAPTER VI. SIFTING AND FINE GRINDING 
 
 ART. 14. FINENESS 45 
 
 Importance of Fineness. Sieves. Methods. Specifications. 
 
 v 
 
vi CONTENTS 
 
 PAGE 
 ART. 15. COARSE PARTICLES IN CEMENT 52 
 
 Effect on Weight, Time of Setting and Tensile Strength. 
 
 ART. 16. FINE GRINDING 58 
 
 Effect on Weight, Time of Setting and Tensile Strength. 
 
 CHAPTER VII. TIMfe OF SETTING AND SOUNDNESS 
 
 ART. 17. SETTING OF CEMENT 65 
 
 Process of Setting. Rate. Variations in Rate. 
 ART. 18. CONSTANCY OF VOLUME 76 
 
 Causes of Unsoundness. Tests. Discussion of Methods. Hot 
 
 Tests for Natural Cements. Conclusions. 
 
 CHAPTER VIII. TESTS OF THE STRENGTH OF CEMENT 
 IN COMPRESSION, ADHESION, ETC. 
 
 ART. 19. TESTS IN COMPRESSION AND SHEAR 89 
 
 ART. 20. TESTS OF TRANSVERSE STRENGTH 90 
 
 ART. 21. TESTS OF ADHESION AND ABRASION 92 
 
 CHAPTER IX. TENSILE TESTS OF COHESION 
 
 ART. 22. SAND FOR TESTS 95 
 
 Value of Tests of Sand Mortars. Uniformity in Sand. Com- 
 parison of Different Kinds. Tests with Natural Sand. Fineness. 
 
 ART. 23. MAKING BRIQUETS 97 
 
 Proportions. Consistency. Temperature. Gaging: Hand and 
 Machine. Methods. Amount of Gaging. Form of Briquets. 
 Molds. Molding. Briquet Machines. Approved Methods 
 of Hand Molding. Marking the Briquets. 
 
 ART. 24. STORING BRIQUETS 117 
 
 Time in Air before Immersion. Moist Closet. Water of Im- 
 mersion. Storing in Air; in Damp Sand. 
 
 ART. 25. BREAKING THE BRIQUETS 
 
 Testing Machines. Clips. Clip-breaks. Comparative Tests of 123 
 Clips. Requirements for a Perfect Clip. Form Recommended. 
 Rate of Applying Tensile Stress. Treatment of Results. 
 
 ART. 26. INTERPRETATION OF TENSILE TESTS OF COHESION .... 137 
 
 CHAPTER X. RECEPTION OF CEMENT AND RECORDS 
 OF TESTS 
 
 ART. 27. STORING AND SAMPLING 144 
 
 Storage Houses. Percentage of Barrels to Sample. Method of 
 
 Taking and Storing the Sample. 
 ART. 28. RECORDS OF TESTS 146 
 
 Value of Records. Marking Specimens. Records at St. Marys 
 
 Falls Canal. 
 
CONTENTS vii 
 
 PART III. THE PREPARATION AND PROP- 
 ERTIES OF MORTAR AND CONCRETE 
 
 CHAPTER XI. SAND FOR MORTAR 
 
 PAGK 
 
 ART. 29. CHARACTER OF THE SAND 154 
 
 Shape and Hardness of the Grains. Siliceous vs. Calcareous 
 Sands. Slag Sand. Sand for Use in Sea Water. 
 
 ART. 30. FINENESS OF SAND 150 
 
 Relation Volume and Superficial Area. Effect of Fineness. 
 
 ART. 31. VOIDS IN SAND 162 
 
 Conditions Affecting Voids: Shape of Grains; Granulometric Com- 
 position. Effect on Tensile Strength of Mortar. Moist Sand. 
 
 ART. 32. IMPURITIES IN SAND 168 
 
 ART. 33. CONCLUSIONS. WEIGHT AND COST OF SAND 170 
 
 CHAPTER XII. MORTAR: MAKING AND COST 
 
 ART. 34. PROPORTIONS OF THE INGREDIENTS 172 
 
 Capacity of Cement Barrels. Equivalent Proportions by Weight 
 and Volume. Richness of Mortars. Effect of Pebbles. Con- 
 sistency. 
 
 ART. 35. MIXING THE MORTAR 177 
 
 Hand Mixing. Machine Mixing. 
 
 ART. 36. COST OF MORTARS 179 
 
 Ingredients Required. Tables of Quantities. Estimates of 
 Cost. Tables of Cost of Portland and Natural Cement Mortars. 
 
 CHAPTER XIII. CONCRETE: AGGREGATES 
 
 ART. 37. CHARACTER OF AGGREGATES 186 
 
 Proper Materials. Screenings in Broken Stone. Foreign In- 
 gredients. 
 
 ART. 38. SIZE AND SHAPE OF FRAGMENTS AND VOLUME OF VOIDS . . 188 
 Conditions Affecting Voids. Effect on Strength oT Concrete. 
 Gravel vs. Broken Stone. 
 
 ART. 39. STONE CRUSHING AND COST OF AGGREGATE 194 
 
 Breaking Stone by Hand. Stone Crushers. Cost of Aggregate. 
 Examples. 
 
 .CHAPTER XIV. CONCRETE MAKING: METHODS 
 AND COST 
 
 ART. 40. PROPORTIONS OF THE INGREDIENTS 200 
 
 Theory of Proportions. Determination of Amount of Mortar 
 Required. Aggregates Containing Sand. Required Strength. 
 
 ART. 41. MIXING CONCRETE BY HAND 203 
 
 Hand vs. Machine Mixing. Method of Hand Mixing; Number of 
 Men and Output; Examples. 
 
viii CONTENTS 
 
 PAGE 
 ART. 42. CONCRETE MIXING MACHINES 207 
 
 General Classification. Description of Machines. Basis of 
 
 Comparison. 
 
 ART. 43. CONCRETE MIXING PLANTS AND COST OF MACHINE MIXING 212 
 ART. 44. COST OF CONCRETE 218 
 
 Ingredients Required for a Cubic Yard. Examples of Actual Cost. 
 
 CHAPTER XV. THE TENSILE AND ADHESIVE STRENGTH 
 OF CEMENT MORTARS AND THE EFFECT OF VARIATIONS 
 
 IN TREATMENT 
 ART. 45. TENSILE STRENGTH OF MORTARS OF VARIOUS COMPOSITIONS . 
 
 AND AGES 227 
 
 ART. 46. CONSISTENCY OF MORTAR AND AERATION OF CEMENT . . 232 
 
 ART. 47. REGAGING OF CEMENT MORTAR 236 
 
 ART. 48. MIXTURES OF CEMENT WITH LIME, PLASTER PARIS, ETC. . 243 
 
 Mixtures of Portland and Natural. "Improved" Cement. 
 
 Ground Quicklime with Cement; Slaked Lime; Plaster of Paris. 
 
 Conclusions. 
 ART. 49. MIXTURES OF CLAY AND OTHER MATERIALS WITH CEMENT. 253 
 
 Effect of Powdered Limestone, Brick, etc. ; Sawdust ; Terra Cotta. 
 ART. 50. USE OF CEMENT MORTARS IN FREEZING WEATHER . . . 260 
 
 Effect of Frost on Set Mortars. Effect of Salt; Heating Materials; 
 
 Consistency ; Fineness of Sand. Conclusions. 
 ART. 51. THE ADHESION OF CEMENT 270 
 
 Adhesion between Portland and Natural. Adhesion to Stone and 
 
 Other Materials. Effect of Consistency; Regaging; Character of 
 
 Surface of Stone. Effect of Plaster of Paris. Adhesion to Brick; 
 
 Effect of Lime Paste. Adhesion to Rods of Iron and Steel. 
 
 CHAPTER XVI. COMPRESSIVE STRENGTH AND MOD- 
 ULUS OF ELASTICITY OF MORTAR AND CONCRETE 
 
 ART. 52. COMPRESSIVE STRENGTH OF MORTARS 288 
 
 Ratio of Compressive to Tensile Strength. 
 ART. 53. CONCRETES WITH VARIOUS PROPORTIONS OF INGREDIENTS 291 
 
 Effect of Consistency; Amount and Richness of Mortar; Methods 
 
 of Storage. 
 ART. 54. CONCRETES WITH VARIOUS KINDS AND SIZES OF AGGREGATES 298 
 
 ART. 55. CINDER CONCRETE AND EFFECT OF CLAY 302 
 
 ART. 56. MODULUS OF ELASTICITY OF CEMENT MORTAR AND CONCRETE 306 
 
 CHAPTER XVII. THE TRANSVERSE STRENGTH AND 
 OTHER PROPERTIES OF MORTAR AND CONCRETE 
 
 ART. 57. TRANSVERSE STRENGTH 313 
 
 Transverse Strength of Mortars Compared to Tensile and Com- 
 pressive Strength. Richness of Mortar; Consistency. Transverse 
 Tests of Concrete Bars: Variations in Mortar Used; Consistency; 
 Mixing ; Aggregate ; Screenings. Deposition in Running Water. 
 Use in Freezing Weather, 
 
CONTENTS ix 
 
 PAGE 
 
 ART. 58. RESISTANCE TO SHEAR AND ABRASION 328 
 
 ART. 59. EXPANSION AND CONTRACTION OF CEMENT MORTAR, AND 
 
 THE RESISTANCE OF CONCRETE TO FIRE 331 
 
 Change in Volume during Setting. Coefficient of Expansion of 
 Mortar and Concrete. Fire-Resisting Qualities of Concrete. 
 Aggregate for Fireproof Work. 
 
 ART. 60. PRESERVATION OF IRON AND STEEL BY MORTAR AND CONCRETE 336 
 Action of Corrosion. Tests of Effect of Concrete. 
 
 ART. 61. POROSITY, PERMEABILITY, ETC 340 
 
 Porosity. Permeability. Waterproof Mortars and Concretes. 
 Washes for Exteriors of Walls. Efflorescence. Pointing Mortar. 
 Cements in Sea Water. 
 
 PART IV. USE OF MORTAR AND CONCRETE 
 
 CHAPTER XVIII. CONCRETE: DEPOSITION 
 
 ART. 62. TIMBER FORMS OR MOLDS 351 
 
 Sheathing. Lining. Posts and Braces. 
 
 ART. 63. DEPOSITION OF CONCRETE IN AIR 358 
 
 Transporting, Depositing, Ramming. Rubble Concrete. Fin- 
 ish ; Plastering; Facing; Bushhammering; Colors for Concrete Finish. 
 
 ART. 64. PLACING CONCRETE UNDER WATER 369 
 
 Laitance. Tremie, Skip, etc. Depositing in Bags; Cost. 
 Block System: Molds; Cost. 
 
 CHAPTER XIX. CONCRETE-STEEL 
 
 ART. 65. MONIER SYSTEM 381 
 
 ART. 66. WUNSCH, MELAN, AND THACHER SYSTEMS 383 
 
 ART. 67. OTHER SYSTEMS OF CONCRETE-STEEL 385 
 
 Hennebique, Kahn, Ransome, Roeblirig, Expanded Metal. 
 
 ART. 68. THE STRENGTH OF COMBINATIONS OF CONCRETE AND STEEL 387 
 
 ART. 69. BEAMS WITH SINGLE REINFORCEMENT 390 
 
 Formulas for Constant Modulus Elasticity; for Varying Modulus. 
 
 Excessive Reinforcement. Tables of Strength. 
 
 ART. 70. BEAMS WITH DOUBLE REINFORCEMENT 403 
 
 ART. 71. SHEAR IN CONCRETE-STEEL BEAMS 405 
 
 CHAPTER XX. SPECIAL USES OF CONCRETE: BUILD- 
 INGS, WALKS, FLOORS, AND PAVEMENTS 
 
 ART. 72. BUILDINGS 410 
 
 Roof; Floor System; Columns. Building Forms. N. Y. Build- 
 ing Regulations. 
 
 ART. 73. WALKS 420 
 
 Foundation; Base; Wearing Surface; Construction; Cost. 
 
 ART. 74. FLOORS OF BASEMENTS, STABLES, AND FACTORIES .... 426 
 
x CONTENTS 
 
 PAGE 
 
 ART. 75. PAVEMENTS AND DRIVEWAYS 428 
 
 Pavement Foundations. Concrete Wearing Surface. Construc- 
 tion. Example. 
 
 ART. 76. CURBS AND GUTTERS 431 
 
 ART. 77. STREET RAILWAY FOUNDATIONS 433 
 
 CHAPTER XXI. SPECIAL USES OF CONCRETE (CONTINUED): 
 SEWERS, SUBWAYS, AND RESERVOIRS 
 
 ART. 78. SEWERS 436 
 
 Methods and Cost. Forms . 
 ART. 79. SUBWAYS AND TUNNEL LINING 443 
 
 Waterproofing. Subways. Tunnels in Firm Earth ; in Soft 
 Ground; in Rock. Examples; Methods; Cost. 
 
 ART. 80. RESERVOIRS: LININGS AND ROOFS 453 
 
 Details of Construction. Groined Arch. Forms. Examples; 
 Cost. 
 
 CHAPTER XXII. SPECIAL USES OF CONCRETE (CONTINUED): 
 BRIDGES, DAMS, LOCKS, AND BREAKWATERS 
 
 ART. 81. BRIDGE PIERS AND ABUTMENTS AND RETAINING WALLS . . 464 
 
 Bridge Piers; Steel Shells. Repair of Stone Piers. Retaining 
 
 Walls and Abutments: Coping; Rules for Use of Concrete. 
 ART. 82. CONCRETE PILES 471 
 
 Building in Place. Concrete-Steel Piles: Molding; Driving. 
 ART. 83. ARCHES 474 
 
 Design; Centers; Construction; Finish and Drainage. Examples 
 
 and Cost. 
 ART. 84. DAMS 484 
 
 Concrete vs. Rubble. Quality of Concrete. Construction. 
 
 Examples. 
 ART. 85. LOCKS 488 
 
 Methods of Building. Examples. 
 ART. 86. BREAKWATERS . 493 
 
PART I 
 CEMENT 
 
 CLASSIFICATION AND MANUFACTURE 
 
 CHAPTER I 
 
 DEFINITIONS AND CONSTITUENTS 
 ART. 1. GENERAL CLASSIFICATION OF HYDRAULIC PRODUCTS 
 
 1. The use of a cementitious substance for binding together 
 fragments of stone is older than history, and it is known that the 
 ancient Romans prepared a mortar which would set under 
 water. So far as our present knowledge of cement manufac- 
 ture is concerned, however, the credit of demonstrating that a 
 limestone containing clay possessed, when burned and ground, 
 the property of hardening under water, is due to Mr. John 
 Smeaton, who announced this as the result of his experiments 
 made in 1756 in seeking a material with which to build the 
 Eddystone Lighthouse. After this discovery by Smeaton nearly 
 sixty years elapsed before M. Vicat gave the true explanation 
 of this action, namely, that the lime during burning combined 
 with the silica to form silicate of lime, the essential ingredient 
 of hydraulic limes and cements. 
 
 In 1796, Parker, of London, obtained a patent for the manu- 
 facture of a cement from septaria nodules, and aptly named his 
 product "Roman Cement." In 1824, Joseph Aspdin of Leeds, 
 England, patented a process of manufacture of "Portland 
 Cement." 
 
 2. The cements in general use in the United States to-day 
 are of two kinds, Portland cements and natural cements, and in 
 what follows our attention will be directed almost entirely to 
 1hese two products. 
 
 Common limes were formerly used largely in engineering 
 construction, but have of late been almost entirely superseded, 
 
2 CEMENT AND CONCRETE 
 
 for this purpose, by cements. Since the hardening of lime 
 mortar depends on the absorption of carbonic acid from the 
 atmosphere, these limes are sometimes called "air limes/' while 
 the hydraulic products which set under water are, for a similar 
 reason, styled "water limes." Hydraulic limes, though playing 
 an important role in foreign countries, are not manufactured or 
 used to any extent in the United States. The European prod- 
 uct known as "Roman" or "Vassy" cement, somewhat re- 
 sembles our natural cement, but is usually inferior to the Ameri- 
 can article. Our chief interest in these products, which are used 
 only abroad, is to know what relation they bear to the cements 
 with which we are familiar. The following classifications are 
 selected as being authoritative: 
 
 3. The conferences of Dresden (1886) and Munich (1884) on 
 Uniform Methods of Testing for Materials of Construction, clas- 
 sified the hydraulic products as follows : - 
 
 (1) Hydraulic limes: made by roasting either argillaceous or 
 siliceous limestones. They slake partially or wholly on the ad- 
 dition of water. 
 
 (2) Roman cements: made from argillaceous limestones hav- 
 ing a large proportion of clay. They do not slake by the addi- 
 tion of water and hence must be mechanically ground to powder. 
 
 (3) Portland cements: obtained by burning to the point of 
 insipient vitrification either hydraulic limestones or mixtures of 
 argillaceous materials and limestones, and afterward grinding 
 the product to fine powder. 
 
 (4) Hydraulic gangues: natural or artificial materials which 
 do not harden alone, but which furnish hydraulic mortars when 
 mixed with quicklime. 
 
 (5) Pozzolana cements produced by an intimate mixture 
 of powdered hydrate of lime and finely pulverized hydraulic 
 gangues. 
 
 (6) Mixed cements: the products of intimate mixtures of 
 manufactured cement with certain materials proper for such a 
 purpose. Mixed cements should always be designated as such 
 and the materials entering into the composition should be stated, 
 but it may be added parenthetically that these things are 
 seldom done. 
 
 4. MM. Durand-Claye and Debray divide cements into six 
 classes, namely, (1), Grappier cements obtained by grinding 
 
LIME 3 
 
 the pieces of hydraulic lime which do not slake; (2), quick-set- 
 ting (Vassy) cements formed by burning very argillaceous 
 limestones at a low temperature; (3), natural Portland cements, 
 or those cements made from natural rock which correspond to 
 artificial Portland in character; (4), mixed cements; (5), arti- 
 ficial Portlands; and (6), slag cements. 
 
 M. H. LeChatelier, an eminent French authority, divides 
 hydraulic products into four classes, namely : l Portland ce- 
 ments, hydraulic limes, natural cements, and mixed cements. Ho 
 subdivides the third class, natural cements, into quick-setting, 
 slow-setting and grappier cements, and includes natural Port- 
 lands among the slow-setting natural cements. Slag cements, 
 which are put in a separate class by MM. Durand-Claye and 
 Debray, are included in "mixed cements" by M. LeChatelier. 
 
 5. Prof. I. 0. Baker gives a classification that is better 
 adapted for use in this country than any of the above. 2 He 
 divides the products obtained by burning limestone, either pure 
 or impure, into lime, hydraulic lime and hydraulic cements. He 
 then sub-divides cement into Portland, Rosendale (preferably 
 
 called natural) and Pozzolana. 
 
 
 
 ART. 2. LIME: COMMON AND HYDRAULIC 
 
 6. Common lime is the product obtained by burning a pure, 
 or nearly pure, carbonate of lime. On being treated with water 
 it slakes rapidly, evolving much heat and increasing greatly in 
 volume. It is now seldom used in engineering construction and 
 will not be considered further. 
 
 7. Prof. M. Tetmajer has thus denned hydraulic limes: Hy- 
 draulic limes are the products obtained by the burning of argil- 
 laceous or siliceous limestones, which, when showered with water, 
 slake completely or partially without sensibly increasing in 
 volume. According to local circumstances, hydraulic limes may 
 be placed on the market either in lumps, or hydrated and pul- 
 verized. The following table gives a classification of hydraulic 
 limes according to M. E. Candlot f who states that the first 
 
 1 "Tests of Hydraulic Materials," by H. LeChatelier. Trans. Am. Inst. 
 Mining Engrs., 1893. 
 
 2 " Masonry Construction," p. 48. 
 
 3 "Ciments et Chaux Hydrauliques," par E. Candlot. 
 
CEMENT AND CONCRETE 
 
 class is seldom used for important work and that the fourth 
 class is quite rare. 
 
 TABLE 1 
 Classification of Hydraulic Limes. E. Candlot 
 
 Class. 
 
 Per Cent, 
 of Clay in 
 Limestone. 
 
 Per Cent, 
 of Silica 
 and Alumi- 
 na in Fin- 
 ished Prod- 
 uct. 
 
 Hydraulic 
 Index, or 
 Ratio of 
 Silica and 
 Alumina to 
 Lime. 
 
 Approx. 
 Time to 
 Set, 
 Days. 
 
 Feebly Hydraulic Lime 
 Ordinary " " 
 Real " " 
 
 5 to 8 
 8 to 15 
 15 to 19 
 
 9 to 14 
 14 to 24 
 24 to 30 
 
 .10 to .16 
 .16 to .31 
 .31 to .42 
 
 16 to 30 
 10 to 15 
 5 to 9 
 
 Eminently " " 
 
 19 to 22 
 
 80 to 33 
 
 .42 to .50 
 
 2 to 4 
 
 Hydraulic limes should be burned slowly, and at such a tem- 
 perature that sintering does not take place. The best hydraulic 
 limes have a composition very similar to that of Portland cement. 
 The comparatively low temperature at which they are burned 
 permits them to slake on the addition of water. They gain 
 strength much more slowly than cements. 
 
 Having considered the classification of hydraulic products as 
 a whole, we may proceed to the discussion of Portland and nat- 
 ural cements, the hydraulic products which have by far the 
 greatest importance here, and the only varieties which will be 
 taken up in detail in the present work. 
 
 ART. 3. PORTLAND CEMENT 
 
 8. As the classification of hydraulic products varies, so do 
 opinions vary as to what shall be included under the name Port- 
 land cement. There seems to be agreement on at least one 
 point, namely, that the burning shall be carried to a point just 
 short of vitrification. Ideas concerning other points are crys- 
 tallizing rapidly. The Association of German Portland Cement 
 Manufacturers has given a definition of Portland cement in a 
 practical manner by binding its members "to produce under 
 the name of Portland cement only such an article as is made by 
 calcining a thorough mixture, consisting essentially of calcare- 
 ous and clayey substances, and then grinding the same to the 
 fineness of flour;" and they further declare that "any article 
 made in a manner differing from the above method, or to which 
 during or after burning any foreign substances have been added/' 
 
PORTLAND CEMENT 5 
 
 is not recognized by them as Portland cement, and the sale of 
 such products under the designation "Portland Cement" is re- 
 garded by them as defrauding the purchaser. This declaration 
 does not apply to such minor additions as are made to regulate 
 the setting time of Portland cement, and which are permitted 
 to an extent of 2 per cent." 
 
 9. M. LeChatelier has given the following limits for the 
 amounts of the materials usually contained in good commercial 
 Portland cements : l - 
 
 Silica 21 per cent, to 24 per cent. 
 
 Alumina 6 
 
 Oxide of Iron 2 
 
 Lime 60 
 
 Magnesia .5 
 
 Sulphuric Acid 5 
 
 Water and Carbonic Acid . 1 
 
 8 
 4 
 
 65 
 2 
 
 1.5 
 3 
 
 The upper limit for lime (65 per cent.) is being exceeded in re- 
 cent years. 
 
 These substances occur as "(1) SiO 2 , 3CaO, the essentially 
 cementitious ingredient; (2) A1 2 O 3 , 3CaO, the substance mainly 
 active during setting and contributing somewhat to the subse- 
 quent hardening; and (3) a fusible calcium silico-aluminate 
 whose chief function is that of a flux during burning to promote 
 the necessary chemical reactions." 2 M. LeChatelier further 
 holds that in good Portland cements the following formulas 
 
 should be true: 
 
 CaO, Mg O < 
 SiO 2 + A1 2 O3- 3 ' 
 and 
 
 CaO, MgO > 
 
 SiO 2 - (A1 2 O 3 , Fe 2 O 3 ) = ' 
 
 in each case the quantities in the formulas being equivalents of 
 the substances, not weights. The ratio of the acid constitu- 
 ents, silica and alumina, and the basic constituents, lime and 
 magnesia, is called the hydraulic index. Although these form- 
 ulas have been quite generally accepted as properly fixing the 
 limit* of the ingredients it maybe noted that they are based on 
 the assumption that SiO 2 , and A1 2 O 3? are equally capable of dispos- 
 
 1 "Tests of Hydr. Materials," Tr. Am. Tnst. Mining Engrs., 1893. 
 
 2 Jour. Soc. Ch. Ind., Mar. 31, 1891, p. 256. 
 
6 
 
 CEMENT AND CONCRETE 
 
 ing of a given quantity of lime and magnesia, and it is thought 
 by some authorities that the assumption is not warranted. 
 
 In the Journal Society Chemical Industry, 1897, Messrs. S. 
 B. and W. B. Newberry give the results of some investigations 
 in this line from which they concluded that the essential in- 
 gredient of Portland cement is a tri-calcium silicate, but that 
 the alumina occurs as a dicalcic aluminate. They therefore 
 considered that the per cent, of lime should equal 2.8 times the 
 per cent, of silica plus 1.1 times the per cent, of alumina. 
 
 10. The following analyses of brands in the market are se- 
 lected from the various sources indicated in the table. They 
 are given here merely to illustrate the proportions obtaining in 
 
 commercial products. 
 
 TABLE 2 
 
 Analyses of Portland Cements 
 
 BRAND. 
 
 Si0 2 . 
 
 Ai 2 O 3 . 
 
 Fe 2 3 . 
 
 CaO. 
 
 MgO. 
 
 Nti 2 
 K 2 O. 
 
 S0 3 . 
 
 H S O & 
 
 Loss. 
 
 1. Alpha 
 
 20.38 
 
 
 
 63.30 
 
 2.86 
 
 
 1.13 
 
 1.75 
 
 2. Atlas 
 
 21.30 
 
 7.65 
 
 '2.85 
 
 60.95 
 
 2.95 
 
 1.15 
 
 1.81 
 
 1.41 
 
 3. Bronson 
 
 22.90 
 
 6.80 
 
 3.60 
 
 63.90 
 
 0.70 
 
 1.10 
 
 0.40 
 
 0.60 
 
 4. Buckeye 
 
 21.30 
 
 6.95 
 
 2.00 
 
 62.30 
 
 1.20 
 
 
 0.98 
 
 4.62 
 
 5. Empire 
 
 22.04 
 
 6.45 
 
 3.41 
 
 60.92 
 
 3.53 
 
 
 2.25 
 
 
 6. Wyandotte 
 
 23.20 
 
 8.00 
 
 2.40 
 
 62.10 
 
 2.00 
 
 
 
 0.80 
 
 7. Omega 
 
 22.24 
 
 7.26 
 
 2.54 
 
 64.96 
 
 2.26 
 
 
 0.41 
 
 0.33 
 
 8. Yankton 
 
 
 7.70 
 
 4.80 
 
 60.00 
 
 0.80 
 
 1.20 
 
 
 
 9. Giant 
 
 23.36 
 
 8.07 
 
 4.83 
 
 58.93 
 
 1.00 
 
 0.50 
 
 0.50 
 
 2.46 
 
 10. Medusa 
 
 23.20 
 
 7.03 
 
 2.41 
 
 64.19 
 
 0.97 
 
 
 
 2.20 
 
 11. Dyckerhoff 
 
 19.35 
 
 7.00 
 
 4.50 
 
 63.75 
 
 
 
 
 5.40 
 
 12. German i a 
 
 21.14 
 
 6.30 
 
 2.50 
 
 66.04 
 
 1.11 
 
 
 
 2.91 
 
 13. Alsen's 
 
 24.90 
 
 8.00 
 
 3.22 
 
 59.38 
 
 0.38 
 
 0.75 
 
 0.98 
 
 2.16 
 
 14. Alsen's 
 
 23.30 
 
 5.85 
 
 4.65 
 
 60.90 
 
 0.90 
 
 0.30 
 
 2.43 
 
 1.40 
 
 3RAND. 
 
 AUTHORITY. 
 
 RAW MATERIALS. 
 
 LOCATION. - 
 
 1 
 
 " Directory Amer'n 
 
 Cement Rock and Limestone 
 
 Alpha, N.J. 
 
 
 Cement Industries" 
 
 
 
 2 
 
 u 
 
 11 11 H tt 
 
 Northampton, Pa. 
 
 3 
 
 II 
 
 Marl and Clay 
 
 Bronson, Mich. 
 
 4 
 
 It 
 
 U it tt 
 
 Bellefontaiue, Ohio. 
 
 5 
 
 tt 
 
 It tt It 
 
 Warners, N.Y. 
 
 6 
 
 1C 
 
 Soda Ash Waste and Clay 
 
 Wyandotte, Mich. 
 
 7 
 
 u 
 
 Marl and Clay 
 
 Jonesville, Mich. 
 
 8 
 
 ti 
 
 Chalk and Clay 
 
 Yankton, S. Dakota. 
 
 9 
 
 U. Cummings 
 
 Cement Rock and Limestone 
 
 Egypt, Pa. 
 
 
 " Amer'n Cements" 
 
 
 
 10 
 
 U It 
 
 Marl and Clay 
 
 Sandusky, Ohio. 
 
 11 
 
 U U 
 
 Limestone, Marl and Clay 
 
 Ainoeneburg, Ger. 
 
 12 
 
 41 (( 
 
 Marl and Clay 
 
 Lehrte, Germany. 
 
 13 
 
 U U 
 
 Chalk and Clay 
 
 Itzehoe, Germany. 
 
 14 
 
 "Richard K. Meade, 
 
 Chalk and Thames Mud 
 
 England. 
 
 
 "Exam, of P. Cem." 
 
 
 
SLAG CEMENT 7 
 
 ART. 4. SLAG CEMENT 
 
 11. Slag cement is manufactured to a considerable extent 
 in Europe and is beginning to assume some importance in the 
 United States. It is a pozzolana cement in which the silica 
 ingredient is supplied by blast furnace slag. Pozzolana ce- 
 ments have been defined as " products obtained by intimately 
 and mechanically mixing, without subsequent calcination, pow- 
 dered hydrates of lime with natural or artificial materials which 
 generally do not harden under water when alone, but do so 
 when mixed with hydrates of lime (such materials being pozzo- 
 lana, Santorin earth, trass obtained from volcanic tufa, furnace 
 slag, burnt clay, etc.), the mixed product being ground to ex- 
 treme fineness." 1 
 
 Slag cement somewhat resembles Portland in its properties, but 
 is more like some of the natural cements in its constituents, while 
 the manner of occurrence of these constituents and the method 
 of manufacture are quite different than in either of these 
 classes. 
 
 12. As this cement is a mixture of lime and pozzolanic ma- 
 terials, its value depends largely upon its extreme fineness and 
 the intimate mixture of the ingredients. Its specific gravity is 
 low, about 2.7 to 2.8, and it sets very slowly, although the 
 setting may be hastened by the addition of certain substances 
 such as caustic soda. On account of the sulphide present, 
 most slag cements are not suited to use in air, as they crack 
 and soften in this medium; neither are they suitable for use in 
 sea water, nor in freezing weather, but when mixed with two 
 or three parts sand and kept constantly wet with fresh water, 
 they give quite satisfactory results. 
 
 Slag cement has an approximate composition of silica, 20 to 
 30 per cent., alumina, 10 to 20 per cent., and lime, 40 to 50 per 
 cent. It usually contains calcium sulphide, the amount some- 
 times reaching three or four per cent. The characteristic green- 
 ish tint which slag cements exhibit when they harden in water 
 is due to this ingredient, as is the odor of hydrogen sulphide 
 sometimes given off by a briquet when broken, especially if it 
 
 1 " Report of Board of Engineers on Steel Portland Cement," Washing- 
 ton, 1900. 
 
8 
 
 CEMENT AND CONCRETE 
 
 has hardened in sea water, 
 a percentage of magnesia. 1 
 
 Some slag cements have also quite 
 
 ART. 5. NATURAL CEMENT 
 
 13. Natural cement, as its name implies, is made from rock 
 as it occurs in nature. Argillaceous limestones, magnesian lime- 
 stones, or argillo-magnesian limestones, having the proper pro- 
 portion of clay, magnesia and lime, may be used for the 
 production of natural cement. The burning is not carried so 
 far as in the manufacture of Portland cement, and the resulting 
 
 TABLE 3 
 Analyses of Natural Cements 
 
 
 
 
 
 
 j 
 
 
 <j K 
 
 
 
 jjj 
 
 
 
 IB 
 
 POTASH 
 
 * * T 
 
 REFER- 
 
 SILICA. 
 
 3 
 
 IRON 
 OXIDE. 
 
 LIME. 
 
 fe 
 
 o 
 
 AND 
 
 | g 1 
 
 ENCE. 
 
 
 1 
 
 
 
 
 SODA. 
 
 *** 
 
 
 c 
 
 d 
 
 e 
 
 / 
 
 9 
 
 h 
 
 i 
 
 1 
 
 24.30 
 
 2.61 
 
 6.20 
 
 39.45 
 
 6.16 
 
 5.30 
 
 15.23 
 
 2 
 
 34.66 
 
 5.10 
 
 1.00 
 
 30.24 
 
 18.00 
 
 6.16 
 
 4.84 
 
 3 
 
 23.16 
 
 6.33 
 
 1.71 
 
 36.08 
 
 20.38 
 
 5.27 
 
 7.07 
 
 4 
 
 26.40 
 
 6.28 
 
 LOO 
 
 45.22 
 
 9.00 
 
 4.24 
 
 7.86 
 
 5 
 
 27.30 
 
 7.14 
 
 1.80 
 
 35.98 
 
 18.00 
 
 6.80 
 
 2.98 
 
 6 
 
 27.98 
 
 7.28 
 
 1.70 
 
 37.59 
 
 15.00 
 
 7.96 
 
 2.49 
 
 7 
 
 27.69 
 
 8.64 
 
 2.00 
 
 42.12 
 
 14.55 
 
 2.00 
 
 3.00 
 
 8 
 
 27.60 
 
 10.60 
 
 0.80 
 
 33.04 
 
 7.26 
 
 7.42 
 
 2.00 
 
 9 
 
 28.02 
 
 10.20 
 
 8.80 
 
 44.48 
 
 1.00 
 
 0.50 
 
 7.00 
 
 
 
 PLACE OF 
 
 REFER- 
 
 BRAND. 
 
 MANUFACTURE. 
 
 ENCE. 
 
 
 
 
 a 
 
 b 
 
 1 
 
 Buffalo 
 
 Buffalo, N. Y. 
 
 2 
 
 Utica 
 
 Utica, 111. 
 
 3 
 
 Milwaukee 
 
 Milwaukee, Wis. 
 
 4 
 
 Louisville 
 
 Louisville, Ky. 
 
 5 
 
 Hoffman 
 
 Rosendale, N. Y. 
 
 6 
 
 Norton High Falls 
 
 Rosendale, N. Y. 
 
 7 
 
 Akron 
 
 Akron, N. Y. 
 
 8 
 
 Utica 
 
 LaSalle, 111. 
 
 9 
 
 Round Top 
 
 Hancock, Md. 
 
 Selected from table compiled by Mr. U. Cuinmings, " Brickbuilder, " May, 1895. 
 
 1 For an excellent resume of the qualities and distinguishing character- 
 istics of slag cements, the reader is referred to " Report of Board of Engi- 
 neers on Steel Portland Cement as used in United States Lock at Plaque- 
 mine, La." Washington, 1900. 
 
NATURAL CEMENT 9 
 
 product is of lighter weight and usually quicker setting, though 
 some natural cements are quite slow setting. The properties of 
 these cements, coming from different localities, vary greatly. 
 In fact, it is difficult to distinguish some natural cements from 
 Portland, and they may be considered to grade into the natural 
 Portlands. Light burning in manufacture, light weight per cubic 
 foot, and slower rate of acquiring strength, may be considered 
 the distinguishing characteristics from a physical point of view. 
 
 14. Analyses. Table 3 gives the results of a number of 
 analyses of natural cement, selected from a table compiled by 
 Mr. U. Cummings. 
 
 Comparing these analyses with those given for Portland ce- 
 ment in Table 2, it is seen that natural cements have a higher 
 percentage of silica, about the same amount of alumina, and a 
 much smaller content of lime, than have Portlands. Many natu- 
 ral cements have a large percentage of magnesia, but the mag- 
 nesia and lime together of natural cements usually do not equal 
 the percentage of lime in Portlands. In other words the hy- 
 draulic index is usually higher than in Portland cements. 
 
CHAPTER II 
 
 MANUFACTURE 
 
 ART. 6. THE MANUFACTURE OF PORTLAND CEMENT 
 
 15. Historical. It is said that as early as 1810 a patent 
 was obtained in England for the manufacture of an artificial 
 product by calcining a mixture of carbonate of lime and clay. 
 This, however, was not called cement, and it was not until 1824 
 that Joseph Aspdin, of Leeds, England, in obtaining a patent 
 for the manufacture of a similar material, called his product 
 " Portland Cement." This name was probably suggested by 
 the fact that the color of the hardened product resembled that 
 of a limestone quarried on the Island of Portland. The industry 
 was introduced into Germany about thirty years later, and has 
 since grown to very substantial proportions in both of these 
 countries, as well as in France, Austria and Russia. 
 
 David O. Saylor was the first to manufacture Portland ce- 
 ment in the United States, at Coplay, Pa., about 1872, and 
 works were established at that point in 1875. These were 
 soon followed by other factories in Pennsylvania and Indiana, 
 and at present cement is successfully manufactured in nearly 
 half of the states of the Union, the production having steadily 
 increased. 
 
 16. MATERIALS REQUIRED. The materials requisite for the 
 manufacture of Portland cement are carbonate of lime and 
 silica. The former may be in the form of limestone, chalk, 
 or calcareous marl, the last two being preferable on account of 
 greater ease of working. The silica may be in the form of 
 shale or clay, the latter to be preferred. The clay need not be 
 entirely free from impurities, but it should not contain any con- 
 siderable amount of sand, for although silica is the most useful 
 constituent of the clay, it must not be in this insoluble form. 
 Although formerly authorities did not agree as to whether the 
 alum'na in the clay was an unwelcome constituent for Portland 
 cement manufacture, it is now considered that the dicalcic or 
 
 10 
 
PORTLAND CEMENT 
 
 11 
 
 tricalcic aluminate formed plays a role in the setting of the 
 cement, and possibly also in the subsequent hardening. 
 
 A few analyses of materials suitable for Portland cement 
 manufacture are given in Table 4. 
 
 TABLE 4 
 
 Analyses of Cement Materials 
 
 MATERIALS. 
 
 SiO,. 
 
 A1 2 3 
 and 
 Fe 2 3 . 
 
 CaC0 3 . 
 
 MgC0 3 . 
 
 S0 3 . 
 
 Water 
 and 
 Loss. 
 
 White Marl, Empire l . . . 
 Clay, Empire l 
 Gray Marl . 
 
 .26 
 
 40.48 
 7.26 
 
 .10 
 20.95 
 1 49 
 
 94.39 
 25.80 
 84 10 
 
 .38 
 .99 
 .91 
 
 . . . 
 
 3.10 
 8.50 
 3 98 
 
 Clay 
 
 53 5 
 
 24.20 
 
 5.15 
 
 2.1') 
 
 
 14.10 
 
 Limestone, Glens Falls 1 . . 
 Clay, Glens Falls 1 .... 
 Gray Chalk Medway Eng 2 
 
 3.30 
 56.27 
 5 45 
 
 1.30 
 28.15 
 3 87 
 
 93.13 
 10.43 
 88 72 
 
 1.58 
 2.25 
 
 0.30 
 0.12 
 
 
 liiver Mud Medway En fr .* 
 
 71 71 
 
 16 70 
 
 4 0") 
 
 
 
 
 
 
 
 
 
 
 
 1 " Manufacture Portland Cement in New York State/' by Mr. Edwin C. 
 Eckel, C. E. 
 
 2 " Cement for Users," Mr. Henry Faija. 
 
 17. The materials for Portland cement manufacture, lime- 
 stone, marl, clay, shale, etc., are widely disseminated, but the 
 suitability of a certain locality for successful commercial manu- 
 facture depends upon the manner of occurrence of these requi- 
 sites. In England the clay is dug from the old beds of the 
 Thames and Medway Rivers, and chalk, which occurs in abun- 
 dance, furnishes the carbonate of lime in most cases, though 
 limestone is sometimes used. In Germany both chalk and 
 marl are used; the chalk being a soft white marl similar to the 
 deposits in this country, and the marl a "more or less hard 
 limestone rock containing clay." In the United States both 
 limestones and marls are used. The most important cement 
 producing egion in the United States is in the Lehigh Valley, 
 where an argillaceous limestone is employed. The factories 
 using marl are situated in New York, Ohio, Indiana, Michigan, 
 etc., where the marl is found overlying beds of clay suitable for 
 cement making. In the Lehigh Valley region many advan- 
 tages are combined. The cement rock of that locality has 
 nearly the correct composition for Portland cement manufac- 
 
12 CEMENT AND CONCRETE 
 
 ture. The supply of this rock is almost inexhaustible, the man- 
 agers of the works have had long experience in the production 
 of cement from these materials, and a market for the product is 
 near at hand. 
 
 Deposits of cement materials are of value only when the 
 limestone or marl, and clay or shale, are found in large quanti- 
 ties and near together, when the physical character of the ma- 
 terials is such as to render them easy of comminution and mix- 
 ture, when coal or other suitable fuel may be had at low prices, 
 and when the market is not too far removed. 
 
 The following estimate of the relative quantities of cement 
 made in the United States in 1902 from the several classes of 
 materials has been made by Mr. E. C. Eckel : * 
 
 Argillaceous limestone and pure limestone .... 68 Per cent. 
 
 Marl and clay 14 " 
 
 Soft limestone and clay 4^ " 
 
 Hard limestone and clay . . 13 " 
 
 18. GENERAL DESCRIPTION OF PROCESSES. The essentials 
 of any method of Portland cement manufacture are that the 
 materials shall be correctly proportioned, very finely comminuted 
 and thoroughly mixed, that the mixture shall be carefully 
 burned to just the proper degree of calcination and the result- 
 ing clinker ground to extreme fineness. How these essentials 
 can be best accomplished depends upon the character of the 
 raw materials and the cost of fuel and labor, so that the de- 
 tails of the method vary with the materials used and with the 
 local conditions. 
 
 In order that the proportions may be accurately determined, 
 it is usually necessary to dry one or both of the raw materials. 
 The ingredients may be ground separately and afterward mixed, 
 though with certain materials the grinding and mixing may be 
 done at the same time. In this mixing, a large amount of water 
 may be used, as in the "wet process," giving a very thin slurry; 
 a moderate amount may be used, as in the semi-wet process, giv- 
 ing a slurry of creamy consistency; or the dry process may be 
 employed, where the amount of water used is no more than 
 sufficient to dampen the materials. The burning may be ac- 
 
 Engineering News, April 16, 1903. 
 
PORTLAND CEMENT 13 
 
 complished in any one of several styles of kiln, the selection 
 depending upon the relative cost of labor and fuel, the relative 
 necessity of economy and rapid production, and, perhaps we 
 should add, the rigidity of the specifications which the finished 
 product must fulfill. The grinding is a simple mechanical prob- 
 lem, to secure the required degree of fineness with least cost. 
 
 19. THE WET PROCESS. Although an excess of water may 
 be used to mix materials that require previous grinding, the wet 
 process is particularly adapted to such raw materials as are easily 
 acted upon by water. This method was developed in England, 
 where it is still employed to some- extent and it has 'been used 
 in this country as well. 
 
 Proper amounts of the raw materials, previously ground if 
 necessary, are placed in a wash mill with a large amount of 
 water. The wash mill is a circular trough in which teeth or 
 arms are made to revolve, agitating the mass. When the mate- 
 rials are so finely divided as to be held in suspension, the thin 
 slurry is run off into " backs, " or shallow settling reservoirs, 
 where the solid matter settles; the clear liquid is then run off, 
 the slurry being allowed to dry further until it can be cut into 
 bricks and placed on drying floors artificially heated. The 
 bricks are then taken to the burning kilns and finally ground to 
 form the finished product. 
 
 The disadvantages of this method are that much space is 
 required for the settling floors, the amount of heat required to 
 dry the brick is excessive, and the process is necessarily slow. 
 These disadvantages are so great that the method above outlined 
 is rapidly falling into disuse. Materials particularly adapted to 
 wet mixing are still treated by this process, but the wet mixture 
 is run directly into very long rotary kilns and is dried in passing 
 through the first half of the length, which is heated by the gases 
 from the lower portion where the burning is completed. 
 
 20. THE DRY PROCESS. This method of manufacture is 
 best adapted to materials, such as limestone and shale, that 
 must be dried and ground before they can be mixed. The rock 
 as it comes from the quarry is first passed through a rock 
 crusher, reducing it to the size of broken stone used for con- 
 crete; then to some other form of crusher, such as heavy rolls, 
 until it is reduced to pieces about one-half inch or less in size. 
 It is then dried by artificial heat. 
 
14 CEMENT AND CONCRETE 
 
 The materials may now be combined in proper proportions 
 and ground together to extreme fineness, thereby becoming 
 thoroughly mixed. If the mixture is to be burned in the old 
 style kiln, it must now be dampened so that it may be pressed 
 into bricks to be charged in the kiln. If a rotary kiln is used, 
 however, the dry mixture may be fed directly into it, or it may 
 be moistened enough so that it will form into little lumps the 
 size of wheat grains, and these fed to the rotary. 
 
 21. THE SEMI-DRY PROCESS. The two processes briefly de- 
 scribed above are extremes admitting many modifications which 
 will not be entered into in detail. What may be called the 
 semi-dry process, however, has been so widely used in the 
 United States that it deserves some special mention, and it may 
 perhaps be best explained by giving the method formerly em- 
 ployed in a well-known American factory which, until a few 
 years ago, was using the vertical kiln. 
 
 The carbonate of lime in the form of marl was found above 
 the clay in beds varying in thickness up to 20 feet. The clay 
 in general contained little sand, and the beds were of such 
 thickness that whenever too much sand was present, the clay 
 might be wasted. The materials were delivered to the factory, 
 about three-quarters of a mile from the deposit, by small cars 
 running on a narrow gage railroad. 
 
 When the clay reached the factory it was put in shallow 
 wooden pans and run into dry kilns on light cars. After dry- 
 ing, which required 36 to 48 hours, the clay was ground and de- 
 livered in weighed quantities to the mixer. The main object 
 of drying the clay was to be able to control the amount added 
 to a given quantity of marl, and the grinding was to facilitate 
 the mixing of the two ingredients. As the cars of marl entered 
 the building, they were brought to a given weight by means 
 of a scale, which was set and locked by the manager. The 
 marl was then dumped directly into the wei pan or mixer. 
 
 The latter consisted of an iron pan, about 12 feet in diameter, 
 in which revolved two cast iron rollers weighing three tons 
 each. These rollers were on opposite ends of a horizontal axis 
 which was attached to a vertical shaft in the center of the pan. 
 This shaft being driven from below, the rollers traveled in a 
 circular path; as the rollers were hung loose on the horizontal 
 axis, they revolved about the latter only when sufficient fric- 
 
PORTLAND CEMENT 15 
 
 tion was developed between their peripheries and the floor of 
 the pan. In front of each roller traveled two blades, one of 
 which pushed the material under the roller from the center, 
 while the other did the same from the circumference. 
 
 A weighed amount of dry, powdered clay was admitted at 
 the side of the mixer, from a hopper scale, at the same time as 
 the marl was dumping into it, and sufficient water was added 
 through a hose to bring the contents of the pan to a pasty 
 mass. Five minutes were allowed for mixing each charge, when 
 a slide was drawn, leaving two holes in the path of the wheels 
 and on opposite sides of the pan. The material, or "mix," 
 was delivered on a belt conveyor and carried to a pug mill, 
 whence it issued in the form of rough bricks, partially cut by 
 wires into six-inch cubes. These cubes, being loaded on cars, 
 were run into the dry kiln, where they remained from two to 
 four days, and were then taken to the kiln room to be filled, by 
 hand, into the burning kilns, which were of the dome type. 
 
 In charging, layers of cement-brick and coke were alter- 
 nated. For convenience, as well as to prevent the bricks being 
 crumbled by a fall, the charging was done from three levels. 
 From 36 to 72 hours were required for burning, a charge. The 
 kiln was then opened at the mouth, and the clinker, which had 
 shrunk in volume about three-quarters, and in weight about 
 one-half, was drawn off as fast as it cooled. The clinker was 
 shoveled from the kilns to a pan conveyor and sorted as shov- 
 eled, only that which appeared properly burned being allowed 
 to pass; the underburned portion was stored for further burn- 
 ing, and the overburned, wasted. Further sorting was done 
 by two men stationed in the kiln room, who watched the 
 clinker as it passed on the conveyor and picked out any pieces 
 defective in burn that might have passed the hands of the 
 shovelers. 
 
 The conveyor delivered the clinker to a Blake crusher, which 
 broke it into pieces the size of pebbles; thence it passed to 
 horizontal millstones, or, to what replaced these, ball and 
 tube mills, for final reduction. The material was then deliv- 
 ered into cylindrical screens having about 2,500 meshes per 
 square inch, that portion retained in the screen being returned 
 to a stone supplied almost entirely with these screenings. The 
 cement was then conveyed to the stock house, which was divided 
 
16 CEMENT AND CONCRETE 
 
 into bins of 1,500 barrels capacity, and finally packed in barrels 
 by means of a screw blade fitting the interior of the barrel. 
 
 22. DETAILS OF THE MANUFACTURE: Preparation and Mix- 
 ing of the Raw Materials. The main points in the preparation 
 of the raw materials for burning are : first, the proper amount of 
 each ingredient must enter the mixture; second, the materials 
 must be reduced to an extremely fine state of division, with no 
 lumps; and third, the mechanical mixing must be as perfect as 
 possible. Unless the ingredients are dried, the first require- 
 ment is difficult to accomplish, especially with marl and clay, 
 as the absorptive power of the materials renders it difficult to 
 properly apportion them. More than three-fourths of the 
 Portland cement manufactured in the United States is made 
 from limestones. These must be ground before they can re- 
 ceive the required addition of clay or of purer limestone, as the 
 case may be, and they are usually dried to facilitate the grind- 
 ing as well as to permit of determining the correct proportions 
 of the ingredients. These hard materials are first crushed in 
 an ordinary stone crusher or between heavy rolls, then dried 
 in rotary driers, or otherwise; next, mixed and ground together 
 to an extreme .fineness in ball or tube mills. When rotary 
 kilns are employed, the mix may be burned dry, but with 
 fixed kilns, it is moistened to form bricks which are charged in 
 the kilns with alternate layers of coke. 
 
 Soft materials, such as marl and clay, are easy of reduc- 
 tion in water, and are naturally treated by the wet or semi- 
 dry process, although they may be prepared by the dry process. 
 In the former method the grinding and mixing are accom- 
 plished by edge runners, pug mills or wash mills. If the ma- 
 terials have not been dried before mixing, the mix or slurry 
 should be sampled and analyzed before it is passed to the kilns. 
 When fixed kilns are employed, it is desirable that the bricks 
 should be as porous as possible, that the fire may more readily 
 reach the interior of the brick. It is claimed by some manu- 
 facturers that by spreading the slurry on a floor to dry, and 
 then cutting into rough cubes when dry enough to be taken to 
 the dry kiln, more porous bricks are obtained. 
 
 23. BURNING: STYLES OF KILNS. The various styles of 
 kilns in use may be divided into four classes, namely: (1) 
 Common dome kilns, (2) Continuous kilns, (3) Chamber and 
 
PORTLAND CEMENT 17 
 
 ring kilns, and (4) Rotary kilns. The dome kiln is the 
 simplest type. The chamber is usually egg shaped. Cement- 
 brick and coke are piled in alternate layers, the use of 
 the proper amount of the latter requiring much skill, as it is 
 a matter of experience. As the draft in the kiln varies with 
 the weather, this method of burning is more or less at the mercy 
 of the winds. When the burning is complete, the kiln is al- 
 lowed to cool before removing the clinker, and thus much heat 
 is lost, and the lining of the kiln is destroyed by alternate heat- 
 ing and cooling. The amount of underburned and over- 
 burned clinker is likely to be large. The output is small, and 
 fuel expense high. 
 
 The Dietsch kiln is one of the best examples of the second 
 type, or continuous kiln. The slurry, in the form of bricks, is 
 introduced at the base of the stack, into what may be called 
 the heating chamber. Below this there is a right angle with a 
 short horizontal section, over which the hot slurry is raked, 
 to fall into the burning chamber. The clinker in the lower 
 part of the latter is cooled by the air entering through the grates, 
 while the slurry in the upper chamber is heated by the gases 
 from the burning zone. At intervals a portion of the clinker, 
 partially cooled, is removed at the bottom; this causes a general 
 settlement in the kiln and leaves a space at the top of the burn- 
 ing chamber, into which the dried clinker from above is raked, 
 and more fuel added. This kiln uses small coal for fuel and is 
 more economical than the dome type. 
 
 The distinguishing feature of the Schofsr kiln is the con- 
 traction of the dome at the point where combustion takes 
 place, concentrating the draft at this point. The air entering 
 the shaft at the bottom cools the clinker already burned, while 
 the gases from the clinker burning in the central section serve 
 to dry the raw bricks above. Several kilns of this type are in 
 successful operation in this country. 
 
 24. Chamber kilns are used largely in England with coke as 
 fuel. The gases from the kiln are made to pass over the slurry 
 spread on brick floors, the kiln proper being at one end of this 
 chamber and the stack at the other. These kilns are inter- 
 mittent, have a comparatively small output, and require con- 
 siderable labor. 
 
 The Hoffman ring kiln consists of a series of compartments 
 
18 CEMENT AND CONCRETE 
 
 built around a large central stack. The chambers communicate 
 by means of flues in such a way that the smoke and hot gases 
 from one may be passed through other chambers before reach- 
 ing the chimney. The kiln may be either "up draft" or "down 
 draft/ 7 according to the direction in which the heat is drawn 
 through the chamber. The compartments* are charged from 
 the sides, and when the moisture has been driven off from the 
 material in the chamber first fired, the gases from this chamber 
 are passed through the adjacent chambers, which have in the 
 meantime been filled with raw materials. Although this kiln 
 is economical of fuel if run continuously, much labor is re- 
 quired to charge and empty it. This type is not used in the 
 United States, though it has been employed to some extent 
 in Germany. 
 
 25. Rotary Kilns. Although rotary kilns for other purposes 
 had been in use for some time, the first patent for a process of 
 manufacture of cement by their use was issued in 1877 to Mr. 
 T. R. Crampton. The method, apparently, did not pass beyond 
 the stage of laboratory experiment until 1885, when Frederick 
 Ransome of England patented a rotary kiln, which, however, 
 required many important modifications to make it a success. 
 
 About 1888 Mr. J. G. Sanderson and Dr. Geo. Duryee made 
 some successful experiments with the rotary kiln for wet mix- 
 tures, and in the following year experiments were begun at the 
 works of the Atlas Portland Cement Co. under Mr. P. Giron, 
 which resulted in the construction of a practical kiln for burning 
 dry mixtures. Prof. Spencer B. Newberry, at about the same 
 time, perfected the rotary process for wet materials at Warners, 
 N. Y., and Sandusky, Ohio. 
 
 A rotary kiln as used for the burning of cement consists of a 
 steel cylinder five feet to six and a half feet in diameter and 
 about sixty feet in length. This cylinder is lined with fire- 
 brick, rests on rollers with its axis slightly inclined to the horir 
 zontal, and is revolved slowly by means of gearing. The mix- 
 ture to be burned is introduced at the upper end of the cylinder, 
 while a jet of gas, crude oil, or more frequently, powdered coal, 
 is injected through a special burner at the lower end. As the 
 cylinder revolves, the material works slowly toward the lower 
 end, the clinkering temperature being maintained throughout 
 about the lower third of the length. In some of the more elab- 
 
PORTLAND CEMENT 
 
 19 
 
 orate styles, the clinker is passed through one or more cooling 
 cylinders before it is conveyed to the grinding machinery. In 
 the Hurry and Seaman rotary, the clinker, after it leaves the 
 first cooling cylinder, is passed between rolls that serve to break 
 any large lumps, and is moistened with water before its passage 
 through the second cooling cylinder, which delivers the clinker 
 warm, moist, and in small pieces. 
 
 The lining of rotary kilns has given much trouble, as the 
 clinker acts upon fire brick lining to form a fusible compound 
 at the high temperatures required in the burning. One method 
 of overcoming this difficulty is to fuse upon the fire brick a coat- 
 ing of clinker which is beaten down while still plastic, so that it 
 adheres to the brick and protects them more or less successfully 
 from further injury. The kind of fuel and the burner giving 
 the best result have also received much attention; while petro- 
 leum was first tried and is still used to some extent, powdered 
 coal is now more commonly employed, and one of the most suc- 
 cessful forms of burner is constructed like an injector, the pul- 
 verized coal being drawn in with the blast of air. 
 
 26. Output and Fuel Consumption of Different Kilns. A 
 comparison of the average output of the several styles of kilns 
 described above, and the approximate fuel consumption, are 
 given in the following table. Where it is necessary to dry the 
 materials before introducing them into the burning kiln, the 
 fuel required in drying is not included. 
 
 STYLE. 
 
 Barrels per Day. 
 
 Fuel as Per Cent, 
 of 
 Weight of Clinker. 
 
 Intermittent dome 
 
 30 
 
 20 to 30 
 
 Hoffman (per chamber) 
 Dietsch and Schofer 
 
 25 
 
 50 to 75 
 
 15 to 20 
 15 to 20 
 
 Chamber .... 
 
 30 
 
 40 to 50 
 
 Rotary .... 
 
 120 to 150 
 
 30 to 40 
 
 
 
 
 27. Advantages of the Rotary Kiln. Although the burning 
 of cement in a rotary kiln requires a somewhat larger fuel con- 
 sumption than with some other types, the ability to use a cheaper 
 form of fuel, and the saving in the amount of labor required, 
 much more than offset this disadvantage. Either wet or dry 
 materials may be fed to the kiln, thereby eliminating the neces- 
 sity of forming the slurry into bricks, drying and stacking 
 
20 CEMENT AND CONCRETE 
 
 them in the kilns. By the rotary process it is possible to so 
 arrange a plant that the material is handled entirely by ma- 
 chinery from raw material to finished product. The control 
 possible in burning with the rotary is much better than with 
 any other style of kiln, as the intensity of the flame and the 
 speed of revolution of the cylinder may both be regulated. On 
 this account, as well as because the pieces of clinker are much 
 smaller, the cement is more uniformly burned. The remarkable 
 development of the Portland cement industry in the United 
 States is due in no small measure to the adoption and perfection 
 of the rotary kiln, for the labor expense in manufacture has been 
 so reduced thereby that we are able to successfully compete 
 with cements made abroad where lower wages prevail. 
 
 28. GRINDING. In grinding it is not sufficient that the 
 cement be so reduced that a certain percentage of it will pass 
 a sieve having, say, 10,000 holes per square inch; but it is de- 
 sired that as large a proportion as possible shall be of the 
 finest floury nature. To accomplish this result it has been 
 claimed that French buhr millstones are the best, but their 
 great consumption of power has led to the introduction of other 
 forms of grinding machinery, so that at present millstones find 
 their chief use in natural cement manufacture. 
 
 It is usually considered that the greatest economy results 
 from a gradual reduction of the clinker as it passes from one 
 form of grinder to another, each machine being supplied with 
 the size of pieces it is best adapted to handle. Large pieces of 
 clinker are first passed through an ordinary rock crusher, such 
 as the Gates or Blake. Where rotary kilns are in use, this step 
 in the process may be omitted, as the clinker comes from the 
 kiln in small, nut-like pieces. 
 
 29. Ball mills may also be used for the first reduction. The 
 ball mill is a short cylinder of large diameter which is partially 
 filled with flint or steel balls. When the cylinder revolves, 
 the balls and the clinker fall upon hard metal surfaces, and as 
 the material is ground to the size of sand grains, it falls 
 through screens in the periphery into a hopper, where it is 
 delivered to a conveyor, or to another form of pulverizer for 
 further reduction. 
 
 30. Tube mills may be used in connection with millstones, 
 but are usually employed for final reduction of the product of 
 
PORTLAND CEMENT 21 
 
 the ball mill. The tube mill is a steel cylinder, about 4 or 5 
 feet in diameter and 15 to 25 feet long, with axis horizontal or 
 nearly so, and revolving on trunnions. The cylinder is lined 
 with hard iron or porcelain, and is half filled with flint pebbles. 
 The material is fed in at one end and is gradually pulverized as 
 it works toward the other end. Some styles are not continuous 
 in their action, but are charged and closed, the material being 
 removed after a certain number of revolutions. 
 
 31. Griffin Mills. The Griffin mill is an American invention 
 that has found much favor, especially in grinding tailings from 
 other mills. A heavy steel roller is attached to the bottom of a 
 steel shaft, which is provided at its upper end with a ball-and- 
 socket joint. When the shaft is given a gyratory motion, the 
 roller presses by centrifugal force against the inside surface of a 
 heavy steel ring where the grinding takes place. The material 
 which drops below the roller is thrown up again by steel blades 
 that are also attached to the shaft, and when finally of sufficient 
 fineness, the powder escapes through screens above the ring into 
 a hopper. 
 
 32. The method of grinding to be adopted at any mill de- 
 pends upon the size and hardness of the particles of clinker, but 
 usually the clinker is passed through at least two machines. 
 
 It has been stated 1 that the power consumed in grinding 
 one ton of cement by the different principles is as follows: 
 
 For millstones . . . . 30 to 32 I. H. P. per ton per hour. 
 
 For ball principle ... 16 to 18 I.H.P. " 
 
 For edge runners ... 12 to 14 I.H.P. " " 
 
 The sifting of the product, which formerly required special 
 revolving or shaking screens of wire cloth, is now usually done 
 by the sieves attached to the grinding machinery. 
 
 33. SAND-CEMENT. This product, which is also called silica 
 cement, is composed of Portland cement and silicious sand mixed 
 in any desired proportion and then ground to extreme fineness. 
 This product is placed on the market by dealers, but rights to 
 use the process may be purchased. In the construction of Lock 
 and Dam No. 2, Mississippi River, between Minneapolis and 
 St. Paul, Major F. V. Abbot 2 used the process, grinding with 
 
 1 Mr. Henry Faija, in Trans. A. S. C. E., Vol. xxx, p. 49. 
 
 2 Report of Mr. A. O. Powell, Asst. Engineer, Report Chief of Engineers, 
 U. S. A,, 1900, p. 2779. 
 
22 CEMENT AND CONCRETE 
 
 a tube mill one part of Portland cement with one part fine 
 sand. The cost, exclusive of plant, is estimated as follows: 
 
 J barrel of Portland cement at $2.85 $1.42 
 
 I " " sand at .05 03 
 
 Cost of grinding 50 
 
 Cost of royalty 05 
 
 Cost of one barrel Silica cement $2.00 
 
 This cement has given remarkably high tests considering the 
 adulteration with sand, and is claimed to be specially useful in 
 making impervious mortar and concrete. 
 
 ART. 7. OTHER METHODS OF MANUFACTURE OF PORTLAND 
 
 CEMENT 
 
 34. Portland Cement from Blast Furnace Slag. The prep- 
 aration of a true Portland cement from blast furnace slag has 
 been followed in Germany and elsewhere in Europe for several 
 years, and recently has been introduced in the United States. 
 As this process utilizes a waste product, its popularity is likely 
 to increase. Whereas, for the manufacture of slag cement only 
 the slag from gray pig iron is available, it is found that in most 
 cases the slag from white pig iron may be used for the produc- 
 tion of Portland cement from slag. 
 
 The method of manufacture is briefly as follows: The slag 
 as it comes from the blast furnace is subjected to the action of 
 a stream of water, which granulates it and changes it chemi- 
 cally, the water .combining with the calcium sulphide, which is 
 injurious to cement, to form lime and sulphuretted hydrogen. 
 The granulated slag is then dried, mixed with the correct 
 proportion of dried limestone, and ground to extreme fineness. 
 The mixture is next burned in rotary kilns, the remainder of the 
 process being the same as that employed when ordinary raw 
 materials are used. While a cement made from slag by this 
 method may have some peculiarities due to the nature of the 
 raw materials used, and should be very carefully tested before 
 it is used in important work, it should not be confounded with 
 slag cement, which is a mixture of granulated slag and hydrated 
 lime subsequently ground, but not burned together. 
 
 35. Portland Cement from By-Products of Soda Manufacture. 
 The Michigan Alkali Company has installed at Wyandotte, 
 Mich., a cement plant to utilize the large amount of limestone 
 
SLAG CEMENT 23 
 
 which they have as waste in the manufacture of soda products. 
 The limestone which has served its purpose in the soda manu- 
 facture is in a finely divided and semi-fluid state; to this is 
 added the proper percentage of clay, which has been dried and 
 pulverized. The two are then very thoroughly mixed by pug 
 mills and wash mills, the slurry corrected by small additions 
 of one or the other of the ingredients, and finally burned in 
 rotary kilns. 
 
 ART. 8. THE MANUFACTURE OF SLAG CEMENT 
 
 36. Slag cement is made by adding calcium hydrate to a 
 granulated basic slag resulting from the manufacture of gray 
 pig iron. The slag must be carefully selected as to its chemical 
 composition, Prof. Tetmajer having found b^ extended experi- 
 ments that slags containing silica, alumina, and lime in the 
 ratio 30 to 16 to 40 are best adapted to the purpose. As the 
 molten slag runs from the blast furnace it is suddenly chilled 
 by being run into water, or is partially disintegrated by being 
 treated with a strong current of water, air, or steam. It is thus 
 reduced to coarse particles resembling sand, or to a spongy or 
 fibrous mass which, after drying, is readily ground to a fine 
 powder. The process of chilling results in a certain chemico- 
 physical change that renders the powder capable of combining 
 more readily with the slaked lime which is subsequently added. 
 Slag which has been allowed to cool slowly will not form an 
 hydraulic product when mixed with the lime, although the 
 chemical composition of the slag may be identical in the two 
 cases. The lime is dipped into water, or treated with steam, 
 until slaked to a fine dry powder, and is then added to the 
 powdered slag in proportions of about one part of the former 
 to three parts of the latter, this proportion depending upon the 
 composition of the slag used. The powdered slag and lime are 
 sifted, then mixed and reground together to an extreme fine- 
 ness, thus insuring an intimate incorporation of the ingredients. 
 Since there is no burning in the process, it is evident that the 
 finished product is merely a mixture, not a chemical compound 
 as is the case with Portland cement. 
 
 37. One of the largest mills for the manufacture of slag 
 cement in the United States is conducted by the Illinois Steel 
 Company, and the following description of the process is con- 
 
24 CEMENT AND CONCRETE 
 
 densed from a statement of Mr. Jasper Whiting, 1 manager of 
 the cement department, and patentee of the process: Slag of 
 the proper composition is chilled as it comes from the furnace 
 by the action of a large stream of cold water under high pres- 
 sure. The slag is thereby broken up, about one-third of its 
 sulphur is eliminated, and it is otherwise changed chemically. 
 A sample of the slag thus granulated is mixed with a proportion 
 of prepared lime, and ground in a small mill whereby actual 
 slag cement is produced. If the tests upon this trial cement 
 are satisfactory, the slag is dried and then ground, first in a 
 Griffin mill and then in a tube mill, where it is mixed with the 
 proper amount of prepared lime and the two materials ground 
 and intimately mixed together. The resulting product is said 
 to be so fine that but 4 per cent, is retained on a sieve having 
 200 meshes per linear inch. The lime is burned from a very 
 pure limestone and stored in bins, beneath which are two 
 screens of different mesh, the coarser at the top. A quantity 
 of lime being drawn on the upper screen is slaked by the addi- 
 tion of water containing a small percentage of caustic soda. 
 The lime passes through the two screens as it slakes and is 
 then heated in a dryer; the slaking being thus completed, the 
 lime may be incorporated with the slag. The purpose of the 
 caustic soda added in the above process is to render the cement 
 quicker setting. 
 
 ART. 9. THE MANUFACTURE OF NATURAL CEMENT 
 
 38. History. The American product called natural cement 
 was first manufactured at Fayetteville, Onondaga County, 
 N. Y., in 1818, and used in the construction of the Erie Canal. 
 Other early dates of manufacture are given as 1823, near Rosen- 
 dale, N. Y., and 1824 at Williamsville, Erie County, N.Y., the 
 products being used in the construction of the Erie and the 
 Delaware & Hudson Canals. Factories were soon started in 
 other states, and at present nearly every State in the Union has 
 one or more natural cement factories, the total annual produc- 
 tion being now about nine million barrels. 
 
 39. Materials Required. The composition of rock from 
 which natural cement may be made, varies within wide limits. 
 As stated in 13, an argillaceous limestone, a magnesian lime- 
 
 1 " Report of Board of Engineers on Steel Portland Cement," Appendix I. 
 
NATURAL CEMENT 25 
 
 stone or an argillo-magnesian limestone may be used. Argilla- 
 ceous limestone makes what is sometimes called an aluminous 
 natural cement, its essential ingredient being a bisilicate, or sili- 
 cate of alumina and lime, while the product made from mag- 
 nesian limestone is called magnesian cement and is composed 
 of a triple silicate of lime, magnesia and alumina. 
 
 The Maryland cements are typical of the former or alumi- 
 nous variety, containing only one to five per cent, of magnesia, 
 while the Rosendale and the Milwaukee are magnesian cements 
 containing 15 to 25 per cent, magnesia. (See Table 3.) 
 
 With a given raw material, the silica and alumina should 
 bear a certain proportion to the lime and magnesia, but close 
 limits cannot be stated for this proportion, as it varies with the 
 chemical and physical character of the rock. The silica should 
 be combined with the alumina, not in the form of sand. 
 
 The materials found at any locality may vary considerably 
 as to chemical composition, especially among the several strata. 
 In some cases the different strata' are utilized to make two or 
 more brands, which differ somewhat in their characteristics as 
 to time of setting, etc. It is common also to mix two or more 
 layers together in the manufacture, with the idea that the in- 
 gredients lacking in one stratum will be supplied by the others. 
 
 40. DESCRIPTION OF PROCESS. As the proper ingredients 
 to produce the cement have been incorporated by Nature, that 
 part of the process of Portland cement manufacture preliminary 
 to the burning is unnecessary. The rock occurs in strata and is 
 either quarried in open cut where the stripping is light, or by 
 means of tunnels. In open cut, a face of twenty feet or more is 
 sometimes worked. As has already been stated, the strata vary 
 in chemical composition, and while two or more brands are 
 sometimes made at the same mill, it is a more general practice 
 to mix the rock from several strata in the production of one 
 brand. The idea is that if one layer contains too much silica, it 
 may be corrected by another containing too much lime or mag- 
 nesia. As the rock is not finely pulverized before it enters the 
 kiln, each lump burns by itself and makes a certain cement; the 
 piece of rock next it must make as distinct a product as though 
 burned in a separate kiln. What is obtained, then, by this 
 method is a mixture of several cements, and it is questionable 
 whether the mere mechanical mixing of an over-limed cement 
 
26 CEMENT AND CONCRETE 
 
 with an over- clayed one will make a well balanced product. 
 This practice may account, in a great degree, for the large vari- 
 ations that occur in the cement from a single factory, variations 
 which are often, however, more noticeable in short-time tests 
 than in the longer ones. 
 
 41. The rock, as quarried, is broken by an ordinary rock 
 crusher or otherwise, into pieces varying in size up to six inches, 
 and is then conveyed, usually by tramway, directly to the kilns. 
 These are of the cylindrical continuous type, built of stone or 
 steel, and lined with fire brick. The kilns are commonly about 
 45 feet high and 16 feet in diameter; the tramway leads to a 
 loading platform on top of the kiln. According to the locality, 
 the fuel may be either bituminous or anthracite coal of about 
 pea size. The rock and fuel are spread in the top of the kiln in 
 alternating layers, the proportion of fuel being usually regulated 
 by the man in charge of the burning, but sometimes a machine 
 is employed which automatically governs the amount of coal 
 used. The temperature in the kilns is much below that required 
 in Portland cement manufacture, but varies of course with 
 the materials. 
 
 42. The calcined rock is conveyed first to some sort of a 
 stone crusher; a common form is known as a " pot-cracker/ 7 and 
 consists of a corrugated conical shell in which works a cast iron 
 core, also corrugated. After passing the cracker, the material 
 may be screened, giving a certain proportion of finished product, 
 and another portion which may go directly to the finishing 
 stones, while the coarsest pieces are conveyed to another form 
 of cracker, such as iron edge runners, which prepares it for the 
 millstones. In many factories ordinary under-run millstones 
 are used, in others rock emery stones are employed, while in 
 some factories stones found locally prove satisfactory. There 
 have been recently installed in some of the natural cement fac- 
 tories, ball and tube mills for grinding as used for Portland 
 cement clinker, and in several factories special forms of grinding 
 machinery are in use that have been perfected by the managers 
 of the works. 
 
 The product passes from the reducing mills *to the " mixers," 
 by means of which the material is thoroughly mixed to promote 
 uniformity. It is now ready for packing, and may be conveyed 
 directly to the chute from which the barrels or bags are filled. 
 
NATURAL CEMENT 27 
 
 In packing, the barrel rests upon a circular disc which is given 
 a vertical jarring motion, and thus the cement is thoroughly 
 settled in the barrel. 
 
 It is seen that the manufacture of natural cement is very 
 similar to that portion of Portland cement manufacture suc- 
 ceeding the preparation of the raw material for burning. In 
 general, less care is requisite with natural cement, the burning 
 is carried on at a lower temperature, and the calcined rock is 
 softer, so that less expense is incurred in grinding. 
 
PART II 
 
 THE PROPERTIES OF CEMENT AND 
 METHODS OF TESTING 
 
 CHAPTER III 
 
 INTRODUCTORY 
 
 43. In the tests of such structural materials as wood and 
 steel it will not usually be difficult to determine the suitability 
 of the material for the intended purpose, provided the test 
 pieces truthfully represent the members to be used. It is known 
 that so long as these members are protected from oxidation and 
 over-loading they will retain their qualities, and there is always 
 a reasonably clear understanding of what these qualities should 
 be. On the other hand, in the testing of cement, one may be 
 perfectly sure that from the moment the cement is manufac- 
 tured until long after it has been in service in the structure its 
 properties will be ever changing; and, further, the qualities 
 which it is desirable the cement should possess are not always 
 clearly in mind. 
 
 44. Desirable Qualities in Cement. The desirable elements 
 in a cement may be stated as follows: 1st, That when treated in 
 the propo&ed manner it shall develop a certain strength at the 
 end of a given period. 2d, That it shall contain no compounds 
 within itself which may, at any future time, cause it to change 
 its form or volume, or lose any of its previously acquired strength. 
 3d, That it shall be able to withstand the action of any exterior 
 agency to which it may be subjected that would tend to decrease 
 its 'strength or change its form or volume. When it is deter- 
 mined that a cement has these three qualities, it is certain that 
 it is safe to use it, but it is further desirable to know that the 
 
UNIFORM METHODS 29 
 
 cement in question will accomplish the given object as cheaply 
 as any other cement. 
 
 The cohesive and adhesive strengths of cement are not usu- 
 ally considered in the design of the structure into which cement 
 enters. The design of a masonry arch does not comprehend any 
 adhesive strength in the cement, except as it may be recognized 
 as an additional factor of safety, and a masonry dam is so de- 
 signed that there shall be no tension at the heel. These facts 
 are due in a large measure to the very imperfect knowledge we 
 have of the behavior of cements in various contingencies. With 
 the increasing use of concrete, as in arches, locks, floors, roofs, 
 etc., the tensile and transverse strengths of cement are coming 
 to be relied on to a certain extent; and as its properties become 
 better known, and as means of recognizing these properties 
 become more certain and widespread in their application, ce- 
 ment will be more extensively employed in a scientific and eco- 
 nomical manner. 
 
 Cement may be compared in one sense to timber and cast 
 iron. A large factor of safety is employed when dealing with 
 these materials because of hidden defects that may exist. The 
 defects which lie hidden in cement may be even greater than 
 these in proportion to its possible strength, and defects in ce- 
 ment are often more treacherous because their development 
 may be deferred for some time. The importance of knowing 
 whether the cement fulfills the second and third requirements 
 noted above is therefore evident. 
 
 45. Having considered the qualities a cement should have, 
 we may proceed to the detailed consideration of the various 
 tests employed to disclose the presence or absence of these qual- 
 ifies. The strength a given cement will develop is investigated 
 by chemical analysis, by obtaining the specific gravity and fine- 
 ness, and by actual rupture tests, whether they be tensile, com- 
 pressive, transverse, or shearing. By tests for change of volume 
 and by chemical analysis, it is sought to determine whether a 
 cement has within itself elements of destruction. For the power 
 to withstand external agencies there are no adequate tests, 
 though chemical analysis is considered an aid. The methods 
 of use, the proportions of the materials, their incorporation and 
 deposition are of great importance in insuring against external 
 causes of injury. 
 
30 CEMENT AND CONCRETE 
 
 46. Uniform Methods of Cement Testing. In order that 
 uniformity should prevail in the methods employed in testing 
 cements, various societies have discussed the subject in detail, 
 usually through committees, and much valuable work has been 
 done along this line. The engineers of public works in many 
 European countries have adopted specifications and laid down 
 more or less detailed rules for testing. The Corps of Engineers, 
 U. S. A., has recently adopted a similar code of rules. 
 
 The International Society for Testing Materials, with which 
 the American Society for Testing Materials is affiliated, has con- 
 sidered the subject and still has committees at work upon it. 
 The New York section of the Society of Chemical Industry has 
 recently formulated a method for analysis of materials for the 
 Portland cement industry. The American Society of Civil En- 
 gineers received a report in 1885 from a committee appointed 
 to consider methods of cement testing, and in order to keep the 
 subject abreast of the latest developments in the manufacture 
 and use of cement, a second committee was appointed several 
 years ago, which has been making a thorough discussion of the 
 subject, and has submitted a preliminary or progress report. 
 
 47. Notwithstanding that so much has been done toward 
 unification of methods, it may never be possible to determine 
 accurately the value of one cement as compared with another 
 tested in a different laboratory; though in tests of iron and 
 steel no such difficulty is experienced. Certainly, as at present 
 carried out, strength tests of cement are purely relative tests 
 and do not show the absolute strength which may be developed 
 in the structures; nor can the results be compared with the re- 
 sults obtained in other laboratories and any fine distinctions of 
 quality drawn. To attempt to carry out acceptance tests fti 
 such a way as to show directly the strength which will be de- 
 veloped in actual construction, is only to introduce causes of 
 irregularity in the tests. 
 
CHAPTER IV 
 
 CHEMICAL TESTS 
 
 ART. 10. COMPOSITION AND CHEMICAL ANALYSIS 
 
 48. Value of Chemical Tests. The definite aid which chem- 
 ical analysis may render in determining the quality of a cement 
 is limited by the following considerations. It is not definitely 
 known just what part is played by each of the compounds that 
 go to make up commercial cement, and chemical analysis does 
 not tell the manner of the occurrence of these compounds. A 
 cement may have a chemical composition that is thought to be 
 perfect, but if the burning has not been properly accomplished, 
 it may be a dangerous product and analysis would show no de- 
 fect. Some of the best authorities say that chemical analysis 
 is useful principally in tracing the cause of defects which, by 
 other tests, have been found to exist. However, there are some 
 constituents which it is fairly well known a cement should not 
 contain in any considerable quantities. An analysis may be of 
 value in estimating quantitatively such constituents, while it 
 may also be of service in detecting adulterations. It is not im- 
 possible, then, that chemical tests may yet play a more impor- 
 tant role in cement testing, especially if the method of analysis 
 can be made more simple and rapid, without too great a sacri- 
 fice of accuracy. 
 
 49. Lime. The proportion of lime in Portland cement may 
 vary from 59 to 67 per cent. A much greater range than this 
 is allowable in natural cement, the percentage usually being 
 from 30 to 45, according to the amount and character of the 
 other active constituents. An analysis of Portland cement which 
 shows a percentage of lime far outside of the limits mentioned 
 above, should be regarded with suspicion and submitted to very 
 thorough tests before acceptance. As already stated, the ratio 
 of the silica and alumina to the lime in a cement is called the 
 hydraulic index. The value of this ratio is usually between .42 
 and .48 for Portland cement. 
 
 31 
 
32 CEMENT AND CONCRETE 
 
 Cement mixtures containing a large percentage of lime re- 
 quire a high temperature for calcination, are difficult to grind, 
 and yield a slow-setting product. The danger in highly limed 
 cements is that they will not be properly calcined and a por- 
 tion of the lime will be left in a free state. The demand for 
 high strength in short-time tests has led manufacturers to 
 make a heavily limed product, and in some cases the limits of 
 safety have probably been overstepped. The introduction of the 
 rotary kiln, however, has so improved the facilities for burning 
 cement that a higher percentage of lime is now possible. 
 
 There is no method known at present for determining quanti- 
 tatively the amount of free lime in a cement, and it seems doubt- 
 ful whether its presence can be detected with certainty by chemi- 
 cal analysis. The method usually employed for this purpose 
 depends on the hydration of the lime and subsequent absorption 
 of carbonic acid. 
 
 50. Magnesia. The detection of magnesia in several con- 
 crete structures that had failed, led to the conclusion that mag- 
 nesia, in quantities exceeding two or three per cent., was a 
 dangerous element in Portland cement. In 1886-87 Mr. Har- 
 rison Hayter 1 mentioned several failures of masonry and con- 
 crete which he considered were due to magnesia, and concluded 
 that cemerut should not contain more than one per cent. Later 
 investigations, however, indicated that such failures could be 
 explained in other ways, and that the magnesia found in the 
 failing structure had come from the sea water and replaced the 
 lime in the cement. Mr. A. E. Carey 2 has considered that "an 
 excess of caustic lime or magnesia causes first, disintegration 
 by expansion due to hydration, and second, being soluble, when 
 conditions permit of their washing out, leave the concrete in a 
 honeycombed state." Notice that this refers to caustic mag- 
 nesia, and Prof. S. B. Newberry 3 has stated that "it is doubtful 
 if magnesia is ever combined in Portland cement. Our own 
 experiments tend to confirm the opinion of many German 
 authorities that magnesia remains free in cement and does 
 not combine with the constituents of clay after the manner 
 of lime." 
 
 1 Proc. Inst. C. E., Part 1, Session of 1886-87. 
 
 2 Ibid., 1891-92. 
 
 3 Municipal Engineering, October, 1896. 
 
COMPOSITION AND ANALYSIS 33 
 
 On the other hand, M. H. LeChatelier l says that the ''acci- 
 dents occasioned by certain magnesian elements, and the similar 
 results obtained in laboratory experiments, have been due to 
 the employment of badly proportioned cements, containing free 
 uncombined magnesia and too small a quantity of clay. Cor- 
 responding mixtures containing lime instead of magnesia would 
 have caused still more serious accidents, yet it would not be con- 
 cluded that there must be no lime in cement." Again, Dr. 
 Erdmenger characterizes magnesia as an adulterant only, and 
 considers that its effect is nil if a greater percentage of lime is 
 added in the manufacture. 
 
 Some authoritative information on the amount of magnesia 
 allowable in Portland cement is contained in the report of the 
 magnesia commission of the Association of German Cement 
 Makers, 1895: Three members of this committee, Messrs. Schott, 
 Meyer and Arendt concluded that "the presence of magnesia 
 up to ten per cent, causes no harmful expansion or cracking of 
 the cement, even after several years." Mr. Dyckerhoff, how- 
 ever, presented a minority report, in which he pointed out that 
 while a large amount of magnesia, not sintered, may not have 
 an injurious effect, yet a content of more than four per cent, of 
 sintered magnesia, whether added or substituted for part of the 
 lime, has an injurious effect after long periods. The committee 
 continued the ruling of 1893 that "a magnesia content of five 
 per cent, in burnt cement is harmless," but held the question 
 open for further investigation, indicating that this limit might 
 be raised. 
 
 In view of the disagreement among such eminent authori- 
 ties it is impossible to arrive at a satisfactory conclusion, but if 
 the effect of magnesia depends upon the manner of its occur- 
 rence, whether free or combined, sintered or unsintered, then 
 chemical analysis can be of but limited value as a test of quality 
 in this regard. Natural cements frequently contain large pro- 
 portions of magnesia replacing lime, and in this case an analysis 
 is of the same value as an analysis for lime. 
 
 51. Alumina and Iron Oxide. The amount of alumina 
 which a cement should contain is not well established. Its 
 presence tends to facilitate the burning, and it renders the prod- 
 
 1 Trans. Amer. Inst. Mining Engrs., 1893. 
 
34 CEMENT AND CONCRETE 
 
 uct quicker setting. Cements containing large percentages of 
 alumina are inferior for use in air or sea water, and it is probable 
 that the percentage of alumina should not exceed eight or ten 
 to obtain the best results in all media. A slag cement may be 
 detected by its large content of alumina. Oxide of iron acts 
 as a flux in burning, but in the finished product is little more 
 than an adulterant. 
 
 52. Sulphuric Acid. French specifications say that Port- 
 land cements shall not contain more than one per cent, of sul- 
 phuric acid or sulphides in determinable proportions. This is 
 doubtless intended for cement to be used in sea water. Adul- 
 terations with blast-furnace slag may sometimes be detected 
 by the amount of sulphides present, but small quantities of sul- 
 phuric acid in the cement may be derived from the coke used 
 in burning and have no injurious effect for use in fresh water. 
 A content of 1.75 per cent, of sulphuric anhydride, S0 8 , is now con- 
 sidered the maximum permissible. Sulphates mixed with the 
 raw materials and burned with the cement may be harmless, 
 while the same amount added after burning would not be per- 
 missible. [For tests on the effect of adding sulphate of lime to 
 cement, see Art. 48.] 
 
 53. Water and Carbonic Acid. The determination of these 
 may give some idea of the deterioration of a product by storage, 
 and they may also indicate defective burning. M. Candlot con- 
 siders that in the case of Portland cement, a loss on ignition 
 (water and carbon dioxide) exceeding three per cent. " indicates 
 that the cement has undergone sufficient alteration to appre- 
 ciably diminish its strength." Natural cements may, however, 
 contain considerable proportions of these ingredients and still 
 give good results. 
 
 54. Conclusions. Finally, then, the determination of silica, 
 alumina, magnesia and lime may be of value, first, in classify- 
 ing a product, and second, as indicating whether the proportions 
 contained in it are such that if properly manufactured it is 
 capable of giving good results. What these proportions should 
 be for Portland cement has already been stated, 9. The de- 
 termination of certain injurious ingredients is also of some 
 value, but it must be remembered that the dangerous elements 
 most commonly occurring, namely, free lime and magnesia, are 
 not determinable by chemical analysis. It has been stated by 
 
COMPOSITION AND ANALYSIS 35 
 
 M. LeChatelier that " neither complete nor partial chemical 
 analysis of the constituents of hydraulic materials can be ranked 
 among normal tests. But chemical analysis may render real 
 service in controlling the classification of a product concerning 
 which there is reason to doubt the declaration of the manufac- 
 turer. Thus, a slag cement can be distinguished from a Port- 
 land by its tenor in alumina and water; certain natural cements, 
 by their contents of sulphuric acid, etc." l 
 
 The methods of analysis for Portland cement are given in 
 considerable detail in a little book, " The Chemical and Physical 
 Examination of Portland Cement," by Richard K. Meade. The 
 method of analysis suggested by the New York Section of the 
 Society of Chemical Industry is published in the Engineering 
 Record of July 11, 1903, and in Engineering News of July 16, 
 1903. 
 
 "Tests of Hydraulic Materials/' H. LeChatelier. 
 
CHAPTER V 
 
 THE SIMPLER PHYSICAL TESTS 
 ART. 11. MICROSCOPICAL TESTS. COLOR 
 
 55. Microscopical examinations are of some interest and 
 value to those who are thoroughly versed in the chemistry of 
 the burning and hardening of cements, as an aid in determining 
 the part played by each compound in the hardening. 
 
 Examinations may be made either of the dry powder, or of 
 thin sections of hardened cement, or clinker. Dry powder of 
 Portland cement appears to be made up of scaly particles, many 
 of which are clearly defined and semi-transparent, while natural 
 cement particles are more nearly opaque and less angular. Thin 
 sections of Portland cement clinker have been found to exhibit 
 colorless crystals somewhat cubical in structure, which are 
 thought to form the essential hardening constituent; thin sec- 
 tions of hardened Portland cement show a clear crystalline 
 structure. Prof. Hayter Lewis found that the particles in good 
 Portland cement were angular in form, consisting of scales and 
 splinters, while the particles of cement of poor quality were 
 rounded or nodular. 
 
 Microscopic examinations have no place at present in ordi- 
 nary tests of quality. 
 
 56. Significance of Color. The color of cement is chiefly 
 derived from its impurities, such as oxides of iron and manga- 
 nese, rather than from its essential ingredients, and the color is 
 therefore of minor importance. Other things being equal, a 
 hard burned Portland cement will be darker in color than an 
 underburned product. An excess of lime may be indicated by 
 a bluish cast, and excess of clay or underburning may give a 
 brownish shade. Gray or greenish gray is usually considered 
 to be indicative of a good Portland. 
 
 57. The colors of natural cements have a wide range, vary- 
 ing from a light yellow to a very dark brown, without reference 
 to quality. Owing to a popular idea that dark color indicated 
 
WEIGHT PER CUBIC FOOT 37 
 
 strength, some manufacturers have been said to add coloring 
 matter to their product, but although this may have been true 
 at one time, the correction of this false idea has doubtless ren- 
 dered such a practice quite unnecessary now. Variations in 
 shade in different samples of the same brand of natural cement 
 may indicate differences in burning or in the composition of the 
 rock; but the interpretation of color for any given brand must 
 be the result of close study, for some cements become lighter 
 on burning and others become darker, while in some cases no 
 variation in shade can be detected for different degrees of 
 burning. 
 
 ART. 12. WEIGHT PER CUBIC FOOT OR APPARENT DENSITY 
 
 58. Significance. Since a hard burned Portland cement 
 will usually be heavier than a light burned one, a test of the 
 weight per cubic foot was once thought to be of great value in 
 judging of the degree of burning. But it has been shown re- 
 peatedly that the weight per cubic foot depends quite as much 
 on the fineness as on the burning. It also depends on the age 
 of the cement, and its chemical composition. As a test for 
 quality, the determination of the apparent density has therefore 
 been discarded. However, it is an aid in classifying a product, 
 since Portland cements weigh from 70 to 90 pounds per cubic 
 foot when loosely filled in a measure, while natural cements 
 weigh from 45 to 65 pounds. A knowledge of the weight per 
 cubic foot is also useful in reducing proportions given by weight 
 to equivalent volumetric proportions, and vice versa. 
 
 59. Method. This test may be made with a very simple 
 apparatus, and the results obtained, though not strictly accu- 
 rate, are sufficient for all practical purposes. A metal tube, 
 
 2 feet 4 inches long, about 6 inches in diameter at the top, and 
 
 3 or 4 inches at the bottom, is supported by a frame resting on 
 four legs. A metal cylinder, 6 inches in diameter and 6j\ 
 inches deep, holding one-tenth cubic foot, is placed on the floor 
 below the tube. A coarse sieve, through which all of the ce- 
 ment will pass, is placed on top of the tube and three fe.et above 
 the bottom of the measure. The cement passes through the 
 sieve, falling freely to the cylinder below, which is struck off 
 level when full. The cement must not be heaped too much, 
 and great care must be taken that the measure is not jarred 
 
38 CEMENT AND CONCRETE 
 
 while it is being filled or struck off. The cement is in such a 
 light condition that a very slight jar is sufficient to cause it to 
 settle. 
 
 The above apparatus is on the same plan as that used by 
 Mr. E. C. Clarke on the Boston Main Drainage Works, and is 
 described here for general use when it is desired to compare 
 the results obtained by operators at different points. Should 
 one wish simply to obtain a series of results on different cements 
 which are to be compared among themselves, it is quite suf- 
 ficient to sift each sample through a coarse sieve, and then with 
 an ordinary scoop carefully fill a measure of any known capac- 
 ity, without other apparatus. 
 
 Mr. Henry Faija has described an apparatus consisting of a 
 funnel with a screw at the mouth which carries the cement 
 horizontally to the point where it falls freely into the measure. 
 Various other devices have been employed, but none seems to 
 have met with universal favor. 
 
 60. To determine the relative accuracy obtainable with the 
 simple form of apparatus first described, the author made a 
 series of tests which may be summarized as follows : 
 
 1st Method. Cement passed a wire mesh sieve, holes .033 
 inch square and fell freely two feet through a 6-inch tube into 
 a measure holding J cu. ft. Five trials with a sample of Dycker- 
 hoff Portland, highest weight per cubic foot, 81 Ibs. 4 oz., 
 lowest, 79 Ibs. 2 oz., difference, 2 Ibs. 2 oz. Three trials with 
 Alsen's Portland, highest weight, 73 Ibs., lowest, 72 Ibs., dif- 
 ference, 1 Ib. 
 
 2d Method. Measure same size filled with scoop without 
 other apparatus, and cement not shaken or jarred in measure. 
 Five trials with Alsen's Portland, highest result, 73 Ibs. 8 oz. 
 per cu. ft., lowest result, 72 Ibs. 12 oz., difference, 12 oz. Five 
 trials with different sample of same cement, highest, 72 Ibs. 
 4 oz., lowest, 72 Ibs., difference, 4 oz. 
 
 3d Method. Measure filled with scoop, and cement well 
 shaken down as filling proceeded. Five trials with Alsen's 
 Portland, highest result, 100 Ibs. 8 oz., lowest, 97 Ibs. 14 oz., 
 difference, 2 Ibs. 10 oz. 
 
 It appears from these tests that when the measure is filled 
 with the scoop, the results are about as uniform as when the 
 apparatus is used, provided the filling is always done by the 
 
SPECIFIC GRAVITY 39 
 
 same person. But the results obtained by different operators 
 with the same sample of cement would probably vary less, one 
 from the other, when the apparatus is employed. In other 
 words, the personal factor is more nearly eliminated when the 
 cement is passed through a sieve and allowed to fall freely 
 from a given height. 
 
 61. As to the effect of age on the weight per cubic foot, it 
 was found in one case that cement which weighed 93^ pounds 
 per cubic foot when freshly ground, weighed but 88 pounds 
 when a few days old, and 78 and 74 pounds after six months 
 and one year, respectively. 1 
 
 Many experiments have been made to show the effect of 
 fineness on the weight per cubic foot, but as this subject will 
 be taken up again under "fineness," it will suffice to quote one 
 series of tests made by Mr. E. C. Clarke, 2 giving the " weight 
 per cubic foot of the same sample of German Portland cement 
 containing different percentages of coarse particles as deter- 
 mined by sifting through the No. 120 sieve." 
 
 Samples containing 0, 10, 20, 30, and 40 per cent, of coarse 
 particles retained on No. 120 sieve gave the following weights 
 per cubic foot: 75, 79, 82, 86 and 90 pounds, respectively. 
 
 It may be repeated that the weight per cubic foot is no 
 longer considered an indication of quality, but should it be 
 desired to specify a given weight, the method by which the 
 test is to be made should also be stated. 
 
 ART. 13. SPECIFIC GRAVITY OR TRUE DENSITY 
 
 62. The apparent density or weight per cubic foot is in- 
 fluenced to such an extent by the degree of fineness of the 
 cement that this test has been almost superseded by the test 
 for specific gravity. Although the true density, or specific 
 gravity, is not affected by the fineness, it is influenced by the 
 composition, the degree of burning, and the age, or amount of 
 aeration of the sample. 
 
 The method commonly employed in this test consists in de- 
 termining the absolute volume of a given weight of the cement 
 
 "Cement for Users," by H. Faija, p. 54. 
 
 2 " Record of Tests of Cements for Boston Main Drainage Works," Trans. 
 A. S. C. E., Vol. xiv, p. 144. 
 
40 CEMENT AND CONCRETE 
 
 powder by measuring the amount of liquid which it will dis- 
 place. A simple form of apparatus may be constructed in 
 any laboratory as follows: In a wide mouth bottle, having 
 straight sides and holding 200 c.c. or more, fit a perforated 
 cork. Through the cork slip a burette graduated in cubic 
 centimeters from to 50, placing the zero end down. Fill the 
 bottle and the tube up to the zero mark, with some liquid such 
 .as turpentine, benzine or kerosene oil, but preferably benzine 
 (62 Baume naptha). By means of a funnel in the top of the 
 burette, add slowly 100 grams of cement; then jar the bottle to 
 remove air bubbles and read the burette. This reading, x, 
 represents the volume of 100 grams of cement; and 100, the 
 volume of 100 grams of water, divided by x gives the specific 
 gravity of the sample. 
 
 63. Among other forms of apparatus which are also of sim- 
 ple construction and tend to facilitate the test, may be men- 
 tioned the following: 
 
 M. Candlot l devised an apparatus consisting of a graduated 
 tube terminating in a bulb at the upper end, the lower end of 
 the tube being ground to fit the neck of a flask. The tube and 
 flask being disconnected, sufficient liquid is placed in the bulb 
 so that when connected with the flask and placed upright, the 
 level of the liquid will be at or near the zero mark on the tube. 
 The actual level of the liquid is read after standing a few minutes ; 
 the apparatus is again inverted and the flask disconnected to 
 allow of the introduction of 100 grams of cement. The flask is 
 then replaced and the contents of the apparatus well shaken to 
 expel air-bubbles. When the latter have been completely ex- 
 pelled, the flask is placed upright, and after standing a short 
 time the level of the liquid is again read, the difference between 
 the two readings indicating the absolute volume of 100 grams 
 of the cement powder. 
 
 The apparatus devised by M. H. LeChatelier 2 consists of a 
 flask of a capacity of about 120 c.c., and having a neck some 
 20 c. in length, halfway up which is a bulb having a capacity 
 
 1 "Ciments et Chaux Hydrauliques," par. E. Candlot. 
 
 2 "Report of Commission des Methods d'Essai des Materiaux de Con- 
 struction," The Engineer (London); Illustrated also in Meade's "Examina- 
 tion of Portland Cement," Spaulding's "Hydraulic Cement," and Engineer- 
 ing News, January 29, 1903. 
 
SPECIFIC GRAVITY 
 
 41 
 
 -ip- 
 
 -31- 
 
 30- 
 
 c/ta 3mm 
 
 of 20 c.c. Near the bottom of the tube, or flask, is the zero 
 mark, and above the bulb the tube is graduated for a length 
 corresponding to a capacity of 3 c.c., each graduation repre- 
 senting .1 c.c. The diameter of the tube is about 9 mm. The 
 zero mark on the tube is below the bulb. The method of opera- 
 tion is similar to that described 
 above. 
 
 64. The following style of 
 apparatus (see Fig. 1) is sug- 
 gested as a very convenient 
 form, and one which may be 
 used for another test soon to 
 be described. In this form, the 
 flask, of a capacity of about 
 200 c.c., has straight sides and 
 a flat bottom. The lower part 
 of the burette is of large diame- 
 ter, about 15 mm., to allow the 
 cement to pass readily, while 
 the upper portion is made 
 smaller, about 8 mm., to per- 
 mit more accurate reading, and 
 is graduated from 30 c.c. to 40 
 c.c., the divisions being 0.1 c.c. 
 Half divisions may be esti- 
 mated. The zero mark is in the 
 larger part of the burette, but it 
 is less difficult to make an ac- 
 curate reading at the zero mark, 
 since at the time of taking this 
 reading the liquid is clear; this 
 mark should entirely surround 
 the burette. The mouth of the 
 bottle and the lower end of the 
 burette should be ground to fit, 
 
 and a ground glass stopper should form a part of the apparatus. 
 A long pipette will be found convenient for adjusting the level 
 of the liquid to the zero mark. 
 
 65. Turpentine is frequently employed for this test, but it 
 is somewhat inconvenient to use, since its volume is so sensi- 
 
 o \//*r/t/e eft*. IS mm. 
 
 / \ 
 
 FIG. 1. SPECIFIC GRAVITY APPA- 
 RATUS 
 
42 CEMENT AND CONCRETE 
 
 live to changes in temperature. This sensitiveness renders it 
 imperative that the temperature at the time of taking the final 
 reading be the same as when the initial reading is taken, or 
 that a correction be applied. To assure this condition the ap- 
 paratus should be immersed in a water bath, and the tempera- 
 ture of the cement should be the same as that of the turpentine. 
 The use of water in the apparatus does not offer this inconven- 
 ience, but it is possible that the hydration of the cement during 
 the experiment might be sufficient to so affect the volume as 
 to change the result, especially with quick-setting cements. 
 Light oils, such as benzine and kerosene, are rather volatile, 
 but the former (62 Baume naptha) is recommended in the 
 preliminary report of the Committee of the American Society 
 of Civil Engineers. With the precautions mentioned above, 
 turpentine may be used with good results; that which has 
 been dried by standing over cement or quicklime is to be 
 preferred. 
 
 66. This test may be extended to give interesting and valu- 
 able results, in the following manner: When the cement has 
 settled in the bottle, leaving the liquid clear, pour off a portion 
 of the latter and replace the burette by a glass stopper. Thor- 
 oughly agitate the remaining liquid and cement until the latter 
 is in suspension; allow the cement to settle again without dis- 
 turbance, and it will be found that it is graded in the bottle 
 according to its fineness, the coarsest particles being at the 
 bottom. With Portland cement, if a portion of the sample is 
 underburned it will appear as the top layer, and be indicated 
 by its yellow color. It will also be interesting to note what 
 proportion of the cement is so fine that the separate grains 
 are indistinguishable. That the bottle should have straight 
 sides and a flat bottom is to accommodate this part of the 
 test, which also dictates the use of some other liquid than 
 water. 
 
 67. Effect of Composition, Aeration, Etc. It has been said 
 above that the composition of a cement affects its specific 
 gravity, a highly limed cement having a higher density. On 
 this account an analysis for lime is valuable in connection with 
 this test, in order to determine whether a high specific gravity 
 is due to a high percentage of lime or to hard burning. 
 
 The age, or aeration of a sample affects its specific gravity 
 
SPECIFIC GRAVITY 43 
 
 because of the absorption of water from the atmosphere. The 
 absorption of two per cent, of water is sufficient to lower the 
 specific gravity from 3.125 to 3.000. The following may be 
 given as illustrating this point: a certain sample of natural 
 cement when taken from the barrel had a specific gravity of 
 3.106; after it had been spread out in the air for two months its 
 specific gravity was 3.000. A quantity of this aerated cement 
 weighing 120 grams was placed in an iron vessel and heated 
 over an oil stove for about one hour; at the end of this time 
 the cement had lost two grains in weight. The specific gravity 
 of the fresh cement being 3.106, 118 grams would have an ab- 
 solute volume of 33 c.c.; two grams of water would occupy 2 
 c.c., hence 120 grams of the aerated cement would occupy 40 
 c.c., and 120 -r- 40 = 3.00, the specific gravity of the aerated 
 cement as found above. It is not always possible to thus 
 drive off all of the water absorbed, since a portion of it may 
 enter into combination with the cement; but a sample should 
 always be heated for at least thirty minutes at a temperature 
 of 100 C. before making the test for specific gravity, and 
 should any appreciable loss of weight occur, it is an indication 
 of aeration. 
 
 68. A determination of the specific gravity is primarily a 
 test for burning, but it may also be of much value in detecting 
 adulterations, as with blast furnace slag or ground limestone. 
 An admixture of 10 per cent, of either of these substances would 
 suffice to lower the specific gravity from 3.15 to about 3.10. 
 The specific gravity of Portland cement ranges from 2.90 to 
 3.25, but a first-class product should not show a lower specific 
 gravity than 3.05. If fresh Portland gives a result below this 
 it is probably either underburned or underlimed, or, perhaps, 
 has been adulterated. 
 
 The specific gravity of natural cements has been found to 
 vary from 2.82 to 3.25. The specific gravity of one sample of 
 underburned natural cement was found to be lower than a 
 sample of the same brand which was overburned, but it seems 
 very doubtful whether this is true of other brands made from 
 rock of a different character. It was also found that the spe- 
 cific gravity of the coarse particles of some natural cements is 
 lower than that of the fine particles (see Table 10, Art. 15), 
 while the opposite is true in the case of Portland cements. 
 
44 CEMENT AND CONCRETE 
 
 No general rules can be given at present for the interpreta- 
 tion of this test that are applicable to all natural cements; it is 
 thought that the test will be of value in comparing samples 
 of the same brand, though it seems doubtful whether it will 
 prove of value in comparing one brand of natural cement with 
 another, since it is quite probable that the interpretation may 
 vary with the variety of rock used in the manufacture. The 
 value of the test for Portland cements is, however, well 
 established. 
 
CHAPTER VI 
 
 SIFTING AND FINE GRINDING 
 
 ART. 14. FINENESS 
 
 69. Importance of Fineness. The fineness of cement is al- 
 ways conceded to be one of its most important qualities, and 
 the determination of fineness is omitted in none but the very 
 crudest tests. Unfortunately, however, sieves that are so coarse 
 as to give delusive results are usually employed. It is very easy 
 to show that grains of cement as large as one-fiftieth of an inch 
 in diameter are practically valueless, but much more difficult 
 to determine the point of fineness at which the particles begin 
 to have cementitious value. 
 
 70. A moderately coarse sieve is easier to operate than a 
 very fine one, less time being consumed in sifting. The impres- 
 sion seems to be quite general also that there is a fixed relation 
 between the proportions of the different sized grains in different 
 samples. Many specifications require that a certain percentage 
 " shall pass a sieve, having 2,500 holes per square inch." Now, 
 there is little doubt that grains of cement larger than .005 inch 
 in one dimension have very little cementitious value, and hence 
 a cement, all of which would pass holes .015 inch square, while 
 but 50 per cent, of it would pass holes .005 inch square, is little 
 better than one which leaves a larger residue on the coarser 
 sieve but the same residue on the finer. 
 
 In America and Germany it is the usual practice in the pro- 
 cess of manufacture to pass the cement through a screen which 
 will reject particles larger than about .015 inch in diameter; the 
 futility in attempting to determine, with a sieve no finer than 
 this, the proportion of the particles which are fine enough to be 
 of value, is therefore apparent. Since the English cement 
 makers have not been so progressive in the practice of screen- 
 ing, they have obtained the reputation of producing a coarse 
 product. In many cases this reputation is probably a just one, 
 but when tested with a very fine meshed sieve, some of the 
 
 46 
 
46 CEMENT AND CONCRETE 
 
 English cements do not compare so unfavorably with those of 
 German manufacture. It is a curious fact in this connection 
 that the English are the most conservative in holding to the 
 use of the coarse sieve in testing, which makes their cement 
 appear so very much coarser than the American or German 
 product. 
 
 71. SIEVES. Sieves for cement testing may be made either 
 of wire or silk gauze, set in metal or wood frames. Sieves of 
 perforated metal plate are sometimes employed for sifting sand, 
 but seldom for cement. It is with considerable difficulty that 
 accurate gauze sieves are obtained. They are usually desig- 
 nated by numbers corresponding to the number of meshes per 
 linear inch; this is in some respects an unsatisfactory method, 
 for the size of the wire, which is quite as important as the 
 number of meshes, is frequently not given at all, or stated 
 in terms of some wire gage which is capable of various 
 interpretations. 
 
 As usually supplied by different manufacturers, sieves pur- 
 porting to have the same number of meshes per linear inch may 
 vary in this regard as much as 10 or 15 per cent. Likewise the 
 size of wire used by different makers, in sieves having the same 
 number of meshes per inch, may vary quite as much. Again, 
 on account of irregularities in the gauze, the holes in a given 
 sieve vary one from another; in some cases an opening may be 
 but 60 or 70 per cent, as large in one dimension as an adjacent 
 one. 
 
 An ideal sieve should conform to the following requirements: 
 (1) holes to be of uniform size and shape throughout, (2) sides 
 of the holes to be very smooth, and (3) the spaces between the 
 holes to be of such size and shape that particles will not easily 
 rest there. 
 
 It is evident that the largest holes determine the character 
 of the sieve. For example, a sieve having half its holes 0.01 
 inch square and the other half 0.02 inch square, would, if used 
 long enough, separate the cement exactly as it would if all 
 the holes had been 0.02 inch square. Hence, if a very small 
 percentage of the holes are larger than the normal, it seriously 
 impairs the accuracy of the sieve by introducing an indeter- 
 mination; but holes smaller than the normal have no greater 
 objection than that, as the sifting proceeds, they become spaces 
 
FINENESS 
 
 47 
 
 between the real or larger holes, and as such do not fulfill the 
 third requirement mentioned above. The shape of the holes, 
 whether round, square or hexagonal, seems of minor importance 
 so long as uniformity is maintained. The second requirement 
 is necessary, because, should particles adhere to the sides of 
 the hole, the size of the latter would be decreased to that ex- 
 tent. The third requirement is for convenience, but would 
 require consideration if the style of the sieve were changed to a 
 punched metal plate. 
 
 72. The Committee of the American Society of Civil Engi- 
 neers, in their report on " A Uniform System for Tests of Cement " 
 in 1885, recommended three sizes of sieves for cement: No. 50 
 (2,500 meshes to the square inch) wire to be of No. 35 Stubbs' 
 wire gage ; No. 74 (5,476 meshes to the square inch) wire to 
 be of No. 37 Stubbs' wire gage; No. 100 (10,000 meshes to the 
 square inch) wire to be of No. 40 Stubbs' wire gage. For sand, 
 two sieves were recommended, No. 20 and No. 30 (400 and 900 
 meshes per square inch) wire to be of No. 28 and No. 31 Stubbs' 
 
 TABLE 5 
 
 Sieves : Number of Meshes per Linear Inch and Sizes of 
 Openings, as Found by Measurement 
 
 
 
 No. OP 
 
 
 
 
 
 MESHES 
 
 PKR 
 
 DIAMETER OF WIRE 
 IN DECIMALS OK 
 
 MEAN SIZE OF OPENING IN DECIMALS 
 OF AN INCH. 
 
 
 f f4 
 
 o 
 
 LINEAR 
 
 AN INCH. 
 
 
 
 a 
 
 INCH. 
 
 
 
 b 
 
 w > 
 
 
 
 
 M 
 
 a w 
 
 ^ 
 
 (M 
 
 
 
 
 . 
 
 
 
 
 
 
 
 
 goo 
 
 0> 
 fcfc. . 
 
 8 
 
 Web. 
 
 Woof. 
 
 8 
 
 C 0> 
 
 c! 
 
 4J 
 
 e 
 
 O> 
 
 
 
 
 
 fc 
 
 ! 
 
 M 
 
 
 
 a 
 9 
 
 Q} M 
 
 VI 
 
 4) 
 
 
 
 c 
 v 
 
 3 
 
 
 
 
 1? 
 
 f- 
 
 
 Diam- 
 
 Diam- 
 
 1 
 
 55 ^ 
 
 ! 
 
 ? 
 11 
 
 
 
 9 
 
 tj 
 
 S3 
 
 IS o 
 
 Remarks. 
 
 
 
 O 
 
 *4 
 
 e 
 
 <J 
 
 eter. 
 
 eter. 
 
 
 
 . m 
 
 W 
 
 s 
 
 af 
 
 
 
 a 
 
 b 
 
 c 
 
 (I 
 
 e 
 
 / 
 
 g 
 
 h 
 
 i 
 
 j 
 
 
 1 
 
 20 
 
 20 
 
 19J 
 
 .0185 
 
 .0169 
 
 .0016 
 
 .0315 
 
 .0337 
 
 .0022 
 
 .93 
 
 
 2 
 
 20 
 
 20 
 
 19 
 
 .0105 
 
 .0108 
 
 .0003 
 
 .0335 
 
 .0358 
 
 .0023 
 
 .93 
 
 
 3 
 
 30 
 
 30 
 
 28 1 
 
 .0119 
 
 .0119 
 
 .0000 
 
 .0214 
 
 .0229 
 
 .0015 
 
 .93 
 
 
 4 
 
 30 
 
 30 
 
 30 
 
 .0118 
 
 .0118 
 
 .0000 
 
 .0215 
 
 .0215 
 
 .0000 
 
 1.00 
 
 
 5 
 
 30 
 
 30 
 
 29\ 
 
 .0116 
 
 .0122 
 
 .0006 
 
 .0217 
 
 .0217 
 
 .0000 
 
 1.00 
 
 
 6 
 
 40 
 
 40 
 
 36 
 
 .0395 
 
 .0095 
 
 .0000 
 
 .0155 
 
 .0183 
 
 .0028 
 
 .85 
 
 
 7 
 
 50 
 
 60 
 
 47 
 
 .0082 
 
 .0083 
 
 .0001 
 
 .0118 
 
 .0130 
 
 .0012 
 
 .90 
 
 
 8 
 
 74 
 
 80 
 
 80 
 
 .0054 
 
 .0054 
 
 .0000 
 
 .0071 
 
 .0071 
 
 .0000 
 
 1.00 
 
 
 9 
 
 100 
 
 101 
 
 88 1 
 
 .0040 
 
 .0040 
 
 .0000 
 
 .0059 
 
 .0073 
 
 .0014 
 
 ' .80 
 
 
 10 
 
 120 
 
 120 
 
 120 
 
 .0037 
 
 .0037 
 
 .0000 
 
 .0046 
 
 .0046 
 
 .0000 
 
 1.00 
 
 
 It 
 
 200 
 
 210 
 
 170 
 
 .0022 
 
 .0022 
 
 .0000 
 
 .0026 
 
 .0037 
 
 .0013 
 
 .70 
 
 Approx. 
 
CEMENT AND CONCRETE 
 
 wire gage, respectively. It seems to be impracticable to com- 
 ply with these sizes of wires, because neither manufacturers nor 
 engineers appear to agree as to what diameters of wire corre- 
 spond to No. 37 and No. 40 Stubbs' wire gage. 
 
 73. The conferences of Dresden and Munich decided that 
 fineness should be determined by sieves of 900 and 4,900 meshes 
 per sq. cm., respectively, for Portland cement, and 900 and 
 2,500, respectively, for other hydraulic products, the size of the 
 wires being as follows: for 4,900, .05 mm.; for 2,500, .07 mm.; 
 and for 900, .10 mm. These sieves would have respectively 
 31,600 (178 X 178), 16,000 (127 X 127), and 5,800 (76 X 76) 
 meshes per square inch, and the sizes of the holes would be 
 approximately .0037 inch square, .005 inch square and .009 inch 
 square, respectively. It was also decided that for sifting sand, 
 punched metal plates were preferable to wire cloth sieves. 
 
 74. In Table 5 are given some of the results obtained by the 
 writer which will serve to show what variations may exist in 
 sieves which have been selected from a considerable number 
 offered for use. 
 
 Table 6 gives the data available concerning certain sieves 
 that have been used or recommended in this country and else- 
 where. 
 
 TABLE 6 
 
 Sizes of Openings in Sieves Recommended or in Use 
 
 
 || ' 
 
 |g 
 
 SIZE 
 
 
 REF. 
 
 ||| 
 
 | 
 
 HOLE, 
 INCH 
 
 REMARKS. 
 
 
 ll 
 
 Is 
 
 SQUARE. 
 
 
 
 a 
 
 b 
 
 c 
 
 
 1 
 
 178 
 
 .00197 
 
 .00366 
 
 Established by Conferences, Dresden & Munich. 
 
 2 
 
 127 
 
 .00276 
 
 .00512 
 
 11 U (I U it 
 
 3 
 
 76 
 
 .00394 
 
 .00920 
 
 U U U U it 
 
 4 
 
 76 
 
 .00437 
 
 .00875 
 
 Present German Standard. 
 
 5 
 
 76 
 
 .00591 
 
 .00721 
 
 Recommended by H. LeChatelier. 
 
 6 
 
 176 
 
 .00162 
 
 .004 
 
 Silk mesh Vyrnwy Reservoir. 
 
 7 
 
 103 
 
 .0022 
 
 .0075 
 
 u u ct u 
 
 8 
 
 170 
 
 .00279 
 
 .00309 
 
 Cornell University, Marx & Mosscrop, 1887. 
 
 9 
 
 80 
 
 .00651 
 
 .00599 
 
 u u u u u 
 
 10 
 
 50 
 
 .00881 
 
 .01119 
 
 H 11 it U U 
 
 11 
 
 30 
 
 .01214 
 
 .02119 
 
 u u u u u 
 
 12 
 
 20 
 
 .01899 
 
 .03101 
 
 t( U <( U it 
 
 13 
 
 200 
 
 .0024 
 
 .0026 
 
 Progress Report, A. S. C. E. Committee, 1903. 
 
 14 
 
 100 
 
 .0045 
 
 .0055 
 
 11 U U U U U 
 
SIFTING AND GRINDING 
 
 49 
 
 75. The time sifting should be continued will depend on the 
 fineness of the meshes, the diameter of the sieve, the amount 
 of cement taken, and the manner of sifting ; it will also depend 
 upon the fineness of the cement, as well as its nature, and its 
 condition as to dryness. But, although some care is necessary 
 concerning these points, very large variations in results due to 
 variations in the time the sifting is continued may easily be 
 avoided. The diameter of the sieve is usually made greater 
 for the finer meshes, but this is not always the case. It is a 
 common practice in America to use one-tenth of a pound of 
 cement in testing the fineness, using a scale weighing in ten- 
 thousandths of a pound. Where the metric system is in use 
 (and it may well be adopted in a cement laboratory), 100 grams 
 of cement are usually taken. 
 
 76. M. H. LeChatelier recommends a sieve having 900 meshes 
 per sq. cm., of wire 0.15 mm, diameter, giving holes 0.18 mm. 
 (.0072 inch) square. He prefers machine screening, but says 
 that for current tests it might be sufficient to screen by hand 
 for ten minutes with a sieve three decimeters (about 12 inches) 
 in diameter. 
 
 Table 7 is taken from experiments made by M. Durand- 
 Claye and M. Candlot, and shows what differences may arise 
 from varying the length of time that a sample is screened. The 
 cements used were not the same in the two cases, but the sieves 
 had each 5,000 meshes per sq. cm. (about 180 per linear inch), 
 and 100 grams of cement were taken in each case. Had a coarse 
 sieve been used, the differences would have been much less 
 after the same lengths of time. 
 
 TABLE 7 
 
 Fineness: Mechanical and Hand Sifting Compared 
 
 M. DURAND-CLAYE. 
 
 M. CANDLOT. 
 
 MECHANICAL SIEVE MAKING 200 
 
 HAND SIEVE, 12 INCHES IN 
 
 REVOLUTIONS PER MINUTE. 
 
 DIAMETER. 
 
 No. 
 Revolutions. 
 
 Per Cent. 
 Retained. 
 
 Ditf. 
 
 After 
 Minutes. 
 
 Per Cent. 
 Retained. 
 
 Diflf. 
 
 500 
 
 41.2 
 
 
 5 
 
 29.6 
 
 
 1,000 
 
 39.4 
 
 1.8 
 
 10 
 
 29.1 
 
 0.5 
 
 1,500 
 
 38.6 
 
 .8 
 
 20 
 
 28.4 
 
 0.7 
 
 2,000 
 
 38.0 
 
 .6 
 
 30 
 
 28.0 
 
 0.4 
 
 2,500 
 
 37.6 
 
 .4 
 
 40 
 
 27.7 
 
 0.3 
 
50 
 
 CEMENT AND CONCRETE 
 
 77. Table 8 gives the results obtained by the author in 
 sifting several samples. The No. 80 sieve was about 6J inches 
 in diameter, and Nos. 120 and 200 about 5J inches. One hun- 
 dred grams of cement were taken in each case, and the sieve 
 was shaken vigorously by hand. It is seen that coarse samples 
 require less time for sifting than fine samples, arid that natural 
 cements require a longer time than Portlands. With the No. 
 80 sieve, five minutes usually suffices to obtain the fineness, 
 
 TABLE 8 
 
 Effect of Time of Sifting on the Result Obtained in Testing 
 
 Fineness 
 
 
 
 I 
 
 
 
 1 
 
 2 
 3 
 4 
 
 5 
 6 
 
 7 
 8 
 
 9 
 
 10 
 11 
 12 
 13 
 14 
 15 
 16 
 
 17 
 
 18 
 19 
 20 
 21 
 22 
 23 
 24 
 
 SIEVE. 
 
 CEMENT. 
 
 PEK CENT. BY WEIGHT THAT HAD 
 PASSED SIEVE AFTER SIFTING. 
 
 * 
 
 2-S.a 
 c o> 
 
 | 
 
 ad a 
 ^fc 
 
 - ^5 
 
 II 
 M 
 
 ^M 
 
 Kind. 
 
 Brand. 
 
 Sam- 
 ple. 
 
 1 
 % 
 
 o> 
 3 
 
 a 
 
 i 
 
 eo 
 
 1 
 
 3 
 
 a 
 
 i 
 
 1O 
 
 I 
 a 
 
 a 
 
 
 t- 
 
 1 
 
 
 
 O 
 
 1 
 1 
 S 
 
 k 
 
 1 
 
 c 
 
 i 
 s 
 
 i 
 
 I 
 
 3 
 
 1 
 8 
 
 a 
 
 b 
 
 c 
 
 d 
 
 e 
 
 f 
 
 <7 
 
 h 
 
 93 
 
 86 
 96 
 
 i 
 
 j 
 
 m 
 
 80 j 
 
 u 
 
 u 
 
 u 
 
 120 j 
 
 it 
 It 
 u 
 1 1 
 (( 
 (( 
 it 
 
 200 j 
 
 u 
 
 t< 
 t( 
 
 ( t 
 
 .0071 
 
 by 
 
 .0071 
 
 (4 
 
 U 
 
 (i 
 (( 
 
 .0046 
 
 by 
 
 .0046 
 t 
 
 i 
 ( 
 
 i 
 t 
 
 .0036 
 by 
 .0037 
 
 i 
 
 ( 
 
 ( 
 
 i 
 
 Port. 
 
 u 
 
 u 
 
 Nat. 
 
 u 
 
 ii 
 u 
 
 Port. 
 
 u 
 (i 
 
 u 
 
 Nat. 
 u 
 
 
 Port. 
 
 1 1 
 
 u 
 ('( 
 
 Nat. 
 
 n 
 
 u 
 it 
 
 X 
 
 Y 
 
 S 
 Z 
 
 Bti 
 
 An 
 In 
 Gn 
 
 X 
 
 Y 
 
 8 
 Z 
 Bn 
 An 
 
 In 
 Gn 
 
 X 
 
 Y 
 
 S 
 Z 
 
 Bn 
 
 An 
 In 
 Gn 
 
 685 
 
 42s 
 34s 
 43s 
 27s 
 G 
 28s 
 108 T 
 
 685 
 
 42s 
 34s 
 43s 
 
 27s 
 G 
 28s 
 108 T 
 
 685 
 
 42s 
 34s 
 43s 
 27s 
 G 
 28s 
 108 T 
 
 91 
 
 82 
 95 
 100 
 68 
 78 
 85 
 75 
 
 45 
 
 41 
 
 72 
 54 
 27 
 45 
 13 
 5 
 
 92 
 
 85 
 
 96 
 
 93 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 71 
 
 80 
 89 
 89 
 
 78 
 
 73 
 89 
 94 
 62 
 73 
 42 
 16 
 
 58 
 
 65 
 72 
 74 
 57 
 67 
 41 
 40 
 
 71 
 82 
 90 
 90 
 
 82 
 
 77 
 90 
 96 
 65 
 76 
 64 
 27 
 
 65 
 
 68 
 76 
 78 
 59 
 69 
 53 
 49 
 
 
 
 
 
 
 82 
 90 
 91 
 
 v 
 
 
 
 
 
 
 
 
 
 91 
 84 
 
 78 
 91 
 97 
 66 
 78 
 82 
 64 
 
 68 
 
 71 
 78 
 81 
 60 
 70 
 69 
 64 
 
 
 
 
 84 
 
 78 
 91 
 98 
 66 
 78 
 83 
 82 
 
 70 
 
 72 
 
 80 
 82 
 60 
 72 
 75 
 76 
 
 
 
 
 
 
 
 
 
 
 
 So 
 71 
 
 72 
 81 
 83 
 
 *72 
 
 77 
 79 
 
 86 
 7*8' 
 
FINENESS 
 
 and with the No. 120 sieve, but little cement usually passed after 
 the sifting had continued ten minutes, though with one brand 
 of natural, Gn ; it appears that the true fineness would not be in- 
 dicated by sifting less than 20 minutes. With the No. 200 
 sieve 20 minutes is usually required, and in the case of two samples 
 of natural cement, a still longer time appears to be necessary. 
 
 78. Conclusions. Until there is a proper standard in the 
 United States concerning sieves and methods of sifting, the best 
 that can be done is to select, from the sieves that manufacturers 
 have to offer, those which appear to be most nearly uniform in 
 size of mesh, and then actually determine the size of the holes. 
 This may be done by counting, under the magnifying glass, the 
 number of meshes per inch each way, and determining the size 
 of wire with a micrometer wire gage. 
 
 As to the time sifting should be continued, one can easily 
 find by trial the time required in using a given sieve in order to 
 confine the error within given limits. A fine natural cement 
 should be selected to determine this, as such a cement requires 
 the longest sifting. Care should be taken that the cement is 
 well dried before making the test for fineness. It will be found 
 that for sieves having holes between .003 inch and .004 inch 
 square (sieves approximating 170 to 200 meshes per linear 
 inch) 20 to 30 minutes are required, while for sieves having 
 holes .007 to .009 inch square (approximately 70 to 100 meshes 
 per linear inch) from five to ten minutes will usually suffice. 
 
 79, Specifications for Fineness. The following table has 
 been compiled to show what are considered reasonable require- 
 ments for fineness. In most specifications there is the usual 
 indeterminatkm concerning the sizes of holes in the sieves. 
 
 TABLE 9 
 Requirements as to Fineness 
 
 SPECIFICATION. 
 
 DATE. 
 
 PERCENT. REQUIRED TO PASH 
 SIEVE HAVING 10,000 
 HOLES PER SQUARE INCH. 
 
 Portland. 
 
 Natural. 
 
 U. S. Army Engineers .... 
 U. S. Navy Department . . . 
 City Pittsburg Pa. . . . 
 
 1901 
 
 1900 
 1897 
 1896 
 1895 
 
 92 
 95 
 90 
 90 
 95 
 85 
 
 80 
 
 '11 
 
 '80* 
 
 New East River Bridge . . . 
 Topeka, Kan., Bridge . . . 
 Master Builders' Exchange, Phila. 
 
CEMENT AND CONCRETE 
 
 ART. 15. COARSE PARTICLES IN CEMENT 
 80. The Effect of Coarse Particles on the Weight of Cement. 
 To remove the coarse particles by sifting will reduce the 
 specific gravity of a sample of Portland cement, as the un- 
 ground particles, are from the harder burned and denser por- 
 tion of the clinker, and to remove these denser particles will, 
 of course, decrease the average density of the sample. This is 
 not always the case with natural cements, as is shown by the 
 following tests: 
 
 TABLE 10 
 
 The Relative Specific Gravity of Coarse and Fine Particles of 
 
 Cement 
 
 
 CEMENT. 
 
 
 
 Kind. 
 
 Brand. 
 
 Fineness. 
 
 SPECIFIC GRAVITY. 
 
 Portland 
 
 R 
 
 As received . 
 
 3.086 
 
 
 if 
 
 50-100 
 
 3 14") 
 
 
 X 
 
 Pass 50 . 
 
 8.039 
 
 
 
 
 Ret. on 50 
 
 3.125 
 
 Nati ral 
 
 Gu 
 
 Pass 100 . 
 
 2.874 
 
 
 u 
 
 Ret on 50 
 
 2.817 
 
 
 An 
 
 Pass 50 . 
 
 2.945 
 
 
 " 
 
 Ret. on 50 
 
 2.817 
 
 The apparent density or weight per cubic foot of Portland 
 will be reduced more than the specific gravity by the removal 
 of the coarse particles; because not only will the true density 
 be decreased, but the packing, which is facilitated by a wide 
 range in the sizes of the particles, will be less perfect than when 
 the coarse particles are present. In 89 a table is given show- 
 ing the changes in specific gravity and weight per bushel oc- 
 casioned by removing the coarse particles by sifting. 
 
 81. Effect of Coarse Particles on the Time of Setting. 
 Table 11 gives the results of a number of tests on Portland and 
 natural cements to determine the relative time of setting of 
 samples from which the coarse particles had been removed by 
 the No. 200 sieve, while Table 12 gives results obtained with 
 a sample of natural cement of varying fineness. 
 
 In Table 11, 30 per cent, of water was used for all Portland 
 cements, and 36 per cent, for all naturals, but the consistency 
 
COARSE PARTICLES 
 
 varied as stated in the table. It is seen that in nearly every 
 case the setting was hastened by removing the coarse particles, 
 though this may have been due in part to the fact that with the 
 same percentage of water the finer cement gave a stiffer paste. 
 For the tests in Table 12, the attempt was made to make all 
 of the mortars of the same consistency by*varying the percent- 
 age of water. As would be expected, the coarse particles are very 
 slow setting. In fact, what hardness they attained was prob- 
 ably due largely to the fine dust that adhered to the grains. 
 These coarse particles may be considered as practically inert, 
 and their presence in a sample would naturally make it slow 
 setting. To show this by actual test, however, is very difficult, 
 
 TABLE 11 
 
 Effect of Coarse Particles 011 the Time of Setting 
 
 CEMENT. 
 
 CKMKNT PASSING No. 20 
 SIEVE. 
 
 CKMKNT PASSING No. liOO 
 SIEVE. 
 
 
 
 Time to bear 
 
 
 Time to bear 
 
 
 Kind. 
 
 Brand. 
 
 i Ib. wire. 
 
 Consistency. 
 
 i Ib. wire. 
 
 Consistency. 
 
 
 
 Minutes. 
 
 
 .Minutes. 
 
 
 Portland 
 
 Y 
 
 39 
 
 Trifle moist 
 
 13 
 
 Trifle dry 
 
 it 
 
 X 
 
 9 
 
 Moist 
 
 4 
 
 0. K. 
 
 ii 
 
 Z 
 
 432 
 
 Trifle moist 
 
 354 
 
 Trifle dry 
 
 it 
 
 S 
 
 550 
 
 (I U 
 
 341 
 
 U li 
 
 Natural 
 
 Gn 
 
 31 
 
 
 29 
 
 
 H 
 
 Bn 
 
 143 
 
 Trifle moist 
 
 161 
 
 Trifle dry 
 
 U 
 
 In 
 
 397 
 
 41 t I 
 
 250 
 
 . 
 
 II 
 
 Un 
 
 256 
 
 . . . 
 
 233 
 
 
 
 NOTE: 30 per cent, water used for all Portlands. 
 
 36 per cent, water used for all natural cements. 
 
 TABLE 12 
 Effect of Coarse Particles on Time of Setting 
 
 Natural Cement, Brand Gn All pastes appeared same consistency, 
 
 
 WATER USED AS 
 
 TIME TO BEAR 
 
 TIME TO BEAR 
 
 FINENESS. 
 
 PER CENT. OF 
 
 i LB. WIRE. 
 
 1 LB. WIRE. 
 
 
 CEMENT. 
 
 
 
 
 
 Minutes. 
 
 Minutes. 
 
 Pass No. 20 sieve 
 
 33 
 
 14 
 
 159 
 
 " 50 " 
 
 36 
 
 29 
 
 219 
 
 " 100 " 
 
 38 
 
 24 
 
 214 
 
 Retained on 50," 
 
 
 
 
 reground to pass 100 
 
 28 
 
 73 
 
 070 
 
 Pass No. 50. retained 
 
 
 
 
 on No. 100 
 
 34 
 
 205 
 
 890 
 
54 CEMENT AND CONCRETE 
 
 as the amount of water required to bring the mortars to the 
 same consistency varies with the amount of coarse particles 
 present, and as there is no very satisfactory method of testing 
 the consistency, the tests for time of setting have in them this 
 indetermination. 
 
 82. Effect of Coarse Particles on the Tensile Strength. A 
 cement having a certain quantity of coarse particles will fre- 
 quently give a higher tensile strength when tested neat than a 
 cement from which the coarse particles have been removed by 
 screening. The reason for this may be found in the fact that 
 a wide range in the sizes of grain of the powder facilitates pack- 
 ing, both when dry and when mixed with water to form a paste. 
 Another reason is that the unground particles are stronger 
 than the hardened mortar, and, considering the broken section 
 of a briquet, the break does not take place through these par- 
 ticles, but they are pulled out of their bed; this virtually in- 
 creases the area of section. Were the same sample of cement 
 reground, so that a certain proportion of the coarse particles 
 was rendered active, it might then give a higher strength, neat, 
 than at first. If so, the reason would be found in the fact that 
 the coarse particles, being the hardest burned, were really from 
 the best part of the cement clinker, and rendering these parti- 
 cles active by fine grinding increased the cohesive properties of 
 the cement so much as to overcome the physical effect of the 
 coarse particles, which, when judged by neat tests, appear to be 
 beneficial. The above serves to illustrate the difference be- 
 tween sifting and fine grinding which are so frequently con- 
 fused in treating this subject. 
 
 83. Among the many tests that have been made to show 
 the effect of sifting on the cohesive and adhesive strength of 
 cements, a few may be given as follows: 
 
 Mr. Maclay l gives a few experiments to show that the pres- 
 ence of coarse particles increases the cohesive strength, neat, 
 seven days. 
 
 Lieut. W. Innes 2 gives two tables of results obtained by ex- 
 perimenting on very coarse cements. The tables show that 
 removing the particles that would not pass through sieves of 
 
 1 Trans. Am. Soc. C. E., Vol. vi. 
 
 2 Minutes Proc. Inst. C. E., Vol. xxv. 
 
COARSE PARTICLES 55 
 
 1,296 meshes and 2,500 meshes per square inch, decreased the 
 strength when tested neat at the ages of three months and six 
 months; but increased the strength when sand mortars were 
 used. The differences at six months were relatively somewhat 
 less than at three months. By separating a sample of cement 
 into two parts, that passing a sieve having 2,500 meshes per 
 square inch and that retained on the same sieve, and then 
 remixing the screenings with the fine portion, he found 
 that the highest strength, neat, six months, was given by the 
 mixture containing the largest amount tried (70 per cent.) of 
 screenings. 
 
 84. Jn the tests of cement for the Cairo Bridge 1 a series of 
 experiments was made to determine the effect of coarse parti- 
 cles on the value of both Portland and natural cements. The 
 cement was separated into two parts, by a sieve having 10,000 
 meshes per square inch. Briquets were made both neat and 
 with sand, the cement used being made of 100, 90, 80, 70 and 
 60 volumes of sifted cement to 0, 10, 20, 30, and 40 volumes, 
 respectively, of cement screenings. The briquets were broken 
 when six months old. 
 
 It was found that in the case of Portland cement, neat, the 
 highest result was obtained with the largest (40) per cent, of 
 screenings, but with one and two parts sand, the strength 
 steadily fell as larger amounts of screenings were used. With 
 Louisville natural cement the presence of screenings seemed to 
 have little effect on neat tests; and with one part of sand to one 
 of cement, the use of as much as 30 per cent, of screenings to 
 70 per cent, of sifted cement did not appear to decrease the 
 strength. With two parts sand to one cement, the results were 
 slowly diminished by successive additions of larger percentages 
 of screenings. 
 
 85. M. R. Feret is said to have replaced with sand the grains 
 of cement retained on sieves having 5,800 and 32,300 meshes 
 per square inch, and found that, except in the case of neat ce- 
 ment mortars, the substitution of sand for coarse particles of 
 cement did not decrease the strength. In experimenting on 
 this subject Mr. Eliot C. Clarke 2 found that the coarse particles 
 
 1 Jour. Assn. Engr. Soc., 1890, and Engineering News, Jan. 31, 1891. 
 
 2 Trans. A. S. C. E., Vol. xiv, pp. 158-162. 
 
CEMENT AND CONCRETE 
 
 TABLE 13 
 Effect of Removing Coarse Particles from Natural Cement 
 
 CEMENT. 
 
 No. PARTS 
 SAND TO 
 ONE 
 CEMENT 
 
 BY 
 
 WEIGHT. 
 
 WATER AS 
 PER 
 CENT. OF 
 WEIGHT 
 DRY 
 INGREDI- 
 ENTS. 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH. 
 
 7 da. 
 
 28 da. 
 
 3 mo. 
 
 6 mo. 
 
 2 years. 
 
 A 
 
 None 
 
 33.3 
 
 120 
 
 275 
 
 
 
 
 B 
 
 u 
 
 35.7 
 
 100 
 
 206 
 
 ' 
 
 
 
 C 
 
 " 
 
 38.5 
 
 83 
 
 253 
 
 . 
 
 
 
 D 
 
 it 
 
 28.3 
 
 202 
 
 264 
 
 . 
 
 . 
 
 
 E 
 
 n 
 
 35.0 
 
 127 
 
 143 
 
 . . . 
 
 . . . 
 
 . . . 
 
 A 
 
 One 
 
 19.0 
 
 121 
 
 253 
 
 
 330 
 
 380 
 
 B 
 
 u 
 
 19.6 
 
 104 
 
 251 
 
 
 344 
 
 398 
 
 C 
 
 tt 
 
 20.0 
 
 94 
 
 261 
 
 
 331 
 
 385 
 
 D 
 
 (( 
 
 16.0 
 
 286 
 
 360 
 
 . . . 
 
 396 
 
 385 
 
 A 
 
 Two 
 
 16.1 
 
 
 168 
 
 215 
 
 223 
 
 210 
 
 B 
 
 it 
 
 16.1 
 
 . 
 
 203 
 
 245 
 
 267 
 
 302 
 
 C 
 
 u 
 
 16.1 
 
 . 
 
 218 
 
 297 
 
 317 
 
 358 
 
 D 
 
 u 
 
 13.9 
 
 . 
 
 227 
 
 230 
 
 245 
 
 262 
 
 E 
 
 u 
 
 15.8-16.1 
 
 . . . 
 
 71 
 
 40 
 
 57 
 
 50 
 
 A 
 
 Three 
 
 14.5 
 
 
 
 
 127 
 
 128 
 
 B 
 
 u 
 
 14.5 
 
 . 
 
 
 
 167 
 
 164 
 
 C 
 
 u 
 
 14 5 
 
 
 . 
 
 . 
 
 205 
 
 234 
 
 D 
 
 l( 
 
 12.1 
 
 
 
 
 125 
 
 118 
 
 Fineness of Cement 
 
 * 
 
 PER CENT. PASSING SIEVE No. 
 
 50 
 
 100 
 
 120 
 
 Cement A . . .... 
 
 82 
 100 
 
 70 
 85 
 100 
 
 64 
 
 78 
 91 
 
 Cement B 
 
 Cement C 
 
 
 NOTE : All cement from same barrel, Brand Bn, Sample 27s. 
 
 Sand, crushed quartz 20-30. 
 
 All briquets made by one molder and stored in one tank. 
 
 All results, mean of 5 briquets, except two which are means of 
 
 ten and two briquets, respectively. 
 
 A Cement passing No. 20 sieve, holes .033 inch square. 
 B " " " 50 " " .012 " 
 
 C- " " 100 " " .0065 
 
 D " retained on No. 50, reground to pass No. 100. 
 E " passing No. 50, retained on No. 100. 
 
COARSE PARTICLES 
 
 57 
 
 of cement were somewhat better, for use in mortar, than fine 
 sand, but very little better than coarse sand. 
 
 86. The tests given in Table 13 were made under the au- 
 thor's direction to determine the effect of sifting and the value 
 of coarse particles. It is seen that in neat tests the strength is 
 slightly diminished by sifting out the coarse particles; in the 
 tests of mortars containing equal parts by weight of sand and 
 cement, there is little difference in the strength of the three 
 samples, though the coarser cement appears to gain its strength 
 a little more rapidly. With two parts sand to one of cement, 
 the greater value of the fine particles is very noticeable, and 
 with one-to-three mortars the difference is still more marked, 
 the sifted cement giving 80 per cent, greater strength than the 
 unsifted. 
 
 87. In Table 14 these results are arranged in a different 
 way. If we assume that the particles that will not pass the 
 No. 120 sieve are not cement at all, but equivalent to sand, 
 
 TABLE 14 
 
 Effect on Tensile Strength of Removing Coarse Particles from 
 Natural Cement 
 
 
 
 
 PARTS SANI> AND 
 
 
 
 PER CENT. 
 
 PARTS SAND 
 
 COAKSE PAR- 
 
 STRENGTH 
 
 CEMENT. 
 
 PASSING SIEVE 
 
 TO 
 
 TICLES TO ONE 
 
 OF MORTAR AFTER 
 
 
 No. 120. 
 
 ONE CEMENT. 
 
 PART 
 
 Two YEARS. 
 
 
 
 
 FINE PARTICLES. 
 
 
 A 
 
 64 
 
 3 
 
 5.2 
 
 128 
 
 B 
 
 78 
 
 3 
 
 4.1 
 
 164 
 
 A 
 
 64 
 
 2 
 
 3.7 
 
 210 
 
 C 
 
 91 
 
 3 
 
 3.4 
 
 234 
 
 B 
 
 78 
 
 2 
 
 2.8 
 
 302 
 
 C 
 
 91 
 
 2 
 
 2.3 
 
 358 
 
 A 
 
 64 
 
 1 
 
 2.1 
 
 380 
 
 B 
 
 78 
 
 1 
 
 1.6 
 
 398 
 
 C 
 
 91 
 
 1 
 
 1.2 
 
 385 
 
 and that all particles passing this sieve are cement, we obtain 
 a new set of proportions of sand to cement. Thus the sample 
 of cement passing No. 20 sieve, sample A, would be composed 
 of 64 parts cement and 36 parts sand, and the 1 to 3 mortar 
 would have in reality the proportion 64 cement to 336 sand, or 
 1 to 5.2. It is seen that the tensile strength bears a closer 
 relation to the richness of the mortar when considered in this 
 way. There is, of course, no abrupt division in size such that 
 
58 
 
 CEMENT AND CONCRETE 
 
 TABLE 15 
 Value of Coarse Particles of Cement, Natural and Portland 
 
 REFERENCE NUMBER. 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH. 
 
 Neat Cement. 
 
 1 Part Standard 
 Sand to 1 Cement 
 
 3 Parts Standard 
 Sand to 1 Cement. 
 
 3 Parts Limestone 
 Screenings, 
 (S8) to 1 Cement. 
 
 3 mos. 
 
 4 mos. 
 
 l yr. 
 
 3 mos. 
 
 lyr. 
 
 3 mos. 
 
 lyr. 
 
 3 mos. 
 
 lyr. 
 
 d 
 
 e 
 
 / 
 
 9 
 
 h 
 
 i 
 
 j 
 
 k 
 
 1 
 
 1 
 
 2 
 3 
 4 
 
 5 
 6 
 7 
 8 
 9 
 10 
 
 330 
 
 259 
 295 
 306 
 309 
 
 
 390 
 336 
 334 
 370 
 343 
 
 
 
 108 
 193 
 203 
 102 
 92 
 378 
 423 
 455 
 357 
 362 
 
 139 
 217 
 224 
 106 
 96 
 395 
 403 
 469 
 399 
 354 
 
 160 
 269 
 251 
 137 
 139 
 472 
 538 
 561 
 425 
 405 
 
 234 
 332 
 319 
 170 
 155 
 589 
 677 
 676 
 5(58 
 506 
 
 ' 
 
 
 
 
 
 
 630 
 550 
 621 
 615 
 591 
 
 706 
 553 
 665 
 651 
 764 
 
 786 
 755 
 745 
 746 
 765 
 
 812 
 838 
 891 
 841 
 837 
 
 Ref. No. 
 
 Natural cement passing No. 20 sieve. 
 " " 2. Natural cement passing No. 80 sieve. 
 " " 3. Natural cement reground before sifting, until all passed 
 
 No. 80 sieve. 
 
 " " 4. Natural cement, 64f per cent, of cement passing No. 80 
 sieve mixed with 35} per cent, of limestone screen- 
 ings retained between Nos. 20 and 80 sieves. 
 
 " " 5. Natural cement, 64f per cent, of cement passing No. 80 
 sieve mixed with 35^ per cent, of crushed quartz 
 retained between Nos. 20 and 80 sieves. 
 " " 6. Portland cement passing No. 40 sieve. 
 " " 7. Portland cement passing No. 80 sieve. 
 " " 8. Portland cement reground (before sifting) until all passed 
 
 No. 80 sieve. 
 
 " " 9. 81 per cent, of cement passing No. 80 sieve mixed with 
 18 per cent, limestone screenings retained between 
 Nos. 40 and 80 sieves. 
 
 " " 10. 81 per cent, of cement passing No. 80 sieve mixed with 
 18 J per cent, crushed quartz retained between Nos. 
 40 and 80 sieves. 
 
 Of the natural cement passing No. 20 sieve, 35| per cent, was retained 
 on sieve No. 80, while 64f per cent, passed the No. 80 sieve. In lines 4 and 
 5 the coarse particles of cement (20-80) were removed and replaced by an 
 equal weight of sand grains, retained between sieves 20 and 80. 
 
 Of the Portland cement passing No. 40 sieve, 18 per cent, was retained 
 on sieve No. 80, while 81 per cent, passed sieve No. 80. In lines 9 and 10 
 the coarse particles of cement (40-80) were removed and replaced by an equal 
 weight of sand grains retained between sieves 40 and 80. 
 
 All briquets made by same molder, each result mean of five specimens, 
 
FINE GRINDING 59 
 
 coarser particles act only as sand, while finer ones enter into 
 combination as cement; part of the coarse particles will have 
 some cementitious value, while some of the finer particles will 
 have somewhat the effect of sand. 
 
 As to the sample composed of coarse particles reground, it 
 must be considered that although this sample was passed through 
 the No. 100 sieve, yet it was in reality much coarser than sam- 
 ple C, because the particles \vere harder, and the grinding 
 in the mortar less thorough than the original grinding. Since 
 this sample of reground cement gives so high a strength neat 
 and with one part sand, it appears that the hard particles from 
 which it was made are of excellent quality if ground fine enough, 
 and the relatively lower results with larger proportions of sand 
 must be attributed to imperfect grinding. 
 
 The coarse particles retained between sieves 50 and 100 gave 
 a higher strength neat than was expected, but much of this 
 strength may be due to the floury portion of the cement that 
 doubtless adhered to the coarse particles instead of passing 
 through the sieve. 
 
 88, The tests in Table 15 were made to determine whether 
 the coarse particles of cement are of greater value in mortar 
 than the same quantity of fine sand. The coarse particles of 
 the cement were sifted out and replaced with sand grains of 
 about the same size. The conclusion drawn from the preced- 
 ing tests would indicate that some of the coarse particles of 
 cement might be replaced by sand without diminishing the 
 tensile strength; but the tests given in this table indicate that 
 this is not the case when it is a question of substituting sand 
 grains of the same size. Although such a substitution has 
 little effect on the strength of rich mortars, it results in a de- 
 creased strength with mortars containing as much as three 
 parts sand to one of cement by weight. (See 85 in this con- 
 nection.) 
 
 ART. 16. FINE GRINDING 
 
 89. Effect of Fine Grinding on the Weight of Cement. Fine 
 
 grinding will decrease the weight per cubic foot, the fine ce- 
 ment not packing as closely as the coarser product. In " Ce- 
 ment for Users," by Mr. Henry Faija, the following results are 
 given, showing the relation between fineness, weight, and spe- 
 
60 
 
 CEMENT AND CONCRETE 
 
 TABLE 16 
 Relation of Fineness to Specific Gravity and Weight per Bushel 
 
 From " Cement for Users " 
 
 SAM- 
 PLE. 
 
 SPECIFIC GRAVITY. 
 
 WEIGHT PER BUSHEL. 
 
 a 
 
 b 
 
 c 
 
 a 
 
 b 
 
 c 
 
 d 
 
 e 
 
 1 
 
 3.00 
 
 2.97 
 
 3.07 
 
 116.5 
 
 107.5 
 
 121.0 
 
 112.0 
 
 115 
 
 2 
 
 3.03 
 
 2.94 
 
 3.04 
 
 116.0 
 
 104.0 
 
 130.5 
 
 109.0 
 
 115 
 
 3 
 
 3^02 
 
 2.91 
 
 3.035 
 
 114.0 
 
 100.0 
 
 128.0 
 
 104.5 
 
 109 
 
 cific gravity: (a), cement as delivered; (b), sif tings that passed 
 through sieve with 2,500 holes per sq. in.; (c), coarse, retained 
 on above sieve; (d), cement all ground to pass above sieve; (<?), 
 coarse particles reground to pass above sieve. 
 
 90. Effect of Fine Grinding on Time of Setting. Since the 
 coarse particles of cement are practically inert, there is every 
 reason to believe Uiat finer grinding will increase the activity 
 of a sample, since it will render some inert particles active. 
 For the reason mentioned in 81, however, it is difficult to show 
 this difference in time of setting by actual tests. 
 
 Tests reported by Mr. David B. Butler 1 showed that several 
 Portland cements which took an initial set in 20 to 30 minutes 
 and hard set in 45 to 120 minutes would, when reground to pass 
 a sieve having 180 meshes per linear inch, begin to set in from 
 1 to 7 minutes and set hard in 5 to 15 minutes. These may be 
 considered extreme results; the rise in temperature of these 
 cements during setting was so great as to indicate they were 
 not normal cements, and variations in consistency of the pastes 
 may have influenced the time of setting. 
 
 91. EFFECT OF FINE GRINDING ON STRENGTH. Since the 
 best burned clinker of Portland cement is the hardest, it follows 
 that the unground particles would, if ground fine enough to 
 become active, form the best portion of the cement. This is 
 not, a priori, true of natural cements, because burning renders 
 some varieties of cement rock softer at first, but when the burn- 
 ing is carried beyond a certain point they become harder again. 
 The coarse particles in a natural cement may thus be either 
 
 1 Proceedings Inst. C. E., 1898. 
 
FINE GRINDING 
 
 61 
 
 from underburned or overburned rock; hence it is possible that 
 in some cases it might be better to leave the hardest particles 
 in an unground state. Thus, while it has been generally ac- 
 cepted that fine grinding improves Portland in a twofold de- 
 gree, by bringing into action the best burned clinker, as well 
 as by rendering a given weight of cement capable of coating a 
 larger number of sand grains, a similar conclusion concern- 
 ing natural cement is not well established. 
 
 TABLE 17 
 
 Effect of Fine Grinding of Natural Cement on the Tensile 
 Strength of Mortar 
 
 REFERENCE. 
 1 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH. 
 
 Neat 
 Cement. 
 
 1 Part Stand- 
 ard Sand 
 to 1 Cement. 
 
 2 Parts Standard Sand 
 to 1 Cement. 
 
 3 Parts 
 Standard 
 Sand to 1 Ce- 
 ment. 
 
 4 Parts 
 Standard 
 Sand to 1 Ce- 
 ment. 
 
 7 da. 
 
 GJ mo. 
 
 7 da. 
 
 28 da. 
 
 28 da. 
 
 3 mo. 
 
 6 mo. 
 
 2yr. 
 
 6 mo. 
 
 2yr. 
 
 6 mo. 
 
 2yr. 
 
 a 
 
 6 
 
 c 
 
 d 
 
 e 
 
 / 
 
 g 
 
 h 
 
 i 
 
 J 
 
 k 
 
 I 
 
 1 
 2 
 
 3 
 4 
 
 5 
 
 268 
 283 
 278 
 392 
 
 538 
 473 
 
 538 
 592 
 
 224 
 
 230 
 307 
 308 
 
 381 
 
 350 
 433 
 538 
 
 207 
 245 
 292 
 271 
 21 
 
 354 
 433 
 469 
 344 
 
 291 
 426 
 406 
 369 
 73 
 
 70 
 102 
 92 
 160 
 45 
 
 202 
 302 
 305 
 274 
 
 48 
 65 
 61 
 110 
 
 156 
 212 
 240 
 205 
 
 49 
 
 78 
 65 
 90 
 
 
 
 
 
 
 
 
 
 
 REFERENCE. 
 
 FINENESS OF CEMENT, 
 PER CENT. PASSING 
 Sieve Number. 
 
 100 
 
 120 
 
 1. Cement as received passed through No. 20 76.5 
 2. Cement as received passed through No. 100 100.0 
 3. Keground in mortar, not sifted .... 9-3.8 
 
 72.4 
 94.6 
 91.5 
 
 Cement; Natural, Brand Jn, 
 No. 1. Passing No. 20 sieve. 
 " 2. Passing No. 100 sieve. 
 " 3. Reground before sifting. 
 
 " 4. Particles retained on No. 50 sieve, reground to pass No. 100 sieve, 
 " 5. Particles retained on No. 50 sieve, reground to pass No. 50 sieve, 
 
 but retained on No. 100 sieve. 
 
 All briquets made by one molder and immersed in one tank. In general, 
 each result is mean of five specimens. 
 
62 CEMENT AND CONCRETE 
 
 92. Some tests bearing upon the value of fine grinding have 
 already been given in Table 15. Samples 3 and 8 were reground 
 with mortar and pestle before being sifted. If we compare the 
 results given by sample 3 with those obtained with samples 1 
 and 2, not reground, it appears that the regrinding diminishes 
 the strength in neat mortars but increases it in mortars con- 
 taining three parts sand to one of cement. Regrinding ap- 
 pears to be no better, however, than sifting. Comparing sam- 
 ple 8 with samples 6 and 7, it is seen that regrinding Portland 
 cement does not diminish the strength in neat mortars to the 
 same extent as sifting does, and in sand mortars regrinding 
 generally results in a greater increase in strength than sifting. 
 
 93. The results in Table 17 were obtained with another 
 sample of natural cement and are of greater practical value as 
 indicating the importance of fine grinding, since in these tests 
 a sample is included obtained by regrinding the original cement 
 without previous sifting. The conclusions concerning the ce- 
 ment retained on No. 50 sieve reground to pass No. 100, and 
 the coarse particles alone retained between sieves 50 and 100, 
 are practically the same as those drawn from Table 13. 
 
 As to the other three samples, the No. 20 sieve removed only 
 a very few coarse particles, and that passing this sieve may be 
 considered to represent the cement as received The No. 100 
 sieve removed about 24 per cent, by weight from the original 
 cement, and the cement that was reground contained but about 
 4 per cent, of particles which would not have passed the No. 
 100 sieve. The third sample, reground cement, may be com- 
 pared with the first to indicate the improvement obtained by 
 finer grinding, and it may be compared with the second to de- 
 termine the difference between removing the coarse particles by 
 sifting and reducing them by finer grinding. In considering 
 these results it will be best to neglect the two-year tests, since 
 all of the samples failed at this age. . A comparison of the re- 
 sults obtained with these three samples indicates that while the 
 advantage of finer grinding is not apparent in neat tests, in 
 sand mortars the value of finer grinding is more marked the 
 larger proportion of sand used, so that with three or four parts 
 sand, the strength with the fine samples is about 50 per cent, 
 greater than with the cement as received. It also appears that 
 the reground sample gains its strength more rapidly than the 
 
FINE GRINDING 63 
 
 sifted sample, though at six months it seems to make little dif- 
 ference whether the coarse particles are removed by sifting or 
 reduced by grinding. 
 
 94. Conclusions as to the Effect of Fine Grinding and Sifting 
 on Tensile Strength. The general conclusions to be drawn 
 concerning fine grinding and sifting may be summarized as fol- 
 lows: According to the tests given, it appears that to remove 
 the coarse particles from a sample of natural cement by sifting, 
 or to reduce them by finer grinding, generally diminishes the 
 strength obtained in tests of neat cement mortars. In one-to- 
 one mortars, the strength of the finer samples is not much 
 greater than when the coarse particles are present; but in mor- 
 tars containing greater proportions of sand, the advantage ob- 
 tained by eliminating the coarse particles is very marked 
 in the case of natural cement, the strength given by the 
 finer samples sometimes exceeding that of the original cement 
 by more than 60 per cent. While the advantages of sifting 
 and finer grinding are also important for Portland cements, 
 there does not result such a large proportionate increase in 
 strength. 
 
 Reground samples of natural cement gain strength more 
 rapidly than resifted samples, but eventually the strength 
 attained is about the same. In Portland cements regrinding 
 seems to be of greater value than resifting. A sample of natural 
 cement made from coarse particles reground gains strength 
 rapidly, and for mortars with small proportions of sand, gives 
 good results. The fact that such samples do not give a high 
 strength with large proportions of sand is doubtless due to the 
 fact that the grinding is not thorough, and the indications are 
 that the material of which such coarse particles are composed 
 would form a valuable part of the cement if ground fine 
 enough. 
 
 The coarse particles of either natural or Portland cement 
 may be replaced by grains of sand of the same size without 
 materially affecting the strength attained by neat and one-to- 
 one mortars, but for mortars containing larger proportions of 
 sand, such a substitution results in a decreased strength. 
 
 95. Finally, it may be said that the process of manufacture 
 and the character of the materials from which cement is made 
 have such an influence on the relative proportions of fine and 
 
64 CEMENT AND CONCRETE 
 
 coarse particles that the percentage of finest particles cannot 
 be determined by testing with a coarse sieve. While it is not 
 known at what point of fineness grains of cement begin to have 
 cementitious value, or what proportion of the cement should be 
 the finest flocculent matter, it is certain that a cement should 
 leave as small a percentage as possible on a sieve having holes 
 .004 inch square, in order to have the greatest sand carrying 
 capacity. 
 
 There is, however, a reason for using a comparatively coarse 
 sieve in connection with the fine one. Overburned lime, which 
 is likely to occur in Portland cements, is more dangerous in the 
 form of coarse particles than an equal quantity in a fine condi- 
 tion, because coarse particles slake more slowly and it is better 
 that expansion should occur early in the process of hardening 
 if it is to occur at all. For the same reason a cement that would 
 be unsound normally may be rendered less dangerous by re- 
 grinding. 
 
 As fine grinding is expensive, it is only a question as to when 
 the increased strength obtained is offset by the extra expense 
 incurred in grinding. There is now little trouble in obtaining 
 either natural or Portland cement of which from 60 to 70 per 
 cent, will pass holes .004 inch square. (See 79.) 
 
CHAPTER VII 
 
 TIME OF SETTING AND SOUNDNESS 
 
 ART. 17. SETTING OF CEMENT 
 
 96. Process of Setting. When cement is gaged with suffi- 
 cient water to bring it to a paste, and is then left undisturbed, 
 it soon begins to lose its plasticity and finally reaches such a 
 condition that its form can no longer be changed without pro- 
 ducing rupture. This change of condition is known as the 
 "setting" of cement and is considered to be, in a measure, dis- 
 tinct from " hardening." Setting usually takes place within a 
 few hours, or perhaps minutes, while the hardening is continu- 
 ous for months or years. 
 
 The precise chemical changes that take place in the setting 
 and hardening of cements are not thoroughly understood. The 
 chief cementitious ingredient in Portland cement is considered 
 to be a tricalcium silicate, 3 CaO, SiO 2 ; in contact with water it 
 forms hydrated monocalcic silicate and calcium hydrate. This 
 process is believed to contribute more to the final hardening of 
 the mortar than to the setting, though the hydration of the 
 finer particles of this important compound also contributes to 
 the first setting. It is considered that the calcium aluminates 
 play an important role in the first setting of cement, as they set 
 rapidly in contact with water, and it has been suggested that 
 they form the chief active constituents of natural cement. 1 
 
 These chemical changes cause the formation of crystals 
 which by their interlocking and adhesion give strength to the 
 new compounds. For a scientific and detailed treatment of 
 this subject, the reader is referred to the articles of M. H. Le 
 Chatelier in Annales des Mines, 11, pp. 413-465, Trans. Am. 
 Inst. Mining Engineers, August, 1893; to the conclusions of 
 S. B. and W. B. Newberry, Cement and Engineering News, 
 1898; and to " The Constitution of Portland Cement from a 
 
 1 S. B. Newberry, " Mineral Resources of the United States," 1892. 
 
66 CEMENT AND CONCRETE 
 
 Fhysico-Chemical Standpoint/' a paper by Mr. Clifford Richard- 
 son read before the Association of Portland Cement Manufac- 
 turers at Atlantic City, June 15, 1904, Engineering Record, 
 August 13 and 20, 1904, Engineering News, August 11, 1904. 
 
 97. THE RATE OF SETTING AND ITS DETERMINATION. The 
 setting of cement being a gradual and continuous process with- 
 out well-defined points of change, it is necessary, in order to com- 
 pare the rates of change in condition of different samples, to 
 adopt an arbitrary standard. The method usually adopted is 
 to determine the resistance of the mortar to the penetration of a 
 wire or needle. The wires used by General Totten and rec- 
 ommended by General Gilmore for this purpose are now in 
 general use in this country. One of the wires is T V inch in diame- 
 ter and is loaded to weigh | pound; the other is J* of an inch 
 in diameter and loaded to weigh one pound. The paste is said 
 to have reached " initial set" and "end of set" when these two 
 wires, respectively, fail to make an impression on the surface. 
 
 98. M. Vicat also suggested a needle test as follows: The 
 cement paste is placed in a conical ring, 4 cm. in height and 7 
 cm. in diameter at the base. The consistency should be such 
 that a rod 1 cm. in diameter and weighing 300 grams does not 
 entirely pierce the mass. This consistency having been ob- 
 tained by trial, a needle of circular cross-section having an area 
 of 1 sq. mm. and loaded to weigh 300 grams, is gently lowered 
 on the paste. The moment when this needle no longer pene- 
 trates the mass is called the beginning of the set, and the time 
 in which it fails to make an impression upon it is called the end 
 of setting. It may be mentioned in passing, that, according to 
 a few comparative tests made by the author, when a cement 
 paste has "set" by Gilmore's "heavy" wire, ^ inch weighing 
 one pound, it requires about 1,100 grams weight on the Vicat 
 1 sq. mm. needle to make an impression on the paste. Vicat's 
 method was indorsed by the Munich Conference and was sug- 
 gested in the recent progress report of the Committee of the 
 American Society of Civil Engineers. 
 
 99. M. LeChatelier has suggested a modification of this 
 method by substituting for the rod 1 cm. in diameter a disc of the 
 same diameter carried by a slender rod, the disc being loaded 
 to weigh 50 grams, the normal consistency being such that the 
 disc will stop midway in the ring, or "vase." The beginning 
 
SETTING OF CEMENT 67 
 
 and end of setting he would define by the penetration of the 
 needle (1 sq. mm. in section) to mid-depth in the ring, the 
 weights being 50 grams and 3,000 grams, respectively. 
 
 100. An approximate method of determining time of setting 
 is also in use as follows: After mixing the cement paste to the 
 proper consistency, place enough of it on a glass plate to form 
 a thin cake, or "pat," about three inches in diameter and one- 
 half inch thick at the center, thinning toward the edges. When 
 the pat is sufficiently hard to bear a gentle pressure of the fin- 
 ger nail, the cement is considered to have begun to set, and 
 when it is not indented by a considerable pressure of the thumb 
 nail, it may be said to have set. 
 
 101. Mr. Henry Faija objected to all methods which are 
 based upon the rates of acquiring hardness, on the ground that 
 there are periods in the early stages of hardening that may be 
 more rationally defined. He considers that the time at which 
 the water leaves the surface of the pat, depriving it of its glossy 
 appearance, is really the beginning of setting, and that this 
 time may or may not correspond to the result obtained by the 
 use of the needle. 
 
 102. Variations in the Rate of Setting. Some of the quali- 
 ties which determine the actual rate of setting of a cement 
 are, its composition, degree of burning, age and fineness. Aside 
 from these qualities of the cement itself, the addition of certain 
 salts subsequent to the manufacture also influences the rate. 
 The observed rate of setting will be influenced by the details 
 of the test, such as the quantity, temperature and composition 
 of the water used in gaging, the amount of gaging, the tem- 
 perature of the cement, and the temperature and character of 
 the medium in which the pat is placed after molding. 
 
 103. An over-limed or highly limed cement is usually slower 
 setting than an over-clayed one. Among natural cements, those 
 of the aluminous variety are usually quick setting. Other 
 things being equal, a well-burned Portland cement will be slower 
 setting than an underburned sample. It is not certain that 
 such is the case for all natural cements, though it probably is 
 true of most of them. It has been said that underburned ce- 
 ments owe their quick setting to their porosity, but the forma- 
 tion of different compounds in the higher temperature may also 
 account for the difference. 
 
CEMENT AND CONCRETE 
 
 104. The effect of the age of cement on its time of setting 
 is very marked, but varies widely with different samples. The 
 idea that cements invariably become slower setting by storage 
 is a false one. The origin of this error may be found in the 
 fact that by the time cement has reached its destination, it 
 has usually passed through the earlier and more rapid changes 
 in characteristics. Dr. Erdmenger l has stated that some Port- 
 land cements become slower setting, while some set more rapidly 
 as a result of storage. Dr. Tomei made experiments on several 
 Portland cements 2 which show that they generally become 
 quicker setting at first (from one to four months after grind- 
 ing), and then become gradually slower setting, until at the 
 end of a year they set in about the same length of time as 
 when fresh. The writer has seen this trait exhibited very 
 
 TABLE 18 
 
 Time of Setting of Five Samples of Natural Cement as Affected by 
 
 Aeration 
 
 
 
 
 
 TIME SETTING 
 
 TIME SETTING 
 
 
 
 
 WATER. 
 
 
 
 CEMENT 
 
 CEMENT AERATED 
 
 
 
 
 
 H 
 
 FROM PACKAGE. 
 
 19 DAYS. 
 
 
 5 
 
 
 
 H 
 
 
 
 
 
 
 
 
 
 
 
 
 w 
 
 w 
 _i 
 
 PH 
 
 42-3 
 
 2 
 
 ^ 
 
 
 
 8 
 
 
 09 
 
 1 
 
 
 
 EFER 
 
 M 
 
 OS 
 
 gc 
 
 
 ^ 
 
 j* 
 
 J 
 
 Diff. 
 
 & 
 
 ^ 
 
 Diff. 
 i-h. 
 
 REMARKS. 
 
 M 
 
 
 PH ^ 
 
 1" 
 
 g 
 
 >-3 
 
 1-3 
 
 
 " 
 
 ^ 
 
 
 
 
 
 j 
 
 
 5 
 
 H-* 
 
 ^ 
 
 
 i-T* 
 
 
 
 
 
 
 
 
 H 
 
 Min. 
 
 Min. 
 
 Min. 
 
 Min. 
 
 Min. 
 
 Min. 
 
 
 
 a 
 
 6 
 
 c 
 
 a 
 
 e 
 
 / 
 
 9 
 
 h 
 
 j 
 
 J 
 
 
 1 
 
 84 R 
 
 32.0 
 
 65 
 
 67 73 
 
 52 
 
 110 
 
 58 
 
 54 
 
 173 
 
 119 
 
 Five samples, 
 
 2 
 3 
 
 83 R 
 82 R 
 
 u 
 (1 
 
 ' 
 
 
 50 
 44 
 
 100 
 100 
 
 50 
 56 
 
 51 
 48 
 
 164 
 166 
 
 113 
 118 
 
 same brand. 
 Uj and O 2 re- 
 quired more 
 
 4 
 
 U 2 
 
 34.7 
 
 
 
 60 
 
 280 
 
 220 
 
 100 
 
 326 
 
 226 
 
 and less wa- 
 
 5 
 
 6 
 
 2 
 
 84 R 
 
 29.3 
 40.0 
 
 
 
 101 
 
 87 
 
 349 
 
 1200 
 
 248 
 1110 
 
 147 
 
 130 
 
 306 
 1241 
 
 159 
 1111 
 
 ter respect- 
 ively than 
 the others to 
 
 7 
 
 83 R 
 
 ii 
 
 
 
 80 
 
 1178 
 
 1098 
 
 122 
 
 1233 
 
 1111 
 
 make same 
 
 8 
 
 82 R 
 
 ii 
 
 
 
 72 
 
 1202 
 
 1130 
 
 125 
 
 1227 
 
 1102 
 
 consistency. 
 
 9 
 
 U 2 
 
 42.7 
 
 
 
 109 
 
 1256 
 
 1147 
 
 202 
 
 1221 
 
 1019 
 
 
 10 
 
 2 
 
 37.3 
 
 
 
 192 
 
 1247 
 
 1045 
 
 234 
 
 1216 
 
 982 
 
 
 plainly by samples of Portland cement of American manufacture, 
 but has not noticed it in natural cements. Table 18 gives the 
 results of some tests on the effect of aeration on the time of 
 
 1 "Notes on Concrete," by John Newman, p. 11. 
 J Trans. A. S. C. E.. Vol. xxx, p. 12. 
 
SETTING OF CEMENT 
 
 69 
 
 setting of five samples of natural cement from the same 
 factory. 
 
 105. The coarse particles in a cement retard the setting be- 
 cause they are inert. Either fine grinding or sifting will doubt- 
 less hasten the rate of setting, but, as has been stated above, 
 the detection of changes in the rate is difficult. Table 11, 
 81, gives the results of a few tests on this subject. 
 
 106. Addition of Salts. The time of setting of a cement is 
 sometimes regulated at the factory by addition of sulphate of 
 lime to the finished product. Such additions are admitted to 
 the extent of two per cent, by the regulations of the Asso- 
 ciation of German Portland Cement Makers, and are now quite 
 generally made by American Portland cement manufacturers. 
 Table 19 gives the results of a few experiments on the effect 
 of plaster of Paris on the time of setting of several cements. 
 
 TABLE 19 
 Effect of Plaster Paris on Time of Setting 
 
 
 a~ 
 
 Time to Bear \ Ib. 
 
 Time to Bear 1 Ib. 
 
 
 Q s .2 
 
 Wire, Minutes, with Plaster 
 
 Wire, Minutes, with Plaster 
 
 CEMENT. 
 
 '" * e{ 
 
 Paris as Certain Percent- 
 
 Paris as Certain Percent- 
 
 
 |jj 
 
 age of Cement and 
 Plaster Paris. 
 
 age of Cement and 
 Plaster Paris. 
 
 Kind. 
 
 Brand. 
 
 |is 
 
 0% 
 
 1% 
 
 2% 
 
 3% 
 
 6% 
 
 0% 
 
 1% 
 
 2% 
 
 3% 
 
 6% 
 
 Portland 
 
 s 
 
 24 
 
 232 
 
 477 
 
 460 
 
 425 
 
 40 
 
 498 
 
 917 
 
 910 
 
 860 
 
 832 
 
 u 
 
 R 
 
 24 
 
 96 
 
 375 
 
 381 
 
 358 
 
 75 
 
 345 
 
 745 
 
 776 
 
 778 
 
 750 
 
 l( 
 
 X 
 
 26 
 
 4 
 
 258 
 
 287 
 
 268 
 
 84 
 
 305 
 
 625 
 
 725 
 
 668 
 
 694 
 
 Natural 
 
 Gn 
 
 34 
 
 38 
 
 106 
 
 107 
 
 86 
 
 42 
 
 543 
 
 414 
 
 527 
 
 671 
 
 632 
 
 " 
 
 An 
 
 34 
 
 93 
 
 179 
 
 302 
 
 295 
 
 93 
 
 193 
 
 439 
 
 592 
 
 726 
 
 698 
 
 It is seen that small percentages retard the initial setting in 
 a marked degree, the maximum effect usually being given by 
 2 per cent, of the plaster. Larger percentages tend to make 
 the cement quicker setting again, so that with 6 to 10 per cent, 
 added, the cement may begin to set quicker than without the 
 addition of plaster. The final set (time to bear one pound 
 wire) does not appear to be thus hastened by large percentages. 
 This might be considered to indicate that the hastening of the 
 initial set is caused by plaster of Paris taking up the water from 
 the cement and obtaining sufficient hardness to bear the light 
 wire. 
 
 The probable explanation of the action of a small amount of 
 
70 CEMENT AND CONCRETE 
 
 sulphate of lime in retarding the setting is that suggested by 
 M. Candlot, l namely, that the aluminate of lime, to which is 
 due the initial setting, dissolves less readily in a solution of 
 sulphate of lime than in pure water. If the aluminate does 
 not commence to hydrate until the silicate of lime has set, the 
 subsequent combination of the sulphate and aluminate may 
 cause the mortar to disintegrate. 
 
 107. Solutions of common salt have been found to retard 
 the setting, but when a large percentage of salt is used, it some- 
 times forms a crust on the top which may resist a light wire and 
 thus make the paste appear to be quicker setting. Sea water 
 generally retards the setting somewhat more than solutions of 
 common salt, probably on account of the magnesian salts pres- 
 ent, but M. Candlot says that cements to which sulphate of 
 lime has been added set more rapidly when gaged with sea water 
 than when gaged with fresh water. 
 
 The effect of calcium chloride on the setting of cements is 
 entered into in detail in M. Candlot's treatise on " Cements and 
 Hydraulic Limes," and may be summarized as follows: A weak 
 solution of calcium chloride renders Portland cement slower 
 setting because the aluminate of lime dissolves more slowly in 
 such a solution than in pure water. On the other hand, the 
 aluminate dissolves rapidly in a concentrated solution of calcium 
 chloride, and therefore such a solution hastens the setting of 
 Portland cement. Aluminous cements, i.e., cements containing 
 a very high percentage of alumina, are not appreciably affected 
 by gaging with a comparatively weak solution of calcium chlo- 
 ride on account of the large excess of aluminate of lime present; 
 and on the other hand, cements containing no alumina are not 
 affected, as in such cements the hardening is due to the silicate 
 of lime. A weak solution of the chloride hastens the hydration 
 of the free lime, and therefore a cement which contains a dan- 
 gerous percentage of the latter may be made sound by gaging 
 with such a solution, as the lime may thus be hydrated before 
 the cement sets. The chloride of calcium test for soundness is 
 based on the supposition that the free lime may be hydrated by 
 the action of the chloride soon after the setting of the cement, 
 and thus the expansive action be hastened. 
 
 "Ciments et Chaux Hydrauliques," par E. Candlot, 
 
SETTING OF CEMENT 
 
 71 
 
 The effect of sugar on the time of setting does not seem to 
 be well known, but it is said l that the presence of saccharine 
 matter may either accelerate or retard the setting of the cement, 
 depending on the amount of sugar present, the character of the 
 cement and the amount of water used. 
 
 108. The quantity of water used in gaging has a most impor- 
 tant influence on the test for time of setting, an increased quan- 
 tity of water retarding the setting. This may be seen from 
 Table 20. 
 
 TABLE 20 
 Effect of Consistency of Mortar on the Time of Setting 
 
 
 ' Water as per cent, of 
 
 
 
 
 
 
 
 J H 
 
 cement by weight . . 
 
 26.7 
 
 2K.O 
 
 30.8 33.3 
 
 36.4 
 
 40.0 
 
 
 gs. 
 
 Minutes to bear -/> inch 
 
 
 
 
 
 
 
 
 H * 
 
 wire weighing \ pound. 
 
 20 
 
 2;> 
 
 30 
 
 42 
 
 4(5 
 
 55 
 
 
 & 
 
 Minutes to bear ^ T inch 
 
 
 
 
 
 
 
 
 
 wire weighing 1 pound 
 
 28 
 
 41 
 
 57 
 
 76 
 
 78 
 
 85 
 
 
 
 Water as per cent, of 
 
 
 
 
 
 
 
 
 a 
 
 H 
 
 cement by weight . 
 
 24 
 
 26 
 
 28 
 
 30 
 
 32 
 
 34 
 
 36 
 
 ll 
 
 Minutes to bear y 1 , inch 
 
 
 
 
 
 
 
 
 | 
 
 \ pound wire . . . 
 
 2 
 
 2 
 
 3 
 
 ~7 
 
 21 
 
 28 
 
 38 
 
 Q 
 
 Minutes to bear ^ inch 
 
 
 
 
 
 
 
 
 
 1 pound wire . 
 
 160 
 
 188 
 
 279 
 
 289 
 
 371 
 
 403 
 
 583 
 
 As might be supposed, this influence varies with different 
 samples, and M. H. LeChatelier 2 has given the following table 
 which illustrates this point. 
 
 TABLE 21 
 
 Effect of Consistency of Mortar on Time of Setting 
 
 CEMEXT. 
 
 PEK CtfXT. 
 WATEK. 
 
 TIME SKTTIXG, 
 Minutes. 
 
 Portland A J 
 
 24 
 
 20 
 
 Portland B . \ 
 
 34 
 25 
 
 85 
 
 7 
 
 Quick settin " Vassy . < 
 
 35 
 
 50 
 
 45 
 5 
 
 
 58 
 
 10 
 
 "Masonry Construction/' I. O. Baker, p. 98. 
 2 "Tests of Hydr. Materials," p. 33. 
 
72 
 
 CEMENT AND CONCRETE 
 
 109. It is necessary, then, in writing specifications and in 
 making tests, where the time of setting is at all carefully con- 
 sidered, to note the consistency of the paste used in the test. 
 Practically, it is preferable to use a paste rather thinner than 
 that usually employed for briquets. 
 
 The consistency is sometimes defined by M. Vicat's apparatus 
 of a rod 1 cm. in diameter, or by M. LeChatelier's modification 
 of the same mentioned above, or by the requirement that it 
 shall be at the point of ceasing to adhere to the trowel. Another 
 definition is that it shall, when placed on a glass plate, flow 
 toward the edges only on repeated jarring of the plate. This 
 last is a very fair approximate method, though giving a rather 
 thin paste. 
 
 That mortars set more slowly than neat cement paste is 
 largely due to the increased amount of water present in the 
 former, this excess of water being required to moisten the 
 grains of sand. The relation between the time of setting of mor- 
 tars and neat cement paste is not definite. M. Candlot found 
 the time of setting of one-to-three mortars to be from two to 
 twenty times as great as that of the paste of neat cement of 
 normal composition. 
 
 110. The temperature of the cement and water also has an 
 important bearing on the observed time of setting. As the 
 temperature of the materials is increased, the time of setting 
 diminishes in about the same proportion. The following table 
 gives a few of the results obtained by M. Candlot * with Port- 
 land cements. 
 
 TABLE 22 
 
 Effect of Temperature of Materials on Time of Setting 
 
 
 TEMPERATURE, 
 Degrees C. 
 
 TIME OF SETTING, 
 Minutes. 
 
 C611161lt No 1 \ 
 
 6 
 15 
 
 60 
 
 25 
 
 Cement No. 2 ] 
 
 25 
 
 7 
 20 
 
 4 
 
 350 
 
 295 
 
 
 30 
 
 190 
 
 Table 23 gives the results of similar tests made under the 
 author's direction. The temperatures of cement and water 
 
 f dments et Chanx Hydrauliques, f> par E. Candlot. 
 
SETTING OF CEMENT 
 
 73 
 
 were varied while the temperature of the room in which the 
 tests were made remained nearly constant, or from 63 to 67 
 Fahr. 
 
 TABLE 23 
 
 Effect of Temperatures of Cement and Water on the Time of 
 
 Setting of Paste 
 
 Temp, cement and water, | 
 Degrees, Fahr. . ) 
 
 40 
 
 50 
 
 60 
 
 70 
 
 80 
 
 90 
 
 100 
 
 110 
 
 Minutes to bear ,-L } 
 
 
 
 
 
 
 
 
 
 inch wire weigh- > Portland 
 
 270 
 
 247 
 
 225 
 
 196 
 
 175 
 
 158 
 
 135 
 
 . . . 
 
 ing * pound. } Nat , ura i 
 
 102 
 
 90 
 
 84 
 
 72 
 
 60 
 
 54 
 
 55 
 
 43 
 
 111. Amount of Gaging. If a cement paste containing a 
 moderate amount of water be insufficiently gaged, it will appear 
 dry, when a more thorough working might make it plastic. 
 Thus an insufficient gaging may make a cement appear quicker 
 setting. It is also the case that when a cement is regaged after 
 having begun to set, the second setting will take place more 
 slowly; this, however, is a somewhat different matter. 
 
 112. The temperature and character of the medium in which 
 the pat is kept during the setting process will have a decided 
 influence on the rate of setting. 
 
 This is clearly shown by the following table, given by M. 
 
 TABLE 24 
 
 Time of Setting as Affected by Temperature of the Water and 
 of the Medium in which Cement Sets 
 
 SAMPLE. 
 
 TEMPERATURE 
 
 TIME REQUIRED TO 
 
 Of water at time 
 of gaging. 
 
 Of air during 
 setting. 
 
 Begin to set. 
 
 Set. 
 
 Degrees C. 
 
 Degrees C. 
 
 Hr. Min. 
 
 Hr. Min. 
 
 1 I 
 
 
 16 
 
 1 
 
 16 
 
 6 47 
 20 
 
 11 
 2 . 23 
 
 2 { 
 
 
 16 
 
 1 
 16 
 
 5 30 
 52 
 
 8 8 
 6 l:i 
 
 3 1 
 
 
 15 
 
 3 
 
 15 
 
 12 
 43 
 
 20 
 3 3 
 
 * 1 
 
 
 15 
 
 3.5 
 17 
 
 24 
 20 
 
 1 3 
 45 
 
74 CEMENT AND CONCRETE 
 
 Paul Alexandre, 1 from which it appears that different samples 
 are affected in very different degrees. It is seen that the 
 higher the temperature, the more rapid the setting. 
 
 113. At temperatures below 32 F. (0 C.), setting seems 
 to be entirely suspended. If a cement paste, which has been 
 submitted to such low temperatures since gaging, is brought 
 into a warm room, the setting process begins as though the 
 mortar had just been gaged. It must not be concluded, 
 however, that freezing has no evil effect on mortars. (See 
 Art. 50.) 
 
 114. Setting in Air and Water. A cement paste sets much 
 quicker in air than in water. This is due to the percolation of 
 water to the interior of the pat, when it is immersed as soon as 
 made, being analogous to using an excess of water in gaging. 
 When a pat sets in dry air, the evaporation of water from the 
 surface hastens the hardening of that portion. If immersed 
 directly after it has set in air, it re-softens, and this is also 
 true of some briquets immersed when twenty-four hours old. 
 The time of setting of cements that are to be deposited under 
 water may well be tested in that medium, when they should 
 be protected by a mold of some form to retain their shape. 
 Ordinarily the time of setting should be tested in moist air. 
 
 Cements are said to set more quickly in compressed air than 
 in free air; this may be partially due to the higher temperature 
 usually existing in the former. 
 
 115. Requirements as to Time of Setting. What is desir- 
 able as to time of setting will, of course, depend on the work 
 in hand; certain purposes requiring that the cement shall be 
 able to retain its shape soon after deposition, while in other 
 cases ability to mix large quantities at a time, without fear of 
 the cement setting before it is in place, may be very convenient. 
 An extremely quick setting cement should be regarded with 
 suspicion until it has proved itself of good quality. It is some- 
 times stated that where a quick setting mortar is desired, nat- 
 ural cement must be used, but this is not true; either Portland 
 or natural may be found with almost any rate of setting de- 
 sired. As a general rule, however, among cements that have 
 been stored several months, the Portlands are slower setting. 
 
 1 "Recherches Experimentales sur les Mortiers Hydrauliques," par Paul 
 Alexandre. 
 
SETTING OF CEMENT 75 
 
 Portland cement will ordinarily begin to set in from twenty 
 minutes to six hours, and natural cement in from ten minutes to 
 two hours, though there are many cements the time of setting of 
 which is outside of these limits. 
 
 116. Conclusions. The purpose aimed at in the test for 
 time of setting will, to a certain extent, regulate the method 
 to be employed. The pressure of the finger nail will be suf- 
 ficient to determine (after a little experience) whether a cement 
 will answer a certain purpose in this regard. But, if one is 
 working to rigid specifications, or pursuing investigations as to 
 the effect of different treatment on time of setting, it becomes 
 very desirable to have a method of determining and defining 
 the consistency of the mortar, and an accurate method of de- 
 termining the rate of setting. 
 
 In the author's experience, the Vicat consistency apparatus 
 as modified by M. LeChatelier (see 99) has proved unsatisfac- 
 tory except for thin pastes of neat cement or mortars contain- 
 ing less than two parts of sand. If the paste is not of such a 
 consistency as to run freely into the ring, or "vase," an error 
 may be introduced in the method of filling the latter. In oper- 
 ating with a natural cement it was found that a neat paste, in 
 which the water used was 32 per cent, of the dry cement, re- 
 quired a gross weight of 640 grams to make the disc (1cm. diam- 
 eter) penetrate midway in the vase; with 33 per cent, water, 
 a weight of 410 grams was required; 34 per cent., about 250 
 grams; 35 per cent., 175 grams; 37 per cent., 155 grams. It 
 would seem that some modification of this apparatus might be 
 made which would not only indicate when a thin, neat cement 
 paste has the assumed "normal" consistency, but which would 
 also define the consistency of a given mortar, whether of neat 
 cement or of sand mixture. 
 
 General Gilmore's wires are very simple, and will perhaps 
 answer the purpose of obtaining the time of setting as well as 
 any method in use. They can be used somewhat more accu- 
 rately if the wires are made to slide vertically in a frame, than 
 when held in the hand. 
 
 The necessity of care in all of the details of this test, tem- 
 perature and amount of water, amount of gaging, character 
 of medium, etc., has been sufficiently emphasized in the preced- 
 ing paragraphs. 
 
76 CEMENT AND CONCRETE 
 
 ART. 18. CONSTANCY OF VOLUME 
 
 117. That a cement should not contain within itself ele- 
 ments which may lead to its destruction, is evidently a most 
 important quality. It is probable that nearly all cements un- 
 dergo a slight change in volume during induration, contracting 
 in air and expanding in water. But it is the detection of those 
 larger changes, which result from bad proportions or defective 
 manufacture, and which cause deterioration or even complete 
 disintegration, that is the object of the tests for soundness. 
 
 118. Causes of Unsoundness. The most frequent cause of 
 unsoundness is considered to be the presence of free lime or 
 magnesia. (See 49 and 50.) Any one of the following causes 
 may account for the presence of free lime in cement: (1) An 
 excessive percentage of lime may have been used in proportion- 
 ing the raw materials; (2) the raw materials may not have been 
 sufficiently mixed to render the mass homogeneous; (3) hard 
 particles of lime, such as shells, may not have been ground fine 
 enough in making the mix to permit them to enter into com- 
 bination with the other ingredients during burning; or (4) the 
 cement may have been underburned, so that part of the lime 
 did not enter into combination. 
 
 The particles of free lime which occur in cements are nat- 
 urally rather difficult to slake on account of their impurity and 
 the high temperature at which they have been calcined, and 
 the same thing is probably true of magnesia. It may thus 
 require weeks or months of exposure to the atmosphere to cor- 
 rect tendencies to expand due to the presence of free lime or 
 magnesia. Likewise when such defective cements are immersed 
 in water of ordinary temperature, the expansion may not occur 
 for a considerable period. This fact has led to the use of hot 
 tests of various kinds to detect such faults, but before touch- 
 ing on these so-called " accelerated tests/ 7 the ordinary cold- 
 water test will be described. 
 
 119. TESTS FOR SOUNDNESS. The Committee of the Amer- 
 ican Society of Civil Engineers on a " Uniform System for 
 Tests of Cement" recommended, in 1885, the following test for 
 soundness: "Make two cakes of neat cement two or three inches 
 in diameter, about one-half inch thick, with thin edges. One 
 of these cakes, when hard enough, should be put in water and 
 
CONSTANCY OF VOLUME 77 
 
 examined from day to day to see if it becomes contorted, or if 
 cracks show themselves at the edges, such contortions or cracks 
 indicating that the cement is unfit for use at that time. In 
 some cases the tendency to crack, if caused by the presence, of 
 too much unslaked lime, will disappear with age. The re- 
 maining cake should be kept in air and its color observed, which 
 for a good cement should be uniform throughout, yellowish 
 blotches indicating a poor quality; the Portland cements being 
 of a bluish-gray, and the natural cements being light or dark, 
 according to the character of the rock of which they are made." 
 For the ordinary cold test this method will probably give as 
 valuable results as any of the forms that are suggested. 
 
 120. The German regulations require a very similar test, 
 except that in the case of slow setting cements the pat is not 
 immersed until twenty-four hours old. While a cement that is 
 decidedly bad may show its defects in from one day to one week 
 by this cold water test, it may be the case that cracks will ap- 
 pear only after several months' immersion. It has therefore 
 been proposed to hasten the destructive action of the free 
 lime or magnesia by submitting the cakes of cement to steam, 
 hot water, or dry heat. 
 
 121. The Kiln Test, recommended by Prof. Tetmajer in 1890, 
 consists in placing in an air bath, pats which have been kept 
 in moist air for twenty-four hours; and then gradually raising 
 the temperature of the air bath to 120 C. This temperature 
 is maintained for at least one-half hour after the disengage- 
 ment of steam has ceased. The pats should show no tendency 
 to expand under this treatment, but if cements fail to pass the 
 test, the results of the ordinary cold water treatment are to be 
 awaited. This test is intended for cements that are to be 
 used in air. 
 
 122. The Boiling Test, which was also recommended by Prof. 
 Tetmajer, consists in placing the pats, twenty-four hours after 
 made, in water of ordinary temperature, and gradually heating 
 the water to bring it to the boiling point in about an hour; 
 five or six hours in the boiling water should develop no defects. 
 This is a severe test, and has been objected to on the ground 
 that cements which have been well proportioned, but which 
 are a trifle underburned, will fail to pass this test while giving 
 good results. in mortars to be used in the air, This test, how- 
 
78 CEMENT AND CONCRETE 
 
 ever, is steadily gaining in favor, and is used in many cement 
 works as a test of quality. 
 
 123. The Warm Water Test. Mr. H. Faija was an early 
 experimenter in accelerated tests for soundness, and about 1882 
 he began the use of a "steamer," using a temperature of about 
 110 Fahr. After eleven years' use he still believed this tem- 
 perature to be high enough to detect tendencies to expand in 
 faulty cements. The apparatus 1 "consists of two vessels, one 
 within the other, a water space being thus maintained between 
 them, which assists in equalizing the temperature of the inner 
 or working vessel." The latter is partially filled with water 
 and is provided with a rack or shelf near the top. A ther- 
 mometer is inserted through the cover of the inner vessel, and 
 the water within is kept constantly at 110 Fahr. As soon as 
 the pat is gaged, it is placed on the rack in the vapor, which 
 will be at about 100 Fahr. After six or seven hours in this 
 moist heat, the pat is immersed in the warm water. "In the 
 course of twenty-four hours it is taken out and examined, and if 
 then found to be quite hard and firmly attached to the glass, the 
 cement may at once be pronounced sound and perfectly safe 
 to use; if, however, the pat has come off the glass and shows 
 cracks or friability on the edges, or is much curved on the 
 under side, it may at once be decided that the cement in its 
 present condition is not fit for use." Mr. Faija also recom- 
 mended, in case of failure in the first test, that the cement be 
 spread out in a thin layer for a few days and a second test 
 made. If the cement passes this second test, it is pronounced 
 sound and fit for use after being stored a sufficient length of 
 time. 
 
 124. The Hot Water Test. The temperature to be used in 
 accelerated tests for soundness is a point which has received 
 much attention and is still under discussion. In 1890 M. Deval 
 described a series of experiments he had made, in which he 
 employed a temperature of 80 C. While this is much more 
 severe than the temperature used by Mr. Faija, it is still mild 
 in comparison to some temperatures that have been advocated. 
 
 125. Mr. W. W. Maclay, who was probably the first engi- 
 neer in this country to introduce a hot test requirement in 
 
 1 "Portland Cement Testing," by H, Faija, Trans, A, S. C. E., Vol. xvii, 
 p. 222. 
 
CONSTANCY OF VOLUME 79 
 
 specifications, gave the results of his experiments in a paper 
 presented to the American Society of Civil Engineers in 1892. 
 The method used " consists in molding six pats of pure cement 
 and water, about one-half inch thick and about three inches in 
 diameter, on thin glass plates, and of the same consistency as 
 for the briquets for tensile strength." The treatment to which 
 these pats are submitted is as follows: - 
 
 No. 1, in steam (vapor) bath, temperature 195 to 200 F., 
 as soon as made. 
 
 No. 2, in same vapor bath when set hard (bear J f inch wire 
 weighing one pound). 
 
 No. 3, ditto, after twice the length of time in air allowed the 
 second pat. 
 
 No. 4, ditto, after 24 hours. 
 
 No. 5, in water of temperature about 60 F. when set hard. 
 
 No. 6, kept in moist air at temperature of about 60 I. 
 
 "The first four pats are each kept in the steam bath three 
 hours, then immersed in water of a temperature of about 200 
 Fahr. for twenty-one hours each, when they are taken out and 
 examined. To pass this test perfectly, all four pats, after being 
 twenty-one hours in hot water, should, upon examination, show* 
 no swelling, cracks, nor distortions, and should adhere to the 
 glass plates. The latter requirement, while it obtains with 
 some cements nearly free from uncombined lime, is not insisted 
 upon; the cracking, swelling and distortion of the pats being 
 much the more important features of this test. The cracking 
 or swelling of No. 1 pat alone can generally be disregarded." 
 
 126. DevaPs Method. Making tests of mortar briquets, 
 which have been kept in hot water, seems to be the most rational 
 accelerated test for soundness. This method was used in Ger- 
 many several years ago, when it was claimed that a definite 
 relation existed between the results thus obtained and the longer 
 time cold water tests. This theory being disproved, threw dis- 
 credit on the hot test, but M. Deval l has since made many 
 experiments showing that it is of much value in detecting 
 bad products. 
 
 The method consists in making briquets with three parts 
 sand to one of cement, and after twenty-four to seventy-two 
 
 1 " Hot Tests for Hydraulic Cements," M. Deval, Bull. Soc. d' Encourage- 
 ment, etc., 1890, pp. 560-583. 
 
80 
 
 CEMENT AND CONCRETE 
 
 hours in moist air, according to the rate of setting, immersing 
 them in water maintained at 80 C., the briquets being broken 
 after an immersion of from two to seven days. These hot 
 water briquets are to be compared with briquets stored in 
 water of the ordinary temperature and broken at seven and 
 twenty-eight days after immersion. 
 
 127. Among other tests M. Deval compared the results ob- 
 tained with six samples of Portland cement as follows: 
 No. 1. Good finely ground cement of modern make. 
 No. 2. Coarsely ground cement of good quality, but partially 
 
 aerated. 
 No. 3. Quick setting cement with low per cent, lime and 
 
 lighter burn. 
 
 No. 4. Made from clinker having property of disintegrating 
 spontaneously while cooling; large proportion of inert 
 m material. 
 
 No. 5. Under-burnt cement; contains free lime. 
 No. 6. Over-limed cement. 
 The results of the tests are given in the following table: 
 
 TABLE 25 
 
 Cold and Hot Tests on Six Samples of Portland Cement 
 (M. Deval) 
 
 
 TENSILE STRENGTH IN KILOS PER SQ. CM. 
 
 CEMENT. 
 
 Cold. 
 
 Hot. 
 
 
 7 days. 
 
 28 days. 
 
 2 days. 
 
 7 days. 
 
 1 
 
 15.0 
 
 23.3 
 
 17.2 
 
 24.3 
 
 2 
 
 6.7 
 
 13.7 
 
 7.6 
 
 11.0 
 
 3 
 
 6.2 
 
 16.5 
 
 7.3 
 
 16.2 
 
 4 
 
 2.9 
 
 3.9 
 
 ) 
 
 5 
 
 6.1 
 
 12.2 
 
 Disintegrated. 
 
 6 
 
 7.6 
 
 20.2 
 
 5 
 
 No. 4, when allowed forty-eight hours to set, gave 3.2 kilos 
 at two days, and 4.3 kilos at seven days, when tested hot. 
 Among the cements which disintegrated in the hot water, the 
 only one that gave a high result cold was No. 6, and this sam- 
 ple, it is stated, would crack and swell badly even in cold water 
 
CONSTANCY OF VOLUME 81 
 
 if mixed neat. It is quite possible, however, that a sample 
 might be found which, not having quite as flagrant defects as 
 No. 6, would pass all the cold tests but be condemned by the 
 hot test. 
 
 128. The conclusions drawn from these experiments have 
 been stated as follows: 
 
 "(1) Tests made cold do not indicate the quality of the 
 cement, inasmuch as cement containing excess of lime, and, in 
 consequence, deplorably bad, may give excellent results." 
 
 "(2) Portland cement of good quality, mixed with normal 
 sand in the proportion of one to three, resists water at 80 C. 
 Its strength at two and seven days after setting is about equal 
 to that which it would have at seven and twenty-eight days 
 in the cold." 
 
 "(3} Poor cement containing much inert material does not 
 resist the action of water at 80 C. unless the setting be allowed 
 to proceed for some days before immersion." 
 
 "(4) Cements containing free lime do not withstand the ac- 
 tion of water at 80 C. if immersed twenty-four hours after 
 setting." Comparison of the strength hot and cold will suffice 
 for the detection of even small quantities of free lime. 
 
 129. Before passing to the comparison of the tests for sound- 
 ness already outlined, a few other tests which have been sug- 
 gested for use may be briefly mentioned. 
 
 The Chloride of Calcium Test depends on the fact that slak- 
 ing of free lime is hastened by feeble solution of chloride of 
 calcium. (See 107.) Concerning this test, Prof. F. P. Spald- 
 ing 1 says he "has found it to give true indications in a number 
 of cases, including some unsound magnesian cements. It con- 
 sists in mixing the mortar for the cakes with a solution of 40 
 grammes chloride of calcium to one liter of water, allowing 
 them to set, immersing them in the same solution for twenty- 
 four hours, and then examining them for checking and soften- 
 ing as in other tests." 
 
 130. M. H. LeChatelier's Method. The method recom- 
 mended by M. H. LeChatelier for testing soundness requires 
 the use of a cylindrical mold, about l inches in diameter and 
 of about the same height, which is made of thin metal and 
 
 1 "Notes on the Testing and Use of Hydraulic Cement," by Fred. P. 
 Spalding. 
 
82 CEMENT AND CONCRETE 
 
 slit along a generatrix. The' mortar is to be placed in the 
 mold as soon as made, and immersed at once in cold water; 
 the mold is held firmly by a clamp, and a flat plate at either 
 end of the mold retains the mortar in shape until set. When 
 setting has taken place, the mold is undamped and the widen- 
 ing of the slit indicates the expansion of the mortar. If de- 
 sired, the swelling may be increased and hastened by transfer- 
 ring the mold and its contents to hot water as soon as the ce- 
 ment is set. The same writer has suggested a modification of 
 the hot test by placing briquets in cold water and gradually 
 heating to near the boiling point, this temperature being main- 
 tained for six hours. 
 
 Various other tests have been suggested, such as the effect 
 of regaging; withstanding immersion as soon as gaged; allow- 
 ing large thin cakes to harden in air and striking them to obtain 
 a musical sound. Most of these tests, however, are worthy of 
 passing notice only. 
 
 131. Discussion. There are but few experiments to show 
 that a cement which will actually fail and disintegrate when 
 properly used, may still pass the cold water neat pat test; yet 
 there is no doubt that inferior cements may pass this test per- 
 fectly, " inferior cements" being those which will not give the 
 best results in practice, though they do not disintegrate. 
 
 Cement is at present used in a very crude way, and it is only 
 in exceptional cases that a poor quality of material may be 
 detected in the completed structure. This is sufficient reason 
 why so few failures can be found in cement work which may 
 be attributed to a poor quality of cement. But in the more 
 economical manner in which this material is, even now, begin- 
 ning to be used, it is absolutely essential to know what its fu- 
 ture behavior will be. That the cement will never be exposed 
 to hot water in actual use, is a weak argument against hot 
 water tests. It must be remembered that the chief object of 
 testing cement is to arrange the various products in their true 
 order of merit, and any system which will effect this result is 
 perfectly legitimate. On the other hand, it is due to the man- 
 ufacturers that a test which will occasionally reject perfect 
 cements should not be adopted when it is possible in any other 
 way to detect poor products with certainty. 
 
 132. It is possible that the temperature used and recom- 
 
CONSTANCY OF VOLVME 83 
 
 mended by Mr. Faija is sufficiently high to detect unsoundness 
 or a tendency to "blow." It has never been clearly proved 
 that it is not, but the higher temperature of 70 to 100 C. has 
 appeared to meet with greater favor. The writer made a few 
 experiments to compare results obtained with mixtures of 
 Portland cement and lime when using the temperature of 110 
 Fahr. (43 C.) with those obtained in water at 190 Fahr. (88 
 C.), and in water at the ordinary temperature of 60 to 65 Fahr. 
 (16 to 18 C.). Quicklime, in proportions varying from one 
 to ten per cent., was added to the cement, and seven pats were 
 made from each mixture of cement and lime. 
 
 These pats were subjected to the following treatment: 
 
 Pat No. 1, placed in vapor of water at 110 F. when made. 
 
 2, 110 F. when set. 
 
 3, 110 F. after 24 hours. 
 " 4, 190 F. when made. 
 
 5, " 190 F. when set. 
 
 6, " 190 F. after 24 hours. 
 Above six pats immersed in the hot water after three hours in 
 
 vapor. 
 
 Pat No. 7, placed in cool water when set. 
 
 When no lime was added, pats 1, 2 and 3 revealed no defects; 
 pats 4 and 5 showed small cracks in two days, but pat No. 6 
 still adhered to the glass after eight days. Pat No. 7 was perfect 
 after two months. With 2 per cent, lime added to the cement, 
 pat No. 1 was slightly warped and cracked, and Nos. 2 and 3 
 were off glass; Nos. 4 and 5 were cracked and warped; No. 6 
 was off glass, and No. 7 became detached from glass after two 
 months, but was otherwise perfect. With 4 per cent, lime, all 
 the pats failed, the one in cool water being off glass, cracked 
 and warped after one day. 
 
 It must be remembered that the free lime occurring in cement 
 is of a different character from the quicklime added in these 
 tests, because the former contains impurities and has been cal- 
 cined at a very high temperature, and would therefore slake 
 more slowly. It has been said that as small an amount as 1 
 per cent of free lime in cement is dangerous. If this is true, 
 and it probably is, the temperature of 110 Fahr. would seem 
 to be inadequate to quickly indicate a tendency to "blow." 
 
84 
 
 CEMENT AND CONCRETE 
 
 133. Some of the results obtained by M. Deval have already 
 been given ( 127). Mr. Maclay made similar tests on several 
 samples of Portland cement, using a temperature of 200 Fahr., 
 but these tests only permit of comparing the strength acquired 
 in cold water in seven and twenty-eight days with the strength 
 in hot water at ages of from two to seven days. Long time 
 tests, showing that the cements which give low results in hot 
 water and normal results in cold water on short time tests, 
 give in reality a low strength at the end of six months or more, 
 have been almost entirely lacking until very recently. 
 
 Table 40, 226, gives some of the results obtained by the 
 author in hot tests and long time cold tests on Portland cement. 
 It is seen that the hot test at 80 C. indicated, in seven days, 
 
 TABLE 26 
 Cold and Hot Tests on Samples of One Brand of Portland Cement 
 
 CEMENT. 
 
 PARTS. 
 SAND 
 
 DATE 
 MADE. 
 1894. 
 
 AGE. 
 
 TENSILE 
 STRENGTH. 
 
 BRIQUETS STORED. 
 
 
 
 Mo. Pa. 
 
 
 
 Moist air. 
 
 Water. 
 
 B' 
 
 2 
 
 4 1C 
 
 5 da. 
 
 8 
 
 1 da. 
 
 80 C. 4 da. 
 
 A 
 
 2 
 
 7 2 
 
 5 da. 
 
 235 
 
 1 
 
 4 
 
 B' 
 
 2 
 
 4 10 
 
 7 da. 
 
 13 
 
 1 
 
 6 
 
 A 
 
 2 
 
 7 2 
 
 7 da. 
 
 229 
 
 1 
 
 ' 6 
 
 B 
 
 3 
 
 7 2 
 
 7 da. 
 
 197 
 
 1 
 
 15 to 18 C. 6 da. 
 
 A 
 
 3 
 
 7 2 
 
 7 da. 
 
 108 
 
 1 
 
 6 
 
 B 
 
 3 
 
 7 2 
 
 28 da. 
 
 298 
 
 1 
 
 < '27 
 
 A 
 
 3 
 
 7 2 
 
 28 da. 
 
 198 
 
 1 
 
 27 
 
 B' 
 
 2 
 
 4 16 
 
 7 mo. 
 
 411 
 
 1 
 
 ' 7 mo. 
 
 A 
 
 2 
 
 7 2 
 
 6 mo. 
 
 465 
 
 1 
 
 6 " 
 
 BEHAVIOR OF PATS MADE JULY 2, 1894 
 
 No. 1 in vapor, when held \$ wire. I Immersed in water 80 C. after three 
 No. 2 in vapor, when held 1# wire. } hours in vapor. 
 No. 3 in tank, when held 1# wire. 
 No. 4 in tank ; two hours after held 1# wire. 
 
 Cement: A, No. 1 off glass in two days; No. 2 warped some in two days. 
 " A, No. 3 O.K. after twenty-one days; off glass and warped in 
 
 fifty-two days. 
 " A, No. 4 loose on glass in twenty-one days; off glass and warped 
 
 in fifty-two days. 
 
 " B, No. 1 off glass and warped some in two days; No. 2 entirely 
 disintegrated in two days; No. 3 loose on glass in twenty- 
 one days; off glass and warped in fifty-two days; No. 4 loose 
 on glass in twenty-one days ; off glass and warped in fifty-two 
 days. 
 
CONSTANCY OF VOLUME 85 
 
 the inferior quality of sample W, although it gave normal re- 
 sults in cold water up to twenty-eight days; the two year tests 
 with mortars containing two parts or more sand, show it to be 
 inferior. If we attempt to carry the analogy too far, however, 
 we fall into the error which placed the hot test in disrepute for 
 several years, that is, we must not expect that the strength in 
 cold water after a long time will be exactly proportional to the 
 strength developed in hot water in a few days. 
 
 134. In Table 26 are given the results of tests by the author, 
 on samples of a single brand of Portland cement. The por- 
 tion marked "A" had been spread out in open air for seventy- 
 seven days in a thin layer. The portion marked "B" was 
 taken directly from the barrel July 2d, and B ' was taken 
 from the same barrel April 16th. Samples B and B' are not 
 identical, because the cement had undergone some change, 
 though stored in the barrel. Each result is the mean of five 
 briquets. 
 
 In the short time cold tests there was nothing to indicate that 
 the cement directly from the barrel was not good, except the 
 very small evidence in the fact that pat No. 3 was loose on glass 
 plate after twenty-one days. In fact, the cold water briquet 
 tests at seven and twenty-eight days unmistakably declare in 
 favor of the sample B. On the other hand, how sharply did 
 the hot tests bring out the defects, two days in hot water being 
 sufficient to entirely disintegrate one of the pats. Although 
 sample B' showed a considerable tensile strength at seven 
 months with two parts sand, yet the pats of neat cement failed, 
 even in cold water, after two months, altogether too late a date 
 to be of any value in preventing the use of the cement. 
 
 135. In a paper read before the American Society for Testing 
 Materials, July, 1903, 1 Mr. W. P. Taylor of the City Testing 
 Laboratory, Philadelphia, gives some very interesting data con- 
 cerning the behavior of cements that failed to pass the boiling 
 test. The method employed was to make cakes of cement in 
 the form of a small egg, keep them in moist air for twenty-four 
 hours, then place them in cold water which is gradually raised 
 to the boiling point and maintained at that temperature for 
 three hours. The results cited show that some unsound ce- 
 
 Proceedings Amer. Soc. for Testing Materials, 1903. 
 
86 CEMENT AND CONCRETE 
 
 ments may be much improved by sifting out the coarse parti- 
 cles, and that a cement failing in the boiling test when fresh 
 may pass it satisfactorily after four or five weeks. 
 
 Examination of the results showed that 96 per cent, of a 
 large number of specimens which did not pass the hot water 
 test failed within three hours, and 99 per cent, in four hours. 
 This fixes a practical limit to the time necessary to continue 
 the test. Some very valuable tests are cited to show the 
 ultimate failure in cold water of samples that failed in the 
 hot tests. Ten cements which passed the cold water pat test 
 of twenty-eight days' duration, but which failed in the boiling 
 test above described, gave normal results in one-to-three mor- 
 tars at twenty-eight days, showing a tensile strength of 217 to 
 252 pounds per square inch, but gave only 47 to 147 Ibs. per 
 square inch at four months. 
 
 Another valuable comparison is given by Mr. Taylor: A 
 compilation of data, covering over a thousand tests on many 
 varieties of cements, showed that "of those samples that failed 
 in the boiling test but remained sound at twenty-eight days (in 
 cold water), 3 per cent, of the normal pats showed checking or 
 abnormal curvature in two months; 7 per cent, in three months; 
 10 per cent, in four months; 26 per cent, in six months and 48 
 per cent, in one year; and of these same samples, 37 per cent, 
 showed a falling off in tensile strength in two months; 39 per 
 cent, in three months; 52 per cent, in four months; 63 per cent, 
 in six months and 71 per cent, in one year." 
 
 136. It may be of interest to introduce here some of the 
 opinions that have been expressed concerning hot tests. M. 
 Candlot 1 says that cements of normal composition, the burning 
 of which has not been carried to the point of vitrification, would 
 be condemned by the hot test of neat cement, although mor- 
 tars made with them show no signs of alteration in sea water, 
 and, when preserved in air ; give entirely satisfactory results. 
 Referring to the tests of one-to-three mortar briquets in water 
 at 80 C., he considers that "cements containing free lime give 
 in hot water, lower resistances than in cold water; cements of 
 good quality give resistances at least equal and nearly always 
 greater in hot water than in cold. Cements well proportioned 
 
 'Ciments et Chaux Hydrauliques," par M. Candlot, pp. 144-145. 
 
CONSTANCY OF VOLUME 87 
 
 and homogeneous, but not having obtained the maximum burn- 
 ing, give satisfactory results with this test." 
 
 In using the slit cylinders mentioned in 130, M. H. Le 
 Chatelier found l that the addition of 5 per cent, of lime could 
 be detected by cold tests in a few hours, while 5 per cent, of 
 magnesia could not be detected in twenty-eight days. The 
 cement containing 5 per cent, lime disintegrated almost at 
 once in hot water, while the sample to which 5 per cent, of mag- 
 nesia had been added, swelled considerably in one day. 
 
 Mr. A. Marichal 2 found that "the percentage of water en- 
 tered in combination, after ten days in hot water, was the same 
 as for six months in cold water, and that the strength of the 
 cement was increasing with the amount of water entered in 
 combination. It was discovered incidentally, that cement con- 
 taining over 5 per cent, of magnesia, or 3 per cent, of uncom- 
 bined lime, would not stand the boiling test." 
 
 137. Hot Tests for Natural Cements. All that has pre- 
 ceded concerning hot tests refers to their use for testing Port- 
 land cements. Very little is known concerning the value of hot 
 tests for natural cements. There are comparatively few natural 
 cements that are absolutely bad, but to distinguish between the 
 first and second quality of this variety of products is much more 
 difficult than to make a similar distinction with Portlands. One 
 point is certain, natural cements must not be expected to with- 
 stand boiling water. Mr. de Smedt experimented with fifteen 
 brands of natural cement, and found that thirteen of them 
 went to pieces in boiling water in two hours, although none of 
 them was thought to contain caustic lime. Prof. Tetmajer 
 has stated that for Roman cements, boiling water, or even 75 C., 
 is not at all conclusive, and recommends 50 C. for trial, but our 
 natural cements are not strictly comparable with Roman ce- 
 ments. 
 
 138. The author has experimented with three temperatures, 
 namely, 50, 60, and 80 C., and is inclined to consider that 
 80 C. is likely to give the most useful information for sand 
 mortar briquets but not for neat cement pastes. Table 41, 
 227, gives the results of hot briquet tests on six brands of 
 
 " Tests Hydr. Materials," by H. LeChatelier. 
 8 Trans. Amer. Soc. C. E., Vol. xxvii, p. 438. 
 
88 CEMENT AND CONCRETE 
 
 natural cement. It is seen that, with two parts sand, brands 
 Jn, Hn, and Bn, give very low results at 80 C.', and these brands 
 are really inferior cements as shown by the two-year cold tests. 
 Brand Jn is the only one that gave a lower result at seven days 
 than at five days when tested at 80 C., and this brand failed 
 entirely at two years, though it gives normal results in cold 
 water up to six months. Neat cement pats of this brand, after 
 being stored in cold water for nearly one year, were found to be 
 cracked, although they had been perfect after one month in 
 cold water. It was also found that neat cement pats of this 
 brand warped and cracked in two days when placed in water of 
 60 C. when set. 
 
 139. CONCLUSIONS. ~ It may be said that although the 
 limits within which the hot tests are reliable have not been well 
 established, and although a strict adherence to them may at 
 times reject a usable product, yet it is believed that sufficient 
 experiments have been made to indicate that they are of much 
 value, and should be made in all cases where the quality of the 
 cement is of high importance. 
 
 The present indications seem to be that Portland cements 
 may well be tested in the form of neat cement pats and sand 
 mortar briquets at a temperature of about 80 C. Natural ce- 
 ments in the form of neat paste should not be called upon to 
 resist a temperature above 60 C., but 80 C. will probably give 
 the most useful information with sand mortars. In either case, 
 the mortar should be allowed to set in moist air of ordinary 
 temperature, then transferred to the vapor, to remain two or 
 three hours before immersion in the hot water. It is not rec- 
 ommended that these hot tests should replace the ordinary 
 cold tests, but simply that in cases where the extra work in- 
 volved is not prohibitive, the hot tests should be made in con- 
 nection with the cold tests. 
 
CHAPTER VIII 
 
 TESTS OF THE STRENGTH OF CEMENT IN COMPRES- 
 SION, ADHESION, ETC. 
 
 140. IN testing the strength of cement the object is three- 
 fold : 1st, to obtain an idea of the strength that may be ex- 
 pected from the cement as used in the structure; 2d, to obtain a 
 basis for comparing the value of different cements in this regard; 
 and 3d, to determine the ability of the cement to withstand 
 destructive agencies, whether these agencies be due to exterior 
 causes or emanate from the character of the cement itself. To 
 illustrate the last point it is only necessary to mention such de- 
 stroying agents as free lime (interior) and frost (exterior). It is 
 evident that the stronger the cement the more effectually will 
 these agencies be resisted. 
 
 The strength of cement may be tested by compression, 
 shearing, bending, adhesion, abrasion and tension. The tensile 
 test is the one most frequently used, but the tests will be con- 
 sidered in the order named. 
 
 ART. 19. TESTS IN COMPRESSION AND SHEARING 
 
 141. Value of Test. In practically all forms of masonry 
 construction, cement is called upon to resist compression. In 
 consequence of this fact, the opinion is somewhat general that 
 the greatest amount of information would be obtained by com- 
 pressive tests. But the compressive strength of cement is so 
 much greater than its tensile strength, that when failures occur, 
 they are likely to be due to other forms of stress. In short, the 
 ratio of the compressive strength to the crushing force it is 
 likely to be called upon to resist, is usually much greater than 
 the corresponding ratio in tensile strength. 
 
 142. There is no doubt that compressive tests are of much 
 interest and value, especially so since the use of concrete and 
 steel in combination has become general, but as yet the facili- 
 ties for making the test are not available without considerable 
 expense. This is on account of the larger force required (the 
 
 89 
 
90 CEMENT AND CONCRETE 
 
 compressive strength being six to ten times the tensile) and be- 
 cause the uniform distribution of the stress over the surface of 
 the specimen, and the accurate recording of the force exerted, 
 are even more difficult than the corresponding operations in 
 tensile tests. Prof. Sondericker, 1 in a paper read before the 
 Boston Society of Civil Engineers, describes an apparatus in 
 which he seems to have overcome a part of these difficulties. 
 
 A convenient specimen for compressive tests is a cube meas- 
 uring two inches on a side. The specimens are prepared and 
 treated in the same way as briquets for tensile tests. Before 
 testing, two opposite faces of the cubes are usually ground so as 
 to be true planes, parallel to each other, or the opposite sides 
 may be faced with plaster of Paris, though this is not recom- 
 mended. Grinding two surfaces to true planes increases very 
 much the work involved in testing, so that several tensile tests 
 may be made in the time required to make one compressive test. 
 143. Conclusions. Although tests of compressive strength 
 are of interest from a scientific point of view, it is not considered 
 that they would give much greater information concerning the 
 relative qualities of cements than is given by tensile tests, and 
 therefore they need not be included in an ordinary series of 
 acceptance tests. 
 
 144. Tests of Shearing Strength. Although cement is fre- 
 quently called upon to withstand a shearing stress, tests of this 
 kind are very seldom made. Some of the difficulties encoun- 
 tered in compressive tests are also present in tests of shearing. 
 Prof. Cecil B. Smith made quite an extended series of shearing 
 tests by cementing together three bricks, the middle one pro- 
 jecting above the other two, and the pressure being so applied 
 as to avoid any transverse stress. It is evident that by this 
 method the adhesive strength is also brought into play. Shear- 
 ing tests need not be included in normal tests of quality. 
 
 ART. 20. TESTS OF TRANSVERSE STRENGTH 
 
 145. It is probable that the earliest rupture tests of cement 
 were made by submitting rectangular prisms to a bending 
 stress; but such tests have long held a place subordinate to 
 trials of tensile strength. A mass of masonry, taken as a whole, 
 is very apt to be subjected to a bending stress, but it is a ques- 
 
 Jour. Assoc. Engr. Soc., Vol. vii, p. 212. 
 
TRANSVERSE TESTS 91 
 
 tion whether a transverse test on a small specimen gives any 
 better idea of the ability of a large beam to carry its load, than 
 do simple tensile and compressive tests. 
 
 In Engineering News of December 14, 1893, appeared an 
 article giving the comparative results obtained in tensile and 
 transverse tests. The tensile specimens had an area of one 
 square inch at the smallest place, and the transverse specimens 
 also had an area of cross-section of one square inch. It was 
 found that the modulus of rupture computed by the common 
 
 3 W I 
 formula/ = . . 2 was from 1.1 to 3.8 times the tensile strength 
 
 developed by the briquets. Some comparative tests made at 
 St. Mary's Falls Canal are discussed in Art. 56. 
 
 146. The objections to transverse tests are: 1st, if the speci- 
 mens are made but one inch in cross-section, it is difficult to 
 handle them without injuring them, and if the section is made 
 much larger than one inch square, a much greater amount of 
 cement is required to make the specimens and more room re- 
 quired to store them; 2d, it would seem that the results ob- 
 tained might be less trustworthy than those in tensile tests 
 because of the greater influence of the outside layers, which are 
 subjected to the greatest accidental variations, on the apparent 
 strength of the specimen. On the other hand, it may be said 
 that, when no testing machine is at hand, the apparatus requi- 
 site to make a crude test may easily be improvised. All that is 
 required is a rectangular wooden mold, three knife edges, and a 
 pail with a quantity of sand or water. 
 
 147. When it is a question of making tests of transverse 
 strength accurately and rapidly, the apparatus required is no 
 more simple than the apparatus for tensile tests. In the con- 
 struction of metal molds in large quantities it makes little dif- 
 ference whether the form requires curved or straight lines. As 
 far as breaking is concerned, there is a certain force to be applied, 
 and a machine that will answer for one test may also be used 
 for the other. In the matter of clips, there may be a slight 
 advantage as to simplicity in a clip designed for transverse 
 breaking. 
 
 In making transverse tests the author has used a form two 
 inches square and eight inches long. By placing the end sup- 
 ports five and one-third inches apart, the modulus of rupture 
 
92 CEMENT AND CONCRETE 
 
 by the formula / = , , 2 becomes equal to W, the center load 
 
 applied. 
 
 148. Finally, it may be said that there is little objection to 
 substituting transverse tests for tensile tests, although no evi- 
 dent advantage would be gained. It would also seem that 
 there is no object in making tests for quality by both trans- 
 verse and tensile tests, though from a scientific standpoint 
 comparative tests of transverse and tensile strength are of great 
 interest. 
 
 ART. 21. TESTS OF ADHESION AND ABRASION 
 
 149. ADHESION. The test for adhesion is also one of long 
 standing, being used during that time when engineers were con- 
 tent with an approximate idea of what might be expected of an 
 hydraulic product. It has been stated above that when failure 
 occurs in a mass of masonry, it is more frequently a failure in 
 tension than in compression; it may be added, that it is also 
 more likely to fail in adhesion than in cohesion. Hence, an 
 adhesive test is a very proper one to make, and will give most 
 valuable results. In fact, it is perhaps the most rational rup- 
 ture test, and were it not for the difficulties involved in its ap- 
 plication, it would doubtless come into general use. 
 
 150. One of the greatest difficulties experienced in making 
 adhesive tests is the preparation of the specimens of stone or 
 other material to which the mortar is to adhere. In early ex- 
 periments common brick were used, or pieces of stone were cut 
 to the same shape as brick, and two or more pieces cemented 
 together. In later methods the flat surfaces of two specimens 
 are sometimes joined with their axes at right angles, thus mak- 
 ing the cemented surface square. The upper brick being held 
 on two supports, a load is applied to the lower brick. 
 
 151. Mr. I. J. Mann, in a paper presented to the Institution 
 of Civil Engineers, 1 described a method of testing adhesion in 
 which are used test pieces 1^ inches long by 1 inch wide by J to 
 | inch thick. These are cemented together in a cruciform shape, 
 and a simple spring balance machine with properly arranged 
 levers pulls them apart. The upper block is supported at its 
 ends and an inverted U-shaped piece bears upon the ends of 
 
 Proc. Inst. C. E., Vol. Ixxi, p. 251. 
 
TRANSVERSE TESTS 93 
 
 the lower block. The stress is applied through a conical shaped 
 pivot bearing on the U-shaped saddle. Mr. Mann states that 
 test pieces may be made either of plate glass or close grained 
 limestone, the latter being sawn into pieces of the right size. 
 
 152. Another method is to make test pieces to fit one end 
 of the mold used for tensile tests, and after placing the piece of 
 stone in the mold, to fill the other end with the mortar to be 
 tested. The objection to this method is the expense of pre- 
 paring pieces of this form. It has been suggested to substitute 
 artificial stone for the cut stone samples. Thus, suppose it is 
 required to test the adhesion of a certain mortar to granite: 
 mold half briquets of a mixture of ground granite with cement, 
 and after these have well hardened, replace them in the mold 
 and fill the other end of the mold with the mortar to be used. 
 It is quite certain that the same result would not be obtained 
 in this way as though the specimens were cut from a piece of 
 solid granite. 
 
 153. One of the simplest methods of applying this test is 
 one which the author has used for some time. The test pieces 
 are in the form of flat plates one inch square and one-fourth 
 inch or less in thickness. These plates being placed in the 
 center of a briquet mold, the ends of the mold are filled with 
 mortar. The plates may be improved by cutting shallow 
 grooves in two opposite sides to make a more perfect fit with 
 the sides of the mold. This may easily be done with a round 
 file. Besides the simple form of the test pieces and consequent 
 ease of making them, this method has the further advantage 
 that a test may be made almost as readily and accurately as a 
 tensile test of cohesion. Also, since the adhesive area is one 
 square inch, the results may be compared with cohesive tests 
 on specimens having the same area of cross-section. 
 
 154. The experiments on adhesive strength made by Mr. 
 Mann were probably more extensive than any others published. 
 His results are useful mainly as showing the lack of cementitious 
 properties in the coarser grains of cement, and this point he 
 proves very clearly by quite a large number of experiments. 
 It was also developed that cement that had been rendered slow 
 setting by aeration or " cooling" gave a lower adhesive strength 
 than samples directly from the makers, which set more rapidly. 
 But the method followed by Mr. Mann, of immersing the speci- 
 
94 CEMENT AND CONCRETE 
 
 mens as soon as cemented together, may have had something to 
 do with this result; the quicker setting samples would earlier 
 resist the injurious action which is likely to follow the immer- 
 sion of such small quantities of mortar before they have set. 
 
 155. All of the things which influence the results in testing 
 the cohesive strength must also be considered as affecting the 
 adhesive test. The consistency of the mortar, the method of 
 gaging, the pressure applied in cementing the specimens, and 
 the conditions of storage until the time of breaking, will all 
 have an influence on the result obtained. In addition to these, 
 the character of the samples as to the kind of stone used, its 
 structure, the physical condition of the surfaces, etc., must all 
 be considered. It is therefore clear that many difficulties must 
 be met before the test for adhesion can ever be included in 
 standard tests. 
 
 156. Special tests directed toward ascertaining the compara- 
 tive adhesion of cement to different varieties of stone, the effect 
 of the various differences in manipulation, the comparative ad- 
 hesion of mortars containing various proportions of sand, etc., 
 are of undoubted value. But, before the adhesive test can be 
 considered a normal one for cement, much of this experimental 
 work will be required. 
 
 The results of a number of adhesive tests made under the 
 author's direction are given in Art. 51. 
 
 157. TESTS OF ABRASION. Abrasion tests of cement are 
 not at all common, and for the ordinary uses to which cement 
 is put, its resistance to such action is of little interest except as 
 it may imply other kinds of strength. Occasionally, however, 
 it may be desired to have a mortar which will withstand wear, 
 as, for instance, in making concrete walk. In such cases, tests 
 for resistance to abrasion have some interest and value. 
 
 The test is usually made on a sample prepared as for tensile 
 or compressive tests, by submitting it to the wearing action of 
 an emery or grindstone, or a cast iron disc covered with sand. 
 The number of revolutions of the stone or disc is recorded, 
 automatically if possible, and the loss of weight is determined 
 after a given number of revolutions. 
 
 A few tests of this kind made at St. Mary's Falls Canal are 
 given in Art. 58. 
 
CHAPTER IX 
 
 TENSILE TESTS OF COHESION 
 
 158. THE testing of cement by applying tensile stress to a 
 previously prepared briquet, containing definite proportions of 
 cement and water, or of cement, sand and water, is the strength 
 test which is now in most general use. The value of this method 
 in comparison with that of other forms of rupture tests has al- 
 ready been briefly discussed. 
 
 That cement fails oftener in tension than in compression is 
 one reason for preferring the tensile test. Its ready applica- 
 bility is a still more important point in its favor. 
 
 ART. 22. SAND FOR TESTS 
 
 159. Whether the tensile test should be applied to neat ce- 
 ment briquets or to those prepared from sand mortars has been 
 a disputed point, but there are now but few authorities who 
 recommend the use of the neat test exclusively. When tests 
 for soundness are not carefully made, the behavior of the cement 
 in neat briquets gives, perhaps, a better idea as to the reliability 
 of the cement than do sand tests, but otherwise the sand test is 
 a better index of the value of the cement. The principal ob- 
 jection to the sand test is that the use of sand introduces another 
 cause of variation in the results obtained by different experi- 
 menters. This objection has considerable weight, because it is 
 impracticable to find sand in widely separated localities which 
 is absolutely the same in composition and physical properties; 
 but two cements which appear to be of equal value when tested 
 neat may exhibit quite different characteristics when used with 
 sand, and it is believed that this fact far outweighs the objec- 
 tion noted. As soon as regularity in sieves is established, the 
 size of the sand grains may be regulated. The chemical and 
 physical properties of the sand and the shape of the grains is a 
 more difficult matter. The crushed quartz that is used in the 
 manufacture of sandpaper was recommended by the Committee 
 
 95 
 
96 
 
 CEMENT AND CONCRETE 
 
 of the American Society of Civil Engineers of 1885, and if some 
 care is taken to select that which is clean and made from pure 
 quartz, there is little difficulty in obtaining a uniform product 
 of this kind. 
 
 160. The German Normal Sand is obtained by washing and 
 drying a natural quartz sand. In various parts of Germany 
 sand answering the purpose may be found. Some tests made 
 in this country to compare the " normal" German sand with 
 American crushed quartz have shown the sand to give a some- 
 what higher strength, while other tests have shown an opposite 
 result. 1 A few of these tests are given in Table 27. 
 
 TABLE 27 
 
 Results Obtained -with German "Normal" and American 
 ard" Sand in Three Laboratories 
 
 Stand- 
 
 SAND. 
 
 AGE. 
 Days. 
 
 STRENGTH OF MORTAR, 
 1 CEMENT, 3 SAND, OBTAINED AT 
 LABORATORY NUMBER 
 
 3 
 
 4 
 
 5 
 
 Normal 
 
 7 
 7 
 28 
 28 
 
 218 
 253 
 317 
 334 
 
 173 
 219 
 341 
 300 
 
 201 
 211 
 281 
 283 
 
 Standard 
 
 Normal 
 
 Standard 
 
 
 PER CENT. OF WATER USED. 
 
 8 
 9 
 
 9 
 10 
 
 . . . 
 
 Standard 
 
 
 
 
 Mr. Max Gary has stated that "the Russian standard sand 
 gives markedly lower, and the Swiss sand considerably higher, 
 strength than the German." 
 
 161. Tests with Natural Sand. It is not to be concluded 
 from what has preceded that one must make mortar tests with 
 a "standard" sand only. On the contrary, one may obtain 
 valuable results by using in tests the sand which it is proposed 
 to use on the work. The only point to be insisted upon is that 
 a cement shall not be rejected on account of the poor quality 
 of the sand used in testing. It is thus very desirable that a 
 certain proportion of the tests be made with a pure quartz 
 sand, and by making parallel tests with the natural sands, the 
 
 Article by Clifford Richardson, Engineering Record, Aug. 4, 1894. 
 
MAKING BRIQUETS 97 
 
 coefficient of the latter may be obtained. In any case it is 
 necessary, in order to obtain comparable results, to sift the 
 sand used for tests. 
 
 162. Fineness of Sand for Tests. The American practice in 
 using crushed quartz is to reject the coarser particles by a sieve 
 having 20 meshes per linear inch (holes about .03 inch square) 
 and to reject the finer particles by a sieve of 30 meshes per linear 
 inch (holes about .02 inch square). The size of grain of German 
 normal sand is practically the same. In using a natural sand 
 it is not necessary to use this size of grain, but it is better to do 
 so, or at least to use some definite size or definite combination 
 of sizes; as, for instance, one-half of 20 to 30 (passing holes .03 
 inch square and not passing holes .02 inch square) and one-half 
 30 to 50 (passing holes .02 inch square and failing to pass holes 
 .012 inch square). Such a method will permit of duplicating a 
 given size of grain at any time, while if the sand is used as it 
 occurs in nature, considerable variations will be found. The 
 effect of the quality of sand on the strength obtained is dis- 
 cussed in Chapter XL 
 
 ART. 23. MAKING BRIQUETS 
 
 163. Proportions. The proportions of the ingredients should 
 always be determined by weight rather than by measure. It 
 will be found more convenient to use metric weights for the 
 dry ingredients. The water should then be measured in cubic 
 centimeters, which is equivalent to weighing it in grams. The 
 proportion of sand to be used for mortar briquets will depend 
 upon circumstances, but for short time (seven day) tests good 
 results are not usually obtained with natural cement if more 
 than two parts of sand by weight are added to one part of ce- 
 ment. Portland cement may be tested at seven days with 
 three parts of sand to one of cement. If too large a proportion 
 of sand is used, the briquets are liable to be injured in handling, 
 and very low strengths are not as accurately recorded by the 
 testing machine. 
 
 164. CONSISTENCY: DETERMINATION. The consistency of 
 the mortar has such a marked influence on the strength obtained 
 that its importance can hardly be overestimated. The difficul- 
 ties attendant upon specifying the consistency of a given mortar 
 have already been touched upon in 116. The Committee of 
 
98 CEMENT AND CONCRETE 
 
 the American Society of Civil Engineers of 1885 recommended 
 the use of a " stiff plastic" mortar, but this phrase has had va- 
 rious interpretations. 
 
 The present Committee in its progress report l recommended 
 the use of the Vicat apparatus: "In making the determination, 
 500 gr. (17.64 oz.) of cement are kneaded into a paste, and 
 quickly formed into a ball with the hands, completing the oper- 
 ation by tossing it six times from one hand to the other, main- 
 tained six inches apart; the ball is then pressed into the rubber 
 ring ( 98) through the larger opening, smoothed off, and placed 
 (on its large end) on a glass plate, and the smaller end smoothed 
 off with a trowel; the paste, confined in the ring resting on the 
 plate, is placed under the rod bearing the cylinder, which is 
 brought in contact with the surface and quickly released. The 
 paste is of normal consistency when the cylinder (1 cm. in di- 
 ameter and loaded to weigh 300 grams) penetrates to a point in 
 the mass 10 mm. (0.39 in.) below the top of the ring. Great 
 care must be taken to fill the ring exactly to the top." 
 
 The following simple test taken from French specifications 
 will determine a good consistency of mortar to 'use for briquets. 
 It should be capable of being easily molded into a ball in the 
 hands, and when dropped from a height of one and a half feet 
 on a hard slab, this ball should retain its rounded form without 
 cracking. The mortar should also leave the trowel clean when 
 allowed to drop from it. Were a smaller quantity of water 
 used, the mortar would be crumbly and the ball would crack 
 when dropped on the slab, while a larger amount of water would 
 cause the mortar to adhere to the trowel and the ball would be 
 flattened by striking the slab. 
 
 165. Another method of determining the proper consistency, 
 which the author believes will prove very satisfactory, is to 
 make several batches of mortar containing the same weights of 
 cement and sand, but having different percentages of water. 
 As each batch is mixed, the volume of the resulting mortar is 
 measured by pressing it lightly into a metal cylinder (a small 
 tin pail will answer the purpose), taking pains to fill the cylinder 
 in the same manner each time. That batch of mortar which 
 
 1 Proc. Amer. Soc. C. E., Jan. 1903; also Engineering News, Jan. 29, 
 1903, and Engineering Record, Jan. 31, 1903. 
 
CONSISTENCY OF MORTAR 
 
 99 
 
 occupies the least volume, when thus lightly packed, is the one 
 in which the amount of water used is most nearly correct. 
 Should either the mortar which contained the least water or 
 that which contained the most water chance to have the least 
 measured volume, then more trials must be made until such a 
 consistency is obtained that either more or less water will in- 
 crease the bulk of the mortar. This method will give a con- 
 sistency somewhat more moist than that which gives the highest 
 results on short time cohesive tests, but it is believed that where 
 briquets are made by hand, more uniform results will be ob- 
 tained when the mortar is a trifle moist. This method is not 
 suited to daily use, as it requires too much time, but is valuable 
 as a check on one's ideas of proper consistency. 
 
 166. EFFECT OF CONSISTENCY ON TENSILE STRENGTH. 
 Tables 28 and 29 give a few of the results obtained by the author 
 
 TABLE 28 
 
 Variations in Consistency of Mortar. Effect oil Tensile Strength, 
 Neat Natural Cement 
 
 
 
 TENSILE STRENGTH, POUNDS PER SQUARE I.\<-H. 
 
 CKMKNT. 
 
 
 WATER USED Kx PRESSED AS 
 
 
 AGE OF 
 
 PER CENT. OF DRY JNUKEDIENTS MY WEHJHT. 
 
 
 BRIQUKTB. 
 
 
 Brand. 
 
 Sample. 
 
 
 25% 
 
 30% 
 
 3T>% 
 
 40% 
 
 4f,% 
 
 Gil 
 
 83 R 
 
 7 days 
 
 1136 
 
 205d 
 
 122/ 
 
 7*0 
 
 6U 
 
 Gil 
 
 84 R 
 
 
 926 
 
 7'2d 
 
 f>8/ 
 
 540 
 
 39A 
 
 All 
 
 G 
 
 
 162c 
 
 165e 
 
 108/ 
 
 log 
 
 54A 
 
 An 
 
 N 
 
 
 1526 
 
 194c 
 
 204e 
 
 134/ 
 
 7% 
 
 Hn 
 
 26 S 
 
 
 226d 
 
 176/ 
 
 900 
 
 56A 
 
 35i 
 
 Gn 
 
 83 R 
 
 28 days 
 
 181)6 
 
 244d 
 
 211/ 
 
 1820 
 
 136ft 
 
 Gn 
 
 84 R 
 
 
 1406 
 
 168<I 
 
 114/ 
 
 107(7 
 
 108/i 
 
 An 
 
 G 
 
 
 210c 
 
 228e 
 
 165/ 
 
 1020 
 
 80A 
 
 An 
 
 N 
 
 
 1736 
 
 28(5c 
 
 254e 
 
 208/ 
 
 150^ 
 
 Hn 
 
 26 S 
 
 
 333d 
 
 309/ 
 
 2170 
 
 121A 
 
 89i 
 
 Ln 
 
 31 S 
 
 1 day 
 
 162c 
 
 148e 
 
 97/ 
 
 630 
 
 S6A 
 
 Ln 
 
 u 
 
 7 days 
 
 178c 
 
 177e 
 
 124/ 
 
 710 
 
 45/t 
 
 Ln 
 
 it 
 
 28 days 
 
 207c 
 
 257e 
 
 202/ 
 
 1400 
 
 88A 
 
 Ln 
 
 n 
 
 3 mos. 
 
 SOOc 
 
 389e 
 
 888/ 
 
 2('Ag 
 
 197/1 
 
 SIGNIFICANCE OF LETTERS 
 
 a barely damp. 
 
 6 very dry ; no moisture shown on surface briquets. 
 
 c dry ; slight moisture shown on surface briquets. 
 
 d trifle dry. 
 
 e about right consistency. 
 
 / trifle moist. 
 
 g moist. 
 
 h very moist ; would just hold shape. 
 
 i extremely moist; would not hold shape. 
 
100 CEMENT AND CONCRETE 
 
 in tests to determine the effect of consistency on the tensile 
 strength of natural cement mortars. All of the briquets were 
 made in the usual manner and stored in fresh water until time 
 of breaking.. Each result given is the mean of from two to ten 
 briquets. The letters affixed to each result indicate the degree 
 of moisture which the mortar appeared to have when mixed, 
 varying from "a," barely damp, to "i," so wet that the mortar 
 could not hold its shape when laid on a glass slab. 
 
 The results in Table 28 were obtained with neat cement mor- 
 tars of several brands of natural cement. The first point to be 
 noted is the variation in the amount of water required by dif- 
 ferent samples to give the same consistency; thus, Brand An, 
 sample N, when mixed with 35 per cent, water, appeared to 
 have about the same consistency as did sample G of the same 
 brand mixed with 30 per cent. It is also apparent that the 
 strength of all samples is not affected alike by given variations 
 in the amount of water used in mixing; comparing the results 
 obtained when 45 per cent, water is used with that given when 
 25 per cent, water is used, it is seen that at seven days the wet 
 mortar gives 42 per cent, of the strength obtained with dry mor- 
 tar for sample 84 R, Brand Gn, while with the sample of Brand 
 Hn the strength of the wet mortar briquet is but 16 per cent. 
 of that given by the dry mortar. Of the six samples tested at 
 seven days and twenty-eight days, three gave the highest 
 strength at seven days when mixed with 25 per cent, water, 
 and five gave the highest strength at twenty-eight days when 
 30 per cent, water was used. The results on Brand Ln show 
 the greater proportionate gain with age of the wet briquets. 
 
 Table 29 shows similar results for mortars made with one, 
 two and three parts sand. With one part sand the wet mortar 
 made from Gn, 21 R, which gave but 22 pounds per square 
 inch at seven days, gave 429 pounds, or nearly the highest 
 strength, at six months. A similar result is shown for sample 
 15 R of the same brand when mixed with two parts sand, the 
 highest strength at one year and two years being given by the 
 mortar containing the greatest per cent, of water. That mor- 
 tars containing three parts sand to one cement may be more 
 easily damaged by an excess of water, is indicated by the re- 
 sults on Brand Ln in this table. 
 
 167. The effect on the strength of Portland cement mortars, 
 
CONSISTENCY OF MQXTAR 
 
 
 
 3 
 
 A 'fi "c^ '(<> 
 
 C5 OS i-H lO 
 !N l^ C<1 Tt< 
 <M CN 
 
 
 
 ^- 
 
 S 
 
 (M ^ 00 OS 
 <N*.3 
 
 
 
 |S 
 
 & 
 >*> 
 
 S 
 
 -<-<-<-< 
 QO >O (M GO 
 ^C M <M t^- 
 
 r-i CO CO 
 
 
 / 
 
 00 
 
 V 
 
 *>. 
 
 CO O 00 C 
 
 cc t^ i-- 
 
 on 
 
 ~~! 
 X 
 
 1 
 
 fc 
 
 ' gl 
 
 -sS^-sj-sS 
 
 ^siS 
 
 <M 
 
 <D 
 
 1 
 
 r. 
 
 >, 
 
 c 
 
 O.j 
 
 is 
 
 ^^Srcc- ^^S?? c? 5 ^^^ 3 
 
 (M CO 00 O O'CCO'-i i-iOO-* 
 1-1 <M CO <* C^ CO CO .-H I-H 
 
 H 
 
 ?, 
 , 
 
 1 
 
 00 
 
 2Z&-3 g>2$g o1;S 
 
 I- I- 1 - I QC >.^ O O O O <M 00 
 
 T-H CC ^ CO CO CO i-H CO CO C^ 
 
 ^ 
 
 
 
 1 
 
 3 
 
 o 
 
 CO 
 
 &5 &S ^5 ^5 **^.**^~,**->. 
 
 Ci co o <N -^:o oo 
 
 CO -> Ci ?O S^l 00 1- I-H 
 
 I-H CO (^ <M r^ Cq 
 
 A 
 
 STRK: 
 
 1 
 
 3 
 
 *IS 
 
 ^^SJ^J ^^^)^J ^^^^ 
 I-H O (7^ t^* ^^ CO t^ OS CO 1^" CO 1^- 
 >O<N'MCO (M^O^OOS COOOCOCO 
 i-i <N CO CO i-HCOCOCO i-l<M<?}<M 
 
 I 
 ' 
 
 o 
 
 ENSILE 
 
 1 
 
 ? 
 
 S* 
 3" 
 
 FO O '-O O *"^ ""^ "^ "^ ^J "^ '"^5 ^^ 
 00(MO^ 'sftDl^-'N COO'NO 
 So-^CO Oi-H^COOoScO'N 
 ^H(M(N i-iCOTOCO'-''N^l(M 
 
 | 
 
 2 
 
 H 
 
 i 
 
 \-> 
 
 I 
 
 ^ 
 
 c-i 
 
 u<juy v%>%>^ 
 ^^OO (M-t-^O 
 OOT^CiCO *<* 
 
 1-H ^ r-i rH l-l (N (^ 
 
 S 
 
 M 
 *-. 
 
 
 a 
 
 
 
 2-t 
 
 S^ 
 
 ^rH^H^H r-(i?5(3qc<l 
 
 1 
 
 
 
 LO 
 
 06 
 
 <M 00 CO O 
 CO O ^ CO 
 
 O 
 
 tC 
 
 a 
 
 
 AGE OF 
 
 BRIQUETS. 
 
 H= r i-- N^i N^ : i s 
 
 T3 a ^ H r3 a >>> T3 S >>>> ^3 a 
 
 t-OOCOCO t-OOCOCO COCOr-l(N QO 50 I-H <N t-OOCOCO 
 C^ (M (M 3^1 <M 
 
 (B 
 
 1 
 1 
 
 
 EH 
 
 
 ft 
 ? 
 
 *s 
 
 ce w 
 
 S 8 ' 
 
 OH 
 
 S B 
 
 4tHo 
 
 3"g 
 
 ^ ^ 
 
 r^ ^H r-i ,_| r-i rH ^-t r-( <N (M (N (N (N -M <N C^ CO CO CO CO 
 
 
 
 ^ 
 
 
 
 
 P5 c p* p* 50 
 
 
 H 
 
 a 
 
 
 j^diuwg 
 
 ?T S S S 
 
 
 
 
 PS 
 
 
 
 a^ 2 2 ^^^^ G - - z x^ - ~ -.. 
 
 
 
 
 
 O >-i O O i-3 " 
 
 
AND CONCRETE 
 
 of variations in consistency, has been investigated by Mr. Eliot 
 C. Clarke/ M. Am. Soc. C. E., and by M. Paul Alexandre, 2 Chief 
 Engineer, Fonts et Chaussees. The results of one series of ex- 
 periments made by M. Alexandre are given in Table 30. The 
 mortars were mixed with fresh water and the samples immersed 
 in sea water. 
 
 TABLE 30 
 
 Variations in Consistency of Mortar 
 
 EFFECT ON TENSILE STRENGTH, PORTLAND CEMENT MORTAR. 
 
 25 pounds cement to 1 cu. ft. sand (about 1 to 4 by weight). 
 
 
 
 RESISTANCE, LBS. PER SQ. IN. AT AGE OF 
 
 
 WATER 
 
 
 
 
 
 
 
 
 
 CONSISTENCY. 
 
 PER 
 CENT. OF 
 
 
 
 >> 
 
 1 
 
 
 y 
 
 i 
 
 E 
 
 
 
 
 SAXD. 
 
 & 
 
 % 
 
 
 
 
 9 
 
 0) 
 
 OJ 
 
 i 
 
 
 
 fl 
 
 ft 
 
 
 3 
 
 !* 
 
 * 
 
 h 
 
 r* 
 
 
 
 eo 
 
 t- 
 
 8 
 
 CO 
 
 n 
 
 <N 
 
 CO 
 
 **i 
 
 Dry . . 
 
 14 
 
 31 
 
 56 
 
 73 
 
 77 
 
 69 
 
 67 
 
 88 
 
 Disintegrated 
 
 Ordinary . 
 
 22 
 
 25 
 
 46 
 
 74 
 
 116 
 
 153 
 
 170 
 
 162 
 
 190 
 
 Wet . . 
 
 30 
 
 16 
 
 35 
 
 55 
 
 89 
 
 126 
 
 136 
 
 180 
 
 189 
 
 From "Recherches Experimentales sur Les Mortiers Hydrauliques," 
 par M. Paul Alexandre, Annales des Fonts et Chausse'es, Sept., 1890 
 
 It is seen that the highest strength at three days and seven 
 days is given by the dryest mortar, at twenty-eight days to two 
 years by that of the ordinary consistency, and at three years by 
 that containing the highest per cent, of water. All of the sam- 
 ples exhibited white spots in the broken section at three years, 
 and at four years the dry mortar briquets had lost their cohe- 
 rence on account of their porosity permitting the sea-water to 
 permeate them. 
 
 168. Conclusions. It may be concluded, then, that the 
 consistency of the mortar has a very marked effect on the ten- 
 sile strength obtained; that different samples of cement are not 
 affected in the same degree by given variations in consistency ; 
 that the effect of consistency is usually shown most plainly in 
 short time tests; and that while the dryer or stiffer mortars give 
 the highest results on short time tests, the moist mortars attain 
 a greater strength after a certain time. 
 
 1 "Records of Tests of Cement made for Boston Main Drainage Works.' 
 Trans. A. S. C. E., Vol. xiv. 
 
 2 Annales des Fonts et Chaussees, Sept., 1890. 
 
TEMPERATURE 103 
 
 169. Temperature of the Ingredients and of the Air where the 
 Briquets are Made. The temperature of the mortar and of 
 the air in which the briquets are prepared is a matter of some 
 moment. In 1877, Mr. Maclay l reported a series of experi- 
 ments on Portland cements from which conclusions may be 
 drawn concerning the effects of the temperature of the mortar. 
 These experiments indicate that mortar having a temperature 
 of 40 Fahr. when gaged, will attain greater strength in from 
 seven days to three weeks than a mortar having an initial 
 temperature of 70 Fahr. One is most likely to work some- 
 where between these two temperatures, but it may be mentioned 
 that according to Mr. Maclay's experiments, it appears that 
 mortars gaged at a temperature of 90 or 100 Fahr. also at- 
 tain a higher strength than those gaged at 70 Fahr. 
 
 Similar experiments made by M. Candlot 2 indicate that mor- 
 tars gaged with cold water give but feeble resistance at first, 
 but in from two weeks to one month, such mortars surpass in 
 strength those gaged with warm water. M. P. Alexandre 3 im- 
 mersed some briquets at a temperature of about 90 C. (194 
 Fahr.) for forty-eight hours and then at 15 to 18 C. (60 to 
 65 Fahr.) until broken, while other briquets were maintained 
 at the latter temperature from the time of molding. The bri- 
 quets that were broken at the age of four days showed that 
 the highest strength had been obtained by the briquets which 
 had been kept hot for forty-eight hours, but at twenty-eight 
 days and three months those briquets which had not been sub- 
 jected to this high temperature gave the highest strength. 
 
 170. Table 31 gives a few of the many experiments on this 
 point made under the author's direction. It appears that the 
 briquets made in a low temperature (34 to 37 Fahr.) are usu- 
 ally stronger than those made in the ordinary temperature of 
 65 to 68 Fahr. In some cases the difference was not very 
 great, and in some of the tests the briquets made in the ordi- 
 nary temperature gave higher results at one day and seven 
 days than those made in the cold; but at twenty-eight days the 
 cold-made briquets were nearly always in the lead, and in many 
 
 1 "Notes and Experiments on the Use and Testing of Portland Cement," 
 Trans. A. S. C. E., Vol. vi, p. 311. 
 
 2 "Ciments et Chaux Hydrauliques" 
 
 3 "Les Mortiers Hydrauliques." 
 
104 
 
 CEMENT AND CONCRETE 
 
 Temperature of Materials and of Air where Made. Effect on Tensile Strength, Natural 
 and Portland Cements 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH. 
 
 a* 
 
 PQ 
 
 1 
 
 0" 
 
 PQ 
 
 o 
 > 
 
 1 
 
 
 pto 
 
 $; 
 
 NOTES: All briquets made by same molder; each result is mean of five to ten specimens. 
 Results in columns headed "warm," temperature of materials used and air where made = 65 to 67 Fahr. 
 Results in columns headed "cold," temperature of materials used and air where made = 34 to 37 Fahr. 
 a. In damp closet 36 hours, except 1 day specimens which were 12 hours in air where made and 12 hours 
 in damp closet. 
 b. In damp closet 24 hours, except 1 day specimens which were 24 hours in air where made. 
 c. In damp closet 19 to 21 hours, except 1 day specimens which were 3 to 5 hours in air where made, 
 to 1\ hours in damp closet and 16^ to 21 hours in tank. 
 
 . . .<N . . 
 
 A 
 
 t~ 
 
 . . .<N . . 
 
 6 Months. 
 
 WO 
 
 oo ;N ca 
 
 CO rH . (M 
 TP CO . S> 
 
 A 
 
 ^53 -^ 
 
 3 Months. 
 
 'PIO 
 
 CO CO Ci O rH CM 
 
 CO * >* ^H "* rH 
 
 CO CN CO <M t^ 1C 
 
 m** 
 
 CO b- t- O 
 
 TH 00 ^ O O CO 
 
 28 Days. 
 
 P,o 
 
 T* CO T* O O rH 
 
 Ut> O rH rH Tfl OO 
 
 HUB AY 
 
 O CO 10 C5 <?} 
 rH CO CO rH Th 
 CM S<l rH t^ CO 
 
 DD 
 
 Q 
 
 t- 
 
 WOO 
 
 O rH 1 ' CO rH 
 <M O5 t- t^ O 
 
 (N rH . CO (M 
 
 HUT? AY 
 
 1-- QO J-- * 1C rH 
 
 rH rH 1 JO ^ 
 
 f 
 
 CO 
 
 IPO 
 
 : is : 
 
 ^M 
 
 : :1 : 
 
 1 
 
 pio 
 
 S CO CO 
 
 uu 
 
 <N CO rH 
 CD 1 O 
 
 cc 
 M 
 
 
 
 W 
 
 ;=SB& 
 
 CO (M 1 rH 
 Oi 
 
 M^KY 
 
 S^32SS 
 
 AHXI 'X.NDrQ 
 
 aaj 'aasn naxvAY 
 
 T^^t-(NOO 
 
 rH rH QO CO O <M 
 CO CO rH rH <M rH 
 
 3*1:0 ox OS-OS zxHvnft 
 
 O O < i CO O CO 
 
 H 
 
 fc 
 
 1 
 
 6 
 
 02 p^ 03 
 
 CO -j ,_, - rH - 
 1 1 
 
MORTAR MIXING 105 
 
 cases this difference held good at three months and six months. 
 Some of the results indicated that if the briquets were allowed 
 to remain twenty-four hours or more in the cold air, it tended 
 to counteract the beneficial effects of cold molding, but this 
 point was not satisfactorily established. 
 
 171. From the foregoing the following conclusions may be 
 drawn: To make briquets of cold materials and allow them 
 to remain some hours in cold air, retards the hardening of the 
 briquets; but when briquets so treated are, after a few hours, 
 placed in a medium of ordinary temperature, they gain strength 
 more rapidly than briquets made of warm materials and kept 
 continuously at the ordinary temperature of 60 to 70 Fahr. 
 After being placed in a warmer medium, the briquets made 
 with cold materials in cold air frequently gain strength at such 
 a rate as to surpass in strength the warm-made briquets at seven 
 days; the former almost invariably surpass the latter at twenty- 
 
 ^ight days. In some cases it appears that this superiority of 
 cold-made briquets is maintained up to six months, but in other 
 cases the difference seems to disappear after three months. 
 
 Although these variations in temperature have not as marked 
 an effect on tensile strength as have many other variations in 
 manipulation, yet in carefully conducted experiments one should 
 always operate in a constant temperature. As a matter of 
 convenience, 65 to 70 Fahr. will commend itself, and this 
 temperature may well be taken. 
 
 172. GAGING BY HAND. The objects to be attained in 
 gaging are to thoroughly incorporate the cement and sand, to 
 evenly distribute the water throughout the mass, and, if pos- 
 sible, to give the mortar a certain tenacity resembling that of 
 putty. This last object is not always possible of attainment 
 with mortars containing a large dose of sand. 
 
 The ordinary method of preparing mortars in the laboratory 
 is to gage with a trowel on a glass, slate, or marble slab. In 
 gaging mortars, the cement and sand are first mixed dry; the 
 materials are then drawn away from the center, leaving a crater 
 to receive the water, which is all added at one time. The dry 
 material is then gradually turned from the edges toward the 
 center until all of the water is absorbed, after which the mass 
 is thoroughly worked with the trowel in such a way as to rub 
 the material between the trowel and plate until the consistency 
 
106 CEMENT AND CONCRETE 
 
 is uniform throughout. A batch of mortar sufficient for five 
 briquets cannot usually be properly gaged by this method in 
 less than five minutes. 
 
 The Committee of the American Society of Civil Engineers, 
 in their preliminary report on methods of manipulation, sug- 
 gested that "as soon as the water has been absorbed, which 
 should not require more than one minute," the mortar should 
 be kneaded with the hands for one and one-half minutes, the 
 process being similar to that used in kneading dough. 
 
 173. HOE AND BOX METHOD. Mr. Alfred Noble used for 
 many years a form of gaging apparatus, consisting of a box 
 with sloping bottom, in which the mortar is worked by means 
 of a hoe. The author has used an iron box made on this prin- 
 ciple (Fig. 2), which has given excellent results. The box is 
 2 feet 7^ inches long, 6 inches wide at the bottom, and at the 
 
 "^v> 
 
 \ -j .-' f 
 
 L 
 
 '- ?' 8* - *! 
 
 Side Elevat/on End- 
 
 FIG. 2. MIXING BOX 
 
 center is 6 inches deep. The level part of the bottom is 3 inches 
 by 6 inches, and from this level part the inclined portions of 
 the bottom slope up toward the ends at an inclination of about 
 22J degrees. The sides of the box extend below these in- 
 clined planes to give a level bearing for the box when in use. 
 It is also well to have the sides flare enough to give a width 
 of 6^ inches at the top to prevent the hoe from becoming 
 wedged. A " German clod hoe," which is strong and heavy, 
 yet a trifle flexible in the blade, is used in connection with the 
 box. 
 
 The weighed quantities of the dry ingredients being put in 
 the box and well mixed, the measured volume of water is added. 
 Two minutes of hard work, in which the operator may put all 
 his strength, is sufficient to bring the mass to plasticity if the 
 amount of water added is correct. A return to the trowel and 
 
/ MORTAR MIXING 107 
 
 slab method of mixing is not likely after a trial of this simple 
 device. 
 
 174. MACHINE FOR MORTAR MIXING. As the mixing by 
 
 hand is a rather slow and tedious method, and the hoe and box 
 method are not very generally known, several machines have 
 been devised to do the work. None of them, however, has 
 given such satisfactory results as to bring it into general use. 
 
 One of the machines is called a "jig," or "milk shake" 
 machine, 1 and consists of a cup which moves rapidly up and 
 down, this motion being imparted by means of a hand wheel, 
 crank and connecting rod. The dry cement and water being 
 placed in the cup and tightly covered, a few rapid turns of the 
 wheel are sufficient to reduce the cement to a paste. This 
 form is only applicable to neat cement mortars, and has been 
 said to give unsatisfactory results even for these, though in 
 some laboratories this machine has been used for all neat 
 mortars. 
 
 Other forms have been made in which the mortar is thoroughly 
 stirred by means of forks or blades projecting into the mortar 
 from a horizontal arm above. The gager devised by Mr. Faija 
 is constructed on this principle, and similar machines may be 
 obtained from manufacturers of testing apparatus. 
 
 175. Steinbriich's Mortar Mixer is a German machine oper- 
 ating on a different principle. It consists of a circular shell 
 having on its upper side and near its outer edge a circular groove, 
 or trough, to receive the mortar to be mixed. In this trough 
 rests a wheel on a fixed horizontal axis, which is above the pan 
 and normal to the axis of the pan. A cross-section of the rim 
 of the wheel is a semicircle fitting the groove in the pan. The 
 gearing is such that the pan is made to revolve about its vertical 
 axis, and the wheel about its horizontal axis, the inner surface 
 of the trough and the under side of the periphery of the wheel 
 where the two are in contact moving in the same direction at 
 a given instant. The mortar is thus rubbed between the two. 
 Small blades, or plows, scrape the sides of the trough as the latter 
 revolves, thus keeping the mortar in the bottom of the trough. 
 The wheel and the plows are mounted on hinged axes, or sup- 
 ports, so that they may be raised from the pan when the mortar 
 
 1 S. Bent Russell, Engineering News, Jan. 3, 1891. 
 
108 
 
 CEMENT AND CONCRETE 
 
 is to be cleaned out. The mixing requires about two and one- 
 half minutes. The price of the machine is about $130. 
 
 176. The amount of gaging which a mortar receives has an 
 important effect on its consistency and the strength it will 
 attain. This was found to be the case in several experiments 
 where mortar gaged eight minutes in the box described above, 
 gave from 15 to 35 per cent, greater strength at one year than 
 that which was gaged but two minutes, the amount of water 
 used being the same in the two cases. Experiments on this 
 point are given in Table 78, 364. It is therefore important to 
 eliminate, if possible, the variations which must follow hand 
 mixing, but as yet no apparatus has seemed to meet with gen- 
 
 FlG. 3. FORM OF BRIQUET 
 USED ON THE CONTI- 
 NENT OF EUROPE. 
 
 FIG. 4. FORM OF BRIQUET 
 USED IN THE UNITED 
 STATES. 
 
 eral approval, though among machine mixers those similar to 
 that used by Mr. Faija seem to have given the best results. 
 The hoe and box method described in 173 partially eliminates 
 the personal equation, and for facility of operation and thor- 
 oughness of mixing leaves little to be desired. 
 
 177. FORM OF BRIQUET. The shape and size of the briquet 
 have been the subject of much discussion and experiment. 
 Mr. John Grant, a pioneer in tensile tests, tried many forms 
 before finally adopting one quite similar to the form afterward 
 recommended by the Committee of the Amer. Soc. C. E. in 
 1885. Mr. Alfred Noble also made a series of experiments on 
 
FORM OF BRIQUET 109 
 
 different styles of molds and clips, and presented the results 
 in a paper read before the American Society of Civil Engineers. 1 
 
 There are two forms of mold that are now in quite general 
 use. On the continent of Europe the form most generally used 
 is that shown in Fig. 3. It has a cross-sectional area of five 
 square centimeters (.775 sq. in.) at the smallest place, and the 
 heads of the briquet are elliptical in form, the major axes being 
 transverse to the briquet axis. The curve forming the side of 
 the briquet in the central portion is of very short radius, giving 
 the effect of a semicircular notch on either side of the briquet 
 at the smallest section. These notches have the effect of con- 
 fining the break to this place. 
 
 The other form of mold is the one mentioned above as recom- 
 mended by the Amer. Soc. C. E. Committee, and used in America 
 and England. A briquet of this form is shown in Fig. 4. The 
 cross-sectional area at the center is one square inch, and the 
 increase of section toward the ends is gradual, the radius of the 
 curve at the side of the briquet being J inch. 
 
 178. Area of the Breaking Section. Formerly a section of 
 2} square inches was more commonly used here and in England, 
 while an area of 16 square centimeters (2.48 sq. in.) was com- 
 mon in France and other continental countries. The larger 
 the area of the breaking section, the smaller will be the com- 
 puted strength per square inch; this point seems fairly well 
 established, although the experiments recorded in a very ex- 
 cellent paper by Mr. Eliot C. Clarke 2 indicate no apparent dif- 
 ference in strength between briquets 1 square inch and 2J 
 square inches in section. 
 
 M. Durand-Claye found that the tensile strength of a briquet 
 varied more nearly as the perimeter than as the area of the 
 section. The experiments of M. Candlot do not point to this 
 conclusion, though they clearly show that the indicated strength 
 per square centimeter is very much greater for a briquet hav- 
 ing an area of five square centimeters at the small section than 
 for a briquet of 16 square centimeters area. 
 
 Mr. D. J. Whittemore 8 experimented with briquets that were 
 
 1 Trans. Amer. Soc. C. E., Vol. ix, p. 186. 
 
 2 Trans. Amer. Soc. C. E., Vol. xiv, p. 141. 
 
 3 "Tensile Tests of Cements," etc. Trans. A. S. C. E., Vol. ix, p. 329. 
 
110 CEMENT AND CONCRETE 
 
 circular in cross-section. He found that while the ultimate 
 strength of a briquet was about proportional to the periphery 
 of the breaking section for the ordinary solid briquet, yet if a 
 core were inserted in the mold, giving the cross-section an annu- 
 lar form, this proportion was not maintained. It was con- 
 cluded from this that the apparent peripheric strength could 
 not be explained by saying that the surface of the briquet had 
 gained a greater strength than the interior, but that the expla- 
 nation must rather be sought in the method of applying the 
 stress in breaking the briquet. The force being communicated 
 to the surface of the briquet, the stress is not uniformly dis- 
 tributed throughout the breaking section, because of the low 
 elasticity of the mortar. 
 
 M. Paul Alexandre showed that the difference in strength 
 per unit area decreased with age, although it did not entirely 
 disappear at one year. It would therefore seem that the expla- 
 nation of this phenomenon may be found in a combination of 
 these two causes; more rapid hardening of the smaller speci- 
 mens, and greater inequality of stress, in breaking the briquets 
 of larger section. 
 
 179. Form of Briquet Suggested. As a result of experi- 
 ments which will be described under the head of " Clips," 1 
 (Art. 25) the following conclusions were drawn as to the desir- 
 able features for a briquet: 
 
 1st. The smallest section should not have an area much less 
 than one square inch. Probably an area of five square centi- 
 meters would represent a minimum. 
 
 2d. The area of the section of the briquet between opposite 
 gripping points should be about one and three-fourths times 
 the area of the smallest section. 
 
 3d. The distribution of stress over the smallest section 
 should be as nearly uniform as possible. 
 
 4th. The curve of the sides at the breaking section should 
 not be very sharp; one-half inch might be taken as a minimum 
 radius. 
 
 5th. The area of the vertical section from the gripping point 
 to the plane of the end of the briquet the section subjected 
 
 1 These experiments were described by the writer in detail in "Municipal 
 Engineering," Dec., 1896, Jan. and Feb. 1897. 
 
FORM OF BRIQUET 
 
 111 
 
 to shear when .the stress is applied should be nearly as great 
 as the area of the neck of the briquet. 
 
 6th. The face and back of the briquet should be parallel 
 planes, to permit of easy storage. 
 
 7th. The total volume should be kept as small as is consis- 
 tent with the other conditions. 
 
 Fig. 5 represents a form of briquet which will, it is thought, 
 satisfactorily fulfill the above requirements, and in which it is 
 believed the full strength of the smallest section may be more 
 nearly developed than with present forms. The curve at the 
 central section has a radius of one inch, and the line of the 
 side of the briquet is con- 
 tinued in a tangent one- ( < 
 
 half inch in length, having 
 an inclination of nearly 
 45 degrees with the axis 
 of the briquet. The total 
 length of the briquet is 
 four inches, the ends be- 
 ing formed by straight 
 lines tangent to the curves 
 forming the corners. If 
 the clip is so formed that 
 the gripping points bear 
 at the centers of the one- 
 half inch tangents form- 
 ing the sides of the briquet, 
 the distance between op- 
 posite gripping points will 
 be 1-J inches. 
 
 180. Comparison with 
 other Forms. Compar- 
 ing this briquet with the forms in common use, the German and 
 the form shown in Fig. 5 both have an area between opposite 
 gripping points about If times the area of the smallest section, 
 but in the form shown in Fig. 4 this ratio is too small to fulfill 
 the second specification. 
 
 The unequal distribution of stress over the breaking section 
 of the briquet has already been mentioned as a probable partial 
 cause why briquets of small cross-section show a greater strength 
 
 FIG. 5. FORM OF BRIQUET SUGGESTED 
 FOR USE 
 
112 CEMENT AND CONCRETE 
 
 per unit area than those having a larger area of cross-section. 
 In Johnson's " Materials of Engineering" is given the theory of 
 the distribution of stress over the breaking section of a briquet, 
 as developed by M. Durand-Claye, and published in Annales des 
 Pouts et Chaussses of June, 1895. Applying the formulas there 
 given to three styles of briquet, the A. S. C. E. form of 1885, 
 the German standard, and the form shown in Fig. 5, it is found 
 that the ratios of the maximum stress to the mean stress are, 
 for the three forms respectively, 1.54, 1.52 and 1.22. From a 
 theoretical point of view, this means that with a total pull of 
 100 pounds on each briquet, the outer fiber of the briquet 
 shown in Fig. 4 would be subjected to a stress of 154 pounds 
 per square inch, while with the form suggested above, the 
 stress on the outer fiber would be but 122 pounds per square 
 inch ; briquets of the latter form should, therefore, theoretically, 
 show a breaking strength 1.27 times the strength given by 
 briquets of the same mortar made in the A. S. C. E. form of 
 1885. 
 
 The German form has too sharp a curve at the sides to fulfill 
 the fourth requirement given above. All of the forms comply 
 with the first, fifth and sixth requirements. 
 
 As to the volume of the briquet, the author's form having a 
 total length of four inches, has about 50 per cent, greater volume 
 than the A. S. C. E. form of 1885. 
 
 181. MOLDS. In the early tests of cement, wooden molds 
 were employed, but they absorb water from the mortar and 
 soon warp out of shape. Iron molds have also been used to a 
 considerable extent, but these are apt to become rusted if not 
 in constant use. Brass, bronze or some similar metal not easily 
 corroded should be used, and molds of this character can be 
 obtained of dealers in testing apparatus. 
 
 The molds may be made single, or in " nests" or "gangs" 
 of three to five. The two halves of the mold may be entirely 
 separable, or may be hinged at one end and fastened by a clip 
 at the other end. The gang molds are somewhat cheaper than 
 the single ones. The hinged molds and those held with patent 
 clip are rather difficult to clean, while the gang molds, if made 
 heavy enough to prevent spreading, are unwieldy, and briquets 
 are removed from them with greater difficulty than from the 
 single molds. It is considered, therefore, that the most con- 
 
MOLDING BRIQUETS 113 
 
 vcnicnt form is the single mold, in which the two halves are 
 held together by a screw clamp of simple design. 
 
 182. To clean these molds, place ten in a row with clamps 
 removed ; scrape the upper faces with a piece of zinc, brush 
 with a stiff " horse-brush/' and wipe with oily waste. Turn 
 them over and repeat the process. Then separate the two 
 halves of each mold, place the twenty halves in line with inner 
 surfaces up, forming a trough twenty inches long. Wipe this 
 trough thoroughly with oily waste, finishing with some that is 
 only slightly oiled. 
 
 183. MOLDING. Methods of molding briquets vary widely 
 and have a considerable effect on the results obtained by differ- 
 ent operators. The mold may be placed on a glass or marble 
 slab, or on a porous bed. This difference in treatment will 
 affect the results chiefly because a porous bed will extract 
 moisture from the briquet, and, unless it is already mixed very 
 dry, will make it give a higher result on a short time test. The 
 use of a porous bed probably originated with a desire to more 
 closely imitate the use of mortar in actual work, but it intro- 
 duces another source of variation in results and should not be 
 followed. 
 
 184. In hand work the whole mold may be filled at once, 
 or small amounts of mortar may be added at a time, and each 
 layer packed; the mortar may be tamped into the mold with a 
 rod, in which case the pressure used may vary widely; or the 
 mortar may be pressed in with the fingers, or with the point of 
 a small trowel; and, finally, the pressure applied on the top of 
 the whole briquet may be light or heavy. It is evident that it 
 is almost impossible to so describe all these details of manipula- 
 tion that another operator may follow the same system and 
 obtain the same results. The practice of ramming the mortar 
 into the mold by means of a metal rod or a stick faced with 
 zinc is objectionable, because of the possible wide variation in 
 the force thus applied. This method is sometimes used by 
 manufacturers, since by making the mortar quite dry and ram- 
 ming it into the molds very hard, a high initial strength is 
 obtained. But the foremost cement makers are now eschewing 
 such methods and are aiming to make fair tests. Some experi- 
 ments made under the author's direction indicate that the 
 pressure applied to the top of the briquet is the salient point in 
 
114 
 
 CEMENT AND CONCRETE 
 
 the process of molding, and that the other details are of minor 
 importance. 
 
 In Germany a heavy trowel or iron plate weighing about 
 250 grams, and provided with a handle, is used in making one- 
 to-three mortar briquets. The mortar is made rather dry 
 (about 10 per cent, water), and after the mold is filled and 
 heaped, the mortar is beaten with the trowel until it becomes 
 elastic, and water appears on the surface. The excess of mor- 
 tar is then scraped off with an ordinary trowel or spatula. 
 
 185. Several machines have been devised for making bri- 
 quets, some of which are said to give good results. Among 
 these the most prominent is the Bohme hammer apparatus, 
 which is much used in Germany, although not employed to any 
 extent in the United States. It consists of a plunger which 
 fits the mold and upon which a given number of blows are 
 struck by a hammer. The mortar is first gaged as for hand 
 molding, and placed in the form. A pinion, turned by a hand 
 crank, is geared to a wheel provided with ten cams. These 
 cams operating on the wrought iron handle of the hammer 
 cause a certain number of blows to be delivered to the plunger. 
 The mechanism is automatically shut off after the proper number 
 of blows has been delivered. The following results were ob- 
 tained by Professor Bohme with his apparatus: 
 
 TABLE 32 
 
 Comparison of Hand Made Briquets with Those Made by Bohme 
 
 Hammer 
 
 
 
 
 MEAN TENSILE 
 
 No. 
 
 METHOD. 
 
 WEIGHT OF 
 BRIQUETS. 
 
 STRENGTH AT 
 7 DAYS IN KGS. 
 
 
 
 
 PER SQ. CM. 
 
 1 
 
 By hand 
 
 160.0 
 
 16 06 
 
 2 
 
 Hammer, 75 blows 
 
 158.0 
 
 12.75 
 
 3 
 
 100 " 
 
 159.5 
 
 13.25 
 
 4 
 
 " 125 " 
 
 159.5 
 
 14.56 
 
 5 
 
 " 150 " 
 
 159.0 
 
 1556 
 
 186. Several American engineers have devised machines for 
 briquet-making, but none of them has been generally adopted. 
 
 An apparatus designed by Prof. Charles Jameson, of Iowa 
 University, is said to work very rapidly. The mortar is packed 
 in the mold by a plunger of the form of the briquet. This 
 
MOLDING BRIQUETS 115 
 
 plunger works in a chamber of the same shape as the briquet 
 mold. The mortar is placed in a hopper at the side of this 
 chamber, and is delivered to the mold automatically when the 
 plunger is raised. The force is applied to the plunger by hand, 
 but it should be so arranged that this be done by a weight, to 
 prevent variations in pressure. In this method the briquet is 
 removed from the mold as soon as made, and this would appear 
 to be an objectionable feature. 
 
 Professor Spalding, of Cornell University, in his excellent 
 little book on "Hydraulic Cement," states that he has found 
 that "a pressure of about 500 pounds upon the surface of the 
 briquet is sufficient to produce a compact and homogeneous 
 briquet, and a crude appliance consisting of a lever arranged to 
 bring a pressure upon the mortar in the mold by means of a 
 weight suspended at the end of the lever, has been found to 
 increase both the rapidity and the regularity of the work, and 
 especially to diminish the variations in results obtained by dif- 
 ferent men." 
 
 A machine which would give more uniform results and work 
 more rapidly than hand molding, would commend itself for 
 general use. 
 
 187. Method Recommended. In making briquets by hand, 
 the mortar may well be packed into the molds by the fingers, 
 which should be protected by rubber tips. When the mold is 
 filled and slightly heaped, the trowel should be placed on top, 
 and the molder put about 60 pounds pressure on the trowel. 
 The excess mortar is then cut off by the trowel and the top of 
 the briquet is smoothed by drawing the trowel across the face. 
 The results obtained by four molders using this method in the 
 same laboratory are given in Table 33. 
 
 188. The recent progress report of the Committee of the 
 American Society of Civil Engineers on uniform tests of cement 
 contains the following, under " Molding": 
 
 " Having worked the paste or mortar to the proper consist- 
 ency, it is at once placed in the molds by hand. 
 
 "The Committee has been unable to secure satisfactory re- 
 sults with the present molding machines; the operation of 
 machine-molding is very slow, and the present types permit of 
 molding but one briquet at a time, and are not practicable with 
 the pastes or mortars herein recommended. 
 
116 
 
 CEMENT AND CONCRETE 
 
 TABLE 33 
 Results Obtained by Different Molders when Using Similar Mortar 
 
 
 
 X 
 
 . 
 
 
 MEAN TENSILE 
 
 c M 
 
 
 
 
 o 
 
 o r 
 
 o w 
 
 
 STRENGTH. 
 
 gw 
 
 
 
 H 
 w 
 
 ^r 
 
 si 
 
 ^O r, fd 
 
 ^ 
 
 
 
 II 
 
 
 pa 
 
 H 
 
 
 
 
 i 
 
 M 
 
 K 
 
 fcO 
 
 l3p 
 
 H H 
 
 ^ 63 
 
 - H 
 
 AGE. 
 
 M 
 
 M 
 
 H 
 A 
 
 
 DATE. 
 
 m 
 
 H 
 
 
 PH 
 
 pj Wfl 
 
 W SB 
 
 
 Q 
 
 Q 
 
 Q 
 
 K "' 
 
 
 W 
 
 jsj 
 
 fc fc 
 
 w ^ 
 
 PH tj 
 
 
 
 _j 
 
 
 33 ^ 
 
 
 K 
 
 rH 
 
 
 H ^ 
 
 s 
 
 
 O 
 
 
 
 
 
 JB K 
 
 
 
 
 CO 
 
 
 
 ^1 
 
 
 * 
 
 S 
 
 S 
 
 P 
 
 
 
 
 a 
 
 6 
 
 c 
 
 d 
 
 e 
 
 / 
 
 9 
 
 /I 
 
 i 
 
 
 1 
 
 
 
 31.6 
 
 62-65 
 
 1 days 
 
 81 
 
 92 
 
 89 
 
 5 
 
 10-22 
 
 Clear 
 
 2 
 
 
 
 '4 
 
 K 
 
 28 " 
 
 197 
 
 213 
 
 220 
 
 5 
 
 
 
 3 
 
 1 
 
 18.7 
 
 67-62 
 
 7 " 
 
 79 
 
 91 
 
 89 
 
 5 
 
 
 
 4 
 
 1 
 
 * 
 
 
 
 28 " 
 
 235 
 
 257 
 
 259 
 
 5 
 
 
 
 5 
 
 1 
 
 
 
 63-68 
 
 3 mo. 
 
 515 
 
 541 
 
 519 
 
 5 
 
 
 
 6 
 
 1 
 
 " 
 
 " 
 
 1 year 
 
 558 
 
 569 
 
 555 
 
 5 
 
 
 
 7 
 
 2 
 
 15.2 
 
 70-65 
 
 28 days 
 
 196 
 
 186 
 
 197 
 
 5 
 
 
 
 8 
 
 2 
 
 u 
 
 u 
 
 3 mo. 
 
 423 
 
 383 
 
 406 
 
 5 
 
 
 
 9 
 
 3 
 
 13.8 
 
 65-61 
 
 3 mo. 
 
 253 
 
 263 
 
 239 
 
 5 
 
 
 
 10 
 
 3 
 
 " 
 
 
 
 1 year 
 
 260 
 
 232 
 
 236 
 
 5 
 
 
 
 11 
 
 Sum of Means 
 
 
 2797 
 
 2827 
 
 2809 
 
 
 
 
 
 
 
 
 
 
 Molder 
 
 Molder 
 
 
 
 
 
 
 
 
 
 
 S. 
 
 T. 
 
 
 
 
 12 
 
 
 
 31.6 
 
 62-65 
 
 7 days 
 
 
 60 
 
 60 
 
 5 
 
 10-28 
 
 Cloudy 
 
 13 
 
 
 
 * 
 
 " 
 
 28 
 
 
 145 
 
 167 
 
 5 
 
 
 
 14 
 
 1 
 
 18.7 
 
 65 
 
 7 " 
 
 . 
 
 67 
 
 71 
 
 5 
 
 
 
 15 
 
 1 
 
 " 
 
 
 
 28 ' 
 
 , 
 
 223 
 
 211 
 
 5 
 
 
 
 16 
 
 1 
 
 1 1 
 
 " 
 
 3 mo. 
 
 , 
 
 435 
 
 449 
 
 5 
 
 
 
 17 
 
 1 
 
 K 
 
 " 
 
 1 year 
 
 . 
 
 504 
 
 491 
 
 5 
 
 
 
 18 
 
 2 
 
 152 
 
 67 
 
 28 days 
 
 . 
 
 182 
 
 179' 
 
 5 
 
 
 
 19 
 
 Sum of Means 
 
 
 
 
 1616 
 
 1628 
 
 
 
 
 Cement, Brand Gn, Sample 21 R. Sand, Crushed Quartz 20 to 40. 
 All briquets in same line received same treatment after made and were 
 immersed in same tank until broken. 
 1 Mean of ten specimens. 
 
 "Method. The molds should be filled at once, the material 
 pressed in firmly with the fingers and smoothed off with a 
 trowel without ramming; the material should be heaped up on 
 the upper surface of the mold, and, in smoothing off, the trowel 
 should be drawn over the mold in such a manner as to exert a 
 moderate pressure on the excess material. The mold should be 
 turned over and the operation repeated. 
 
 "A check upon the uniformity of the mixing and molding is 
 afforded by Weighing the briquets just prior to immersion, or 
 
STORING BRIQUETS 
 
 117 
 
 upon removal from the moist closet. Briquets which vary in 
 weight more than 3 per cent, from the average should not be 
 tested." 
 
 189. Marking the Briquets. The briquets made in a given 
 laboratory should be numbered consecutively, so that no con- 
 fusion can arise, and this one number is all that should be placed 
 on the briquet. The record of the brand of cement, the pro- 
 portions used, etc., should be placed in a book opposite the 
 briquet number. The briquets should be numbered on the 
 face, near the end. Steel stamps furnish a ready means of 
 numbering, and when mortar contains more than two parts of 
 sand to one of cement a thin strip of neat cement paste plastered 
 across one end of the briquet will aid in making the numbers 
 legible. 
 
 ART. 24. STORING BRIQUETS 
 
 190. The Time in Air before Immersion. As soon as the 
 briquets are molded they should be covered with a damp cloth 
 
 TABLE 34 
 
 Variations in Length of Time Briquets are Left in Moist Air before 
 Immersion Natural Cement 
 
 
 
 
 TENSILE STRENGTH, POUNDS PER 
 
 
 
 
 SQ. INCH. 
 
 CEMENT. 
 
 PARTS CRUSHED 
 QUART/, 20-30 
 
 TO 
 
 1 OEMKNT 
 
 AGE 
 WHEN 
 BROKEN. 
 
 
 HOURS IN MOIHT AIR BEFORE 
 IMMERSION. 
 
 
 
 
 8 
 
 12 
 
 24 
 
 48 
 
 72 
 
 168 
 
 Brand. 
 
 Sample. 
 
 Gn 
 
 15 K 
 
 
 
 7 days 
 
 123 
 
 
 139 
 
 151 
 
 161 
 
 237 
 
 
 t 1 
 
 1 
 
 7 days 
 
 91 
 
 . 
 
 106 
 
 114 
 
 114 
 
 182 
 
 
 16 R 
 
 
 
 28 days 
 
 110 
 
 
 106 
 
 109 
 
 89 
 
 113 
 
 
 
 1 
 
 28 days 
 
 142 
 
 . 
 
 138 
 
 139 
 
 152 
 
 175 
 
 
 
 2 
 
 28 days 
 
 102 
 
 
 105 
 
 112 
 
 113 
 
 115 
 
 An 
 
 G 
 
 
 
 7 days 
 
 
 168 
 
 181 
 
 194 
 
 185 
 
 238 
 
 
 
 
 
 28 days 
 
 . 
 
 200 
 
 210 
 
 224 
 
 241 
 
 243 
 
 
 
 1 
 
 7 days 
 
 
 108 
 
 137 
 
 141 
 
 157 
 
 160 
 
 
 
 1 
 
 28 days 
 
 
 278 
 
 283 
 
 297 
 
 297 
 
 301 
 
 
 
 3 
 
 28 days 
 
 
 120 
 
 130 
 
 137 
 
 139 
 
 152 
 
 NOTE : All briquets made by same molder. Each result is mean of ten 
 specimens. 
 
 until they are ready to be removed from the molds, when they 
 should- be transferred to a "damp closet," lined with zinc or 
 other non-corroding metal. It was formerly the practice to 
 immerse the briquets as soon as they were considered to be 
 
118 
 
 CEMENT AND CONCRETE 
 
 sufficiently set; but for the sake of uniformity, they are now 
 left in moist air for twenty-four hours before immersion, whether 
 the cement is quick or slow setting. Briquets which are to be 
 broken at twenty-four hours, however, are usually immersed 
 as soon as set hard. 
 
 Table 34 gives the results obtained by allowing natural 
 cement briquets to remain in moist air different lengths of time 
 before immersion. In general, the strength is greater for seven 
 and twenty-eight day tests the longer the briquets are allowed 
 to remain in the moist air. It appears that, while the time in 
 moist air should be made as nearly uniform as possible, a varia- 
 tion of a few hours will not cause an important difference in 
 strength. 
 
 TABLE 35 
 Gain or Loss in Strength of Natural Cement Briquets by Immersion 
 
 
 
 
 TENSILE STRENGTH, POUNDS 
 
 
 
 
 PER SQ. INCH. 
 
 TIME IN 
 MOIST AIR. 
 
 TIME IN TANK. 
 
 AGE \\ H K N 
 
 BROKEN. 
 
 
 One Part Stand- 
 
 
 
 
 Neat Cement. 
 
 ard Sand to 
 
 
 
 
 
 One Cement. 
 
 20 hours 
 
 
 20 hours 
 
 151 
 
 94 
 
 18 hours 
 
 6^ days 
 
 7 days 
 
 147 
 
 153 
 
 2 days 
 
 
 2 days 
 
 192 
 
 126 
 
 2 days 
 
 5 days 
 
 7 days 
 
 160 
 
 158 
 
 3 days 
 
 
 3 days 
 
 *205 
 
 141 
 
 3 days 
 
 4 days 
 
 7 days 
 
 177 
 
 155 
 
 4 days 
 
 . 
 
 4 days 
 
 218 
 
 165 
 
 4 days 
 
 3 days 
 
 7 days 
 
 191 
 
 165 
 
 5 days 
 
 .... 
 
 5 days 
 
 230 
 
 175 
 
 5 days 
 
 2 days 
 
 7 days 
 
 192 
 
 169 
 
 NOTE : All briquets made by same molder. Each result is mean of 
 five specimens. 
 
 Table 35 shows the early action of the water on the briquets. 
 These tests were made in sets of ten; five briquets of a set were 
 immersed after twenty hours, forty-eight hours, etc., while the 
 other five of the same set were broken at the time the first five 
 were immersed. With this sample of natural cement, it appears 
 that the briquets lose part of their strength by immersion, and 
 that some time is required to regain this lost strength. Thus, 
 with neat cement mortar the briquets broken at twenty hours 
 without immersion were as strong as those broken at seven 
 days which had been immersed the last six and one-fourth days. 
 With briquets of one-to-one mortar, it appears that if immersed 
 
STORING BRIQUETS 119 
 
 at the end of four days, the gain in strength during the last 
 three days (in water) is about equal to the loss of strength due 
 to immersion. If immersed earlier than this, the gain is greater 
 than the loss, but if immersed later, the loss is greater than 
 the gain. 
 
 191. For storing briquets the required twenty-four hours 
 before immersion a moist closet is very convenient, tends to 
 promote uniformity of treatment, and may be very easily 
 made. The use of a damp cloth for covering briquets is incon- 
 venient, as the cloth may dry out. If it is used, the end of 
 the cloth should rest in a pail of water, so it will keep wet by 
 capillarity; it should also be kept from touching the briquets by 
 a wire screen or by wooden slats. 
 
 A moist closet may be made of slate, glass or soapstone, or 
 of wood lined with metal. In the bottom of the box is a pan of 
 water, or a sponge kept constantly wet. The shelves may well 
 be of glass, and should be so arranged that any shelf may be 
 removed without disturbing the others. 
 
 192. Water of Immersion. When the briquets are ready to 
 be immersed, i.e., usually, twenty-four hours after made, they are 
 placed in a tank, containing water that is kept fresh by frequent 
 renewals. The water in the tank should also be maintained at 
 a nearly constant temperature. It is sometimes the case that 
 briquets are subjected to considerable variations of temperature 
 while in storage. It also frequently occurs that the water is 
 allowed to become stale. A few of the many experiments 
 made at St. Marys Falls Canal to show the effect, on the tensile 
 strength of natural cement briquets, of variations in the tem- 
 perature of the water of immersion, are given in Table 36. The 
 details of these experiments, as well as other tests on the same 
 point, may be found in the Annual Report, Chief of Engineers, 
 U. S. A., for 1894, page 2314. 
 
 The very marked effect which the temperature of the water 
 may have on the rate of hardening of natural cements is clearly 
 shown. When broken at the age of one day or seven days, 
 the effect on the strength may not be evident, or the briquets 
 stored in cold water may develop a greater strength, but the 
 more rapid hardening of the briquets stored in warm water is 
 usually very evident at twenty-eight days, and increases up to 
 two or three months. Some samples of cement are affected 
 
120 
 
 CEMENT AND CONCRETE 
 
 less than others, and a few experiments indicated that the 
 differences in strength due to the temperature of water of im- 
 mersion decrease after three months and become almost nil at 
 one year. 
 
 193. The conclusion drawn from these tests may be briefly 
 stated as follows: Between certain limits the early strength of 
 natural cement mortars is usually developed faster in cool 
 
 TABLE 36 
 
 Variations in Temperature of Water in which Briquets are 
 
 Immersed 
 
 
 
 a 
 
 
 
 6 
 
 
 
 ( 
 
 TENSILE STRENGTH, POUNDS PER SQUARE 
 
 ft 
 
 NATURAL 
 
 "^ Sr H 
 
 2 
 
 INCH, WHEN IMMERSED IN 
 
 H 
 
 CEMENT. 
 
 9 Q E-TtC 
 
 
 WATER OF APPROXIMATE TEMPERATURE, 
 
 U 
 
 
 sill 
 
 & y 
 
 DEGREES FAHR. 
 
 1 
 
 Brand. 
 
 Sample. 
 
 A 
 
 oj 
 
 O 1 
 
 38 
 
 40 
 
 50 
 
 55 
 
 60 
 
 65 
 
 70 
 
 80 
 
 PH 
 
 
 
 OH 
 
 
 
 
 
 
 
 
 
 
 j 
 
 Gn 
 
 15 R 
 
 o 
 
 Tdavs 
 
 146 
 
 
 137 
 
 125 
 
 
 
 126 
 
 154 
 
 2 
 
 
 
 
 
 LIC*^ 
 
 14 days 
 
 144 
 
 . 
 
 131 
 
 125 
 
 131 
 
 150 
 
 168 
 
 208 
 
 3 
 
 
 
 
 
 28 days 
 
 166 
 
 . . . 
 
 178 
 
 . . . 
 
 184 
 
 . . . 
 
 247 
 
 280 
 
 4 
 
 
 
 1 
 
 7 days 
 
 83 
 
 . . . 
 
 88 
 
 84 
 
 89 
 
 98 
 
 97 
 
 121 
 
 5 
 
 
 
 1 
 
 14 days 
 
 84 
 
 . . . 
 
 111 
 
 
 123 
 
 . . . 
 
 150 
 
 191 
 
 . 6 
 
 
 
 1 
 
 28 days 
 
 96 
 
 . . . 
 
 156 
 
 187 
 
 . . . 
 
 221 
 
 243 
 
 288 
 
 7 
 
 Ln 
 
 31 S 
 
 
 
 1 day 
 
 . . 
 
 143 
 
 . . . 
 
 124 
 
 120 
 
 . . . 
 
 109 
 
 109 
 
 8 
 
 
 
 
 
 7 days 
 
 . 
 
 204 
 
 201 
 
 . . 
 
 183 
 
 . . . 
 
 193 
 
 186 
 
 9 
 
 
 
 
 
 14 days 
 
 . 
 
 184 
 
 203 
 
 . . . 
 
 204 
 
 . . 
 
 229 
 
 245 
 
 10 
 
 
 
 
 
 28 days 
 
 . 
 
 221 
 
 245 
 
 . . . 
 
 254 
 
 . . . 
 
 281 
 
 303 
 
 11 
 
 
 
 
 
 2 mos. 
 
 . 
 
 261 
 
 292 
 
 . . . 
 
 348 
 
 . . . 
 
 382 
 
 429 
 
 12 
 
 An 
 
 G 
 
 1 
 
 7 days 
 
 . . 
 
 134 
 
 140 
 
 . . . 
 
 150 
 
 . . . 
 
 154 
 
 158 
 
 13 
 
 
 
 1 
 
 14 days 
 
 . . 
 
 149 
 
 162 
 
 . . . 
 
 189 
 
 . . . 
 
 182 
 
 216 
 
 14 
 
 
 
 1 
 
 28 days 
 
 . . . 
 
 198 
 
 223 
 
 
 250 
 
 . . . 
 
 281 
 
 296 
 
 15 
 
 
 
 1 
 
 2 mos. 
 
 . . . 
 
 251 
 
 286 
 
 
 337 
 
 
 386 
 
 403 
 
 16 
 
 
 
 3 
 
 14 days 
 
 . . . 
 
 50 
 
 58 
 
 . . . 
 
 69 
 
 
 73 
 
 100 
 
 17 
 
 
 
 3 
 
 28 days 
 
 . . . 
 
 67 
 
 87 
 
 
 100 
 
 
 102 
 
 157 
 
 18 
 
 
 
 3 
 
 2 inos. 
 
 
 104 
 
 127 
 
 
 147 
 
 . . . 
 
 194 
 
 231 
 
 water, but after the first seven days, and sometimes after a 
 shorter time, the strength is developed more rapidly in warm 
 water, and the strength at any time between seven days and 
 three months is approximately proportional to the temperature. 
 After three months, the effect of the temperature seems to 
 diminish, and may entirely disappear in time. 
 
 M. Paul Alexandre l made quite a number of experiments 
 on this point with Portland cement. In these experiments the 
 
 " Recherches Experimentales sur les Mortiers Hydrauliques." 
 
STORING BRIQUETS 121 
 
 gaging was done in about the same temperature as that at which 
 the water of immersion was maintained, so that a double cause 
 of variation was present. However, it was found that in all 
 cases the higher strength was attained at seven days by the 
 briquets made and stored in the higher temperature (15 to 
 18 C., 60 to 65 Fahr.) while at twenty-eight days the briquets 
 of the lower temperature (0 to 5C., 32 to 40 Fahr.) were 
 ahead in the case of neat cement, and nearly as high as the 
 warm briquets in the case of mortar. At three months the 
 differences seemed to disappear. 
 
 194. Stale Water. Some experiments made to compare the 
 strength of briquets which were alike in all other respects, but 
 were immersed in different tanks in which the water had not 
 been frequently renewed, showed very clearly the possible varia- 
 tions from this source. Natural cement briquets, neat, and with 
 one and two parts sand, gave, when immersed in one of the 
 tanks, only from 40 to 60 per cent, of the strength attained 
 in another tank by briquets entirely similar. 
 
 To store briquets in running water is going to the other 
 extreme; this appears to be the best method, at least for short- 
 time acceptance tests, provided the temperature can be regu- 
 lated. However, in some cases where this has been adopted, 
 the strength of the briquets is said to have fallen off very much 
 after four or five years. Whether this is due to the action of 
 running water is a very interesting point, and a valuable one 
 from the practical standpoint of the use of cement, but it has 
 not yet been thoroughly investigated. 
 
 195. It appears from the foregoing that variations in the 
 temperature and freshness of the water in which the briquets 
 are immersed is an uncertain contingent, and therefore that all 
 such variations should be carefully avoided. As a matter of 
 convenience, the tanks may well be maintained at 60 to 70 
 Fahr., but if one does not care for a comparison of his results 
 with those obtained in other laboratories, then any other con- 
 stant temperature between 40 and 75 may be adopted. The 
 water in the tanks should be renewed at least once a month, 
 and preferably once a week. 
 
 196. Storing Briquets in Sea Water. When the cement 
 under test is to be usod for constructions in the sea, some of 
 the briquets should be stored in sea water to indicate the be- 
 
122 CEMENT AND CONCRETE 
 
 havior in this medium. Many tests have been made in this 
 way by several experimenters, but the varied results obtained 
 only indicate the different effects of such treatment on different 
 samples of cement. One of the effects of storing in sea water 
 has been touched upon under the head of consistency of mortar, 
 where it is shown that porous briquets may disintegrate in this 
 medium. A small specimen like a briquet will of course be 
 more quickly affected than a large mass of concrete, but on the 
 other hand, the concrete in work is likely to be more porous 
 than the briquet. The effect of sea water upon cement will be 
 taken up in another place. 
 
 197. Other Methods of Storing Briquets. It has been 
 thought that briquets, made to test cement that is to be used 
 in air, should be hardened in the same medium in order that 
 the tests should more nearly approach the conditions of use. 
 Several points, however, should be borne in mind in interpret- 
 ing the results obtained with air-hardened specimens. In actual 
 work the mortar is usually in a large mass, or is protected from 
 the influence of a warm, dry atmosphere, so that it remains 
 moist for a long time, whereas a briquet placed in the open 
 air is much more affected by changes in atmospheric condi- 
 tions. If the briquets are allowed to harden in a room, such a 
 small quantity of mortar may become quite dry in a few days, 
 and, unless the amount of moisture in the air is regulated, an- 
 other source of variation is introduced in the tests. 
 
 It has been found impossible to obtain uniform results from 
 briquets made as nearly alike as possible and stored side by 
 side in the air of the laboratory. The regular acceptance tests 
 should, therefore, it is thought, be made in the ordinary man- 
 ner, but if cement is to be used in locations where it is likely 
 to become very dry, a few special tests should be made to assure 
 one that the brand of cement in question is one that will yield 
 good results in such exposure. It may be found that certain 
 kinds or brands should be entirely avoided for use in such lo- 
 cations. A few tests of this character are given in Tables 72 
 and 73, 359, 360. The results in any given line of the table 
 are from briquets made the same way but treated differently 
 in the method of storing. It is seen that these brands harden 
 well in dry air. The effect of the amount of water used in 
 gaging appears to follow somewhat the same law, whether the 
 briquets are stored in air or water, 
 
BREAKING THE BRIQUETS 123 
 
 A method more nearly approaching conditions that fre- 
 quently prevail in practice is to bury the briquets in damp 
 sand. Table 120, 409, gives the results obtained with a large 
 number of briquets stored in this way. While the results are 
 somewhat more irregular than those for water-hardened speci- 
 mens, since the conditions cannot be made so nearly uniform, 
 yet this method gives better results than dry air storage. 
 
 ART. 25. BREAKING THE BRIQUETS 
 
 198. THE TESTING MACHINE. The function of the testing 
 machine is simply to furnish a means of applying the tensile 
 stress, and of measuring the amount of force required to break 
 the briquet. Aside from the clips, which hold the briquet, 
 any contrivance which may be conveniently operated, and 
 which will accurately measure the force applied, may be used 
 for this purpose. 
 
 There are several forms of testing machines on the market, 
 all designed on the lever principle, though differing slightly in 
 the method of application. The force is applied either by al- 
 lowing water or shot to run into or out of a vessel suspended 
 at the end of the longer arm of a lever, or a weight is made to 
 run along the lever arm, which is graduated so that the force 
 applied may be read from the beam. 
 
 199. In machines of the first class the delivery of shot is 
 cut off automatically the instant the briquet breaks. The ad- 
 vantage of this style is that the flow of shot may be so adjusted 
 as to approximately regulate the rate of applying the stress; 
 but little skill is required to operate it, and, since in its best 
 form two levers are used, the shorter arm of one acting on the 
 longer arm of the other, the machine occupies but little space. 
 This machine does not permit rapid operation, since the shot 
 must be weighed ea.3h time a briquet is broken. One of the 
 main disadvantages of this form has been that in the case of 
 strong briquets, a certain initial strain had to be applied in 
 order that the stretch of the briquet and the slipping of the 
 clips should not allow the shot to be cut off before the briquet 
 broke. This objection, however, has recently been met by the 
 makers, who have provided means of taking up this slip by a 
 hand crank. 
 
 200. Another objection urged against the short-lever shot 
 
124 CEMENT AND CONCRETE 
 
 machines is the fact that as the stream of shot flowing into the 
 scale pan is cut off by the breaking of the briquet, a certain 
 amount of shot on its way to the pan falls into the pan after 
 the briquet breaks, and is weighed, although not acting on the 
 briquet at the time of the break. A form of shot machine is 
 now on the market, however, in which this objection has been 
 overcome. The load is applied by means of a weight hanging 
 from one end of a lever. This weight is at first counterbalanced 
 by a pail of shot at the other end of the lever, but as the shot 
 is allowed to run out of the vessel, the unbalanced portion of 
 the weight acts, through suitable levers, upon the briquet. 
 The flow of shot is shut off automatically by the breaking of 
 the briquet, and the shot that has escaped is weighed on a 
 spacial scale to determine the load acting on -the briquet. 
 
 201. In the other form of machine the weight is made to 
 move along the arm by means of a cord and hand-wheel. This 
 style may be operated much more rapidly, but some skill is 
 required to use it properly, and as now made it occupies too 
 much space. These machines are preferable for laboratories, 
 while the shot machines may well be used in cement factories 
 and small works where a foreman does the testing. 
 
 202. It would seem that a machine could easily be made 
 which would combine the desirable features of both of these 
 forms, by placing a heavy weight provided with rollers upon 
 the upper lever arm of the shot machine, and using it in the 
 same way that the hand power machine is now used. This 
 would involve placing a hand wheel and cord upon the machine 
 to operate the moving weight, the shot attachment being re- 
 moved. Such a machine would combine the compactness of 
 the shot machine, with the accuracy and speed of the single 
 lever machine; the graduations on the beam could represent 
 five pounds each, instead of two pounds, the value of the grad- 
 uations now on the single lever machines. 
 
 203. FORMS OF CLIP. Since cement has been tested by 
 tensile strain, it has ever been a problem to obtain a clip which 
 would give a perfectly true axial pull on the briquet. Various 
 forms of clips have been used from time to time, but none of 
 them has proved satisfactory in all respects. To trace the his- 
 tory of the development of the clip is not warranted by its 
 interest, but.it may be said that in some of the early forms the 
 
BREAKING THE BRIQUETS 
 
 125 
 
 head of the briquet was held between two plates and clamped 
 tight enough to develop sufficient friction to transmit the stress. 
 The later forms of briquets are made with a shoulder or with 
 wedge-shaped ends to allow the clip to grasp them. Mr. John 
 Grant, Mr. Alfred Noble, General Gilmore, Mr. J. Sondericker 
 and Mr. D. J. Whittemore have each designed or adapted dif- 
 ferent forms, and more recently Mr. S. Bent Russell and Mr. 
 
 I" ** H * I Z 3//T. 
 
 FlO. 6. RIEHLE "ENGINEERS' STANDARD" CLIP 
 
 W. R. Cock have each devised a clip which will be mentioned 
 below. 
 
 204. Form in Most General Use. The clip in most general 
 use in the United States is of the general style shown in Fig. 6. 
 It differs only in detail from the form recommended by the 
 Amer. Soc. C. E. Committee of 1885, which has been called 
 
126 CEMENT AND CONCRETE 
 
 the "Engineers' Standard.". The general form is pear shaped; 
 the briquet is grasped at the points of reverse curve at the side 
 of the briquet, giving an area between opposite gripping points 
 of about one and a quarter square inches. The gripping points 
 are rather too sharp, when new, as they have a tendency to 
 crush the briquet locally. The width of the bearing increases 
 with the amount of wear the clip sustains. The clip is provided 
 with a conical pivot, which rests in a cone-shaped cavity at- 
 tached to the machine, so that the two parts of the clip are 
 free to swing. In a form which was previously used to a con- 
 siderable extent, each bearing surface was designed to be about 
 an inch square, the jaw being made to conform to the outline 
 of the briquet. This form, however, did not give satisfactory 
 results; a particle of sand between the briquet and the bearing 
 surface of the clip would give an eccentric pull, and strong 
 briquets, would sometimes break in the head of the briquet 
 transverse to the axis, in several curved layers joining opposite 
 gripping surfaces. 
 
 205. CLIP-BREAKS. When a briquet is inserted in the or- 
 dinary clip, the gripping points will not, in general, grasp the 
 briquet symmetrically. The gripping points have a tendency 
 to slide on the surface of the briquet in order to assume a sym- 
 metrical position; there is friction to resist this sliding, and 
 when this resistance overcomes the tendency to motion, the two 
 clips and the briquet become a rigid system, and bending strains 
 may be introduced. Again, if the briquet is not too badly ad- 
 justed in the clips, it is apt to break in a line joining two op- 
 posite gripping points, instead of at the smallest section; this 
 is called a " clip-break." The tendency to form clip-breaks is 
 greater if the gripping points are very narrow or have sharp 
 edges; neat cement briquets exhibit this tendency much more 
 than briquets from sand mortars, and some samples of cement 
 are much more likely to give clip-breaks than others. 
 
 206. Cause of Clip-breaks. When a briquet breaks in this 
 manner, the broken section is usually about normal to the side 
 of the briquet at the point where the jaw was in contact. This 
 indicates that a clip-break is caused by compression at that 
 place: there is evidently compression along the plane joining 
 the two opposite gripping points, and tension at right angles 
 to that plane, and the briquet fails here as a result of the two 
 
BREAKING THE BRIQUETS 127 
 
 stresses. If the briquet is not properly adjusted in. the clips, 
 but is so placed that its longest axis is at one side of the line 
 joining the points of application of the forces (in the " Engi- 
 neers' Standard" clip, the line joining the pivot points), then 
 the bending strain that is introduced is greatest at the central 
 section of the briquet; this may cause the briquet to break at 
 the smallest section, when if it were properly adjusted in the 
 clips it would develop a clip-break. The bearing surfaces of 
 the clip should not be too small, as this increases the intensity 
 of pressure, but on the other hand there appears to be no prac- 
 tical advantage in making this area more than j\ to inch 
 wide (the length being limited by the thickness of the briquet, 
 one inch). 
 
 207. Prevention of Clip-breaks. -- The method most fre- 
 quently adopted to prevent clip-breaks is to cushion the grip- 
 ping points with some compressible material, such as thin rub- 
 ber or blotting-paper. This device prevents clip-breaks, but 
 the result of about three hundred tests made under the author's 
 direction showed clearly that it also lowered the apparent 
 strength very materially. 1 Briquets broken with the bare clips 
 showed a mean strength of 606 pounds per square inch, while 
 the cushioned clips gave an apparent strength of but 521 pounds, 
 or 86 per cent, of the strength without the cushion; of the bri- 
 quets broken with the bare clips, 33 per cent, were clip-breaks; 
 with the cushioned clips no clip-breaks occurred. The rubber 
 was applied by slipping two rubber bands over each end of the 
 briquet, giving cushions about T V inch thick. 
 
 208. Strength of Briquets that Develop Clip-breaks. It was 
 also found in breaking 277 briquets with two styles of clips 
 without cushions that 129 of them that gave clip-breaks aver- 
 aged 611 pounds per square inch, while 148 which did not de- 
 velop clip-breaks had a mean strength of 590 pounds. This 
 result is easily accounted for by saying that some of the bri- 
 quets that broke in the small section were made to do so by the 
 cross-strain introduced by imperfect adjustment in the clips. 
 
 When a briquet breaks at other than the smallest section, it 
 is certain that the smallest section has a greater strength per 
 
 1 For a report of these tests in detail, see Annual Report Chief of Engi- 
 neer's, U. S. A., 1895, p. 2913. Also "Municipal Engineering; 1 Dec., 1896, 
 Jan., Feb., 1897. 
 
128 
 
 CEMENT AND CONCRETE 
 
 square inch than is shown by the result obtained ; how much 
 greater cannot be told. But it follows that if clip-breaks could 
 be eliminated in a proper way, one which would not cause center 
 breaks by the introduction of cross-strains or other undesirable 
 conditions, the strengths thus obtained would be greater than 
 when clip-breaks occur. -The fact that the use of a rubber 
 cushion gives lower strengths, shows that this is not the proper 
 method of preventing clip-breaks. 
 
 209. Mr. W. R. Cock has devised a clip, with rubber-covered 
 gripping points, which has attracted some attention. It has 
 
 f Vt ft I 2 
 
 FIG. 7. RUSSELL CLIP 
 
 sometimes been assumed that because this clip eliminated clip- 
 breaks it must give a higher apparent strength than the rigid 
 form. No extensive series of experiments have been published 
 which permit of comparing this clip with other forms, but 
 from the results obtained above, in using rubber cushions, it 
 would appear that the Cock clip may give lower apparent 
 strengths. 
 
BREAKING THE BRIQUETS 
 
 129 
 
 210. The form of clip designed by Mr. S. Bent Russell is 
 constructed on the "evener" principle, each clip having free- 
 dom of motion imparted by four pin-connected joints (see Fig. 
 7). It is sought to prevent any but an axial pull being ap- 
 plied to the briquet. On account of details of construction, 
 into which it is not necessary to enter here, the clip must be 
 in its normal position when the briquet is inserted, in order 
 that the possibility of cross-strain shall be effectually removed. 
 As a result of many tests with this form and the ordinary "En- 
 
 it o a 
 
 FlO. 8. SINGLE GIMBAL CLIP 
 
 3 in. 
 
 gineers' Standard," it was found that they gave very nearly 
 the same strength. But that the evener motion itself was of 
 some value was shown by a series of experiments in which part 
 of the briquets were broken by this form of clip without modi- 
 fication, while part were broken by the same clip when it had 
 been changed to a rigid form by means of a clamp that elimi- 
 nated the evener motion. It is believed that with some modifi- 
 cations this clip will give good results, and it may be used al- 
 most as rapidly as the ordinary rigid form. 
 
130 CEMENT AND CONCRETE 
 
 211. Several experiments were made with a clip in which 
 the gimbal principle was applied, the stress passing from the 
 machine to the gripping points through knife edges placed in 
 the line joining opposite gripping points and midway between 
 them 1 (Fig. 8). Higher results were obtained with this form, 
 the " Single Gimbal," than with any of the styles with which it 
 was compared, but it was made only for experimental purposes, 
 and unless modified is not convenient enough to be recom- 
 mended for general use. 
 
 212. In the course of these experiments it was shown that to 
 increase the distance between gripping points, grasping the bri- 
 quet nearer the head, increased the apparent strength and 
 diminished the number of clip-breaks. With the Russall clip, 
 increasing this distance from l^V inches to ! T 7 g- inches gave an 
 increase of about six per cent, in the apparent strength; and a 
 similar increase in the width between jaws of the Gimbal clip, 
 from lf\ to 1-& inches, gave an increase in apparent strength 
 of about five per cent. It was found later that Mr. J. Son- 
 dericker had previously arrived at similar results, 2 and as the 
 form of briquet used by the latter had permitted extending 
 the experiment, he found that when the points were about If 
 inches apart (making the area of the briquet about If square 
 inches between opposite gripping points), nearly all the frac- 
 tures occurred at the smallest section. 
 
 213. Effect of Improper Adjustment. The effect of not 
 properly adjusting the briquets in the clip was also investigated. 
 In some cases the briquets were placed in the proper position 
 as nearly as possible. In the other cases they were in a de- 
 cidedly distorted position, much worse than they would be 
 placed with the most careless manipulation. It was found that 
 if the briquet was so placed in the " Engineers' Standard" 
 clip that the gripping points on one side of the briquet were 
 farther apart than those on the other side, the decrease in break- 
 ing strength was very marked (about 35 per cent.), while if the 
 planes determined by the lines of contact of the gripping points 
 of each clip were parallel, there appeared to be no effect. The 
 
 1 This clip was devised by the author at the suggestion of Mr. E. S. 
 Wheeler, M. Am. Soc. C. E. 
 
 2 Jour. Assoc. Eng. Soc., Vol. vii, p. 212. 
 
BREAKING THE BRIQUETS 131 
 
 reason of this is evident: in the former case the line of force, 
 joining the two pivot points, does not pass through the center 
 of the smallest section of the briquet, and transverse stresses 
 are introduced, while in the latter case the line of force does 
 pass through the center of the smallest section, though not at 
 right angles to its plane. With the Russell and Gimbal clips 
 the distortion seemed to have little effect, provided, that in 
 the case of the former, the clip was itself in its normal position 
 when the briquet was inserted. 
 
 214. Conclusions Derived from Tests of Several Styles of 
 Clips. From the tests described above, 1 the following conclu- 
 sions may be drawn: - 
 
 1st. When using the ordinary form of clips with metal 
 gripping points, the briquets which break at the places of con- 
 tact of the jaws give higher apparent strengths than those 
 which break at the smaller sections. 
 
 2d. A rubber cushion between the briquet and the jaw of 
 the clip prevents clip-breaks, but materially lowers the stress 
 required to break the briquet. 
 
 3d. The form of clip designed by Mr. S. Bent Russell gives 
 somewhat less irregular results than are obtained with the 
 Riehle " Engineers' Standard" rigid clip. Although the results 
 given by the Russell clip in its present form are a trifle lower 
 than those given by the Riehle, it seems probable that these 
 lower results are due to defects in detail which may readily 
 be eliminated. 
 
 4th. By the application of the gimbal principle to cement 
 testing clips, higher, as well as more nearly uniform, results 
 may be obtained. 
 
 5th. In using the rigid form of clip, careless manipulation 
 in adjusting the briquet may result in serious error due to the 
 introduction of cross-strains, while with either the Single Gim- 
 bal or Russell clip slight deviations in adjustment are not im- 
 portant. 
 
 6th. With the form of briquet recommended by the commit- 
 tee of the American Society of Civil Engineers in 1885, the break- 
 ing stress may be somewhat increased, and the number of clip- 
 
 1 These tests were described in greater detail and discussed by the 
 writer in "Municipal Engineering," Dec., 1896, Jan. and Feb., 1897. 
 
132 
 
 CEMENT AND CONCRETE 
 
 breaks may be very materially decreased, by such a modifica- 
 tion of the clip as to allow grasping the briquet nearer the head. 
 
 215. Requirements for a Perfect Clip. As a logical result 
 of these conclusions, the ideal clip should fulfill the following 
 requirements : 
 
 1st. It should impart a true axial pull to the briquet with- 
 out subjecting it either to cross-strains or to compressive forces 
 
 liilnl 
 
 r A YL ft o i 2 
 
 FlG. 9. FORM OF ARTICULATED CLIP SUGGESTED FOR USE 
 
 sufficient to cause it to break at other than the smallest section. 
 2d. The bearing surfaces of the gripping points should not 
 be more than about one-fourth of an inch wide, since this is 
 sufficient to prevent crushing the briquet at these places, and 
 too wide a jaw will not usually bear uniformly over its whole 
 surface. 
 
BREAKING THE BRIQUETS 
 
 133 
 
 3d. Its parts should have sufficient strength and stiffness, 
 so that they will not bend appreciably when in use. 
 
 4th. It should permit rapid operations, and 
 
 5th. It should be as light as consistent with the above 
 requirements. 
 
 216. Form Suggested. Fig. 9 shows a style of clip which 
 closely conforms to the above specifications. The evener 
 form devised by Mr. Russell has been selected for modification. 
 The S. G. clip would more nearly meet some of the requirements, 
 and, so far as the principle is concerned, this form is considered 
 quite the equal of the evener clip. But no method of applying 
 the gimbal principle has commended itself as affording such 
 rapid manipulation as does the evener motion, and since it is 
 thought that either form will obviate cross-strains in a plane 
 parallel to the face of the briquet, the evener form has been 
 adopted on account of convenience. 
 
 The defeats in detail of the Russell clip which have already 
 been mentioned have been obviated in the present form. The 
 gripping points are made one-fourth of an inch wide, and a little 
 more material has been used between the gripping points and 
 the first pin to stiffen the clip. This form is designed for use 
 with the briquet shown in Fig. 5 (see 179). 
 
 217. RATE OF APPLYING THE TENSILE STRESS. Table 37 
 gives the results of several hundred experiments made by Mr. 
 
 TABLE 37 
 Relation of Apparent Tensile Strength to Rate of Applying Stress 
 
 RATE OF APPLYING 
 
 TENSILE STRENGTH 
 
 STRESS, 
 POUNDS PER MINUTE. 
 
 OBTAINED, 
 POUNDS PER So,. INCH. 
 
 50 
 
 400 
 
 100 
 
 415 
 
 200 
 
 430 
 
 400 
 
 450 
 
 6,000 
 
 493 
 
 Henry Faija * to show the effect on tensile strength of varying 
 the rate of applying the stress. 
 
 A few of the results obtained from nearly 900 tests, made 
 
 1 "Cement for Users," by Mr. Henry Faija. 
 
134 
 
 CEMENT AND CONCRETE 
 
 under the author's direction to illustrate this point, are given 
 in Table 38. Some of these results accord very well with those 
 given in Table 37, but the results in the latter table were doubt- 
 less obtained from neat Portland briquets only, while the ex- 
 periments given in Table 38 were made with briquets neat 
 and with two parts sand, and on natural as well as Portland 
 cement mortars. 
 
 TABLE 38 
 
 Relation of Apparent Tensile Strength to the Rate of Applying the 
 
 Stress 
 
 
 
 
 TENSILE STRENGTH, POUNDS PER 
 
 
 
 
 SQUARE INCH, FOR STRESS APPLIED AT 
 
 CEMENT. 
 
 PROPORTIONS. 
 
 AOK OF 
 
 BRIQUETS. 
 
 RATE OF POUNDS PER MINUTE. 
 
 
 
 
 100 
 
 300 
 
 500 
 
 700 
 
 900 
 
 Portland 
 
 Neat cement 
 
 7 and 14 days 
 
 453 
 
 485 
 
 521 
 
 520 
 
 528 
 
 it 
 
 Neat cement 
 
 3 months 
 
 
 
 590 
 
 617 
 
 622 
 
 640 
 
 n 
 
 1-2 
 
 3 months 
 
 445 
 
 467 
 
 487 
 
 507 
 
 510 
 
 Natural 
 
 Neat cement 
 
 7 days 
 
 150 
 
 169 
 
 186 
 
 . . . 
 
 202 
 
 K 
 
 Neat cement 
 
 3 months 
 
 309 
 
 351 
 
 363 
 
 378 
 
 390 
 
 U 
 
 1-2 
 
 3 months 
 
 255 
 
 299 
 
 327 
 
 329 
 
 354 
 
 218. It appears from all these results that the increase in 
 the breaking strength due to increasing the rate of applying 
 the stress is considerable in the case of low rates of speed, but 
 when a rate of 500 or 600 pounds per minute has been reached, 
 a further increase in rapidity does not make a material increase 
 in the apparent strength. Since certain variations in rate are 
 sure to occur, until some device is used to automatically regu- 
 late it, a rate should be adopted which would allow of slight 
 variations without materially changing the result of the test. 
 A rate of 600 pounds per minute would fulfill this requirement, 
 and, with certain machines at least, would be still more con- 
 venient than the rate of 400 pounds per minute which has here- 
 tofore been quite generally used. 
 
 An analysis of the -experiments made to determine the de- 
 gree of uniformity obtained by using each of the given rates, 
 showed there was but little difference in this regard, but if 
 any choice could be made on this basis it seemed to lie with 
 the more rapid rate. 
 
 219. With the shot machines it is not difficult to approxi- 
 
BREAKING THE BRIQUETS 135 
 
 mately regulate the rate at which the stress is applied. In 
 operating a machine in which a handwheel moves a weight 
 along the graduated beam, it must be remembered that the rate 
 at which the weight moves is the controlling factor, and not 
 the movement of the lower wheel, which simply serves to take 
 up lost motion, the stretch of the briquet under strain, and 
 the slipping of the briquet in the jaws of the clip. A mistaken 
 idea concerning this matter has sometimes led to the adoption 
 of a device to regulate the motion of this lower wheel. Until 
 one is accustomed to applying the stress at a given uniform 
 rate, he will find it an aid to hang near the machine a pendulum 
 of such a length that a certain number of vibrations correspond 
 to a complete revolution of the handwheel. 
 
 220. Treatment of the Results. The number of briquets 
 which are made to test the strength of a given sample of cement 
 will depend on the accuracy which it is desired to attain. If 
 but two briquets are made, neither of the results may be re- 
 jected; however widely they may differ one from the other, 
 the mean of the two must be considered the result of the experi- 
 ment when nothing is known as to their comparative value. 
 But if several briquets are made from the same sample, and 
 they vary one from another, the final result is sometimes ob- 
 tained by rejecting certain of the observations. In some cases 
 if five or six specimens are made, the highest and the lowest 
 ones are omitted, while sometimes the two lowest are rejected, 
 and the mean of the three or four highest is taken. 
 
 221. While the absolute mean of all of the observations will 
 ordinarily be quite sufficient, and should usually be considered 
 the result of the test, yet where tests are very carefully made 
 to compare two samples, or two methods of manipulation, it 
 may be desired to reject certain observations that appear to 
 be abnormal. The beginner in cement testing, unfamiliar with 
 observations of this character, may not feel confidence in his 
 own judgment as to what observations may be rejected, and the 
 criteria sometimes used in more accurate work are entirely too 
 complicated for this purpose. To serve as a guide in such 
 cases, the writer would suggest the following simple method 
 which, though entirely arbitrary, is more justifiable than either 
 of the methods mentioned above. As the experimenter be- 
 comes more familiar with the work, he will doubtless prefer to 
 
13G 
 
 CEMENT AND CONCRETE 
 
 depend on his own judgment in the rejection of observations, 
 taking into account the general accuracy of the work. 
 
 First obtain the absolute mean and the difference between 
 this mean and each individual result; let us call this difference 
 the "error" for each result. Reject any observations whose 
 error is, say, ten per cent, of the absolute mean, and obtain the 
 mean of the remaining observations as the true result. 
 
 222. For example, suppose that we have broken ten bri- 
 quets obtaining the strengths given below, and wish to deter- 
 mine the result of the test. The absolute mean is found to be 
 213.9 pounds, or, the nearest whole number, 214 pounds. 
 
 TABLE 39 
 Rejection of Observations 
 
 NUMBER 
 
 OF 
 
 BRIQUETS. 
 
 OBSERVED 
 STRENGTH. 
 
 ERROR. 
 
 OBSERVED 
 STRENGTH. 
 
 NEW 
 ERROR. 
 
 1 
 
 209 
 
 5 
 
 209 
 
 1 
 
 2 
 
 226 
 
 12 
 
 226 
 
 16 
 
 3 
 
 227 
 
 13 
 
 227 
 
 17 
 
 4 
 
 184 
 
 30 
 
 
 
 5 
 
 217 
 
 3 
 
 217 
 
 7 
 
 6 
 
 252 
 
 38 
 
 
 
 7 
 
 200 
 
 14 
 
 200 
 
 10 
 
 8 
 
 195 
 
 19 
 
 195 
 
 15 
 
 9 
 
 193 
 
 21 
 
 193 
 
 17 
 
 10 
 
 236 
 
 22 
 
 
 . . . 
 
 Sum . 
 
 2,139 
 
 177 
 
 1,467 
 
 83 
 
 Mean . . 
 
 213.9 
 
 17.7 
 
 209.6 
 
 11.9 
 
 The " errors" are given in the third column, and it is seen 
 that three of them are greater than ten per cent, of the mean. 
 Omitting the results having these large errors, we obtain a new 
 mean of 209.6 pounds, which is to be considered the result of 
 the test. An inspection of the first column of errors shows that 
 the mean of the errors is 17.7 pounds; if we divide this by the 
 mean of the tensile strengths, we obtain 17.7 * 213.9 = .0827. 
 Expressing this as a percentage, we may call 8.27 per cent, 
 the " average error." The same result is, of course, obtained 
 by dividing the sum of errors by the sum of the strengths. 
 Now if we consider column five, we see that the new average 
 error will be but 83 -*- 1467 = 5.66 per cent. 
 
I INTERPRETATION 137 
 
 223. In giving the results of a series of tests, it is a common 
 practice to state only the absolute mean, but it is of considerable 
 interest to know the variations that occurred in breaking in 
 order that one may judge of the reliability of the results, or, 
 in other words, to make a rough approximation as to the prob- 
 able error. For this purpose the highest and lowest result may 
 be given, but a much better index to reliability would be tc 
 give the " average error" as explained above. However, in 
 reporting a large number of tests, the extra labor involved in 
 obtaining this " average error" is usually considered too great 
 to be attempted, and in such cases the absolute mean and the 
 highest and lowest results must serve the purpose. 
 
 224. Accuracy Obtainable. When an operator has become 
 expert and is working under good conditions, he may expect 
 to obtain results within the following limits: The extreme varia- 
 tions between the results in a set of ten briquets (the difference 
 between the highest and lowest) not exceeding 20 per cent, of 
 the mean strength of the set, the maximum variation from the 
 mean not exceeding 12 per cent, of the mean, and the " aver- 
 age error," as explained above, not exceeding 8 per cent. 
 
 ART. 26. THE INTERPRETATION OF TENSILE TESTS OF 
 COHESION 
 
 225. One of the problems presented in the inspection of 
 cement is to foretell the ultimate relative strengths of two 
 samples from the results of short time tests. Formulas have 
 been presented purporting to solve this problem, such formulas 
 being based on the assumption that the strength gained at the 
 end of months or years is a function of that developed in a few 
 days. In fact, the raison d'etre of tensile or other short-time 
 strength tests for the acceptance of cement, rests, in a sense, 
 upon this same assumption. 
 
 The value of strength tests as one of the guides in determin- 
 ing in a short time the probable quality of a cement is unques- 
 tioned. One is apt, however, to seek too close an agreement 
 between the results of such tests and the actual quality of the 
 cement. It would be easy to select examples illustrating the 
 harmony between short and long time tests; but it will be of 
 greater value to show, rather, some of the many exceptions to 
 such a rule, and thereby emphasize the fact that it is only by 
 
138 
 
 CEMENT AND CONCRETE 
 
 a close analysis of all of the information obtainable concerning 
 a sample, and a general knowledge of the behavior of the dif- 
 ferent grades of cement, that one may hope to arrive at a tol- 
 erably accurate opinion. 
 
 226. Comparative Tests of Portland Cements. In Table 
 40 are given the results of tests on four brands of Portland 
 cement at seven days, twenty-eight days and two years. From 
 the tests at two years it appears that T and U are the best 
 cements, V is nearly as good, but W gives a much lower result. 
 Turning now to the seven and twenty-eight day tests of bri- 
 quets maintained at the ordinary temperature, it is seen that 
 W gave in every case' higher results than T, and nearly as high 
 as U or V. Among the short time tests it is only the results 
 of briquets maintained at 80 C. that indicate the inferiority 
 of Brand W. 
 
 TABLE 40 
 
 Interpretation of Short Time Tests of Portland Cement, Several 
 
 Brands 
 
 
 
 
 TENSILE STRENGTH, POUNDS 
 
 PARTS 
 SAND TO 1 
 CEMENT 
 
 BY 
 
 TEMPERA- 
 TURE WATER 
 
 OF 
 
 AGE OF 
 BRIQUETS. 
 
 PER SQUARE INCH. 
 
 Brand. 
 
 WEIGHT. 
 
 IMMERSION. 
 
 
 
 
 
 
 
 
 
 T 
 
 U 
 
 V 
 
 W 
 
 2 
 
 Hot, 80 C. 
 
 7 days 
 
 339 
 
 278 
 
 284 
 
 222 
 
 3 
 
 a a 
 
 ' 
 
 221 
 
 191 
 
 180 
 
 134 
 
 3 
 
 " 60 C. 
 
 ( 
 
 144 
 
 142 
 
 169 
 
 144 
 
 
 
 Ordinary 
 
 ' 
 
 420 
 
 510 
 
 487 
 
 565 
 
 1 
 
 
 c 
 
 327 
 
 425 
 
 400 
 
 396 
 
 2 
 
 
 ( 
 
 172 
 
 275 
 
 256 
 
 236 
 
 3 
 
 
 t 
 
 73 
 
 150 
 
 160 
 
 150 
 
 1 
 
 
 28 days 
 
 526 
 
 577 
 
 557 
 
 556 
 
 2 
 
 
 ' 
 
 312 
 
 394 
 
 387 
 
 332 
 
 3 
 
 
 
 
 142 
 
 241 
 
 223 
 
 247 
 
 1 
 
 
 2 years 
 
 719 
 
 753 
 
 763 
 
 654 
 
 2 
 
 
 ' 
 
 554 
 
 553 
 
 513 
 
 407 
 
 3 
 
 
 
 380 
 
 373 
 
 340 
 
 287 
 
 227. Comparative Tests of Natural Cements. From the 
 nature of natural cements a much greater variation in strength 
 among different brands, and even among different samples of 
 the same brand, is to be expected. With Portland cements 
 made in accordance with ordinary methods, the variations in 
 strength among ten or twenty brands will usually be compara- 
 tively small. One of them may possibly prove unsound, and 
 
1NTERPRETA T10N 
 
 139 
 
 one or two others may give inferior strength, but the variations 
 in strength among three-fourths of the samples will not gener- 
 ally exceed 20 per cent. With the same number of brands of 
 natural cements, variations of 50 to 200 per cent, may be 
 expected. 
 
 TABLE 41 
 
 Interpretation of Short Time Tests of Natural Cement, Several 
 
 Brands 
 
 
 
 TENSILE STRENGTH, POUNDS 
 
 
 
 
 PKK SQUARE INCH. 
 
 PARTS 
 SANI> TO 1 
 CEMENT. 
 
 TEMPERA- 
 TITRE WATKK 
 
 OF 
 
 A<;rc <K 
 BRIQUETS. 
 
 
 Brand. 
 
 
 I M M K K S I ( ) N . 
 
 
 Jn 
 
 Hn 
 
 Bn 
 
 Mn 
 
 Nn 
 
 Kn 
 
 2 
 
 Hot, 50 C. 
 
 7 days 
 
 152 
 
 192 
 
 84 
 
 133 
 
 160 
 
 277 
 
 2 
 
 Hot, 60 C. 
 
 
 170 
 
 270 
 
 79 
 
 154 
 
 164 
 
 254 
 
 2 
 
 Hot, 80 C. 
 
 
 58 
 
 136 
 
 128 
 
 179 
 
 16(5 
 
 221 
 
 
 
 Ordinary 
 
 
 174 
 
 203 
 
 130 
 
 189 
 
 210 
 
 189 
 
 1 
 
 
 
 125 
 
 108 
 
 103 
 
 164 
 
 169 
 
 164 
 
 
 
 
 28 days 
 
 208 
 
 344 
 
 25>3 
 
 203 
 
 316 
 
 289 
 
 1 
 
 
 
 237 
 
 342 
 
 247 
 
 247 
 
 252 
 
 385 
 
 2 
 
 
 
 132 
 
 223 
 
 148 
 
 158 
 
 184 
 
 217 
 
 3 
 
 
 
 64 
 
 113 
 
 85 
 
 93 
 
 104 
 
 101 
 
 1 
 
 
 2 years 
 
 177 
 
 271 
 
 358 
 
 631 
 
 065 
 
 532 
 
 2 
 
 
 
 106 
 
 157 
 
 195 
 
 515 
 
 550 
 
 561 
 
 3 
 
 
 
 99 
 
 130 
 
 117 
 
 340 
 
 328 
 
 372 
 
 In Table 41 six brands of natural cement are compared by 
 tests at seven days, twenty-eight days and two years. These 
 six brands have been arranged in the table according to their 
 value as shown by the two year tests, and it is seen that the 
 first three, Jn, Hn and Bn, are especially poor, while the last 
 three, Mn, Nn and Kn, are exceptionally good. In the short 
 time tests of briquets maintained at ordinary temperature, Jn 
 and Bn gave low results and Nn and Kn gave fairly high results, 
 in harmony with the long time tests; but Hn, which proved to 
 be one of the poorest samples, gave in every case the highest, 
 or next to the highest, result in seven and twenty-eight day 
 cold tests. In this table we find again that the results of the 
 briquets maintained at 80 C. for seven days gave, in a general 
 way, the best indication of the relative values of the six brands. 
 
 228. Several Samples of One Brand. To show that short 
 time tests do not always indicate the relative values of several 
 samples of cement, even when all of the samples are of the 
 
140 
 
 CEMENT AND CONCRETE 
 
 same brand, Tables 42 and 43 are given. All of the results in 
 these tables are from samples of the one brand of natural ce- 
 ment. 
 
 TABLE 42 
 
 Comparison of Short and Long Time Tests of Samples of One 
 Brand of Natural Cement 
 
 SERIES. 
 
 SAND. 
 
 AGE. 
 
 
 TENSILE STRENGTH, LBS. 
 PER SQUARE INCH. 
 
 Kind. 
 
 Parts to 
 1 Cement. 
 
 A 
 
 
 
 
 
 28 days 
 6-7 months 
 
 Number 
 of Samples 
 Tested. 
 
 3 
 
 7 
 
 2 
 
 3 
 
 5 
 
 
 
 84 
 121 
 
 123 
 
 186 
 
 177 
 241 
 
 220 
 301 
 
 297 
 381 
 
 B 
 
 Std. 
 
 1 
 1 
 
 7 days 
 6 months 
 
 Number 
 Samples. 
 
 17 
 
 20 
 
 74 
 
 4(58 
 
 50 
 
 17 
 
 16 
 
 
 
 62 
 462 
 
 86 
 442 
 
 146 
 
 367 
 
 
 C 
 
 P.P. 
 
 u 
 
 u 
 
 1 
 
 1 and 2 
 
 2 
 
 7 days 
 6 months * 
 7 days 2 
 
 Number 
 Samples. 
 
 50 
 
 19 
 
 19 
 
 
 
 
 49 
 473 
 
 273 
 
 54 
 426 
 249 
 
 73 
 381 
 
 265 
 
 128 
 321 
 283 
 
 
 
 D 
 
 P.P. ' 
 
 
 2 
 
 2 
 
 7 days 
 1 year 
 
 7 days 2 
 
 Number 
 Samples. 
 
 13 
 
 48 
 
 38 
 
 18 
 
 
 
 
 66 
 473 
 
 257 
 
 80 
 422 
 234 
 
 95 
 377 
 
 277 
 
 147 
 325 
 215 
 
 . . 
 
 E 
 
 . . . 
 
 
 
 2 
 
 7 days 
 6 months 
 
 Number 
 Samples. 
 
 12 
 
 21 
 
 18 
 
 9 
 
 
 
 
 74 
 535 
 
 83 
 
 477 
 
 120 
 424 
 
 167 
 373 
 
 
 
 F 
 
 ( Cr.Qtz. ) 
 \ 20 to 40 ( 
 
 
 2 
 
 28 days 
 ( 6 months ) 
 ( and 1 year ] 
 
 Number 
 Samples. 
 
 287 
 
 170 
 
 41 
 
 
 
 
 
 135 
 565 
 
 191 
 454 
 
 235 
 
 367 
 
 
 . * 
 
 1 Mean one-to-one and one-to-two mortars. 
 
 2 Briquets immersed six days in water maintained at 60 C. 
 
INTERPRET A TION 141 
 
 In Table 42 the results are selected from a large number of 
 tests of this brand, and are arranged in groups according to the 
 strength shown at a certain age. For instance, in Series A 
 the results of twenty samples are given, arranged according to 
 the strength at twenty-eight days. Three of the samples gave 
 less than 100 pounds per square inch, neat, at twenty-eight 
 days; the same three samples gave a mean strength of 121 
 pounds per square inch, neat, six to seven months. Seven 
 samples, the strength of which fell between 100 and 150 pounds 
 at twenty-eight days, gave a mean strength of 186 pounds at 
 six to seven months. The results of this series show the 
 harmony between short and long time tests when it is a question 
 of comparing neat cement mortars. 
 
 In Series D of this table the samples are arranged in order 
 according to the strength developed by one-to-two mortars 
 one year old. Thirteen samples had a strength at this age of 
 between 450 and 500 pounds, average 473 pounds. The same 
 samples gave but 66 pounds, neat, seven days. Forty-eight 
 samples, giving between 400 and 450 pounds, average 422 
 pounds, gave but 80 pounds, neat, seven days, while eighteen 
 samples that developed only 300 to 350 pounds mean, 325 
 pounds at one year, showed a mean strength of 147 pounds, neat, 
 seven days. 
 
 A little study of this table will show that the samples which 
 were comparatively weak in seven and twenty-eight day tests, 
 either neat or with sand, gave the best results in the long time 
 tests of sand mortars. Series A shows that the neat tests at 
 seven days and at six months are consistent, but in all cases 
 where sand mortars are tested at six months to one year, the 
 highest results are given by the samples showing the lowest 
 strength in the short time tests in cool water. It is very sel- 
 dom that this conclusion has not been indicated by the author's 
 tests of this brand. It is not invariably true, however, for 
 some samples which were selected as being defective in burn, 
 gave low results both in short and long time tests. The con- 
 clusion stated above must therefore be understood to have 
 limits even for this brand, and may not apply at all to many 
 brands. 
 
 As to the results of short time tests of briquets stored in hot 
 water, Series C and D indicate that such results are more nearly 
 
142 
 
 CEMENT AND CONCRETE 
 
 consistent with the long time tests, yet it is evident that even 
 with hot tests one could not readily and accurately differen- 
 tiate the best from the mediocre samples. 
 
 TABLE 43 
 
 Natural Cement : Rate of Increase in Strength, Hardening in "Water 
 
 and Dry Air 
 
 
 
 TENSILE STRENGTH PER SQ. IN., OF SAMPLES. 
 
 SAND, PARTS 
 TO ONE 
 CEMENT. 
 
 AGE OF 
 BRIQUETS 
 WHEN BROKEN. 
 
 Hardened in Water. 
 
 Hardened in Air of 
 Itoom . 
 
 
 
 84 
 
 U' 
 
 O' 
 
 84 
 
 U' 
 
 0' 
 
 1 
 
 7 days. 
 
 74 
 
 53 
 
 103 
 
 107 
 
 68 
 
 187 
 
 1 
 
 28 days. 
 
 228 
 
 189 
 
 228 
 
 188 
 
 95 
 
 256 
 
 1 
 
 3 111 os. 
 
 415 
 
 345 
 
 331 
 
 158 
 
 100 
 
 248 
 
 1 
 
 mos. 
 
 500 
 
 381 
 
 307 
 
 425 
 
 161 
 
 359 
 
 1 
 
 2 years. 
 
 446 
 
 383 
 
 209 
 
 151 
 
 147 
 
 403 
 
 3 
 
 28 days. 
 
 99 
 
 97 
 
 64 
 
 112 
 
 61 
 
 180 
 
 3 
 
 3 mos. 
 
 244 
 
 241 
 
 129 
 
 153 
 
 81 
 
 194 
 
 3 
 
 mos. 
 
 255 
 
 232 
 
 162 
 
 92 
 
 69 
 
 173 
 
 3 
 
 1 year. 
 
 274 
 
 264 
 
 186 
 
 229 
 
 70 
 
 144 
 
 3 
 
 2 years. 
 
 258 
 
 268 
 
 167 
 
 274 
 
 152 
 
 228 
 
 Sample 
 
 Fineness : Per cent, passing Sieve No. 120, Holes 
 
 .0046 inch square 
 
 Time Setting to bear ^" Ib. Wire, min. . . . 
 Specific Gravity 
 
 84 U' O' 
 
 80.5 87.8 89.7 
 
 54 23 97 
 3.012 2.950 3.145 
 
 U', underburned, O', overburned. All samples same brand, Gn. 
 
 229. The results in Table 43 will serve to illustrate the same 
 point by showing the very different rates of increase in strength 
 of three samples when the briquets are stored in water and in 
 dry air. One of these samples, 84, was taken at random from a 
 shipment, while U' and O' were supposed to be defective in 
 burn. Of the water-hardened specimens, No. 84 gained in 
 strength up to six months or one year and then suffered only 
 a slight falling off. The underburned sample showed a con- 
 tinuous gain, but the overburned cement showed a marked 
 decrease in strength after six months or one year. The air- 
 hardened specimens were very irregular in strength, but the 
 underburned sample gave very low results throughout. 
 
 Table 44 gives similar results obtained with several samples, 
 the briquets being hardened in water as usual, 16 R is a fair 
 
INTERPRETATION 
 
 143 
 
 sample of the best cement of this brand, and its rate of increase 
 in strength with one to three parts sand is shown. Samples 
 M and L were tested together, as were CC and DD. M and 
 CC are of the class giving comparatively high results at seven 
 days, while L and DD give high results at seven days, but 
 develop only a moderate ultimate strength. 
 
 TABLE 44 
 
 Natural Cement: Difference in Rates of Increase in Strength of 
 Several Samples of the Same Brand 
 
 H 
 
 CKMKXT. 
 
 SAX i). 
 
 TEXSILK STKKXGTH, POUNDS IKK Scj. Jx. 
 AT AGK OF 
 
 s 
 
 
 
 
 !*> . 
 
 
 
 
 
 
 
 
 
 H 
 
 -3 
 
 p2 
 
 
 4- ~ 
 
 03 
 
 
 
 s 
 
 
 
 V) 
 
 i* 
 
 2 
 
 
 H 
 *v 
 
 el 
 
 w 
 
 3 
 
 CC 
 
 Kind. 
 
 |l! 
 
 ce 
 o 
 
 c 
 ^ 
 
 a 
 
 S 
 
 CO 
 
 I 
 
 ! 
 
 CO 
 
 1 
 
 ( 11 
 
 16 R 
 
 Crushed Qtz. 20-30 
 
 1 
 
 04 
 
 142 
 
 334 
 
 300 
 
 430 
 
 500 
 
 445 
 
 
 2 
 
 
 u 
 
 u 
 
 2 
 
 50 
 
 101 
 
 280 
 
 341 
 
 335 
 
 386 
 
 354 
 
 
 3 
 
 
 (t 
 
 u 
 
 3 
 
 
 73 
 
 204 
 
 243 
 
 252 
 
 268 
 
 262 
 
 248 
 
 4 
 
 
 1ST 
 
 
 
 
 118 
 
 100 
 
 
 ^56 
 
 ?48 
 
 300 
 
 
 
 5 
 
 
 T, 
 
 
 
 
 40 
 
 88 
 
 
 148 
 
 146 
 
 167 
 
 
 
 6 
 
 
 M 
 
 Point aux Pins 
 
 2 
 
 63 
 
 155 
 
 
 216 
 
 941 
 
 
 
 
 7 
 
 
 L 
 
 
 2 
 
 30 
 
 150 
 
 
 206 
 
 415 
 
 360 
 
 
 
 8 
 
 
 CC 
 
 t 
 
 1 
 
 123 
 
 232 
 
 
 276 
 
 260 
 
 
 317 
 
 
 9 
 
 
 DD 
 
 i 
 
 1 
 
 77 
 
 218 
 
 
 327 
 
 337 
 
 
 474 
 
 
 10 
 
 
 CC 
 
 i 
 
 2 
 
 
 185 
 
 
 268 
 
 242 
 
 270 
 
 270 
 
 
 It 
 
 
 DD 
 
 i 
 
 2 
 
 
 180 
 
 
 326 
 
 303 
 
 373 
 
 350 
 
 
 230. Conclusions. From the above tables one should not 
 draw the conclusion that all strength tests are valueless be- 
 cause likely to be misleading. Some lessons, however, seem to 
 be plain; conclusions drawn from the results of short time tests 
 of strength alone are likely to be far from infallible. This is 
 especially true of natural cements. The correctness of one's 
 conclusions concerning the value of a sample is likely to de- 
 pend very much upon his knowledge of the behavior of that 
 particular brand, and the beginner in cement testing should 
 not have too great confidence in his early conclusions. Samples 
 under inspection should be tested in comparison with other 
 samples of known quality, and the results of the strength tests 
 studied in connection with all the information obtainable from 
 the other tests of quality already outlined. 
 
CHAPTER X 
 
 
 
 THE RECEPTION OF CEMENT AND RECORDS OF TESTS 
 
 ART. 27. STORING AND SAMPLING 
 
 231. STORAGE. The storage houses provided for the ce- 
 ment should be such as will effectually preserve it from damp- 
 ness, the floor being dry and strongly built. A circulation of 
 air under the floor will insure dryness. 
 
 In building houses for storage, due regard should be given 
 to the ease of getting the cement in and out, and facilities pro- 
 vided for the use of block and tackle in tiering. 
 
 When the cement is received, whether in sacks or barrels, it 
 should, if possible, be so tiered in the warehouse that any pack- 
 age is accessible for sampling. In the case of barrels this may 
 readily be attained by tiering in double rows, the barrels lying 
 on the side. It has been found that ordinary cement barrels 
 will withstand the pressure if tiered five high with a " binder" 
 row on top; and when so piled, a warehouse 32 feet wide and 
 100 feet long will readily hold 2,200 barrels, an allowance of 
 about one hundred fifty square feet of floor space for one hun- 
 dred barrels. 
 
 232. Where storage space is limited, the barrels may be 
 numbered and sampled before they are placed in the warehouse, 
 and they may then be piled solid, but this should be avoided 
 if practicable. Sacks cannot be quite so neatly stored, and 
 since a smaller quantity is contained in a sack, they may be 
 tiered so that every third or fourth sack is accessible. It is 
 desirable where work is executed with the greatest care that 
 every package be numbered for future identification, but this 
 may sometimes prove impracticable, especially when the ce- 
 ment is in sacks, and in such cases the sampled packages only 
 may receive numbers. 
 
 233. Percentage of Barrels to Sample. The amount of ce- 
 ment which shall be accepted on the test of a single sample 
 must be determined by each user of cement according to his 
 
 144 
 
STORING AND SAMPLING 145 
 
 knowledge as to the uniformity and reliability of the brand in 
 use, and according to the character of the work in which the 
 cement is to be used. In a few isolated cases every barrel is 
 tested, while sometimes several tons of cement are accepted on 
 a single test. As the improvements in methods have decreased 
 the work involved in making the simpler tests, the tendency 
 has been to test a larger percentage of the packages. 1 
 
 The report of the committee of the Amer. Soc. C. E. in 
 1SS5, contains the following concerning sampling: " There is no 
 uniformity of practice among engineers as to the sampling 
 of the cement to be tested, some testing every tenth barrel, 
 others every fifth, and others still every barrel delivered. Usu- 
 ally, where cement has a good reputation, and is used in large 
 masses, such as concrete in heavy foundations, or in the back- 
 ing or hearting of thick walls, the testing of every fifth barrel 
 seems to be sufficient; but in very important work, where the 
 strength of each barrel may in great measure determine the 
 strength of that portion of the work where it is used, or in 
 the thin walls of sewers, etc., every barrel should be tested, 
 one briquet being made from it." 
 
 234. Taking the Sample. The sample should be taken in 
 such a manner as to fairly represent the package, and for this 
 purpose a " sugar trier" may be used, by which is obtained a 
 core of cement about one inch in diameter and eighteen inches 
 long. As any tool used for boring cement barrels soon becomes 
 dull, and as a sugar trier is somewhat difficult to sharpen, the 
 author prefers to use an ordinary bit and brace to penetrate 
 the barrel head, and then extract the sample with a " trier," 
 or a long, slender scoop of similar form provided with a handle. 
 
 For storing the sample until it is tested, it has been found 
 convenient to use covered tin cans holding about one pint, 
 the cover of the can being labeled with the number of the pack- 
 age from which the sample is taken. 
 
 1 In a paper read before the Institution of Civil Engineers in 1865-66, Mr. 
 John Grant states that " after using, during the last six years, more than 
 70,000 tons of Portland cement, which has been submitted to about 15,000 
 tests, it can be confidently asserted that none of an inferior or dangerous 
 character has been employed in any part of the work in question. " (The 
 Metropolitan Main Drainage, London.) This is an average of one test to 
 twenty-five barrels. 
 
146 CEMENT AND CONCRETE 
 
 ART. 28. RECORDS OF TESTS 
 
 235. Value of Records. In conducting work in which the 
 use of cement enters as a prominent factor, it is not only neces- 
 sary to know that the cement used is of a good quality, but also 
 to be able to show at any future time what tests were made to 
 establish its value. This fact, as well as the convenience of 
 the work, demands that a record shall be kept of all the tests 
 made. These records may be more or less elaborate, according 
 to the kind and amount of the work in hand, but in any case, 
 enough detail should be given to make them intelligible to other 
 engineers. 
 
 236. Marking Specimens. There is sometimes a tempta- 
 tion, in making tensile specimens, to stamp upon them many 
 details of the test, and for this purpose an elaborate cipher 
 system has sometimes been used. But this method is to be 
 strongly deprecated. Each briquet should receive its proper 
 consecutive number, as mentioned in 189, and the details 
 concerning it should be placed in the record book. 
 
 237. RECORDS KEPT AT ST. MARYS FALLS CANAL. In the 
 tests of cement at St. Marys Falls Canal, during the construc- 
 tion of the Poe Lock, a system of records was used that gave 
 entire satisfaction. At the time the largest amount of cement 
 was being used three molders were employed, each making 
 fifty briquets per day of eight hours. Over one hundred thou- 
 sand briquets were made in five and one-half years. Although 
 the system of records used at this point may be more elaborate 
 than is often necessary, yet the system will be described, and 
 certain modifications will be suggested for places requiring less 
 complete records. 
 
 238. Barrel Records. The barrels receive consecutive num- 
 bers after they are tiered up in the warehouse. The " receiv- 
 ing book" is a simple transit book in which are entered the 
 date of the receipt of each cargo, the name of the boat (or the 
 car number, if shipped by rail), the brand of cement, the num- 
 ber of barrels, the first and last barrel number of the cargo and 
 the warehouse in which the cement is placed. The next book 
 to be used is the " barrel book," in which the numbers of the 
 barrels are entered consecutively in a column at the left, each 
 barrel being given one line. This book is also of transit size, 
 but might well be larger. The headings are given below. 
 
RECORDS OF TESTS 
 
 147 
 
 SAMPLE PAGE OF "BARREL RECORD" 
 
 No. 
 BBL. 
 
 SAMPLED. 
 
 DEFECTS. 
 
 AC- 
 CEPTED. 
 
 RK- 
 
 JECTED. 
 
 ISSUED. 
 
 RE.MABKS. 
 
 88251 
 
 2 
 3 
 
 4 
 
 5 
 6 
 
 8 
 o 
 
 88260 
 
 1 
 
 
 t-j 
 
 3 
 
 4 
 5 
 
 M. D. 
 
 5 10 
 
 M. 1). 
 
 6 5 
 
 S7 = 4Sf 
 
 M. D, 
 
 G 13 
 
 M. D. 
 
 July 25 
 July 25 
 July 25 
 
 Sept. 27 
 
 July 25 
 July 25 
 July 25 
 July 25 
 July 25 
 
 67 = 65# 
 
 5 19 
 
 G 6 
 G 5 
 
 
 G 13 
 
 6 12 
 
 S7 = 102$ 
 j Removed by 
 I Contractor 
 
 ( 87= 32# I 
 I S7 = 35$ ] 
 
 
 
 
 
 
 
 5 19 
 
 
 
 5 26 
 
 - 
 
 
 
 
 
 
 
 
 
 
 
 5 19 
 
 
 
 5 26 
 
 
 
 July 25 
 July 25 
 July 25 
 
 Sept. 28 
 
 July 25 
 July 25 
 
 
 
 5 19 
 
 6 
 6 5 
 
 
 6 13 
 
 6 12 
 
 S7 = 105# 
 \ Removed by 
 I Contractor 
 
 
 
 j S7= 40# I 
 
 } S7 = 321 ( 
 
 
 
 
 \ : : : 
 
 
 
 
 
 
 
 When the barrels are sampled and briquets made, the date 
 sampled is entered in the second column of the barrel record 
 book. The other columns will be explained later. 
 
 239. Holders' Records. Separate sheets of paper ruled and 
 headed as shown on page 148 are used by the molders to record 
 the details concerning the making of briquets. 
 
 Separate sheets properly ruled and headed are also given to 
 the assistants who test time of setting and fineness. These 
 record sheets, when filled in by the assistants, are copied the 
 following day by the bookkeeper, into the permanent " record 
 book." At the end of the month these separate sheets, con- 
 taining original records of work done, are folded and filed for 
 future reference. 
 
 240. Briquet Record. The briquets are made in sets of 
 ten for convenience. Each set is given a page in the " record 
 book/' as is indicated on page 149 where the form for this book 
 is given. The size of page is 9 by 12 inches. Paper having the 
 same ruling and column headings is convenient for reporting 
 tests to the chief engineer. 
 
 241. Summary Book. The data for each set of briquets 
 are copied from the record book, in a condensed form, into the 
 " summary book," one line of the latter containing a page of 
 the former. In the summary book each brand is given a few 
 
148 
 
 CEMENT AND CONCRETE 
 
 axvci 
 
 to ^ 
 
 '8681 k 9 
 
RECORDS OF TESTS 
 
 149 
 
 g 
 
 so" 
 
 fl 
 
 'HXOX3HXS 
 
 IN MORTAR 
 GHT. 
 
 TIME OF SETTING NE 
 CEMENT. 
 
 S3 aT- 
 
 H 
 
 fi.pnop JdiifDdjft 
 'utoou, 
 
 001 ' 
 
 HA3Ig 
 
 '3A3IS 
 
 x oxissvj; 
 
 CEM 
 
 ir iHa jo ' 
 
150 
 
 CEMENT AND CONCRETE 
 
 8-Hoixvoi.si03.ig 
 
 PROPORTIONS IN 
 MORTAR BY WT. 
 
 
 to GO o 
 
 sxanoiaa 
 
 sxanorag 
 
 CORD 
 OOK. 
 
RECORDS OF TESTS 
 
 151 
 
 pages by itself, so that this book corresponds to a ledger in form. 
 By this means a large number of tests on the same brand may 
 be looked over at once. The summary book might be omitted 
 where a smaller number of tests are to be made, or it might 
 be slightly modified and take the place of the record book. A 
 sample page is given below. 
 
 242. Records of Fineness, Time of Setting and Soundness. - 
 Although provision is made in the record book for recording 
 time of setting and fineness, it has been found that where a 
 large amount of cement is being tested it is more convenient 
 to have separate books for each test. Especially is this true 
 as it has been judged necessary to test but a very small per- 
 centage of the barrels for fineness, while a larger percentage of 
 the barrels are tested for time of setting and soundness. The 
 " fineness book" is as simple as possible and need not be illus- 
 trated. A sample page of the "pat book" is given below. 
 
 Sample Page of "Pat Book" or Record of Time of Setting and 
 
 Soundness 
 
 Lagerdorfer Portland Cement Pats, Two from Ever// Third liarrel 
 
 
 r.> 
 
 M 
 
 2* 
 
 & 
 
 |2 
 
 u 
 
 
 
 
 
 
 x* 
 
 i* ^. 
 <~ 
 
 cc 
 
 
 
 TREATMENT 
 OF PATS. 
 
 EXAMINKH. 
 
 RI:MOVEI>. 
 
 
 No. 
 BBL. 
 
 ^ *<1 
 
 J(^ 
 
 *1 /( 
 
 - 
 
 S a" 
 
 O 'A 
 H 3S 
 
 H Si 
 
 
 
 
 RE- 
 MARKS. 
 
 
 a 10 
 
 3* 
 
 O3 
 
 s l 
 
 Is 
 
 |J 
 
 Stmr. 
 
 Tank. 
 
 Date. 
 
 Condi- 
 tion. 
 
 Date. 
 
 Condi- 
 tion. 
 
 
 
 
 
 
 
 
 
 Mo. J). 
 
 
 Mo. J). 
 
 
 
 110274 
 
 7 
 
 9:23 
 
 777 
 
 557 
 
 
 
 8 1 
 
 O.K. 
 
 7 15 
 
 O.K. 
 
 Water 
 
 
 # 
 
 
 
 
 
 
 
 
 8 17 
 
 Ll 
 
 24%. 
 
 7 
 
 7 
 
 9:28 
 
 57 
 
 232 
 
 S^ 
 
 
 8 1 
 
 O.K. 
 
 7 15 
 
 a 
 
 
 
 
 
 
 
 
 S o 
 
 
 
 
 8 17 
 
 it 
 
 
 80 
 
 7 
 
 9:32 
 
 75 
 
 268 
 
 73"^ 
 
 1 
 
 S ' 1 
 
 O.K. 
 
 7 15 
 
 it 
 
 
 
 1? 
 
 
 
 
 * < 
 
 < 
 
 
 
 S 17 
 
 it 
 
 
 3 
 
 7 
 
 9:34 
 
 7^5 
 
 383 
 
 \\ 
 
 1 
 S 
 
 S 1 
 
 O.K. 
 
 7 15 
 
 f Surface 
 -' cracked 
 
 
 
 f 
 
 
 
 
 ~ w 
 
 ^ 
 
 
 
 8 17 
 
 l o.k e . ' 
 
 
 6 
 
 7 
 
 9:41 
 
 24 
 
 259 
 
 'i s 
 
 S 
 
 8 1 
 
 'O.K. 
 
 7 15 
 
 * 
 
 
 . . . 
 
 2 
 
 
 
 
 ?* ** 
 
 's 
 
 
 
 8 17 
 
 
 
 
 9 
 
 1 
 
 9:50 
 
 15 
 
 250 
 
 g^c 
 
 S 
 
 8 ' 1 
 
 O.K. 
 
 7 15 
 
 u 
 
 
 
 & 
 
 
 
 
 ^ to 
 
 Is 
 
 
 
 8 17 
 
 4 ^ 
 
 
 '92 
 
 7 
 
 9:53 
 
 12 
 
 247 
 
 J 
 
 
 8 1 
 
 O.K. 
 
 7 15 
 
 I ; 
 
 
 . 
 
 # 
 
 
 m 
 
 
 ^ ** 
 
 o 
 
 
 
 8 17 
 
 It 
 
 
 95 
 
 7 
 
 9:57 
 
 123 
 
 363 
 
 o2 
 
 ^i 
 
 8 1 
 
 O.K. 
 
 7 15 
 
 K 
 
 
 
 
 
 
 
 
 [> 
 
 . 
 
 
 
 8 17 
 
 (1 
 
 
 '98 
 
 7 
 
 10:28 
 
 '97 
 
 367 
 
 . 8 
 
 
 
 8 1 
 
 O.K. 
 
 7 15 
 
 41 
 
 
 
 * 
 
 
 
 
 e ^ 
 
 
 
 
 8 17 
 
 (I 
 
 
 301 
 
 7 
 
 10:32 
 
 103 
 
 365 
 
 H S 
 
 
 S 1 
 
 O.K. 
 
 7 15 
 
 t t 
 
 
 
 
 * 
 
 
 
 
 
 
 
 
 8 17 
 
 t i 
 
 
 
 
152 
 
 CEMENT AND CONCRETE 
 
 t 
 
 A 
 
 * 1 
 
 O I 
 
 s * 
 
 a> 
 
 M I 
 
 . I 
 
 M S 
 
 o 
 
 o ^* 
 
 B ^ 
 
 S f 
 
 H 5' 
 
 ID 
 
 ** 
 
 H , 
 
 jracnoj\[ 
 
 cc 
 
 QQ 
 
 1894101 
 
 fe- - - - 
 
 8UQ <*) S!aB J 
 
 a 
 
 XX 
 
 Xlft 
 
 Sf 8 
 
 U5 ^C) 
 
RECORDS OF TESTS 153 
 
 243. The Diary. When the bookkeeper has copied the 
 data contained on the record blanks into the record book and 
 summary book, he turns to the proper page in the diary and 
 records the briquets to be broken. Thus, if briquets made 
 May 17th are to be broken at three months, he enters the num- 
 bers of these briquets and the tank in which they have been 
 placed under the date Aug. 17th. This leaves no chance of 
 allowing briquets to go beyond the proper time of breaking. 
 
 244. Acceptance or Rejection. If all of the tests on a given 
 sample are satisfactory, the date of acceptance is placed in the 
 proper column of the barrel book. It only remains then to 
 mark the barrels "O. K.," and issue them when needed, placing 
 the date of issue in the column indicated. If, however, some 
 of the tests have given unsatisfactory results, the failure is 
 noted in the " defects" column of the barrel book, and the 
 barrel is resampled to determine whether the failure was due to 
 faulty manipulation. If finally rejected, the barrel is promi- 
 nently marked to prevent its being issued for use. 
 
 It is seen from the above that the history of each barrel is 
 given in the barrel book, and the record of any brand is given 
 in a condensed form in the summary book. 
 
 245. Special Tests. When special tests are made to inves- 
 tigate the effects of variations in manipulation, or for any 
 other special purpose, such as to test the value of certain kinds 
 of sand, it becomes convenient to have still another form which 
 may be called a " series book." In this the results are so ar- 
 ranged that they may be studied for conclusions, and tables 
 for reports may be copied directly from it. A sample form is 
 given on preceding page. Should extra rulings be needed, 
 they may be placed at the right in the "remarks" column. 
 
PART III 
 
 PREPARATION AND PROPERTIES OF 
 MORTAR AND CONCRETE 
 
 CHAPTER XI 
 
 SAND FOR MORTAR 
 
 246. Mortar. When cement is mixed with sand and water, 
 the resulting paste is called mortar. The term. " neat cement 
 mortar" is sometimes used to designate a cement paste with- 
 out sand, but when the term mortar is not qualified, it refers 
 to the mixture containing sand. The primary function of mor- 
 tar is to bind together pieces of stone of greater or less size, 
 though it is sometimes used alone to prevent the percolation of 
 water, to make a smooth exterior finish, or in places too confined 
 to permit of placing concrete. 
 
 There are comparatively few cases in which it is judicious 
 to use cement without the addition of sand, for such an ad- 
 mixture not only cheapens the mortar, but actually improves 
 it for nearly all purposes. The quality of sand used is only 
 second in importance to the quality of the cement. Indeed, if 
 one does not know how to select either a good cement or a good 
 sand, he is in greater danger of going amiss in the selection of 
 the latter than the former; for the cement has been placed upon 
 the market by a manufacturer who has a reputation to estab- 
 lish or maintain. 
 
 ART. 29. CHARACTER OF THE SAND. 
 
 247. Various kinds of rock are capable of producing sand of 
 good quality. The natural sands are usually siliceous in char- 
 acter, but calcareous sands are also met with and may give 
 excellent results in mortar. Good artificial sand may be made 
 from almost any kind of rock that is not liable to chemical 
 
 154 
 
CHARACTER OF SAND 
 
 155 
 
 decay, even though it be only moderately hard. One of the 
 most essential features of a good sand is that the grains should 
 be perfectly sound. Evidences that chemical decay is going 
 on in the grains would indicate that the sand is of very inferior 
 quality. 
 
 248. SHAPE AND HARDNESS OF THE GRAINS. it is gener- 
 ally believed that the grains of sand should be angular in order 
 to give the best results; this is probably true, although in test- 
 ing three varieties of calcareous sand, M. Paul Alexandre * ob- 
 tained results which seemed to indicate that if rounded grains 
 are disadvantageous, the other properties of the sand may 
 readily counterbalance this disadvantage. 
 
 M. Alexandre used three sands which were reduced to the 
 same fineness by sifting into different sizes and then remixing 
 them in fixed proportions (equal parts of five sizes). The three 
 sands were, 1st, white marble, very hard with sharp corners; 
 2d, moderately hard limestone; and 3d, chalk, very soft with 
 rounded grains. The proportions used were 400 kg. of cement 
 to one cubic meter of sand, the amount of water varying from 
 twenty-five to thirty per cent, of the sand, according to the 
 amount required to produce plasticity. The tensile strength of 
 the mortars, in pounds per square inch, is given in Table 45. 
 
 TABLE 45 
 Results Obtained with Three Varieties of Calcareous Sand 
 
 CHARACTER OF SAND. 
 
 TENSILE STRENGTH, POUNDS 
 PEU SQUARE INCH AT 
 
 7 da. 
 
 28 da. 
 
 6 mo. 
 
 H y. 
 
 1. Marble 
 
 45 
 72 
 
 86 
 
 107 
 148 
 120 
 
 171 
 222 
 
 205 
 
 220 
 
 256 
 252 
 
 
 3 Chalk 
 
 
 As these sands varied in the structure and hardness as well 
 as in the shape of the grains, it cannot be concluded that rounded 
 grains are as good as sharp and angular ones for mortar-making. 
 There is little question that if two samples of pure quartz sand, 
 
 1 "Recherches Experimentales sur Les Mortiers Hydrauligues 
 
156 
 
 CEMENT AND CONCRETE 
 
 differing in sharpness but alike in all other respects, including 
 the percentage of voids, were tested side by side, the rounded 
 grains would be found inferior. (See also 253.) 
 
 M. Alexandre also made tests on sands differing both in 
 chemical and physical characteristics, but having the same fine- 
 ness, namely, twenty per cent, each of five sizes of grain. Some 
 of the results are given in Table 46. 
 
 TABLE 46 
 
 Results Obtained with Various Sands 
 
 SAND. 
 
 WATER 
 PER 
 CENT. OF 
 VOLUME 
 OF SAND. 
 
 TENSILE STRENGTH, IN LBS. PER 
 SQ. IN., OF MORTARS CONTAIN- 
 ING 400 KG. OF CEMENT TO 1 Cu. 
 METER SAND, AT AGES OF 
 
 7 da. 
 
 lyr. 
 
 3 yrs. 
 
 Sea Sand . . 
 
 21 
 28 
 21 
 20 
 20 
 28 
 
 69 
 78 
 65 
 63 
 79 
 35 
 
 165 
 198 
 168 
 174 
 178 
 99 
 
 245 
 
 267 
 201 
 215 
 244 
 132 
 
 Calcareous (Renville stone) . 
 Granitic 
 
 Siliceous (cliff quartz) 
 Siliceous (Cherbourg Quartzites) . 
 Coke 
 
 
 249. Siliceous vs. Calcareous Sands. The above tests 
 would seem to show that sand to be used in mortar need not be 
 siliceous. In experimenting on different varieties of sand, both 
 natural and artificial, the author has obtained results that 
 point to a similar conclusion. Some of these tests are given in 
 Tables 47 to 50. 
 
 Table 47 gives the results obtained with four varieties of 
 siliceous sand. The first was an artificial sand made by crush- 
 ing sandstone, the second and third were natural sands con- 
 taining a large percentage of quartz grains, and the fourth 
 appeared to be almost pure quartz. Only the fine particles of 
 the sands were used in the tests given in this table. The dif- 
 ferences in strength at the end of two years are not great, but 
 the two natural sands appear to give somewhat lower results. 
 
 In Table 48 the two natural sands were again compared, 
 but this time in connection with a calcareous sand formed by 
 crushing limestone. The latter gave the best results. Only 
 the finer grains were used in these tests. 
 
 250. Tables 49 and 50 are more valuable in this connection, 
 
CHARACTER OF SAND 
 
 157 
 
 TABLE 47 
 
 Values of Different Varieties of Fine Siliceous Sand for Use in 
 Portland Cement Mortar 
 
 Two PARTS SAND TO ONE CEMENT BY WEIGHT 
 
 REFERENCE. 
 
 SAND. 
 
 FINENESS. 
 
 WATER, 
 
 Per 
 Cent. 
 
 TENSILE 
 STRENGTH, LBS. 
 PER SQ. IN. AT 
 
 6 Mo. 
 
 2Yr. 
 
 a 
 
 b 
 
 c 
 
 d 
 
 
 
 1 
 2 
 
 ( Screenings from ( 
 crushing Pots- j 
 ( dam sandstone ( 
 
 Pass 40 sieve . 
 Pass 40 sieve, 
 retained on 100 
 
 18.5 
 17.5 
 
 388 
 478 
 
 470 
 539 
 
 3 
 
 Bank sand, siliceous 
 
 Pass 40 sieve . 
 
 13.3 
 
 433 
 
 445 
 
 4 
 
 River sand, siliceous 
 
 Pass 40 sieve . 
 
 12.1 
 
 382 
 
 437 
 
 5 
 
 Clean quartz 
 
 Pass 40 sieve . 
 
 13.3 
 
 398 
 
 606 
 
 NOTE. Holes in No. 40 sieve 0.015 inch square, holes in No. 100 sieve 
 about 0.0065 inch square. 
 
 TABLE 48 
 Different Varieties of Fine Sand for Portland Cement Mortar 
 
 
 
 
 
 TENSILE STRENGTH, POUNDS 
 
 w 
 
 
 
 PER CENT. 
 
 PER SQUARE INCH. 
 
 
 
 to 
 
 SAND. 
 
 FINENESS. 
 
 
 1 Part Sand to 1 
 
 2 Parts Sand to 
 
 
 
 
 
 
 Cement by Wt. 
 
 Cement by Wt. 
 
 
 
 
 1 tol 
 
 1 to2 
 
 6 mo. 
 
 13 
 mo. 
 
 3yr. 
 
 6 mo. 
 
 18 
 mo. 
 
 3yr. 
 
 
 a 
 
 6 
 
 c 
 
 d 
 
 e 
 
 / 
 
 9 
 
 h 
 
 i 
 
 J 
 
 1 
 
 River sand, sili- 
 
 
 
 
 
 
 
 
 
 
 
 ceous . . . 
 
 Pass 40 sieve 
 
 14.0 
 
 12.4 
 
 715 
 
 725 
 
 776 
 
 491 
 
 575 
 
 581 
 
 2 
 
 Bank sand, sili- 
 
 
 
 
 
 
 
 
 
 
 
 ceous . . . 
 
 Pass 40 sieve 
 
 14.5 
 
 12.6 
 
 664 
 
 699 
 
 759 
 
 442 
 
 502 
 
 5?4 
 
 3 
 
 Calcareous sand 
 
 
 
 
 
 
 
 
 
 
 
 from crushing 
 limestone . . 
 
 Pass 40 sieve 
 
 18.2 
 
 17.7 
 
 721 
 
 770 
 
 788 
 
 531 
 
 632 
 
 (580 
 
 4 
 
 Calcareous sand 
 from crushing 
 limestone . . 
 
 Pass 40, re- 
 tained on 100 
 
 17.5 
 
 17.0 
 
 753 
 
 783 
 
 844 
 
 597 
 
 659 
 
 727 
 
 since the coarser particles of the sand were used with the fine. 
 The sand was separated into four sizes by sifting, and then 
 remixed in equal proportions. Table 49 gives the results ob- 
 
158 
 
 CEMENT AND CONCRETE 
 
 tained with natural cement, and Table 50 refers to Portland. 
 The superiority of the screenings is very clearly shown, the 
 limestone giving especially good results. Indeed, the strength 
 obtained with three parts limestone screenings to one part of 
 either Portland or natural cement is remarkably high. The 
 mortar made from such sand is peculiarly plastic when fresh, 
 and soon gains a high strength which it appears to maintain. 
 
 TABLE 49 
 Values of Different Varieties of Sand for Natural Cement Mortar 
 
 
 
 
 . ^ 
 
 TENSILE STRENGTH, Lus. 
 
 
 
 02 
 
 fc E 
 
 PER SQ. IN., 3 PARTS SAND 
 
 H 
 M 
 
 SAND. 
 
 B 
 
 
 
 TO 1 CEMENT BY WT. 
 
 
 
 
 
 
 
 
 
 
 
 28 Da. 
 
 6Mos. 
 
 1 Yr. 
 
 2 Yrs. 
 
 
 a 
 
 
 
 c 
 
 d 
 
 e 
 
 / 
 
 ff 
 
 1 
 
 Clean crushed quartz . . . 
 
 MX. 
 
 15.4 
 
 117 1 
 
 344 
 
 356 
 
 332 
 
 2 
 
 River sand, siliceous . 
 
 MX. 
 
 13.3 
 
 93 
 
 297 
 
 339 
 
 308 
 
 3 
 
 Limestone screenings . 
 
 MX. 
 
 16.7 
 
 143 
 
 467 
 
 526 
 
 601 
 
 4 
 
 Potsdam sandstone screenings 
 
 MX. 
 
 18.2 
 
 113 
 
 316 
 
 416 
 
 462 
 
 5 
 
 Clean crushed quartz . 
 
 20-30 
 
 12.5 
 
 118 
 
 330 
 
 342 
 
 324 
 
 1 13.6 per cent, water, trifle dry. 
 
 NOTE. Fineness MX. means 25 psr cent, each of 20-30, 30-40, 40-50 
 
 and 50-80. 
 Expression 20-30 means passing No. 20 sieve and retained on 
 
 No. 30 sieve. 
 
 TABLE 50 
 
 Values of Different Varieties of Sand for Portland Cement Mortar 
 
 
 
 . 
 
 . . 
 
 TENSILE STRENGTH, LBS. 
 
 
 
 CO 
 
 fc w 
 
 PER SQ. IN., 3 PARTS SAND 
 
 fe 
 
 
 g 
 
 w ^ 
 
 TO 1 CEMENT BY WT. 
 
 H 
 
 PH 
 
 SAND. 
 
 B 
 
 fc 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 28 Da. 
 
 6Mos. 
 
 lYr. 
 
 2 Yrs. 
 
 
 a 
 
 6 
 
 c 
 
 d 
 
 e 
 
 / 
 
 9 
 
 1 
 
 Clean crushed quartz . . 
 
 MX. 
 
 12.5 
 
 255 
 
 327 
 
 359 
 
 335 
 
 2 
 
 River sand, siliceous . 
 
 MX. 
 
 11.1 
 
 206 
 
 284 
 
 329 
 
 324 
 
 8 
 
 Limestone screenings . 
 
 MX. 
 
 12. 5 1 
 
 407 
 
 574 
 
 667 
 
 665 2 
 
 4 
 
 Sandstone screenings . . 
 
 MX. 
 
 12. 5 1 
 
 321 
 
 438 
 
 495 
 
 492 3 
 
 5 
 
 Clean crushed quartz . 
 
 20-30 
 
 11.1 
 
 259 
 
 344 
 
 369 
 
 335 
 
 1 Trifle dry, plastic. 2 13.3 per cent, water. 3 14.3 per cent, water. 
 NOTE. Fineness MX. means 25 per cent, each of 20-30, 30-40, 40-50 
 and 50-80. 
 
FINENESS OF SAND 159 
 
 251. Slag Sand. To turn to good account some of the 
 immense quantities of blast furnace slag produced yearly, the 
 use of granulated slag in place of ordinary sand has been ad- 
 vocated. In a paper read before the Engineers' Society of 
 Western Pennsylvania, in March, 1904, Mr. Joseph A. Shinn 
 described some experiments he had made, in which it was shown 
 that "slag sand," with Portland cement, natural cement, or 
 common lime, gave a higher strength than the sample of river 
 sand used in the comparison. 
 
 The "slag sand" is produced by projecting two flat jets of 
 water into the stream of molten slag, the resulting sand being 
 heavier, finer and more nearly uniform in size of grain than the 
 ordinary slag granulate. 
 
 252. Sand for Use in Sea Water. It has been said that 
 granitic sands when used in sea water do not give good results 
 on account of the felspar of the granite being attacked by the 
 cement when the concrete is impregnated with sea water. M. 
 Paul Alexandre would proscribe the use of argillaceous sands 
 in sea water, but he found that sands containing calcareous 
 marl gave excellent results in the sea, and others have stated 
 that the mixture of crushed limestone with concrete has been 
 known to hinder the action of sea water upon it. Since porous 
 and permeable mortars are most liable to disintegration by 
 sea water, it is evident that it is especially desirable to employ 
 a sand in which the proportion of voids is small. 
 
 ART. 30. FINENESS OF SAND 
 
 253. The size and shape of the grains are important ele- 
 ments in the quality of sand. Considering grains of the same 
 shape but differing in size, the larger grain will have a smaller 
 surface area in proportion to the volume than the smaller grain, 
 since the volume varies approximately as the cube of one di- 
 mension while the surface varies as the square. Since, in order 
 to obtain the best results in mortar, each grain of sand must be 
 coated with cement, it follows that, other things being equal, 
 the coarser grained sands will give the best results, because 
 they will be more thoroughly coated; this will be especially true 
 when the amount of sand in the mortar is relatively large. 
 
 Following the same reasoning given above as to the relative 
 volume and superficial area of sand grains, it would appear 
 
100 
 
 CEMENT AND CONCRETE 
 
 that spherical grains would be better than cubical or angular 
 ones (see 248). This, however, is not thought to be the case, 
 for the better bond obtained with angular grains seems to coun- 
 terbalance the advantage which the small superficial area would 
 appear to give to the spherical grains. For this reason a len- 
 ticular shaped grain, while having a very large area relative to 
 its volume, will give excellent results in mortar if otherwise 
 suited to the purpose. 
 
 It is usually desirable to have all of the voids in the sand 
 filled by the cement paste, as this renders the mortar less por- 
 ous, and makes it more certain that all the grains are coated 
 with cement. On this account a mixture of fine and coarse 
 particles is excellent. 
 
 TABLE 51 
 
 Effect on Tensile Strength of Varying Fineness of Limestone 
 Screenings Used with Portland Cement 
 
 AGE 
 BRIQUETS WHEN 
 BROKEN. 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH 
 FINENESS OF SCREENINGS. 
 
 10-20. 
 
 20-30. 
 
 30-40. 
 
 40-50. 
 
 40-80. 
 
 PASS 50. 
 
 6 months . 
 
 718 
 
 657 
 
 633 
 
 516 
 
 . . . .. 
 
 403 
 
 2 years . 
 
 812 
 
 754 
 
 656 
 
 . . . 
 
 516 
 
 488 
 
 4 years . . . 
 
 845 
 
 782 
 
 714 
 
 ... 
 
 571 
 
 516 
 
 SIGNIFICANCE OF FINENESS 
 
 
 SIEVE NUMBER. 
 
 
 
 
 APPROXIMATE 
 
 
 Passing. 
 
 Retained on. 
 
 GRAIN. 
 
 
 
 
 Inch. 
 
 10-20 
 
 10 
 
 20 
 
 .057 
 
 20-30 
 
 20 
 
 30 
 
 .028 
 
 30-40 
 
 30 
 
 40 
 
 .020 
 
 40-50 
 
 40 
 
 50 
 
 .015 
 
 40-80 
 
 40 
 
 80 
 
 .012 
 
 Pass 50 
 
 50 
 
 . . . . 
 
 .008 
 
 NOTES. Three parts screenings to one cement by weight. 
 
 All briquets made by one molder and immersed in one tank. 
 Variations in consistency were slight, the largest percentage of 
 water being used for the finest particles. 
 
FINENESS OF SAND 
 
 161 
 
 254. TESTS ON EFFECT OF FINENESS OF SAND. Many of 
 the experiments made to show the effect of the fineness of sand 
 on the strength of the mortar are defective, because the sand 
 used varies in the shape of the grains and in chemical charac- 
 teristics as well as in fineness. The experiments given in Table 
 51 were made with screenings obtained in crushing limestone, 
 and thus all causes of variation aside from the fineness of the 
 sand were absent,, except the differences in consistency of the 
 mortar, the uniformity in consistency depending on the judg- 
 ment of the operator. The results show quite clearly the su- 
 periority of the coarser sand. 
 
 255. The Relative Effect of Fine Sand on Portland and Nat- 
 ural Cement. The tests in Table 52 were made to determine 
 
 TABLE 52 
 
 Coarse and Fine Sand, Relative Effects with Portland and 
 
 Natural Cement 
 
 
 
 TENSILE 
 
 
 Q 
 
 TENSILE 
 
 
 
 
 STRENGTH, 
 
 
 K 
 
 STRENGTH, 
 
 
 AGE OK 
 BRIQUETS 
 
 WHEN 
 
 m 
 
 POUNDS PER 
 SQ. IN. WHEN 
 SAND is 
 
 IP! 
 
 g5K 
 
 i$. 
 
 POUNDS PER 
 SQ. IN. WHEN 
 SAND is 
 
 2SS 
 
 *%~? 
 
 BROKEN. 
 
 
 
 
 5 2 ^ 
 
 *; 
 
 
 
 * H*^ 
 
 
 
 20-30 
 
 40-80 
 
 a 92 H 
 
 i. 
 
 20-30 
 
 40-80 
 
 a 02 p 
 
 
 o 
 
 
 
 ft* 
 
 O 
 
 
 
 ftn 
 
 28 days . . j 
 
 Bn 
 
 In 
 
 197 
 
 89 
 
 145 
 
 57 
 
 74 
 
 64 
 
 A 
 
 U 
 
 406 
 352 
 
 337 
 275 
 
 83 
 78 
 
 6 mouths . . | 
 
 Bn 
 In 
 
 216 
 364 
 
 188 
 267 
 
 87 
 73 
 
 A 
 U 
 
 520 
 499 
 
 446 
 415 
 
 86 
 83 
 
 2 years . . j 
 
 Bu 
 In 
 
 256 
 450 
 
 250 
 
 419 
 
 98 
 93 
 
 A 
 U 
 
 546 
 567 
 
 451 
 
 496 
 
 83 
 89 
 
 NOTES. Sand, limestone screenings; three parts to one cement by 
 
 weight. 
 20-30 means sand passing sieve with 20 meshes per linear 
 
 inch, and retained on sieve with 30 meshes per linear 
 
 inch. 
 Columns 5 and 9 show percentage that strength with finer 
 
 sand is of the strength with coarser sand. 
 
 the relative effects of fine sand on Portland and natural cements. 
 Limestone screenings of two sizes of grain were used in con- 
 nection with two brands of each kind of cement. At twenty- 
 eight days the natural cement shows the decrease in strength 
 due to the use of fine sand more than Portland cement does. 
 
162 CEMENT AND CONCRETE 
 
 At six months the fine sand seems to have about the same 
 effect on Portland and natural, but the two-year results in- 
 dicate that the ultimate effect is less on the natural cement 
 than on the Portland; the mean ratio of the strength obtained 
 with fine sand to that given by coarse sand being ninety-six 
 in the case of natural, and only eighty-six in the case of Port- 
 land. The effect of fine sand appears to decrease with age, 
 especially with natural cement. 
 
 The fineness of sand will be treated further in the following 
 article relating to voids. 
 
 ART. 31. VOIDS IN SAND 
 
 256. Conditions Affecting Voids. The voids present in a 
 given mass of sand will depend upon the shape of the grains, 
 the degree of uniformity in size of grains, the amount of moisture 
 present, and the amount of compacting to which the mass has 
 been subjected. If all of the grains in a given mass of sand are 
 of uniform size, the percentage of voids will be independent of 
 what that size may be. In other words, the percentage of 
 voids in a cubic foot of buckshot will be the same as in a cubic 
 foot of bird shot; but if we take a cubic foot of a mixture of 
 buck and bird shot we will find that the voids are much 
 less. 
 
 257. Effect of Shape of Grain. M. Feret has published in 
 France the results of a large number of experiments made by 
 him as to the voids in sand and broken stone. 1 Table 53 gives 
 the results he obtained concerning the effect of the shape of 
 the grains on the percentage of voids present. He first divided 
 each sand into three parts by means of three sieves, which we 
 will call A, B and C. Sieve A had four meshes per sq. cm. 
 (about five meshes per linear inch), sieve B had 36 meshes per 
 sq. cm. (about fifteen meshes per linear inch), and sieve C had 
 324 meshes per sq. cm. (about forty-five meshes per linear 
 inch). The grains that passed A and were retained on B were 
 designated G, the grains that passed B and were retained on C 
 were designated M, and the grains that passed C were desig- 
 nated F. These different sizes were then recombined by tak- 
 ing five parts of G, three parts of M and two parts of F, and 
 
 Abstracted in Engineering News, Vol. XXVII, p. 310. 
 
VOIDS IN SAND 
 
 163 
 
 the resulting sand was designated G 5 M 3 F 2 . Thus, all of the 
 sands tested had the same "granulometric" composition. 
 
 TABLE 53 
 
 Voids in Sands Having Different Shaped Grains 
 FROM M. FERET 
 
 NATURE OF SAND. 
 
 VOLUME OF VOIDS REMAINING IN 
 ONE LITER OF SAND. 
 
 Unshaken. 
 C.C. 
 
 Shaken to Refusal. 
 C.C. 
 
 Natural sand with rounded grains. 
 Cherbourg quart/ite, angular grains. 
 Crushed shells, flat grains. 
 Residue of Cherbourg quartzite crushed 
 between jaws, laminated grains. 
 
 359 
 
 421 
 44:) 
 
 475 
 
 256 
 274 
 
 318 
 
 346 
 
 It is seen that the rounded grains have the smallest percent- 
 age of voids, or about thirty-six per cent, unshaken, while the 
 laminated grains gave the largest percentage. It may also be 
 noticed that the angular grains were compacted more by shak- 
 ing than any of the others. 
 
 258. Effect of Granulometric Composition of Sand on the 
 Percentage of Voids. To determine the effect of uniformity of 
 size of grain upon the percentage of voids and the strength of 
 mortars, the author has experimented with an artificial sand 
 formed by crushing limestone. That portion of the product 
 that passed the coarse screen of the crusher varied in fine- 
 ness from particles three-eighths of an inch in one dimension to 
 a very fine powder, the particles of which were less than .0065 
 inch in one dimension. Such material admits of division into 
 parts that differ widely in fineness, but which are essentially 
 of the same composition, and it is therefore excellent for an 
 experiment of this kind. 
 
 The four sieves used in first separating the material into 
 parts had, respectively, 10, 20, 40 and 80 meshes per linear inch, 
 the sizes of the holes being, respectively, about as follows: 0.08 
 inch, 0.033 inch, 0.017 inch, and 0.007 inch square. The sev- 
 eral sizes of grain are designated as follows: 
 
 "C," Coarse, passing No. 10, retained on No. 20. 
 "M," Medium, " " 20, " " 40. 
 
 "F," Fine, " " 40, " " 80. 
 
 "V," Very fine, " " 80. 
 
164 
 
 CEMENT AND CONCRETE 
 
 M. Feret's method of designating the granulometric compo- 
 sition, namely, to represent by exponents the number of parts 
 of each size of grain, has been adopted. 
 
 259. The voids were obtained by first weighing a given 
 volume of the sand; dividing the weight by the specific gravity 
 of the limestone, as previously determined, gives the amount 
 of solid material in the measure, and this subtracted from the 
 volume of the measure, gives the voids. This method is con- 
 sidered more nearly accurate than the usual one of measuring 
 the amount of water required to fill the voids in a measure of 
 sand, especially so for a sand of uniform character and one 
 which absorbs water quite freely. 
 
 TABLE 54 
 
 Voids in Limestone Screenings, Showing Effect of Variations in 
 GFranulometric Composition 
 
 FINENESS OF 
 GRAN ULOMET RIO 
 COMPOSITION. 
 
 WEIGHT OF 
 ONE LITER OF 
 SAND, DRV, 
 GRAMS. 
 
 VOLUME SOLID 
 SAND IN 
 ONE LITER 
 
 (SP. GR. = 2.667) 
 Cu. CENT. 
 
 PER CENT. 
 VOIDS 
 IN SAND. 
 
 
 Loose. 
 
 Shaken. 
 
 Loose. 
 
 Shaken. 
 
 Loose. 
 
 Shaken. 
 
 a 
 
 b 
 
 c 
 
 d 
 
 
 
 / 
 
 9 
 
 C = Coarse 10 to 20 
 
 1126 
 
 1358 
 
 422 
 
 509 
 
 57.8 
 
 49.1 
 
 M = Medium 20 to 40 
 
 1140 
 
 1362 
 
 428 
 
 511 
 
 57.2 
 
 48.9 
 
 F = Fine 40 to 80 
 
 1150 
 
 1392 
 
 431 
 
 522 
 
 56.9 
 
 47.8 
 
 V = Very fine, pass 80 
 
 1165 
 
 1609 
 
 437 
 
 603 
 
 56.3 
 
 39.7 
 
 C 
 
 
 1395 
 
 . . . 
 
 523 
 
 
 47.7 
 
 M 
 
 
 1439 
 
 . 
 
 540 
 
 
 46.0 
 
 F 
 
 
 1459 
 
 . 
 
 547 
 
 
 45.3 
 
 V 
 
 
 1656 
 
 
 621 
 
 
 37.9 
 
 C56, M25, F* 5 , V* 
 
 
 1606 
 
 . 
 
 602 
 
 . 
 
 39.8 
 
 C 40 , M 30 , F2o, V 10 
 
 
 1732 
 
 
 649 
 
 . 
 
 35.1 
 
 C 25 , M 25 , F 25 , V 25 
 
 
 1912 
 
 
 717 
 
 
 28.3 
 
 C 30 , M 25 , Fis, V 30 
 
 
 1850 
 
 . 
 
 694 
 
 , 
 
 30.6 
 
 C 50 , M, F>, V 5 
 
 
 1991 
 
 
 746 
 
 
 25.4 
 
 The results obtained are given in Table 54. Comparing the 
 voids in C, M, F and V, it is seen that the first three have nearly 
 the same percentage, but V has less voids than the others. 
 This is explained by the fact that this sample was made up of 
 all sizes smaller than the holes in No. 80 sieve, down to the 
 fine powder. Comparing the mixed sands, it is seen that the 
 sample made up of equal parts of coarse and very fine had 
 
VOIDS IN SAND 
 
 105 
 
 the least voids, the percentage being only a little more than 
 half of that obtained with coarse particles alone. The next 
 lowest percentage was given by the sample having equal parts 
 of four sizes. 
 
 It is apparent that the granulometric composition has a 
 very important effect on the percentage of voids. When one 
 desires to make a compact mortar with as small a quantity of 
 cement as possible, similar tests might well be made with the 
 materials available for use. 
 
 260. Effect on Strength of Mortars of Varying the Granulo- 
 metric Composition of Sand. Table 55 gives the results of 
 tensile tests of mortars made with limestone screenings of vari- 
 ous granulometric compositions. The differences in strength 
 are not very great, but it appears that with one-to-three mor- 
 tars the highest strength is developed at six months, with the 
 coarse grains alone, but when poorer mortars are in question 
 the result is affected by the percentage of voids in the sand. 
 
 TABLE 55 
 
 Limestone Screenings with Portland Cement. Effect on Tensile 
 Strength of Variations in Granulometric Composition of Sand 
 
 
 
 TENSILE STRENGTH AT 
 
 
 (i RANULOMETRIC COMPOSITION 
 
 
 6 Mos. POUNDS PER 
 
 WEIOHT OP 
 
 OK SANI>. PER 
 
 
 SQ. IN. WITH PARTS SAND 
 
 BRIQUETS IN 
 
 CENT. OF EACH SIZE GRAIN. 
 
 VOIDS. 
 
 TO ONE CEMENT BY 
 
 GRAMS. 
 
 
 % 
 
 WEIGHT. 
 
 
 c 
 
 M 
 
 F 
 
 V 
 
 
 3 
 
 5 
 
 3 
 
 5 
 
 
 
 100 
 
 
 
 
 
 46 
 
 609 
 
 324 
 
 1465 
 
 1438 
 
 40 
 
 30 
 
 20 
 
 10 
 
 35 
 
 505 
 
 392 
 
 1466 
 
 1480 
 
 25 
 
 25 
 
 25 
 
 25 
 
 31 
 
 470 
 
 356 
 
 1445 
 
 1455 
 
 30 
 
 25 
 
 15 
 
 30 
 
 28 
 
 496 
 
 391 
 
 1448 
 
 1470 
 
 60 
 
 
 
 
 
 50 
 
 25 
 
 487 
 
 349 
 
 1455 
 
 1460 
 
 CEMENT. Portland, Brand R. 
 see text. 
 
 For significance of composition of sand, 
 
 261. Table 56 gives the results of similar tests of both Port- 
 land and natural cement with Point aux Pins sand dredged 
 from St. Marys River and containing a very large percentage 
 of quartz grains. The sand was divided into but three parts 
 by sifting, and was then remixed, the proportion of each size 
 being indicated in the table. The results verify the conclusions 
 
166 
 
 CEMENT AND CONCRETE 
 
 already drawn that the coarser sands give the higher strength. 
 It appears that not more than one-half of the grains should be 
 very fine if the best results are desired. 
 
 TABLE 56 
 
 Varying the Granulometric Composition of River Sand. Effect on 
 Value of, for Use in Cement Mortar 
 
 COMPOSITION OF SAND AS 
 TO FINENESS. 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH. 
 
 Parts Used 
 that Passed 
 No. 20 Sieve 
 and Re- 
 tained on 
 No. 30. 
 
 Parts 
 Used, 
 30-40 
 
 Parts 
 Used that 
 Passed 
 No. 40 
 Sieve. 
 
 Portland Cement Avith 
 Two Parts Sand to One 
 Cement by Weight, at 
 age of 
 
 Natural Cement with 
 Three Parts Sand to One 
 Cement by Weight, at 
 age of 
 
 M 
 
 F 
 
 V 
 
 28 da. ! 6 mo. 
 
 lyr. 
 
 2 yr. 
 
 28 da. 
 
 6 mo. 
 
 lyr. 
 
 2 yr. 
 
 10 
 
 
 
 
 
 342 
 
 471 
 
 560 
 
 591 
 
 77 
 
 267 
 
 348 
 
 341 
 
 4 
 
 1 
 
 5 
 
 300 
 
 448 
 
 515 
 
 507 
 
 77 
 
 237 
 
 304 
 
 319 
 
 2 
 
 4 
 
 4 
 
 290 
 
 425 
 
 494 
 
 503 
 
 79 
 
 278 
 
 291 
 
 325 
 
 1 
 
 3 
 
 6 
 
 246 
 
 384 
 
 455 
 
 442 
 
 46 
 
 222 
 
 234 
 
 251 
 
 1 
 
 2 
 
 7 
 
 271 
 
 366 
 
 456 
 
 438 
 
 07 
 
 226 
 
 247 
 
 251 
 
 NOTE. River sand, mostly quartz, obtained at Point aux Pins. Each 
 result mean of five briquets, all made by one molder. 
 
 262. Effect of Moisture. The effect of a small amount of 
 moisture on the bulk of a given weight of sand is not usually 
 appreciated, but it may easily be shown that it is very marked. 
 The results in Table 57 were obtained by adding small amounts 
 of water to a given bulk of dry sand. Each time, after the 
 water was added, the sand was stirred up and the weight of a 
 given volume of the moist sand was obtained. It appears that 
 the finer sands are affected more than coarse ones. 
 
 In the case of the limestone screenings 40-80, if we add but 
 3.7 per cent, water to a given quantity of dry sand, the bulk 
 of the sand is so increased that if we take 1,000 c.c. of the moist 
 sand it will contain but 720 c.c. of dry sand. The voids are, 
 of course, correspondingly increased from 54.5 per cent, to 
 67.2 per cent. 
 
 The cause of this increase in bulk is that each grain of sand 
 is surrounded by a film of water which prevents the grains 
 from lying close together after they have been disturbed. A 
 large amount of air is also imprisoned in the mass. It may be 
 noticed that the 'difference in bulk between moist and dry sand 
 is greater when, the measurements are made " loose." 
 
VOIDS IN SAND 
 
 167 
 
 TABLE 57 
 Volume of Sand and Voids as Affected by the Addition of Water 
 
 ERENCE. 
 
 SAND. 
 
 EXPRESSED 
 t CENT. OF 
 SAND BY 
 
 Kir.HT. 
 
 WEIGHT OF 
 DRY SAND IN 
 ONE LITER 
 OF MOIST 
 SAND. 
 
 VOLUME OF 
 DRY SAND 
 IN ONE 
 LITER OF 
 MOIST 
 SAND. 
 
 PERCENT. 
 VOIDS IN 
 SAND BY 
 VOLUME. 
 
 
 
 ^ 
 
 Kind. 
 
 g 
 
 *#'* 
 
 T2 
 
 Cx 
 
 - r 
 
 gl 
 
 
 
 a 
 
 V 
 
 
 
 E 
 
 4 *^ 
 
 c p 
 
 11 
 
 ~ 
 
 H^ 
 
 
 
 5 
 
 
 
 
 
 
 -^ 
 
 " '3 
 
 C 
 
 JS 3 
 
 * 
 
 33 
 
 
 a 
 
 b 
 
 c 
 
 rf 
 
 e 
 
 f 
 
 9 
 
 h 
 
 i 
 
 
 Crushed 
 
 
 
 
 
 
 
 
 
 1 
 
 Limestone. 
 
 10-20 
 
 0.0 
 
 1288 
 
 1489 
 
 1000 
 
 1000 
 
 . 
 
 
 2 
 
 i 
 
 " 
 
 4.8 
 
 1094 
 
 1367 
 
 849 
 
 919 
 
 
 
 3 
 
 i 
 
 " 
 
 7.7 
 
 1023 
 
 1295 
 
 794 
 
 869 
 
 
 
 4 
 
 i 
 
 " 
 
 11.9 
 
 996 
 
 1276 
 
 773 
 
 857 
 
 . 
 
 
 5 
 
 
 
 40-80 
 
 0.0 
 
 1214 
 
 1481 
 
 1000 
 
 1000 
 
 54.5 
 
 44.6 
 
 6 
 
 c 
 
 " 
 
 0.85 
 
 1124 
 
 1489 
 
 920 
 
 1005 
 
 57.9 
 
 44.2 
 
 7 
 
 I 
 
 u 
 
 1.5 
 
 1059 
 
 1470 
 
 872 
 
 993 
 
 603 
 
 44.9 
 
 8 
 
 ' 
 
 (1 
 
 2.2 
 
 950 
 
 1383 
 
 782 
 
 934 
 
 64.4 
 
 48.2 
 
 9 
 
 ' 
 
 (i 
 
 3.7 
 
 875 
 
 1298 
 
 720 
 
 877 
 
 67.2 
 
 51.4 
 
 10 
 
 1 
 
 u 
 
 6.3 
 
 824 
 
 1274 
 
 679 
 
 860 
 
 69.1 
 
 52.3 
 
 11 
 
 ' 
 
 " 
 
 7.8 
 
 799 
 
 1266 
 
 658 
 
 855 
 
 70.0 
 
 52.6 
 
 12 
 
 ' 
 
 " 
 
 12.3 
 
 817 
 
 1280 
 
 672 
 
 864 
 
 69.4 
 
 52.0 
 
 13 
 
 1 
 
 (I 
 
 16.8 
 
 829 
 
 1306 
 
 683 
 
 881 
 
 69.0 
 
 51.1 
 
 14 
 
 
 
 u 
 
 20.2 
 
 836 
 
 1274 
 
 689 
 
 860 
 
 68.6 
 
 52.3 
 
 15 
 
 ' 
 
 u 
 
 25.3 
 
 891 
 
 1357 
 
 783 
 
 916 
 
 66.6 
 
 49.1 
 
 16 
 
 1 
 
 " 
 
 30.3 
 
 1049 
 
 1270* 
 
 864 
 
 858* 
 
 < 
 
 t 
 
 17 
 
 ' 
 
 Pass 80 
 
 0.0 
 
 1185 
 
 1500 
 
 1000 
 
 1000 
 
 
 
 18 
 
 ' 
 
 " 
 
 2.4 
 
 1038 
 
 1394 
 
 873 
 
 929 
 
 
 
 19 
 
 1 
 
 " 
 
 5.1 
 
 835 
 
 1281 
 
 704 
 
 854 
 
 
 
 20 
 
 ' 
 
 " 
 
 12.2t 
 
 806 
 
 1310 
 
 680 
 
 873 
 
 . 
 
 
 21 
 
 1 
 
 " 
 
 17.7t 
 
 806 
 
 1260 
 
 680 
 
 840 
 
 
 
 
 Point aux 
 
 
 
 
 
 
 
 
 
 22 
 
 Pins. 
 
 c 
 
 0.0 
 
 1725 
 
 
 1000 
 
 
 
 
 23 
 
 i 
 
 
 2.0 
 
 1405 
 
 
 815 
 
 
 
 
 24 
 
 t 
 
 
 4.0 
 
 1400 
 
 
 810 
 
 
 
 
 25 
 
 i 
 
 t J 
 
 6.0 
 
 1400 
 
 
 810 
 
 
 
 
 26 
 
 i 
 
 \ 
 
 10.0 
 
 1415 
 
 
 820 
 
 
 
 
 27 
 
 i 
 
 
 11.6 
 
 1425 
 
 
 825 
 
 
 
 
 28 
 
 i 
 
 
 18.4 
 
 1485 
 
 
 860 
 
 
 
 
 
 
 
 * Not jarred down in measure as much as usual. Water rose to surface, 
 t Sand crumbled like damp earth. 
 
 J Fineness of Point 
 aux Pins Sand 
 
 r Sieves No. 20 30 40 50 80 
 J Approx. size 
 
 holes - .033 .022 .017 .012 .007 
 
 [Percent, passing 96.0 82.3 46.6 6.7 1.2 
 
 NOTE. 10-20 = passing No. 10 sieve (holes about .08 in. sq.) and retained 
 on No. 20 sieve. 
 
1G8 CEMENT AND CONCRETE 
 
 263. This subject is of great importance in proportioning 
 mortars, because, in construction, the amounts of cement and 
 sand are usually measured. Suppose it is desired to use a mix- 
 ture of one hundred pounds of cement to four hundred pounds 
 of sand, and for convenience we will suppose the packed cement 
 and dry sand each weigh one hundred pounds per cubic foot. 
 If now we use damp sand, containing about 3.5 per cent, water, 
 instead of dry sand, and measure the materials, we would have 
 four cubic feet of damp sand to one cubic foot of cement; but 
 damp sand would contain only about 4 X 75 = 300 pounds of 
 dry sand, and we would really have a one-to-three mixture 
 instead of a one-to-four. 
 
 ART. 32. IMPURITIES IN SAND 
 
 264. The usual specification for sand is that it shall be 
 " clean, sharp and siliceous." We have shown that it need not 
 be siliceous, and we have also noted that one authority con- 
 siders that it need not be sharp, though this latter does not 
 appear to be proven; let us see what interpretation should be 
 given to the word " clean" if it must be retained in all speci- 
 fications for sand. 
 
 Mr. E. C. Clarke, in the tests for the Boston Main Drainage 
 Works, showed that "clay in moderate amounts" (ten per 
 cent, to thirty per cent, of the sand) "does not weaken cement 
 mortars." Calcareous marl might be considered an impurity, 
 but we have seen that M. Alexandre found that sands contain- 
 ing this material gave excellent results. On the other hand, 
 there seems to be no doubt that loam, peaty matter or humus 
 will very materially decrease the strength of mortars, or even 
 destroy them entirely. Likewise, decayed particles of some 
 kinds of stone, or grains which readily break up into thin scales, 
 should be strenuously avoided. 
 
 265. Detection of Impurities. Clean sand when rubbed in 
 the hand will not leave fine particles adhering to it, but should 
 the sand not prove to be clean, the character of the impurities 
 should be investigated before finally rejecting it. When there 
 is not time for making proper tests, it will, of course, be safest 
 to use only such sand as has no foreign matter whatever; but 
 when strictly pure sand can only be obtained at great cost, tests 
 may show that a small percentage of impurities may be tolerated. 
 
1 IMPURITIES 169 
 
 Another simple test, beside the one of rubbing in the hand, 
 is to place a little of the sand in a test tube filled with water; 
 if any impurities are present, they may separate from the sand 
 on account of their lighter weight, or if in a very fine state of 
 division, the water may be rendered murky in appearance. 
 This test is not absolute, however, especially for calcareous 
 sand, as the fine particles of limestone will give the murky 
 appearance to the water, although not objectionable except on 
 account of their extreme fineness. 
 
 The use of poor sand will result in a larger proportionate 
 decrease in strength for a mortar containing a large amount 
 of sand than for one made with a small amount. The effect of 
 incorporating various foreign substances in cement mortar is 
 treated in Art. 49. As some of these substances may occur 
 in sand, the article referred to should be read in connection 
 with this subject. 
 
 266. SAND WASHING. When impurities occur, they may 
 sometimes be removed by washing, but such work must be 
 carefully inspected if the foreign matter be of a really danger- 
 ous character. 
 
 In the construction of the Canal at the Cascades, Columbia 
 River, Oregon, quite an elaborate concrete plant was estab- 
 lished, which had in connection a sand and gravel washer and 
 separator. 1 This consisted of a tube about two and one-half 
 feet in diameter and seventeen feet long, made of one-quarter- 
 inch boiler iron and revolving about an axis slightly inclined to 
 the horizontal. Angle irons were riveted on the inside of the 
 tube to carry the material up on the side and drop it again, 
 while a spray of water issued from a perforated pipe inside the 
 tube. The materials were separated by screens near the lower 
 end of the tube. The material contained considerable earthy 
 matter and is said to have been fairly well washed by this pro- 
 cess. 
 
 Another style of sand washer was designed by the contract- 
 ors for the construction of Lock No. 3, improvement of Alle- 
 ghany River. 2 The sand contained earthy matter and some 
 coal, the latter being hard to remove by ordinary processes. A 
 . * 
 
 1 Report of Lt. Edw. Burr, Report Chief of Engineers, 1891, p. 3334. 
 2 \V. H. Rober, Engineering News, Feb. 16, 1899. 
 
170 CEMENT AND CONCRETE 
 
 large barrel or tank, nine feet in diameter and seven feet high, 
 was provided with double floor, the upper one being pierced 
 with one-inch holes. Paddles were attached to a vertical shaft 
 in the axis of ^he tank and revolved by suitable gearing, while 
 water was forced into the space between the two floors. The 
 water finding its way through the holes in the upper floor, passed 
 up through the sand and overflowed at the top, carrying with 
 it the coal and sediment. The cost of washing is said to have 
 been about seven cents per cubic yard, but it is evident that 
 methods of handling would have to be quite perfect to keep the 
 cost at so low a figure. 
 
 ART. 33. CONCLUSIONS. WEIGHT. COST 
 
 267. REQUIREMENTS FOR GOOD SAND. In conclusion, then, 
 we may say that good sand may consist of grains of almost 
 any moderately hard rock that is not liable to future alteration 
 in the work. The grains may be of any shape, but preferably 
 should be sharp and angular or lenticular in form, not rounded 
 and smooth. The sand should not contain such impurities as 
 loam or humus, but for most purposes a small percentage of 
 clay or fine rock dust is not objectionable. Clay should not, 
 however, be permitted in sand for use in sea water. 
 
 Coarse grained sands are better than fine grained ones, but 
 a mixture of fine and coarse is excellent, especially where but 
 a small amount of cement is used, because such a mixture con- 
 tains less voids and will make a less permeable mortar, while giv- 
 ing a good strength. As might be expected, the deleterious effect 
 of poor sand is more apparent the larger the dose of sand used. 
 
 268. Weight of Sand. It is evident from what has pre- 
 ceded that the weight of sand per cubic foot will vary greatly, 
 not only with the character of the rock from which it came, 
 but also with its physical condition. Natural sand, as it or- 
 dinarily occurs, will weigh about as follows, according to its 
 condition : 
 
 Moist and loose . . . 70 to 90 pounds per ru. ft. 
 
 Moist and shaken 75 to 100 " 
 
 Dry and loose 75 to 105 " 
 
 Dry and shaken 90 to 125 " " 
 
 When settled in water, weight of wet 
 
 sand, voids full ...."".... 100 to 140 " " 
 
. 
 WEIGHT AND COST 171 
 
 If the rock from which the sand is made weighs, say, one 
 hundred sixty pounds per cubic foot solid (specific gravity, 
 2.56), then the sand will weigh per cubic foot 120, 100, and 
 80 pounds, for voids of 25, 37.5 and 50 per cent., respectively. 
 
 269. Cost of Sand. The cost of sand will, of course, vary 
 with the locality. In exceptional cases where it is found di- 
 rectly at the works, it may not cost more than twenty to thirty 
 cents per cubic yard to deliver it on the mixing platform. If 
 it has to be pumped from the bed of a river or lake and can be 
 conveyed to the work in scows with a tow of not more than 
 ten miles, it may be delivered at the work for from forty to 
 sixty cents per cubic yard. If it must be hauled in wagons for 
 some distance, it may cost from fifty cents to one dollar per 
 yard; and again, if sand is so difficult to obtain that it must be 
 made by crushing rock, it may cost from one dollar to three 
 dollars per yard. Usually from sixty cents to a dollar is a fair 
 price for sand. Several examples of cost of sand will be given 
 in connection with the subject of cost of concrete. 
 
CHAPTER XII 
 
 MORTAR: MAKING AND COST 
 ART. 34. PROPORTIONS OF THE INGREDIENTS 
 
 270. CAPACITY OF CEMENT BARRELS. Since there i* no 
 standard size for cement barrels, the capacities vary consider- 
 ably, Portland cement barrels ranging from 3.1 to 3.6 cu. ft., while 
 natural cement barrels contain from 3.4 to 3.8 cu. ft. In Ger- 
 many cement is packed to weigh three hundred ninety-six 
 pounds per barrel, gross, the net weight being about three hun- 
 dred seventy-five pounds. American Portland usually weighs 
 four hundred pounds gross or about three hundred eighty 
 pounds net. 
 
 In 1896 the Boston Transit Commission had a number of 
 measurements made of the capacity of Portland cement bar- 
 rels, and these have been compiled by Mr. Sanford E. Thompson. 1 
 Table 58 presents some of the averages obtained from this series 
 of tests. It is seen that the capacity of the barrels varied from 
 3.12 to 3.50 cu. ft., the mean volume being 3.29 cu. ft. The 
 difference between the capacity of the barrel and the volume 
 of the packed cement contained is due to the fact that there 
 is usually a small space beneath the head not filled with cement. 
 A barrel of packed cement makes about 1.25 barrels, measured 
 loose . 
 
 271. Natural cements made in the East are packed to 
 weigh three hundred pounds net, while some of the Western 
 natural cements weigh but two hundred sixty-five pounds per 
 barrel net. Any of the natural cement factories will doubtless 
 pack their cement to suit customers on large orders, and there 
 seems to be little reason for this variation in weight between 
 the West and the East. There would perhaps be some trouble 
 in getting three hundred pounds of a very light, finely ground, 
 natural cement in the ordinary sized barrel, but two hundred 
 
 1 Engineering News, Oct. 4, 1900. 
 
 172 
 
PROPORTIONS 
 
 173 
 
 TABLE 58 
 Capacity of Portland Cement Barrels 
 
 
 HIGHEST. 
 
 LOWEST. 
 
 MEAN. 
 
 Height of barrel between heads, feet .... 
 Capacity between heads, cubic feet 
 Volume of packed cement in barrel, cubic feet . 
 Volume of loose cement in barrel, cubic feet . 
 Net weight of cement in barrel, pounds . 
 Weight per cubic foot of cement as packed in 
 barrel pounds 
 
 2.19 
 3.50 
 3.48 
 4.10 
 387.0 
 
 123 16 
 
 2.01 
 3.12 
 3.03 
 3.75 
 370.7 
 
 113 81 
 
 2.00 
 3.21) 
 3. IK 
 4.07 
 377.4 
 
 I ]g 79 
 
 Weight per cubic foot, loose, pounds .... 
 
 100.40 
 
 88.r,2 
 
 92.63 
 
 NOTE. Results are averages of thirty-one tests with seven brands, four 
 of which were American. The above data compiled by Sanford E. Thomp- 
 son and published in Engineering News of Oct. 4, 1900. 
 
 eighty pounds may be put in a barrel without difficulty, and it 
 would seem that a compromise might be made on this weight. 
 
 272. QUANTITY OF SAND. The amount of sand to be used 
 in mortar will depend entirely on the character of the work 
 and the quality of the cement and sand. If it is merely a 
 matter of strength to be developed, no special care need be 
 taken to have the voids in the sand filled with cement, but if 
 an impervious mortar is desired, the mortar must not be too 
 poor in cement, even though only a moderate strength is re- 
 quired. 
 
 In France the proportions of cement and sand are usually 
 given in terms of kilograms of cement to one cubic meter of 
 sand. In England and America the proportions are usually 
 given by volume, as so many parts of cement to one of sand, 
 while in Germany the proportions are given by weight. The 
 bulk of cement varies so much according to the degree of pack- 
 ing, and the volume of sand is so varied by the amount of mois- 
 ture contained, that the German method of stating proportions 
 by weight seems to be the most logical one to adopt. 
 
 273. Proportions by Volume. It has been shown that the 
 volume of a given quantity of cement may vary twenty-five 
 per cent, according as it is measured packed or loose, and that 
 likewise the volume of sand may vary twenty per cent, accord- 
 ing to the amount of moisture contained. This makes it ne- 
 cessary to take great precaution in proportioning mortars by 
 
174 CEMENT AND CONCRETE 
 
 volume if the desired richness of the mortar is to be assured. 
 Nevertheless, mortars for use in actual construction are usually 
 proportioned by volume. The usual method is to taste the 
 proportions as one part of packed cement (as it comes in the 
 barrel or bag) to so many parts of loose sand, but proportions 
 are sometimes stated as volumes of loose sand to one volume 
 of loose cement. 
 
 274. Equivalent Proportions by Weight and Volume. As 
 cement is now so frequently sold in sacks of one-fourth barrel 
 each, in which the cement is not so compact as in a barrel, we 
 have assumed the contents of a barrel to be 3.45 cu. ft. for 
 Portland, and 3.75 cu. ft. for natural, which are somewhat 
 higher than the mean actual capacities of stave barrels as 
 shown by tests. At three hundred eighty pounds and two 
 hundred eighty pounds net weight respectively for Portland 
 and natural, this is equivalent to one hundred ten pounds per 
 cubic foot and seventy-five pounds per cubic foot packed. If 
 we also assume that loose cement weighs eighty-five pounds per 
 cubic foot for Portland and sixty pounds per cubic foot for 
 natural; and that loose, dry sand weighs one hundred pounds 
 per cubic foot, while loose, damp sand weighs eighty pounds per 
 cubic foot, we may obtain the following comparisons, Table 59. 
 
 275. It is evident that in all specifications and in reports 
 of tests, as well as in the use of cement, the method of stating 
 proportions should be made clear, and in interpreting the re- 
 sults of tests this must be borne in mind. For instance, in 
 tests to compare the value of limestone screenings with quartz 
 sand, proportions by weight will favor the quartz, while pro- 
 portions by volume will favor the screenings, since the latter 
 are lighter. 
 
 276. Richness of Mortar. Mortars containing small amounts 
 of sand are often stronger than neat cement mortars. Es- 
 pecially is this true of most natural cements. Some of these 
 will give as high strengths when mixed with two parts sand by 
 weight as when neat, and usually the one-to-one mortars are 
 stronger than the neat mortars. These remarks refer to tensile 
 tests where a good quality of sand is used and the mortars are 
 three months old or more. The neat cement mortars gain 
 their strength more rapidly, short time tests usually not show- 
 ing the results mentioned, Portland cements of good quality 
 
PROPORTIONS 
 
 175 
 
 TABLE 59 
 
 Comparison of Proportions by Weight and Volume 
 
 
 EQUIVALENT PARTS SAND, PROPORTIONS STATED BY VOLUME. 
 
 
 PORTLAND CEMENT. 
 
 NATURAL CEMENT. 
 
 PARTS DRY 
 
 
 . 
 
 
 
 ^VH 
 
 o ** 
 
 >>^ 
 
 2a 
 
 SAND 
 
 
 o-2 1 
 
 s> "* 
 
 f** s 
 
 2 o c 
 
 g 5 g 
 
 al 
 
 
 TO ONE 
 CEMENT BY 
 WEIGHT. 
 
 """ v 2 
 
 i-s^s 
 
 on** S 
 
 ."2*. - 
 
 III 
 
 I!* 
 
 J CK 5 
 
 ^^1 
 
 ^3 2 2 
 05 
 33^ 
 *~ 
 
 CC u 
 
 ill 
 
 i-5 *-"3 
 
 5 3 C U 
 
 ^03gU 
 
 jp 
 
 ** O 
 
 
 
 ~ 
 '2 5 2 
 * 
 
 ^1 1 
 
 CLi 
 
 1-2 
 
 ^o 3 
 
 ^ll 
 
 Ill 
 
 *ai 
 
 HI 
 
 ^ J 
 
 1 
 
 1.10 
 
 1.38 
 
 085 
 
 1.06 
 
 0.75 
 
 0.94 
 
 0.60 
 
 0.75 
 
 2 
 
 2.20 
 
 2.75 
 
 1.70 
 
 2.12 
 
 1.50 
 
 1.88 
 
 1.20 
 
 1.50 
 
 3 
 
 3.30 
 
 4 12 
 
 2.55 
 
 3.19 
 
 2.25 
 
 2.81 
 
 1 80 
 
 2.25 
 
 4 
 
 4.40 
 
 5.50 
 
 3.40 
 
 425 
 
 3.00 
 
 3.75 
 
 2.40 
 
 3.00 
 
 5 
 
 5.50 
 
 0.88 
 
 4.25 
 
 5.31 
 
 3.75 
 
 4.6J) 
 
 3.00 
 
 3.75 
 
 6 
 
 6.60 
 
 8.25 
 
 5.10 
 
 6.38 
 
 4.50 
 
 5.62 
 
 3.60 
 
 4.50 
 
 In preparing the above table the following assumptions are made : 
 
 MATERIAL. 
 
 WEIGHT 
 
 IN A 
 
 BARREL. 
 
 VOLUME 
 
 OF A 
 
 BARREL. 
 
 WEIGHT TER CUBIC FOOT. 
 
 Packed. 
 
 Loose 
 Dry. 
 
 Loose 
 Damp. 
 
 Portland cement . 
 Natural cement . 
 Sand .... 
 
 380 
 
 280 
 
 3.45cu. ft. 
 3.75cu. ft. 
 
 110 
 75 
 
 85 
 60 
 100 
 
 80' 
 
 
 
 
 usually give about the same tensile strength neat and with 
 one part sand by weight. Tests showing the rate of decrease 
 of strength with added sand are discussed in 363 to 365. 
 
 Portland cements are usually mixed with from one to three 
 parts sand by weight, and natural cements are mixed with 
 from one to four parts by weight (or three-fourths part to 
 three parts by measure). For certain special purposes poorer 
 mortars are sometimes employed. To arrive at the proper 
 proportion to use in mortar for a given purpose, the tables of 
 strength given in Chapter. XV will be of value. 
 
 277. Effect of Pebbles. If the sand contains pebbles, the 
 proportions should be considered in a little different way. 
 Suppose we make a one-to-three mortar with sand that con- 
 tains ten per cent, of pebbles. We have in reality, then, 3 X .90 
 = 2.7 parts of sand to one of cement, and .3 part pebbles em- 
 bedded in this richer mortar. This point is of special signifi- 
 cance in making concrete from gravel containing some sand, or 
 
176 CEMENT AND CONCRETE 
 
 from broken stone from which the fine particles or screenings 
 have not been removed. Such fine particles serve to weaken 
 the mortar by increasing the dose of sand, while the pro- 
 portion of aggregate is diminished. In using aggregates con- 
 taining some fine material, then, or in using sand containing 
 pebbles or fine gravel, one should not permit himself to be de- 
 ceived as to the actual richness of the resulting mortar or 
 concrete. 
 
 278. AMOUNT OF WATER FOR MORTAR. The amount of 
 
 water required for mortar will vary with the proportion of sand 
 to cement, the character and condition of the ingredients, the 
 weather, and the purpose which the mortar is to serve. If the 
 water is stated as such a percentage of the combined weight 
 of cement and sand, the amount required for a rich mortar 
 will be greater than for a poor one, since the cement requires 
 more water than the sand. Fine cement will require more 
 water than coarse; the same is true of sand. Sand from ab- 
 sorbent rock will require a larger amount of water. On a hot, 
 dry day, more water must be used to allow for evaporation; 
 and again, if the mortar is to be placed in contact with brick 
 or porous stone, the mortar must be more moist than when 
 used in connection with metal, or with hard rocks such as 
 granite. All of these points must be borne in mind when 
 determining the proper consistency for a given purpose. 
 
 279. We may arrive at the approximate amount of water 
 required in the following manner: find what proportion of 
 water is required for the neat cement. This will vary among 
 different samples, and especially between Portland and natural 
 cements; the former requiring twenty to twenty-eight per cent, 
 of water (by weight), and the latter thirty to forty per cent. 
 Then find the amount of water required to bring the sand alone 
 to the consistency of mortar. This will vary considerably, fine 
 sand requiring much more water than coarse, etc., as men- 
 tioned above. Having these two quantities, we may find the 
 amount of water required for a mortar having any given pro- 
 portions of these samples of cement and sand. Thus, suppose 
 we find that the neat cement requires twenty-five per cent, 
 water and the sand ten per cent, water to bring them to the 
 proper consistency. If we wish to make a one-to-three mortar 
 from these ingredients, using one hundred pounds of cement, the 
 
MIXING 177 
 
 required amount of water is (100 X .25) + (100 X 3 X .10) 
 = 25 + 30 = 55 pounds. 
 
 280. However, it will usually be better to experiment di- 
 rectly upon the mixture which it is proposed to use, and for 
 this purpose the following rule will be found of value. For 
 ordinary purposes, that amount of water should be used which 
 for given weights of the dry ingredients will give the least 
 volume of mortar with a moderate amount of packing. In the 
 actual use of mortars it is not practicable to state that a cer- 
 tain definite amount of water shall always be used with given 
 quantities of the dry materials. ' It is the resulting consistency 
 of the mortar that must be specified and insisted upon, while 
 the amount of water required to produce this consistency will 
 vary from day to day and must be left to the discretion of the 
 inspector or foreman. For a discussion of the relation of con- 
 sistency to the tensile strength of the mortar, see Art. 46. 
 
 ART. 35. MIXING THE MORTAR 
 
 281. Having decided upon the proportions of cement, sand 
 and water, it remains to incorporate these into a plastic, homo- 
 geneous mass. The size of the batch should be so adjusted, if 
 possible, that a full barrel of cement shall be used, and for 
 careful work the amount of sand should be weighed instead of 
 measured. Where this is impracticable, the condition of the 
 sand from day to day, as regards the amount of moisture con- 
 tained, should be taken into account (see 262 and 263). 
 
 Mortar is usually mixed by hand, but where large amounts 
 are to be used, machine mixers may profitably be introduced. 
 
 282. HAND MIXING. For hand mixing, a water tight plat- 
 form or shallow box should be used, of such a size that the given 
 batch will not cover the bottom more than four inches deep. 
 
 If the sand is measured, a bottomless box, provided with 
 two handles at each end, will be found more convenient than 
 the bottomless barrel which is often employed for this purpose. 
 When the sand is delivered on the mixing platform in barrows, 
 the latter may be fitted with rectangular boxes to avoid . re- 
 measuring. A two-wheel cart, the box of which may be in- 
 verted to discharge the contents on the mixing platforjn, will 
 also be found very serviceable when the runway is suited to 
 such a vehicle. 
 
178 CEMENT AND CONCRETE 
 
 The proper amount of sand is evenly spread on the plat- 
 form, the cement is then dumped on top of the sand and spread 
 out over it to an even thickness. With either hoes or shovels 
 the dry materials are then thoroughly mixed, until, when a 
 small amount is taken in the hand, it will appear of uniform 
 color throughout. From two to five turnings of the materials, 
 according to the expertness of the workmen, will be required to 
 produce this result. The dry mixture is then drawn to the 
 edges of the platform to form a ring, and the requisite amount 
 of water is added at one time in the center. The mixture is 
 then gradually incorporated with the water, and the mass is 
 thoroughly worked until plastic and homogeneous. Should it 
 be found that too little water has been used, a small amount 
 may be added from a sprinkling pot or rose nozzle, but the 
 mass should always be worked over again after such addition. 
 Four shovels may be used at one platform, but if the mixing is 
 done by hoes, not more than two can be used to advantage 
 with a batch of ordinary size. 
 
 Some engineers prefer one method and some the other, but 
 in whatever manner done, the mixing should not be stinted. 
 From two to four turnings of the mass are usually considered 
 sufficient, but as a general rule it will be found that further 
 mixing, beyond that required to just give the mass a uniform 
 appearance, will be amply repaid in the strength of the result- 
 ing mortar. (See Art. 47.) 
 
 283. MACHINE MIXING. Where large quantities of mortar 
 are required, machine mixers are sometimes used. A very 
 complete plant for mortar-making was used in building the 
 Titicus Dam. 1 In this case machinery was used in measuring 
 the proportions of cement and sand as well as in making the 
 mortar. The measuring apparatus consisted of two cylindrical 
 troughs, one for cement and one for sand. Each trough was 
 divided, by means of six radial vanes and four discs, into eigh- 
 teen equal compartments. These cylinders revolved in cast 
 iron boxes which were so constructed as to serve as hoppers 
 for filling the compartments. Three compartments were pre- 
 sented to the hoppers at once, and slides were provided by 
 which any of the hoppers could.be cut off at will. The cylin- 
 
 Engineering Record, August 3, 1895. 
 
INGREDIENTS REQUIRED 179 
 
 ders being geared to the same pinion, it was possible, by means 
 of the slides, to make any desired proportion of cement and sand 
 from neat cement, to three parts sand to one cement. 
 
 The mixing machine "consisted essentially of a semi-cylindri- 
 cal wrought-iron trough with extended flaring sides, with ele- 
 ments slightly inclined to the horizontal, and in its axis a re- 
 volving shaft with oblique radial blades set at an incline of 
 ninety degrees to each other and of a length to just clear the 
 bottom of the trough." 
 
 284. Another form of machine that is sometimes employed 
 consists of a semi-cylindrical trough in which rotates an axis 
 carrying a blade in the form of a screw. The materials are 
 fed to the mixer at one end and the screw mixes them while 
 working the mass toward the other end. 
 
 ART. 36. COST OF MORTAR* 
 
 285. INGREDIENTS REQUIRED FOR ONE CUBIC YARD OF 
 MORTAR. The character of the ingredients used in making cement 
 mortar varies so much that it is difficult to accurately deter- 
 mine the quantities of materials required for a proposed mortar 
 except by experimenting with the materials that are to be 
 employed. It has been shown that the weights per cubic foot 
 of both cement and sand vary greatly according to the condi- 
 tions of packing, the moisture, etc. The percentage of voids 
 in the sand is one of the most important variations affecting 
 the amount of mortar made with certain materials mixed in 
 given proportions. The consistency of the mortar also has a 
 marked effect, and different cements show a considerable varia- 
 tion in the volume of mortar that a given weight will yield. 
 In any general treatment of the question, then, we may expect 
 only approximate results, and the tables given in this connec- 
 tion must be considered in this light. 
 
 286. Results of Experiments. The tests from which Tables 
 60 and 61 were derived, were made with a natural sand weigh- 
 ing about one hundred pounds to the cubic foot, dry, and having 
 about three-eighths of the bulk voids. The grains varied in 
 size from 0.01 in. to 0.1 in. in diameter with a few grains out- 
 
 1 Portions of this article were contributed to " Municipal Engineering," 
 and appeared in that magazine, Feb., 1809. 
 
180 
 
 CEMENT AND CONCRETE 
 
 TABLE 60 
 Ingredients Required for One Cubic Yard of Mortar, Portland Cement 
 SAND WEIGHS ABOUT 100 LBS. PER CUBIC FOOT. VOIDS THREE-EIGHTHS OF VOLUME 
 
 Proportions by Volume Dry 
 Loose Sand to Loose Cement, 
 Loose Cement Assumed at 
 85 Ibs. per Cu. Ft. 
 
 n 
 
 
 
 O CO 00 00 Si OS ' ' 
 
 Cement. 
 
 * *y 
 
 - 
 
 O OS O * CO (M . . 
 1-' CO G^i i-i i-i i-i ' 
 
 Pounds. 
 
 rij 
 
 O O O O *-O iO * 
 
 oo ^ cs 5 o ^JH 
 
 Proportions by Volume Dry 
 Loose Sand to Packed Cement, 
 Cement Assumed at 114 Ibs. per 
 Cu. Ft. or 380 Ibs. per Bbl. of 
 3.33 Cu. Ft. 
 
 
 
 *-> 
 
 T* ^ ?M r- OS T-H co 
 O lr^ OO X OO OS OS 
 
 Cement. 
 
 pq S= 
 
 .. 
 
 BS1B2222 
 
 Pounds. 
 
 - 
 
 00 CC i i CO SO O ><ti -<*l 
 
 Proportions by Volume Dry 
 Loose Sand to Packed Cement, 
 Packed Cement Assumed 
 at 104 Ibs. per Cu. Ft. or 380 ibs. 
 per Bbl. of 3.65 Cu. Ft. 
 
 O 
 
 * 
 
 O t H 00 O 
 O t- 00 00 OS ' 
 
 o o o o o o 
 
 Cement. 
 
 1 
 
 - 
 
 o^ gco3 : 
 
 t^^CNCN^^^ 
 
 Pounds. 
 
 
 
 i i CO GO OO rH O CO 
 00 O O l>- SO >O rfi 
 
 <N 1-1 !-t 
 
 PROPORTIONS BY 
 WEIGHT, DRY SAND AND 
 CEMENT. 
 
 ll 
 
 V 
 
 t OO Tjn 00 i-H CO 
 _ U5 i>. 00 00 OS OS ' 
 
 Cement. 
 
 . te 
 
 w 
 
 00 CO 00 OS O 
 *< O l^ O O G<l T-H 
 !>' * <N <M' i-i i-I r-J ' 
 
 Pounds. 
 
 - 
 
 o >o p p o o o 
 
 ^ s s ^ o ^ ^ . 
 
 AO I OJ. CTKVS SXHVJ 
 
 e 
 
 o A-.-.- 
 
INGREDIENTS REQUIRED 
 
 181 
 
 1 H 
 
 & i 
 
 I s 
 
 
 
 60 - 
 
 s * 
 
 Ctt 
 
 Proportions by Volume Dry 
 Loose Sand to Loose Cement, 
 Loose Cement Assumed at 
 G0# per Cubic Foot. 
 
 Proportions by Vol- 
 ume Dry Loose 
 Sand to Packed Ce- 
 ment, Packed 
 Cement Assumed at 
 75# per Cu. Ft. or 
 280# Net perBbl. 
 
 ol- 
 e 
 Ce- 
 
 pe 
 bl. 
 
 Proportions by 
 ume Dry Loo 
 Sand to Packed 
 ment, Cemen 
 Assumed at 71 # 
 Cubic Foot o 
 265# Net per 
 
 PROPORTIONS 
 WEIUHT, DRY SAN 
 CEMENT. 
 
 
 spanoj 
 
 11 
 
 spunoj 
 
 spanoj 
 
 spunoj 
 
 III 
 
 spanoj 
 
 SXHVJ 
 
 d o o 6 
 
 r~ O O i ' ' 
 
 ^ ;o 3q o . . . 
 
 t^ co 'N' f-J . . . 
 
 8:0 -c co 
 oo c? c> . . . 
 
 00 CO -N -H' . . . 
 
 o GO ;r. t- 
 
 rtl O * 1-- . . . 
 00 Tt< 7<l -^ . . . 
 
 O 00 00 t^ 
 O >O t> GO QO O6 . 
 
 o o c o o o . 
 
 t~ t^ O O t^ O 
 
 Tt< 1-1 t~- O O CO . 
 
 i-' "*<' c<i <M' -' -H . 
 
 o o o o 
 
 CO -^ C^ (^ r-l . . 
 
 O O O O O ' 
 
 -* 1-1 X t O 
 
 n r i^ o <* ' 
 
 1-1 X OS 
 _ to i x x . . 
 
 O O CO * 
 
 t<Tt<xo 
 
 x $< n r* 
 
 s 
 
 o o o o 
 
 Tt< OC Tt 
 
 1 ' 'M O Tf t^- ^ 
 _ o r-- x x x x 
 
 -^ co oi i-< r-J i-J 
 
 70 O t- CC X CO 
 OS ^ "O C: ?O * 
 
 x' ^' co c<i ?i ^' ^' 
 
 O ^ 3^ CO'* >O 
 
182 CEMENT AND CONCRETE 
 
 side of these limits. The consistency of the mortar was such 
 that when struck with the shovel blade the moisture would 
 glisten on the smooth surface thus formed. In the experiments 
 the proportions were determined by weight, and the results for 
 proportions by volume were deduced from them. The results 
 for neat natural cement mortar and for the natural cement 
 mortars containing more than four parts sand by weight were 
 derived by analogy. 
 
 287. Explanation of Tables. The first section of Table 60 
 gives the amount of materials required for Portland cement 
 mortar when the proportions are stated by weight; the second 
 and third sections refer to proportions by volume of loose sand 
 to packed cement when the size of the cement barrel is as- 
 sumed at 3.65 cu. ft. and 3.33 cu. ft., respectively. The fourth 
 section gives the materials required when the proportions are 
 given in terms of loose sand to loose cement. Likewise, the 
 first section of Table 61 for natural cement refers to proportions 
 by weight; the second, third and fourth sections, to propor- 
 tions by volume of loose sand to packed cement when the 
 cement weighs 265 pounds, 280 pounds and 300 pounds, net r 
 per barrel, respectively; while the fifth section refers to propor- 
 tions of loose sand to loose cement. 
 
 As has been shown, the method of stating proportions by 
 weight is the most accurate, but when the sand does not ap- 
 proximate the weight of 100 pounds per cubic foot when shoveled 
 dry into a measure, the sections of the tables referring to weight 
 proportions may require a correction, and it may be simpler 
 to use the sections giving proportions by volume of loose sand 
 to packed cement. The method of stating proportions by vol- 
 umes of loose sand to loose cement is to be deprecated, but since 
 it is occasionally used, provision is made for it in the tables. 
 
 In using those portions of the tables where the proportions 
 are stated by volume, it should be borne in mind that if the 
 sand is damp when used it will weigh less per cubic foot, and 
 hence more, by measure, will be required to make a cubic yard 
 of mortar. 
 
 288. Estimating Cost of Mortar. With the data given in 
 Tables 60 and 61 and a knowledge of unit prices of the mate- 
 rials used in the mortar, one may estimate the cost of the ma- 
 terials in a given quantity of mortar. The cost of the mixing 
 
COST OF MORTAR 
 
 183 
 
 will, of course, depend upon the cost of labor, the method em- 
 ployed, etc., and may vary from fifty cents to a dollar and 
 fifty cents per cubic yard. If we assume, for illustration, that 
 natural cement can be delivered on the mixing platform for 
 $1.10 per barrel of 280 pounds net, that sand costs 60 cents 
 per cubic yard, and the mixing costs $1.00 per yard of mortar, 
 then we have for the cost of a mortar composed of one part 
 cement to two parts sand by weight: - 
 
 3.46 bbls. cement at $1.10 $3.80 
 
 0.72 cu. yd. dry sand at .60 43 
 
 Cost of mixing per cu. yd 1.00 
 
 Total cost of one cu. yd. of mortar $5.23 
 
 289. For approximate results, Tables 62 and 63 give the 
 cost of the materials used in a cubic yard of mortar for different 
 prices of cement. In Table 62 the proportions by weight only 
 are indicated, since for Portland the proportions by volume of 
 loose sand to packed cement vary so little from proportions 
 by weight. 
 
 TABLE 62 
 Cost of Portland Cement Mortar 
 
 COST OF CEMENT AND SAND IN ONE CUBIC YARD OK PORTLAND CEMENT 
 MORTAR. SAND, 75 CENTS PER Criuc YARD 
 
 COST OF PORT- 
 LAND 
 CEMENT PER 
 BARREL 
 OF 380 POUNDS 
 NET. 
 
 COST OF INGREDIENTS IN MORTAR, IN DOLLARS. 
 Proportions in Mortar by Weight, Parts Sand to One of Cement. 
 
 
 
 l 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 $1.20 
 
 8.90 
 
 5.33 
 
 3.89 
 
 3.03 
 
 2.56 
 
 2.23 
 
 2.02 
 
 1.30 
 
 9.62 
 
 5.73 
 
 4.17 
 
 3.23 
 
 2.72 
 
 2.36 
 
 2.13 
 
 1.40 
 
 10.36 
 
 6.14 
 
 4.44 
 
 3.43 
 
 2.88 
 
 2.49 
 
 2.24 
 
 1.60 
 
 11.10 
 
 6.55 
 
 4.72 
 
 3.63 
 
 3.03 
 
 262 
 
 2.35 
 
 1.60 
 
 11.84 
 
 6.96 
 
 5.00 
 
 3.83 
 
 3.19 
 
 2.75 
 
 2.46 
 
 1.70 
 
 12.58 
 
 7.37 
 
 5.27 
 
 4.03 
 
 3.35 
 
 2.88 
 
 2.57 
 
 1.80 
 
 13.32 
 
 7.77 
 
 5.55 
 
 4.23 
 
 3.51 
 
 3.01 
 
 2.68 
 
 190 
 
 14.06 
 
 8.18 
 
 5.82 
 
 4.43 
 
 3.67 
 
 3.13 
 
 2.79 
 
 2.00 
 
 14.80 
 
 8.59 
 
 6.10 
 
 4.63 
 
 3.82 
 
 3.26 
 
 2.90 
 
 2.10 
 
 15.54 
 
 9.00 
 
 6.38 
 
 4.83 
 
 3.98 
 
 3.39 
 
 3.01 
 
 2.20 
 
 16.28 
 
 9.41 
 
 6.65 
 
 5.03 
 
 4.14 
 
 3.52 
 
 3.12 
 
 2.30 
 
 17.02 
 
 9.81 
 
 6.93 
 
 5.23 
 
 4.30 
 
 3.65 
 
 3.23 
 
 2.40 
 
 17.76 
 
 10.22 
 
 7.20 
 
 5.43 
 
 4.46 
 
 3.78 
 
 3.34 
 
 2.50 
 
 18.50 
 
 10.63 
 
 7.48 
 
 5.63 
 
 4.61 
 
 3.91 
 
 3.45 
 
 2.60 
 
 19.24 
 
 11.04 
 
 7.76 
 
 5.83 
 
 4.77 
 
 4.04 
 
 3.56 
 
 2.70 
 
 19.98 
 
 11.45 
 
 8.03 
 
 6.03 
 
 4.93 
 
 4.17 
 
 3.67 
 
 2.80 
 
 20.72 
 
 11.85 
 
 8.31 
 
 6.23 
 
 5.09 
 
 4.30 
 
 3.78 
 
 2.90 
 
 21.46 
 
 12.26 
 
 8.58 
 
 6.43 
 
 5.25 
 
 4.42 
 
 3.89 
 
 3.00 
 
 2220 
 
 12.67 
 
 8.86 
 
 6.63 
 
 5.40 
 
 4.55 
 
 4.00 
 
184 
 
 CEMENT AND CONCRETE 
 
 TABLE 63 
 Cost of Natural Cement Mortar 
 
 COST or CEMENT AND SAND IN ONE CUBIC YARD OF NATURAL CEMENT 
 MORTAR. SAND, 75 CENTS PER CUBIC YARD 
 
 METHOD OF STATING 
 PROPORTIONS, AND 
 WEIGHT OF CEMENT IN 
 ONE BARREL. 
 
 BTS SAND TO! 
 F CEMENT. | 
 
 COST OF CEMENT PER BARREL, DOLLARS. 
 
 
 
 
 
 
 
 
 
 
 
 
 fc- 
 
 0.60 
 
 0.70 
 
 0.80 
 
 0.90 
 
 1.00 
 
 1.10 
 
 1.20 
 
 1.30 
 
 1.40 
 
 1.50 
 
 
 
 
 5.07 
 
 5.92 
 
 6.76 
 
 7.60 
 
 8.45 
 
 9.30 
 
 10.14 
 
 10.98 
 
 11.83 
 
 12.68 
 
 
 1 
 
 3.50 
 
 4.02 
 
 4.54 
 
 5.06 
 
 5.59 
 
 6.11 
 
 6.63 
 
 7.15 
 
 7.67 
 
 8.19 
 
 
 2 
 
 2.74 
 
 3.10 
 
 3.47 
 
 3.83 
 
 4.20 
 
 4.57 
 
 4.93 
 
 5.30 
 
 5.66 
 
 6.03 
 
 
 3 
 
 2.23 
 
 2.50 
 
 2.78 
 
 3.05 
 
 3.32 
 
 3.59 
 
 3.86 
 
 4.14 
 
 4.41 
 
 4.68 
 
 
 4, 
 
 1.90 
 
 2.12 
 
 2.33 
 
 2.54 
 
 2.76 
 
 2.98 
 
 3.19 
 
 3.41 
 
 3.62 
 
 3.83 
 
 
 
 
 4.48 
 
 5.23 
 
 5.98 
 
 6.72 
 
 7.47 
 
 8.22 
 
 8.96 
 
 9.71 
 
 10.46 
 
 11.20 
 
 
 1 
 
 3.14 
 
 3.60 
 
 4.06 
 
 4.52 
 
 4.98 
 
 5.44 
 
 5,90 
 
 6.36 
 
 6.82 
 
 7.28 
 
 
 2 
 3 
 
 2.48 
 2.04 
 
 2.80 
 2.28 
 
 3.12 
 2.52 
 
 3.45 
 2 76 
 
 3.77 
 3.00 
 
 4.09 
 3.24 
 
 4.42 
 3.48 
 
 4.74 
 3.72 
 
 5.06 
 3.96 
 
 5.38 
 4.20 
 
 . 
 
 4 
 
 1.75 
 
 1.94 
 
 2.13 
 
 2.32 
 
 2.51 
 
 2.70 
 
 2.89 
 
 3.08 
 
 3.27 
 
 3.46 
 
 By volume ; parts dry I 
 loose sand to packed 1 
 
 
 1 
 
 5.07 
 3.13 
 
 5.92 
 3.58 
 
 6.76 
 4.02 
 
 7.60 
 4.46 
 
 8.45 
 4.91 
 
 9.30 
 5.36 
 
 10.14 
 
 5.80 
 
 10.98 
 6.24 
 
 11.83 
 6.69 
 
 12.68 
 7.14 
 
 cement. Cement as-J 
 
 2 
 
 2.29 
 
 2.57 
 
 2.85 
 
 3.14 
 
 3.42 
 
 3.70 
 
 3.99 
 
 4.27 
 
 4.55 
 
 4.84 
 
 sumed 265 Ibs. per bbl. 
 
 3 
 
 1.83 
 
 2.04 
 
 2.24 
 
 2.45 
 
 2.65 
 
 2.86 
 
 3.06 
 
 3.26 
 
 3.47 
 
 3.67 
 
 of 3.75 cu. ft. 
 
 4 
 
 1.63 
 
 1.79 
 
 1.95 
 
 2.11 
 
 2.27 
 
 2.43 
 
 2.59 
 
 2.75 
 
 2.91 
 
 3.07 
 
 By volume; parts dry ( 
 
 1 
 
 2.94 
 
 3,36 
 
 3.77 
 
 4.19 
 
 4.61 
 
 5.03 
 
 544 
 
 5.86 
 
 6.28 
 
 6.70 
 
 loose sand to packed 1 
 cement. Cement as { 
 sumed 300 Ibs. per bbl. 
 
 2 
 3 
 
 2.22 
 1.82 
 
 2.50 
 2.02 
 
 2.77 
 2.22 
 
 3.05 
 2.42 
 
 3.32 
 2.62 
 
 360 
 2.82 
 
 3.87 
 3.02 
 
 4.15 
 3.22 
 
 4.42 
 3.42 
 
 4.70 
 3.60 
 
 of 3.75 cu. ft. t 
 
 4 
 
 1.59 
 
 1.75 
 
 1.91 
 
 2.06 
 
 2.22 
 
 2.38 
 
 2.53 
 
 2.69 
 
 2.85 
 
 3.00 
 
 In Table 63 the cost of materials in one cubic yard of natural 
 cement mortar is given, 1st, for various parts of sand to one 
 of cement by weight when the cost of cement refers to a barrel 
 of 265 pounds; 2d, when this cost is for a barrel of 300 pounds 
 net; 3d, for various parts sand to one cement when the propor- 
 tions are expressed as parts by volume, dry loose sand to one 
 volume of packed cement weighing 265 pounds per barrel; and 
 4th, when the proportions are expressed as parts of dry loose 
 sand to one volume of packed cement weighing 300 pounds per 
 barrel. The quantities in the table are based upon the as- 
 sumption that the sand used is similar to that used in the ex- 
 periments from which Tables 60 and 61 were derived, and that 
 the cost of sand is seventy-five cents per cubic yard. 
 
! COST OF MORTAR 185 
 
 290. Example. To indicate the use of these tables, let us 
 determine the cost per cubic yard of natural cement mortar 
 composed of one volume of packed cement to three volumes of 
 loose dry sand when the cement weighs 300 pounds per barrel, 
 net, and costs $1.25 per barrel, while sand costs $1.00 per cubic 
 yard. In the fourth section of Table 63, opposite three parts 
 sand and under $1.20 and $1.30, we find, respectively, $3.02 and 
 $3.22; then with cement costing $1.25 and sand $0.75, we should 
 have cost of mortar per cubic yard $3.12. But in our example 
 sand is assumed to cost $1.00 per cubic yard, or twenty-five 
 cents more than the price for which the tables are computed, 
 and from Table 61 we find that for this mortar 0.83 cubic yard 
 of sand is required. We must therefore add to $3.12, .83 X 25 
 = 21 cents, giving $3.33 as cost of materials in one cubic yard 
 of the mortar. The cost of mixing the mortar must be added 
 to obtain the total cost per cubic yard. 
 
CHAPTER XIII 
 
 CONCRETE : AGGREGATE 
 
 291. Cement concrete is composed of a mixture of cement 
 mortar and fragments of stone, brick or other moderately hard 
 substances to which the mortar may adhere. Put in place while 
 plastic, it soon obtains a strength and hardness equal to good 
 building stone. This property, combined with its cheapness 
 and adaptability to monolithic construction, renders it one of 
 the most useful of engineering materials. 
 
 ART. 37. CHARACTER OF AGGREGATE 
 
 292. MATERIAL FOR AGGREGATE. Many of the points men- 
 tioned concerning the selection of a good sand are also applicable 
 to broken stone. The latter may be produced from almost 
 any moderately hard rock, provided it is not subject to decay. 
 The best material for broken stone is a rather hard and tough 
 rock, which breaks into angular fragments with surfaces that 
 are not too smooth. 
 
 Gravel makes a good aggregate, although its surfaces are too 
 smooth and rounding to give the best results. Coarse gravel 
 may be improved by running it through a rock crusher to render 
 some of the fragments angular and rough. A mixture of gravel 
 and broken stone gives excellent results (see 454). The gravel 
 assists the compacting of the mass, and the fragments of broken 
 stone furnish a good bond. A mixture of this kind also leaves 
 but a small percentage of voids in the mass, and this decreases 
 the amount of mortar required. 
 
 293. Sandstones are sometimes said to be better than lime- 
 stones, but this will depend on their relative hardness and 
 structure, and the use to which the concrete is to be put; no 
 general rule will apply. Some limestones seem to be particu- 
 larly adapted to concrete-making, as the cement adheres to the 
 surface so firmly. Granite, syanite and trap are excellent for 
 the purpose. Fragments of brick and of other burnt clay 
 
 186 
 
CHARACTER OF AGGREGATE 187 
 
 products give good results up to the limit of the strength of the 
 pieces, but this limit is not high. Table 155 gives the results 
 of transverse tests of concrete bars made under the author's 
 direction, to show the comparative value of different kinds of 
 stone. The results of these tests are discussed in 454. 
 
 Mr. E. L. Ransome l has pointed out that "for fireproof 
 work, care should be taken to avoid such aggregates as contain 
 feldspar," and that limestone should not be used if the con- 
 crete is likely to be subjected to a long continued red heat. 
 The same writer mentions the fact that finely crushed granite 
 may be inferior to finely crushed limestone for use in concrete; 
 one reason for this being that, " owing to the brittle quality of 
 granite, in crushing it is not only broken into small pieces, but 
 many of these pieces are so bruised or contused that upon a 
 little pressure being exerted upon them, such, for instance, as 
 can be applied by the finger or thumb, they will crumble." 
 
 294. The care required in the selection of a proper quality 
 of broken stone or gravel will depend upon the required strength 
 of the concrete. If a strong concrete is required, rich mortar will 
 not be able to make up a deficiency in the strength of the stone; 
 but if a low strength is sufficient, and consequently a poor mor- 
 tar is to be used, but little will be gained by having a very 
 strong rock from which to obtain broken stone. In this case 
 a rock which presents a good surface to which mortar may 
 adhere is the principal requirement, and a very hard rock need 
 not be insisted upon. 
 
 295. PRESENCE OF SCREENINGS IN BROKEN STONE. it is 
 frequently required that the broken stone shall be freed from 
 all fine material, resulting from the crushing of the stone, before 
 the mortar is added to form concrete. The wisdom of this 
 requirement is not always clear and depends upon the kind of 
 stone. It has already been stated that some forms of crusher 
 dust or screenings give, if not too fine, most excellent results 
 in mortar; this is especially true of limestone screenings. Again, 
 to retain in the broken stone all of the screenings, will result in 
 diminishing the percentage of voids in the aggregate, and thus 
 decrease the amount of mortar necessary. 
 
 On the other hand, if a stone is covered with a layer of 
 
 Engineering Record, Nov. 17, 1894. 
 
188 CEMENT AND CONCRETE 
 
 moistened, floury dust, it cannot be so readily brought in direct 
 contact with the mortar, and if the mortar does reach the 
 stone it is made less rich by the dust, which acts as so much 
 fine sand. It must be said, however, that so far as our ex- 
 periments go, they do not confirm this latter theory when a 
 moderate amount of fine material is in question, especially with 
 crushed limestone. There is a reason, however, in some cases 
 why the very fine material which acts as sand should be screened 
 out of broken stone, even if it is again used in the mortar for 
 the concrete; the fine material collects in certain parts of the 
 bin or pile, making the proportions irregular, so that one batch 
 of concrete may have a rich mortar with a comparatively large 
 amount of stone, while another may have a poor mortar with 
 but little stone. If, therefore, all of that portion of the broken 
 stone finer than, say, one-eighth of an inch, be screened out 
 and used as so much sand in making the mortar, the resulting 
 concrete will be better and more nearly uniform in quality. 
 
 296. Impurities. Material that is really foreign, such as 
 vegetable mold or loam, will be detrimental to the strength of 
 the concrete. Even clay is not permissible here if it adheres to 
 the stone, because if the surface of a piece of stone is smeared 
 with clay, the mortar will not be able to adhere as well to that 
 surface. Clay in a granulated form and not adhering to the 
 stone may be permitted, however, in small amounts, possibly 
 as much as ten per cent., without seriously injuring the concrete 
 for many uses. 
 
 When old masonry is torn down, the stones are sometimes 
 crushed for use in concrete, but such stones, having particles 
 of mortar adhering to the surfaces, will not be of first quality 
 for the purpose; their cheapness, however, will frequently out- 
 weigh such objections. 
 
 ART. 38. SIZE AND SHAPE OF THE FRAGMENTS AND THE 
 VOLUME OF VOIDS 
 
 297. As in the case of sand, the shape of the fragments and 
 the degree of uniformity in size have an important effect on 
 the proportion of voids in the mass, and all of these elements 
 affect the value of broken stone for use in concrete. As in 
 mortar each grain of sand should be completely covered with 
 cement, so in concrete should each piece of stone be completely 
 
SIZE OF FRAGMENTS 
 
 189 
 
 covered with mortar. As the pieces in a given volume of broken 
 stone will have a smaller total superficial area when the frag- 
 ments are large than when they are small, we should conclude 
 that the larger fragments will require less mortar or be more 
 thoroughly coated with a limited amount. From the same point 
 of view we should expect that round fragments would require 
 less mortar than those of irregular shape. 
 
 It is found however, in practice, that these theoretical con- 
 siderations must be modified to correspond with the facts. 
 
 TABLE 64 
 
 Voids in Broken Stone and Gravel Varying in Granulometric 
 
 Composition 
 
 
 
 WEWIIT OK 
 
 
 CHARACTER STONE. 
 
 GR AN ULOMETRIC COMPOSITION. 
 
 BROKEN 
 STONE, LBS. 
 
 PER 
 
 PER CENT. 
 VOIDS. 
 
 
 
 
 Cu. FT. 
 
 
 Limestone . . . 
 
 K 
 V 
 
 83 
 89 
 
 47 
 44 
 
 
 F 
 
 90 
 
 43 
 
 
 M 
 
 91 
 
 42 
 
 
 C 
 
 85 
 
 46 
 
 
 F, M*> 
 
 91 
 
 42 
 
 
 F C* 
 
 94 
 
 40 
 
 
 K*>, F, M& 
 
 102 
 
 35 
 
 
 K>, V 20 , F 20 , M*\ C 20 
 
 104 
 
 34 
 
 
 C, 120i Ibs., K, 33* Ibs. 
 
 11H 
 
 29 
 
 Potsdam sandstone 
 
 V 
 
 86 
 
 45 
 
 u 
 
 M 
 
 84 
 
 44 
 
 (( 
 
 V33, F67 
 
 88 
 
 43 
 
 u 
 
 V 83 , F 83 , M 88 
 
 92 
 
 40 
 
 (( 
 
 K 25 , V 25 , F 26 , M 25 
 
 97$ 
 
 36^ 
 
 Gravel .... 
 
 V 
 
 110 
 
 32 
 
 t 
 
 F 
 
 108 
 
 33 
 
 i 
 
 M 
 
 106 
 
 34 
 
 c 
 
 V8, F M 88 
 
 112 
 
 30 
 
 c 
 
 V*>, M 50 
 
 114 
 
 29 
 
 Potsdam sandstone 1 
 
 P. 6 cu. ft. on in. screen ) 
 G. 2 " " | in. to Jin. " J 
 
 100 
 
 39 
 
 and gravel . . ] 
 
 P. 4 cu. ft. on \ in. screen ) 
 G.4 ' " ^in.to Jin. " J 
 
 109 
 
 33 
 
 NOTE . Stone jarred down in measure for all trials. 
 K passed holes 1 inch square, failed to pass holes T V inch square. 
 V "I " i " 
 
 F "1 " " " 
 
 M "2 " " 1 " 
 
 C " 3 " " 2 " 
 
190 
 
 CEMENT AND CONCRETE 
 
 When the pieces of broken stone are too large, they do not 
 bed themselves well in the matrix of mortar, but become wedged 
 one against another, leaving voids in the concrete. While 
 round fragments have a small superficial area in relation to 
 their volume, have a small percentage of voids, and pack to- 
 gether readily, yet they are lacking in ability to form a good 
 bond, and hence do not give the best results. 
 
 298. Relation of Size of Stone to Volume of Voids. As illus- 
 trating the effect of the size of fragments and granulometric com- 
 position of stone on the volume of voids, Table 64 gives a number 
 of results obtained at St. Marys Falls Canal. 
 
 Table 65 gives some of the results obtained by M. Feret in 
 similar tests. 1 
 
 TABLE 65 
 Size of Stone and Volume of Voids 
 
 COMPOSITION, BY WEIGHT, OF SMALL STONE. 
 
 PER CENT. OF VOIDS BY VOLUME. 
 
 Fragments Passing a Ring of 
 
 
 
 90 mm. 
 
 60 mm. 
 
 40 mm. 
 
 20 mm. 
 
 
 
 
 
 
 
 
 TJ 
 
 and Retained on a Ring of 
 
 
 
 60 mm. 
 
 40 mm. 
 
 20 mm. 
 
 10 mm. 
 
 
 
 1 
 
 
 
 
 
 
 
 41.4 
 
 52.1 
 
 
 
 1 
 
 
 
 
 
 40.0 
 
 53.4 
 
 
 
 
 
 1 
 
 
 
 38.8 
 
 51.7 
 
 
 
 
 
 
 
 1 
 
 41.7 
 
 52.1 
 
 
 
 1 
 
 
 
 1 
 
 35.6 
 
 47.1 
 
 1 
 
 4 
 
 1 
 
 1 
 
 33.5 
 
 48.8 
 
 1 
 
 1 
 
 1 
 
 4 
 
 356 
 
 46.4 
 
 The percentage of voids in a mass of broken stone of uniform 
 size should be independent of what the size may be, and the 
 first few lines in Table 65 show this to be nearly the case with 
 the four samples tested. It is seen from both tables that the 
 more complex mixtures give smaller percentages of voids, and 
 that for all sizes the voids are much less in the gravel than in 
 the broken stone. 
 
 1 " Sand and Stone Used for Cement Mortar and Concrete, " by M. Feret. 
 Abstracted in Engineering News, March 26, 1892. 
 
SIZE OF FRAGMENTS 
 
 191 
 
 299. M. Feret's Experiments. To show the effect of the 
 variation in sizes of fragments on the strength of the concrete 
 made, M. Feret experimented with four mixtures of three sizes. 
 The proportions used in the mortar were one part by weight of 
 Portland cement to three parts of Boulogne gravel, gaged with 
 an amount of water equal to seventeen per cent, of the total 
 weight of cement and sand. The volume of mortar used in 
 each case was made equal to the volume of the voids in the 
 stone. The concrete was thoroughly mixed and then rammed 
 into a large cylindrical mold. After four months' exposure to 
 the air, twelve cubes were cut from the cylindrical block, four 
 cubes being cut from each of three consecutive horizontal 
 layers. These cubes were placed in sea water and crushed 
 after one month, being then five months old. The results of 
 the tests are given in Table 66. 
 
 TABLE 66 
 Strength of Concrete. Varying Size Stone 
 
 
 
 
 
 MEAN RESISTANCE IN KG. 
 
 
 
 
 
 PER SQ. CM. 
 
 
 
 VOL. OF 
 
 WEIGHT 
 
 
 GRANULOMETRH; 
 COMPOSITION OF 
 BROKEN STONE. 
 
 VOL. OF 
 VOIDS PER 
 Cu. METER. 
 
 MORTAR 
 PER Cu. 
 MKTER OF 
 
 OF CON- 
 CRETE 
 PER Cu. 
 
 OF 4 CUBES FROM 
 
 lid 
 
 
 
 
 
 
 STONE. 
 
 METER. 
 
 a 
 
 0> 
 
 3 
 
 Bottom. 
 
 w sJ2 
 
 G 
 
 M 
 
 F 
 
 Cu. Meter. 
 
 Cu. Meter. 
 
 Kg. 
 
 
 
 
 
 4 
 
 1 
 
 1 
 
 0.492 
 
 0492 
 
 2296 
 
 144 
 
 143 
 
 173 
 
 153 
 
 1 
 
 4 
 
 1 
 
 0.494 
 
 0.494 
 
 2272 
 
 141 
 
 141 
 
 154 
 
 145 
 
 1 
 
 1 
 
 4 
 
 0.486 
 
 0.48(5 
 
 2276 
 
 106 
 
 121 
 
 133 
 
 120 
 
 2 
 
 2 
 
 2 
 
 0.478 
 
 0.478 
 
 2264 
 
 115 
 
 132 
 
 151 
 
 133 
 
 Size "G" of broken stone passed a ring 60 mm. (2.4 inches) 
 in diameter and was held by a ring 40 mm. (1.6 inches) in 
 diameter; "M" passed 40 mm. (1.6 inches) ring and was held 
 by a ring 20 mm. (0.8 inch) in diameter; while "F" passed 
 the 20 mm. (0.8 inch) ring and would not pass a ring 10 mm. 
 (0.4 inch) in diameter. 
 
 The following conclusions are drawn from this table: (1) In 
 each block the lower layers, which had been submitted to 
 longer continued ramming than the upper layers, offered a 
 
192 CEMENT AND CONCRETE 
 
 greater resistance. (2) The mean resistance varied according 
 to the granulometric composition of stone used, and was greater 
 with the increasing proportion of large stone in each block. 
 Since the amount of mortar used was in all cases equal to the 
 volume of voids in the stone, the effect of voids on the strength 
 was not noticeable. 
 
 300. Further Experiments. Tables 153 and 155 give the 
 results of some experiments made under the author's direction 
 to test the effect of size and character of broken stone. In 
 these tests the proportions are generally 35 pounds of cement 
 to 105 pounds of sand and 3.75 cubic feet of broken stone, the 
 stone being measured after jarring it down in the vessel. The 
 amount of mortar made was sufficient to fill the voids in the 
 stone when the latter did not exceed about thirty-three per 
 cent (452, 454). 
 
 It is seen that, in general, a higher result was given by mix- 
 tures of various sizes than by any one size alone, and the fine 
 stone gave higher results than the coarse. In these tests the 
 effect of voids is shown, since in some cases there was not suf- 
 ficient mortar to fill the voids. 
 
 301. GRAVEL vs. BROKEN STONE AS AGGREGATE. The ele- 
 ments entering into the analysis of the superiority of one kind 
 of aggregate over another are given above, but since the ques- 
 tion of the relative merits of gravel and broken stone is so 
 frequently discussed, a word may be added here to show the 
 special points involved in such a comparison. 
 
 Gravel is composed of hard, rounded pebbles, the surfaces 
 of which are usually quite smooth. On account of the manner 
 of its formation and occurrence, the sizes of the pebbles are 
 usually graded from coarse to fine. Occasional beds of gravel 
 are found, however, in which the sizes of the several fragments 
 are nearly the same. In broken stone the fragments are angular 
 and usually have rough surfaces, though the degree of rough- 
 ness depends upon the kind of stone. The sizes of the frag- 
 ments as they come from the stone crusher vary from coarse 
 to fine, but by regulating the crusher jaws and by screening, 
 any desired size may be obtained. 
 
 302. In determining the value of a certain material for 
 aggregate, at least six characteristics are to be considered, the 
 strength and durability of the stone, the size and shape of the 
 
GRAVEL VS. BROKEN STONE 193 
 
 fragments, the volume of the voids, and the character of the sur- 
 face to which the cement must adhere. As gravel is usually from 
 the igneous rocks, its strength and durability are not often open to 
 question. This may or may not be so in the case of broken 
 stone, but the question of relative value of gravel and broken 
 stone, which is so frequently conclusively settled either one way 
 or another, seldom hinges on this point. As to the average size 
 of the fragments, it is evident that as a general proposition it 
 must be allowed that by proper screening either broken stone 
 or gravel may be obtained of any desired size. 
 
 Of the three remaining characteristics, the shape of the 
 fragments, volume of voids, and character of surface, the first 
 is probably the least important and the third of the greatest 
 moment. The round pebbles of the gravel slide readily one on 
 another, and do not interlock to give a good bond. The angular 
 fragments of broken stone give a better bond, but on the other 
 hand, if not thoroughly tamped, are likely to bridge, or arch, 
 and thus leave holes in the mass. On account of the shapes of 
 the fragments and because the sizes are usually more varied in 
 gravel, the latter has generally a smaller percentage of voids; 
 thirty to thirty-seven per cent, voids in gravel, and forty to 
 fifty per cent, in broken stone, may be considered to give, in 
 a general way, some comparative figures. Coming now to the 
 character of surface, cement will not usually adhere so firmly 
 to the smooth surface of the gravel as to the freshly broken 
 surface of the fragments of stone, but this cannot be con- 
 sidered a universal rule, for the strength in adhesion is not 
 simply a matter of smoothness or roughness as it appears to 
 the eye or the touch. The adhesion to limestone may be 
 very much stronger than to a sandstone which has a rougher 
 appearance. 
 
 303. Summing up the relative advantages, we find that the 
 gravel is suitable for concrete because, first, it is not likely to 
 bridge and leave holes in the concrete; if mixed rather wet, very 
 little tamping is required to compact it; and second, the usual 
 smaller percentage of voids makes it possible to secure a com- 
 pact concrete with a smaller amount of mortar than would be 
 required for broken stone. On the dther hand, the angular 
 fragments of broken stone will knit together, as it were, to form 
 a strong concrete if properly tamped, and the very important 
 
194 CEMENT AND CONCRETE 
 
 question of a suitable surface for adhesion is usually in favor 
 of the broken stone. It is evident, then, that this matter must 
 resolve itself into a question of relative cost and suitabil- 
 ity, and a general statement that either gravel or broken stone 
 is superior, is not tenable. One experimenter using a small 
 percentage of mortar in the concrete, so that the voids in the 
 broken stone are not nearly filled, may conclude that gravel is 
 the better, while another experimenter using a larger amount 
 of mortar, filling the voids in the broken stone but giving a 
 large excess of mortar for the gravel, will conclude that broken 
 stone is much to be preferred. 
 
 ART. 39. STONE CRUSHING AND COST OF AGGREGATE 
 
 304. Breaking Stone by Hand. When but a small quantity 
 of concrete is to be made, and broken stone cannot be pur- 
 chased in the vicinity, the stone for concrete may be broken by 
 hand. This is an extremely tedious process, however, and is 
 generally avoided, since broken stone prepared in this way will 
 cost from two dollars and a half to four dollars per cubic yard. 
 In the reconstruction of the breakwater at Buffalo, the cost of 
 breaking stone by hand was two dollars and eighty-six cents 
 per cubic yard, and loading on boat cost thirty-nine cents, 
 making total cost about three dollars and twenty-five cents per 
 cubic yard. 1 
 
 305. STONE CRUSHERS. The most common forms of rock 
 crushers are the gyratory and the movable jaw types. The 
 jaw breaker, or Blake crusher, consists of one fixed plate or 
 jaw and one movable one. The latter is hinged at the upper 
 end and the lower end is moved backward and forward through 
 a short space by means of a toggle joint or other mechanism. 
 The jaws are several inches apart at the upper end, depending 
 on the size of the machine, and converge toward the bottom. 
 The distance between the jaws at the bottom regulates the size 
 of fragments delivered, and this distance may be adjusted at 
 will. 
 
 The Gates crusher is of the gyratory type and consists of 
 a corrugated cone of chilled iron, called the breaking head, 
 
 1 Report of Capt. F. A. Mahan in Report Chief of Engineers, U.S.A., 
 1888, p. 2034, 
 
STONE CRUSHING 195 
 
 wittiin a larger inverted cone, or shell, which is lined with chilled 
 iron pieces. The vertical shaft bearing the breaking head is 
 pivoted at the upper end while the lower end travels in a small 
 circle ; an eccentric motion is then imparted to the head, so that 
 it approaches successively each element of the shell. The size of 
 opening can be regulated by raising or lowering the breaking head. 
 
 Stone crushers are made of various sizes having capacities 
 up to one hundred tons per hour. The cost of running a stone 
 crusher is not great, the principal expense being incurred in 
 breaking the stone into pieces of proper size to feed the crusher, 
 the delivery of the stone to the crusher, and taking it away 
 when broken. 
 
 Crushing plants are usually provided with revolving screens 
 into which the broken stone is delivered from the crusher. 
 These screens are usually made of perforated steel plate, 
 the holes being such as to separate the material into the sizes 
 desired. 
 
 Where large amounts of concrete are required, and the stone 
 is to be crushed on the work, the arrangement of the crusher 
 plant should receive careful study to facilitate the transporta- 
 tion of the rock to and from the crusher. The broken stone 
 should be discharged from the crusher into bins, from which 
 the carts or cars may be filled by gravity, or from which the 
 material may be led directly to the mixer through a chute or 
 other form of conveyor. In quarries preparing aggregate for 
 sale, and on important works, very complete stone crushing 
 plants are erected. 1 
 
 306. COST OF AGGREGATE. The cost of aggregate varies 
 greatly according to the proximity of the stone to the crusher, 
 the character of the stone, and the amount required. In ex- 
 ceptional cases gravel suitable for use in concrete is so near 
 at hand that it may be delivered on the mixing platform for 
 from twenty-five to forty cents per cubic yard. When it must 
 be brought from a distance, the cost is correspondingly in- 
 creased. Where a considerable quantity of stone is to be broken, 
 the cost of crushing, aside from transportation of the materials 
 
 1 The stone crushing and sand and gravel washing plant used in the con- 
 struction of the Canal at the Cascades of the Columbia, Ore., is described 
 and illustrated in Report of Chief of Engineers, 1891, p. 3332. 
 
196 CEMENT AND CONCRETE 
 
 to and from the site of the work, would usually be from thirty 
 to forty cents per cubic yard. 
 
 In one case where the stone was delivered to the crusher in 
 carts after having been sorted from spoil banks containing 
 much poor stone that had to be handled over, the cost per 
 cubic yard of crushed stone was approximately as follows for 
 about six thousand cubic yards crushed in one season: 
 
 Labor, including sorting and delivering to crusher, per cubic 
 
 yard of crushed stone $.67 
 
 Rent of power plant 04 
 
 Fuel . 05 
 
 Tools, supplies, breakages, etc 12 
 
 Interest and depreciation of plant 12 
 
 Total cost per cubic yard $1.00 
 
 307. The following data concerning the cost of breaking a 
 large amount of stone for road material are given by Messrs. 
 Spielman and Brush. 1 "The stone was broken by a ten-inch 
 Blake stone crusher at the rate of about twenty cubic yards in 
 ten hours. The size of the stones as they came from the crusher 
 was: fifty per cent., two inches size; twenty-five per cent., one 
 and one-half to one inch size; twenty-five per cent., screenings 
 and pea dust. The cost of the crusher, engine, boiler, etc., 
 set up complete, was about twenty-five hundred dollars. The 
 cost of working per day independent of the original cost of the 
 machinery and interest thereon, and also independent of any 
 royalty on the stone, was found by the contractor to be as 
 follows: 
 
 Repairs, lubricants, wear and tear on crusher and engine, about $6.00 
 1 engineman, $2.50; 1 feeder, $1.50; 1 screener, $1.50; 5 laborers 
 
 quarrying and breaking up stone at $1.00 10.50 
 
 1 team hauling stone 5.00 
 
 $ ton coal 2.50 
 
 Cost of preparing and crushing 20 cu. yds. of stone, $24.00 
 Cost of one cubic yard, $1.20. 
 
 308. The cost of breaking trap on the Palisades is given as 
 follows: 2 "Two crushers deliver thirty-five cubic yards of two- 
 
 1 Trans. Am. Soc. C. E., April, 1879. 
 
 2 "Construction and Maintenance of Roads," by Mr. Edward P. North, 
 M. Am. Soc. C. E., Trans. A. S. C. E., April 16, 1879. 
 
COST OF CRUSHING STONE 197 
 
 inch stone per day. when working well, the stone being sledged 
 to go into the jaws readily; fifteen per cent, of the time is lost 
 by breakdowns: 
 
 1 engineman and fireman $2.50 
 
 2 laborers feeding, at $1.25 2.50 
 
 2 laborers screening, at $1.25 2.50 
 
 Coal, 1 ton 3.50 
 
 Oil and waste 1.00 
 
 Breakages 5.00 
 
 $17.00 
 or about fifty-seven cents per cubic yard. 
 
 "On Snake Island, three crushers were arranged in a row, 
 and the broken stone was carried by an endless belt to the 
 revolving screen, whence it fell into the bins, so that no screen- 
 ers were employed. The engine had one cylinder, eight inches 
 by twenty-four inches, and was running with eighty pounds of 
 steam. The product was said to be one hundred eighty cubic 
 yards per day when there was no breakdown." The cost was 
 as follows: 1 
 
 1 engineman and fireman $2.50 
 
 3 laborers feeding, at $1.25 3.75 
 
 1\ tons coal, at $3.50 8.75 
 
 Oil, etc 2.00 
 
 Breakages 15.00 
 
 $32.00 
 
 "Allowing for the fifteen per cent, lost by breakdowns, the 
 cost would be about twenty-one cents per cubic yard." 
 
 At another place on the Hudson, two crushers, set face to 
 face, nine-inch by fifteen-inch jaws, could deliver at the rate of 
 one hundred twenty cubic yards per day when no trouble 
 occurred, but one hundred cubic yards was a fair average. 
 
 COST. 
 
 1 engineman and fireman $2.50 
 
 3 feeders 3.75 
 
 2 screeners 2.50 
 
 \\ tons coal, at $4.00 6.00 
 
 Oil, etc 2.50 
 
 Repairs 10.00 
 
 $27.00 
 or twenty-seven cents per cubic yard." 
 
 1 "Construction and Maintenance of Roads," by Mr. Edward P. North, 
 M. Am. Soc. C. E. Trans. A. S. C. E., April 16, 1879. 
 
198 
 
 CEMENT AND CONCRETE 
 
 It is noticeable that in all the above cases the item for 
 repairs is very large. The wages paid are lower than at 
 present. 
 
 309. The following data concerning the cost of quarrying 
 and crushing about five thousand six hundred. yards of broken 
 stone at Baraboo, Wis., is taken from an article by Mr. W. G. 
 Kirchoffer, C. E. 1 
 
 Cost per Cubic Yard of Crushed Rock 
 
 ITEMS. 
 
 1901. 
 
 1902. 
 
 Stone in quarry 
 
 $ .040 
 
 $ .027 
 
 Dynamite, at 24 to 27 cents pound . 
 Tools, repairs, depreciation, supplies and 
 improvements . 
 
 .056 
 
 .200 
 
 .110 
 .218 
 
 Labor, quarrying and tending crusher . 
 Fuel, at 4.60 per ton, and oil .... 
 Rent of engine 
 
 .714 
 
 .078 
 .085 
 
 .544 
 .053 
 .006 
 
 Superintendence, including livery . 
 
 .086 
 .500 
 
 .165 
 .500 
 
 
 
 
 Total cost per cubic yard . 
 
 11.76 
 
 $1.68 
 
 The cost of common labor was fifteen cents an hour, quarry- 
 men and drill runners, seventeen and one-half to twenty cents, 
 engineers and engine, thirty-five cents, and team and driver, 
 thirty cents. 
 
 310. The cost of crushing cobble stone with a rented plant 
 at Port Huron, Mich., is given by Mr. Frank F. Rogers, C. E. ; 
 from which the following data have been derived. 2 
 
 
 JULY AND 
 AUGUST. 
 
 OCTOBER 
 
 AND 
 
 NOVEMBER. 
 
 Hours run 
 
 171.5 
 
 94 
 
 Stone crushed cubic yards 
 
 1145. 
 
 522.0 
 
 Average cubic yards crushed per hour . . 
 Average rental cost per cubic yard . . . 
 Average fuel cost per cubic yard .... 
 Average labor cost per cubic yard . 
 Average total cost of crushing per cu. yd. 
 
 6.67 
 11. 6 cents 
 3.7 " 
 22.2 " 
 37.5 " 
 
 5.55 
 16.1 cents 
 7.1 " 
 27.9 " 
 51.1 " 
 
 1 Engineering News, Jan. 15, 1903. 
 
 2 Michigan Engineers' Annual, 1902, abstracted in Engineering News, 
 
 March 6, 1902. 
 
COST OF CRUSHING STONE 199 
 
 In the construction of the defenses at Portland, Me., 1 a No. 
 5 Champion Crusher was used, driven by a thirty horse-power 
 portable engine. Granite was purchased at one dollar per ton 
 on the wharf. Hauling to crusher cost thirteen cents per ton. 
 Cost of crushing, twenty cents per cubic yard of crushed stone, 
 making total cost of crushed stone in bin at crusher one dollar 
 eighty-three cents per cu. yd. 
 
 1 Report of Charles P. Williams; Officer in charge, Maj. Solomon W. 
 Roessler, Corps of Engineers, U.S. A.; Report Chief of Engineers, 1900, p. 757. 
 
CHAPTER XIV 
 
 CONCRETE MAKING: METHODS AND COST 
 
 ART. 40. PROPORTIONS OF THE INGREDIENTS 1 
 
 311. Concrete is simply a class of masonry in which the 
 stones are small and of irregular shape. The strength of the 
 concrete largely depends upon the strength of the mortar; in 
 fact, this dependence will be much closer than in the case of 
 other classes of masonry, since it may be stated as a general 
 rule, that the larger and more perfectly cut are the stone, the 
 less will the strength of the masonry depend upon the strength 
 of the mortar. 
 
 In deciding, then, upon the proportions of ingredients to use 
 in a given case, the quality of the mortar should first be con- 
 sidered. If the concrete is to be subject to but a moderate 
 compressive stress, the mortar may be comparatively poor in 
 cement; but if great strength is required, the mortar must be 
 of sufficient richness; while if imperviousness is desired, the 
 mortar must also possess this quality and be sufficient to thor- 
 oughly fill the voids in the stone. 
 
 312. THEORY OF PROPORTIONS. The usual method of stat- 
 ing proportions in concrete is to give the number of parts of 
 sand and aggregate to one of cement. These parts usually 
 refer to volumes of sand and stone, measured loose, to one vol- 
 ume of packed cement. However, there is no established prac- 
 tice in regard to this and a " 1-2-5 concrete" may mean five 
 volumes of loose stone to two volumes loose sand to one volume 
 loose cement, or any one of several combinations. 
 
 This method of stating proportions leads to confusion unless 
 one is careful to explain what is meant by such an expres- 
 sion as " 1-3-6 concrete." The evils of similar methods of 
 stating proportions in mortars, and the desirability of fixing 
 upon some standard system of weight or volume, have already 
 
 1 Portions of this article were contributed to Municipal Engineering by 
 the author, and appeared in that magazine, May, 1899. 
 
 200 
 
PROPORTIONS OF THE INGREDIENTS 201 
 
 been pointed out. The only circumstances under which such 
 expressions as the above may be used with propriety are when 
 one wishes to give only an approximate idea of the character 
 of concrete used. 
 
 From tests of strength it is known that to obtain the strong- 
 est concrete with a given quality of mortar the quantity of the 
 latter should be just sufficient to fill the voids in the aggregate. 
 The strength is notably diminishe i if the mortar is deficient, 
 and is also impaired by a large excess of mortar. This last 
 statement is subject to one exception: if the mortar is stronger 
 than the stone, then an excess of mortar does not weaken the 
 concrete. This case, however, should never be allowed to occur, 
 since it is evident that the strength of the stone should be at 
 least equal to the required strength of the concrete. Further, 
 the ordinary uses of concrete are generally best served by a 
 compact mixture containing as few voids as possible. 
 
 For these reasons, then, one should consider concrete not as 
 a mixture of cement, sand and stone, but rather as a volume 
 of aggregate bound together by a mortar of the proper strength. 
 The volume of voids in the aggregate, the per cent, of this 
 volume filled with mortar, and the strength of this mortar be- 
 come then the important considerations in proportioning con- 
 crete. When thus considered, it is an easy matter to determine 
 the required volume of mortar for a given volume of stone, 
 and the amount of cement and sand required for a given volume 
 of mortar has already been considered. 
 
 313. Determination of Amount of Mortar to Use. The bulk 
 of a given quantity of broken stone is not so variable as the 
 volume of sand. The volume of the stone, and consequently 
 the voids, will vary with the degree of packing, but the packing 
 is not influenced appreciably by the amount of moisture present. 
 
 The proportion of voids in the broken stone may be obtained 
 as follows: Find the weight per cubic foot of the broken stone 
 in the condition in which the volume of voids is sought, being 
 careful to use a measure holding not less than two or three cubic 
 feet. Also obtain the specific gravity, and hence the weight 
 per cubic foot of the solid stone. Then one, less the quotient 
 obtained by dividing the weight per cubic foot of the broken 
 stone by the weight per cubic foot of the solid stone, will be 
 the proportion of voids in the aggregate. 
 
202 CEMENT AND CONCRETE 
 
 For example, suppose the weight per cubic foot of the broken 
 stone is 102 pounds. The specific gravity of the solid stone 
 determined in the ordinary manner is found to be 2.724. Then 
 weight per cubic foot of solid stone is 62.4 X 2.724 =170 
 
 102 
 pounds and 1 - = .40, voids in stone. 
 
 Another method is to fill a vessel of known capacity with 
 the. stone to be used, and to pour in a measured quantity of 
 water until the vessel is entirely filled. The volume of water 
 required indicates the necessary amount of mortar to use. The 
 stone should be moistened before placing in the vessel, to approxi- 
 mate more nearly its condition when used for concrete, and to 
 avoid an error from absorption of the waterjised to measure voids. 
 
 314. As to the degree of jarring or packing to which the 
 stone should be subjected in filling the measure, if the stone 
 is filled in loose, and it is proposed to ram the concrete in place, 
 the amount of mortar indicated will be a little more than the 
 required quantity. If the concrete is to be placed without 
 ramming (as in submarine construction), the amount of mortar 
 indicated will not be too great. On the other hand, if the stone 
 is shaken down in the vessel to refusal, the voids obtained will 
 be less than the amount of mortar which should be used, be- 
 cause it is not possible to obtain a perfect distribution of mor- 
 tar in a mass of concrete, and because the concrete will usually 
 occupy a greater space than did the stone when shaken down. 
 And again, for perfect concrete, pieces of stone should be sepa- 
 rated one from another by a thin film of mortar, and hence the 
 volume of the concrete will be greater than the volume of the 
 stone measured in a compact condition without mortar. A 
 deficiency of mortar is usually more detrimental than an excess. 
 It is safer, therefore, to measure the voids in the stone loose, 
 or when but slightly packed, and make the amount of mortar 
 equal to, or a trifle in excess of, the voids. so obtained. 
 
 315. Aggregates Containing Sand. If in the case of broken 
 stone all of the fine particles are used, or if gravel which con- 
 tains a considerable amount of sand is employed, then this 
 fine material or sand must be considered as forming a part of 
 the mortar. This will not change the method of obtaining the 
 amount of mortar required for such broken stone or gravel, 
 but it will change the composition of the mortar used. Thus, 
 
MIXING BY HAND 203 
 
 suppose we have a gravel ten per cent of which is sand (grains 
 smaller than one-tenth inch in diameter) and we find the voids 
 to be thirty-three and one-third per cent. To three cubic yards 
 of this gravel we will add one cubic yard of a one-to-three 
 mortar. The voids will be filled, but instead of having three 
 cubic yards of stone imbedded in one cubic yard of a one-to- 
 threc mortar, we will in reality have a little less than that 
 amount of stone imbedded in a mortar composed of one part 
 of cement to about three and three-tenths parts sand. 
 
 316. Required Strength. In the paragraphs just preced- 
 ing, an attempt has been made to indicate the general principles 
 to be applied in proportioning the materials in concrete. To 
 decide on the actual proportions of the ingredients to use for a 
 given purpose, one must have clearly in mind the strength that 
 will be demanded and any special condition to which the con- 
 crete is to be subjected. A reference to Art. 57 concerning the 
 strength of concrete, will be of service in deciding on the proper 
 proportions to use in a given case. 
 
 ART. 41. MIXING CONCRETE BY HAND 
 
 317. Necessity of Thorough Mixing. Too much stress can 
 hardly be laid upon the necessity of thoroughly mixing the 
 concrete if the best results are to be attained. It has already 
 been shown that thoroughness in mixing mortar is repaid by 
 greatly increased strength, and the result is even more marked 
 in the case of concrete. Every grain of sand should be coated 
 with cement, and every piece of stone should be covered with 
 mortar. In general, the cost of mixing is from one-tenth to 
 one-fifth of the total cost of the concrete in place. If by doub- 
 ling the cost of mixing we can increase its strength more than 
 one-tenth or one-fifth in these respective cases, or permit a 
 corresponding decrease in the amount of cement necessary for 
 a given result, the additional labor in mixing is justified. 
 
 318. Concrete may be mixed by hand or by machine. Opin- 
 ions vary as to the comparative merits of the two systems, but 
 as a machine properly installed usually furnishes much the 
 cheaper method of mixing, it is usually employed. The saving 
 by this method, however, will evidently depend upon the cost 
 of labor, the total amount of work to be done, and the degree 
 of concentration of the work, or facilities for distributing the 
 
204 CEMENT AND CONCRETE 
 
 concrete. In certain sections where cheap labor is abundant, 
 the cost of hand mixing may be as low as machine mixing. 
 
 With proper supervision, hand mixing may be thorough, and 
 the chief argument against it, aside from its cost, is that such 
 hard work is likely to be slighted. The best forms of mixers 
 now on the market, however, give results quite equal to the best 
 hand work. 
 
 319. METHOD OF HAND MIXING. We will assume that the 
 materials have been brought within easy reach of the mixing 
 place. If the concrete is to be mixed near the point where it 
 is to be deposited, the mixing platform must be made portable. 
 Three platforms, each 8 by 14 feet, built of two-inch plank or 
 of two layers of one-inch boards, nailed to four 2x6 inch longi- 
 tudinal scantlings laid flat, will b3 suitable for such a case. 
 The platforms should be made without vertical sides, though if 
 desired a narrow piece of one-inch board may be laid flat around 
 the edges and nailed. A short piece of rope attached to each 
 corner of the platforms, or to the ends of the longitudinal scant- 
 lings, will be found convenient in moving them. These mixing 
 boards are placed side by side. 
 
 The sand, which may be delivered to the mixing platform 
 in wheelbarrows, is first dumped on the board and spread 
 evenly over the surface. If the sand is measured, the barrows 
 should be so arranged as to hold the required amount after 
 " striking" with a straight edge. This will make the measure- 
 ment independent of the judgment of the shoveler. If the sand 
 is delivered in cars, bottomless boxes of two or three barrels 
 capacity, according to the proportions used, will be found 
 more convenient for measuring than barrels. If the sand is 
 determined by weight, which as has been shown is the more 
 accurate method, the scales should be set at a weight which is 
 a factor of the total weight, and but little time will be required 
 to bring the scales to a balance for each barrow. 
 
 If it is possible, the batch should be of such a size as to 
 take either one or two full barrels, or a certain number of full 
 sacks of cement. This will obviate the necessity of measuring 
 or weighing the cement. The sand having been spread over the 
 surface of the mixing board, the cement is dumped upon it and 
 spread evenly over the sand. These ingredients are then mixed 
 dry, the required amount of water is added at one time in the 
 
i MIXING BY HAND 205 
 
 center of a ring formed of the dry materials, and the whole is 
 thoroughly mixed as described under the head of mortar-making. 
 
 320. The mortar having been spread evenly over the board, 
 the broken stone is dumped upon it and evenly distributed 
 over the surface. Four shovelers then mix the concrete. Each 
 shoveler starts at a corner of the board and turns each shovel- 
 ful completely over, casting toward the end and spreading the 
 mortar a little as he draws the shovel toward him. The two 
 shovelers at each end work toward each other, and meeting at 
 the axis of the platform, return to the side and repeat. When 
 the four shovelers meet at the center of the board, they turn the 
 mass again by casting toward the center in a similar manner. 
 If in mixing the concrete it is found that sufficient water has 
 not been used, more may be added from a rose nozzle, or sprink- 
 ling pot, previous to the last turning of the mass. The shovel 
 should always be used at right angles to either the side or the 
 end of the board, never diagonally; and it should always scrape 
 the mass clean from the board, never cut it at mid-depth. From 
 three to five turnings are required to thoroughly mix the concrete. 
 
 The mode of mixing has been thus minutely described, be- 
 cause if a gang of men are started properly they will soon be- 
 come expert, working in unison; whereas if each man is allowed 
 to mix according to his notion, confusion is sure to result. It 
 is sometimes preferred to spread the stone on a separate board 
 and- cast the mortar upon it, but this necessitates one handling 
 of the mortar which does not appear to contribute much to the 
 incorporation of the ingredients. 
 
 While the shovelers are engaged in mixing the concrete on 
 one platform, the mortar mixers have proceeded to the next 
 platform to mix another batch of mortar, and the cement and 
 sand are being placed upon the third platform. Thus the work 
 proceeds in regular progression without delays. The shoveling 
 of concrete is hard work, and it will be found necessary not 
 only to pick good men for this duty, but to cull them until the 
 evolution results in the proper men for the work. An extra 
 compensation for men who perform satisfactory service in the 
 mixing of concrete will usually be repaid in the character and 
 quantity of the output. 
 
 321. With the method described above, a working gang 
 would consist of the following men under ordinary conditions: 
 
206 CEMENT AND CONCRETE 
 
 Measuring and supplying cement and sand 1 
 
 Mixing mortar 2 
 
 Delivering stone from bin, one man with horse and cart, or two 
 
 men with barrows 2 
 
 Shovellers to mix concrete and cast or wheel to place .... 4 
 
 Water boy 1 
 
 Spreading and tamping concrete 1 
 
 Total men required 11 
 
 If it is found impracticable to mix the concrete near the 
 place of deposition, it may be necessary to put on two or more 
 extra men to wheel the concrete to place. This gang of eleven 
 men may be doubled and still work on the same three platforms 
 when so desired. 
 
 With a moderate length of wheel for the materials and the 
 finished concrete, a gang of eleven picked men, working ac- 
 cording to system, will be able to make from twenty-five to 
 thirty cubic yards per day of ten hours, or about two and a 
 half yards per man. The double gang of twenty-two men may 
 not work to quite as good advantage, and will probably not 
 put in more than from forty to fifty cubic yards per day. It 
 would therefore be somewhat more economical to work two 
 gangs of eleven men each on separate sets of platforms, espe- 
 cially as in this way a rivalry is created. Lack of room, however, 
 will frequently preclude this arrangement. 
 
 322. COST OF MIXING BY HAND. The amount of concrete 
 stated above, two and a half yards per man, may be taken as 
 a maximum. With wages at $1.75 per day this would corre- 
 spond to a cost of about seventy cents per yard, exclusive of 
 the wages of a foreman. Numerous examples might be cited 
 where the mixing costs more. Colonel MendeLl, in writing of 
 the fortifications at Fort Point, California, 1 states that a fore- 
 man (at $4 per day) and twenty laborers (at $2 per day) made 
 forty-five cubic yards per day of eight hours, the cost of mixing 
 being thus about $1 per cubic yard. It is stated that "the" 
 circumstances were exceptionally favorable." 
 
 As an instance where hand mixing was done at a very low 
 cost, the Lonesome Valley Viaduct 2 may be mentioned. At 
 
 1 Jour. Assn. of Engr. Soc., March, 1895. 
 
 2 Construction of Substructure for Lonesome Valley Viaduct, Gustave R. 
 Tuska, Trans. A. S, C. E., Vol. xxxiv, p. 24?, 
 
MACHINES FOR MIXING 207 
 
 this point colored labor was used at a cost of $1 for eleven hours' 
 work. A gang of men, distributed as follows, would mix and 
 lay forty cubic yards of concrete per day: 
 
 Filling sand barrows and handling water 1 
 
 Filling rock barrows 2 
 
 Mixing sand and cement 4 
 
 Mixing stone and mortar 4 
 
 Wheeling concrete 2 
 
 Spreading concrete in the molds 1 
 
 Tamping concrete in the molds 1 
 
 Foremen 1 
 
 Total 4(5 
 
 Fifteen men at $1 per day, and foreman at $2.50 per day, 
 makes a cost of $17.50 for forty yards of concrete, or at the 
 rate of forty-four cents a yard for mixing. Had the laborers 
 received $1.75 per day, however, the cost would have been 72 
 cents per yard. 
 
 323. In the construction of the Forbes Hill Reservoir and 
 stand pipe at Quincy, Mass., 1 all concrete was mixed and placed 
 by hand. "The ordinary concrete gang was made up of a 
 sub-foreman, two men gaging materials, two men mixing mor- 
 tar, three men turning the concrete, three men wheeling con- 
 crete, one man placing, and two men ramming. Two gangs 
 were ordinarily employed, placing about twenty cubic yards 
 per day each, or about 1.43 cubic yards per man. The con- 
 crete was turned at least three times before placing." With 
 labor at $1.75 per day, this would give the cost of mixing and 
 placing $1.22 per cubic yard. The actual cost of mixing and 
 placing varied from $0.97 to $1.53, according to the character 
 of the work. 
 
 ART. 42. CONCRETE MIXING MACHINES 
 
 324. General Classification. Concrete mixing machines may 
 be divided into two general classes, batch mixers and continu- 
 ous mixers. In the former, sufficient materials are proportioned 
 to make a convenient sized batch for the mixer. They are then 
 charged into the machine at once, given a certain amount of 
 mixing, and then discharged at once. In the continuous mixers 
 
 1 Described by C. M. Saville, M. Am. Soc. C. E., Engineering News, Mar. 
 13, 1902. 
 
208 CEMENT AND CONCRETE 
 
 the materials are dumped on a platform, and after being prop- 
 erly proportioned, are delivered gradually to the mixer, and if 
 fed uniformly, the concrete is discharged continuously by the 
 machine. In the latter method care must be taken to feed 
 the cement, sand and stone together and at a uniform rate. 
 If one man shovels cement, two men shovel sand and four men 
 handle the stone, and the cement man stops to fill his pipe, 
 there is likely to be a poor streak of concrete. It is therefore 
 desirable in feeding a continuous mixer to spread the measured 
 quantity of stone on the platform, and on top of this place the 
 weighed quantities of sand and cement. Then if each shoveler 
 gets his shovel blade under the whole mass, he will have some 
 of each ingredient. 
 
 325. There are many styles of concrete mixers of both classes 
 on the market. One of the oldest, as well as one of the best, 
 is the cubical box mixer which consists of a box four or five 
 feet on a side, supported by trunnions at opposite corners, and 
 made to revolve about this axis. A hinged door is provided 
 near one corner of the box by which the latter is charged and 
 emptied. The dry materials ma}' be first charged and mixed 
 and the water added later, either through the door or through 
 a perforated pipe in the axis, or the water may be added with 
 the dry materials; after from ten to thirty revolutions of the 
 box, the mixed concrete is discharged into a skip or on a car, 
 to be conveyed to the place of deposition. 
 
 The great merit of this mixer is that the materials are thrown 
 back and forth from one side of the cube to another and a 
 thorough commingling results. The chief disadvantage is the 
 difference in elevation between the receiving hopper and point 
 of delivery, making it necessary to elevate the materials; one 
 other defect is that the batch is not in view while being mixed, 
 so that the amount of water cannot be regulated according to 
 slight variations that may occur in the moisture of the sand and 
 stone when charged. 
 
 326. To obviate this latter difficulty as well as to facilitate 
 to some extent the charging and dumping of the batch, a form 
 of box mixer is made in which the corners of the box in the axis 
 of revolution are truncated, and the trunnions are replaced by 
 collars which support the box, and through which the materials 
 may be fed and discharged. The collars are supported in 9 
 
MACHINES FOR MIXING 209 
 
 tilting cradle which permits the delivery end to be depressed 
 after the batch is mixed. The advantage of having the batch 
 visible during mixing is perhaps somewhat offset by the greater 
 difficulty of thoroughly cleaning the box when discharged. 
 
 Mixers working on the same, principle are sometimes made 
 in other forms than the cube. One of these is the cylindrical 
 mixer, which is made of boiler plate and may be four or five feet 
 in diameter and five or six feet long. This is rotated about a 
 diagonal axis. It is said to be more easily and cheaply made 
 than the cubical mixer, and dumps more quickly and cleanly, 
 while the cost of operation is about the same, and the mixing 
 is as satisfactorily done as in the cubical form. 
 
 327. The so-called " Dromedary Mixer " l is a batch mixer 
 specially designed for use on street work. The mixing chamber 
 is a cylindrical steel drum with closed ends, mounted between 
 two wheels. It is hinged along an element of the cylinder so 
 that it opens into two halves like a clam shell bucket, to dis- 
 charge. A trap door is provided for filling. The cart is drawn 
 by a horse, and the chamber may be thrown in or out of gear 
 with the cart wheels. The cement and sand being first added 
 and the trap door closed, the horse draws the cart to the stone 
 pile. The stone and water are here added and the cart is drawn 
 to the work; the concrete, mixed on the way, is dumped by the 
 driver, who merely raises a lever which not only separates the 
 two halves of the mixer, but throws it out of gear so that it 
 stops revolving. The chamber may be thrown out of gear at 
 any time without dumping if desired. 
 
 328. The Ransome Concrete Mixer 2 " consists of a hollow 
 rotary dome, having upon the inner surface of its periphery 
 directing guides or flanges, and hinged shelves, by means of 
 which the materials are thrown together and perfectly com- 
 mingled. A discharge chute, or spout, is arranged to deliver the 
 material into the barrow or cart when properly mixed." The 
 mixer is also provided with an automatic device for proportion- 
 ing the materials, and a conveyor to carry them to the mixer. 
 Water is supplied to the mixer through a pipe with facilities 
 for regulating the supply. 
 
 1 Fisher and Saxton, 123 G St., N. E., Washington, D. C. 
 
 2 Ransome Concrete Machinery Co., 11 Broadway, N. Y. 
 
210 CEMENT AND CONCRETE 
 
 329. The Smith Mixer l is a batch machine made of two 
 truncated cones placed base to base, and provided on the in- 
 terior with deflecting plates designed to throw the materials 
 from one end of the mixer to the other as the machine is re- 
 volved. At the junction of the two cones, on the outer cir- 
 cumference, is a spur gear by which the chamber is actuated. 
 The latter rests upon rollers in a swinging frame, so arranged 
 that the machine may be tilted for dumping while the drum is 
 revolving. In operating this mixer it has been found advanta- 
 geous to charge the broken stone or gravel first, and give one or 
 two revolutions before adding the cement and sand, as this cleans 
 the mortar from the corners. This form seems to be particularly 
 adapted for a portable machine. They may be had mounted 
 on trucks, with or without an engine, as desired. 
 
 330. The McKelvey Mixers 2 are made in two styles, con- 
 tinuous and batch. Both styles are cylinders revolving on 
 friction rollers, and having, on the interior, deflecting blades 
 and a patent " gravity shovel" which lies against the rising 
 side of the drum and casts the materials downward when the 
 cylinder has revolved far enough to overturn the blade. The 
 batch mixer has a shorter cylinder and can be discharged at 
 will. These mixers may be fed by shovels, or they may be 
 provided with a hopper into which the materials may be dumped 
 from carts or barrows. They discharge directly into wheel- 
 barrows. The mixer, and an engine and boiler to run it, are 
 mounted compactly on a truck, or the mixer is furnished on a 
 steel frame without an engine. 
 
 331. The pan mixer 3 consists of a large shallow pan into 
 which may be lowered a framework carrying a series of plows. 
 The materials are spread in the pan in layers, the plows are 
 lowered into it, and the pan is revolved about its vertical axis, 
 the plows remaining stationary. The plows are so arranged as 
 to move the materials radially toward and away from the cen- 
 ter of the pan. The water may be added from a rose nozzle. 
 For dumping, an opening is made in the bottom of the pan by 
 withdrawing a slide. Were the plows made to revolve in a 
 
 1 Contractors' Supply Co., 232 Fifth Ave., Chicago. 
 
 2 McKelvey Concrete Machinery Co., N. Y. Life Bldg., Chicago. 
 
 Clyde Iron Works, Duluth, Minn. 
 
MACHINES FOR MIXING 211 
 
 stationary pan, the concrete would be more conveniently 
 dumped in a pile, or in a car, instead of being scattered about 
 under the pan. 
 
 332. The Cockburn, 1 a continuous mixer, is in the form of a 
 long box square in cross-section, surrounded at either end by 
 circular rings supported on friction rollers. By suitable gear- 
 ing the mixer is revolved about its longest axis, which has a 
 slight inclination toward the discharge end. The materials are 
 added through a hopper at one end, and fall from one side of 
 the box to the adjacent side as the machine revolves, working 
 gradually toward the delivery end, which is open. The water 
 is added through a pipe at about one-third of the length of the 
 box from the feed end. While this machine has no complicated 
 system of blades to become clogged, the mortar has a tendency 
 to stick in the corners of the mixer, making the interior cylin- 
 drical, and thus much less effective in mixing. Striking the 
 sides of the box with a heavy hammer will detach the mortar, 
 and this requires occasional attention. 
 
 333. A common form of continuous mixer consists of a screw 
 working in a cylinder. The materials are fed to the cylinder 
 near one end and are mixed while being gradually worked toward 
 the other end by the screw. The water is added through a 
 fixed perforated pipe at a point about one-third of the distance 
 from the feed end of the cylinder, and the mixed concrete falls 
 from the outlet at the other end. This style is frequently 
 made in a light form and mounted on wheels, and is then con- 
 venient in the laying of concrete for pavements. 
 
 A modification of the screw mixer consists of a semi-cylin- 
 drical trough, in which revolves a shaft carrying blades set at 
 right angles to the shaft and to each other. The trough is 
 sometimes given a slight inclination to the horizontal, and the 
 blades are so shaped as to assist in working the materials toward 
 the delivery end. 
 
 334. The Drake Mixer 2 is of the general form just described. 
 One of the machines made by this company is a semi-cylindrical 
 trough in which revolve in opposite directions two shafts, each 
 carrying some thirty blades. Most of the blades are straight, 
 
 1 Cockburn Barrow and Machine Co., Jersey City, N. J. 
 
 2 Drake Standard Machine Works, 298-302 W. Jackson Boul., Chicago. 
 
212 CEMENT AND CONCRETE 
 
 but some of them are curved to work the material toward the 
 delivery end. 
 
 335. Gravity Mixer. An appliance recently devised, which 
 is called a concrete mixer, consists of a steel trough provided 
 with staggered pins and deflecting plates. The trough is sup- 
 ported in an inclined position and has a hopper at its upper 
 end. Water is supplied through spray pipes at the side of the 
 trough. The materials, stone, sand and cement, are spread in 
 layers on the mixing platform, with the stone at the bottom. 
 The materials are then thrown into the hopper; they are mixed 
 as they descend through the pins, and the product is caught in 
 barrows or carts at the bottom. 
 
 336. In a very able article on concrete mixers, 1 Mr. Clarence 
 Coleman, M. Am. Soc. C. E., makes an analytical discussion of 
 the relative efficiencies of the several forms. In this analysis 
 he gives the following weights to the several requirements for 
 a perfect mixer. That the entire mass of concrete shall be so 
 commingled that the cement shall be uniformly distributed 
 throughout the batch. is given a weight of forty; that the amount 
 of water shall be subject to control is given a weight of twenty- 
 five; perfect dry mixing and relative time of mixing, each ten; 
 and receiving materials, discharging concrete and self-cleaning, 
 are each given a weight of five. 
 
 The first three requirements, with a combined weight of 
 seventy-five, relate to the production of good concrete, while 
 the remaining requirements, with a combined weight of twenty- 
 five, pertain to economy in use. In short, the first requisite 
 is that a machine shall be capable of producing a perfect mix- 
 ture; then the machine that accomplishes this result at the 
 lowest cost per cubic yard is the best. The choice of a machine 
 will depend frequently on the character of the work to be done, 
 as some machines can only be used economically where large 
 quantities of concrete are to be used in a restricted area, while 
 others are particularly adapted for portable plants. 
 
 ART. 43. CONCRETE MIXING PLANTS AND COST OF MACHINE 
 
 MIXING 
 
 337. Coosa River Improvement. The concrete plant used 
 at Lock No. 31, Coosa River Improvement, 2 was erected in a 
 
 1 Engineering News, Aug. 27, 1903. 
 
 2 Major F. A. Mahan, Corps of Engineers, U. S. A., in charge. 
 
CONCRETE PLANTS 213 
 
 three-story shed. The top story served as a cement storage 
 room and two hoppers were arranged in the floor to receive 
 the cement for the mixers below. Level with the floor of the 
 second story were two other hoppers immediately below the 
 cement hoppers, to receive the sand and broken stone, while in 
 the first story or basement the mixers were suspended at a 
 height sufficient to allow concrete cars to pass under them. 
 The following description is from the report of the designer, 
 Mr. Charles Firth, U. S. Asst. Engineer: 1 
 
 "The cars used in handling the sand and broken stone are 
 of the side dump pattern and are brought into the charging 
 room on either side of the hoppers. The cement is drawn from 
 the cement room overhead In proper quantities, through verti- 
 cal chutes arranged somewhat on the principle of the old- 
 fashioned powder flask. 
 
 "The water is added to the materials as they enter the mix- 
 ers, and the quantity, which will probably be variable with 
 the temperature, is controlled by valves on the mixing floor, 
 the operators being governed by indicators, which show the 
 quantity used. The mixers are cubical boxes four feet on each 
 side, inside measurement, made of steel plate five-sixteenths of 
 an inch thick, with 2 by 2J-inch angle irons. Each mixer is 
 provided with a door in one corner, twenty-two inches square, 
 fastened with a tempered steel spring catch, and held open 
 when required with a hinged screw bolt. The shaft which 
 revolves the mixers is three inches square. It is securely 
 fastened to them by trunnion castings at diagonally opposite 
 corners. The whole is driven by a 10 by 16 inch horizontal 
 engine, and thrown in and out of gear by ordinary friction 
 gearing with friction and brake levers. 
 
 "After a sufficient number of revolutions in the mixers, the 
 concrete is dumped into the concrete cars below, which are of 
 the center dump pattern." 
 
 The method given of measuring the cement is not recom- 
 mended, as the charge of cement, if not a full barrel, should 
 always be weighed. The three-story arrangement by which 
 the materials were handled almost entirely by gravity was 
 made possible by the high bank at the side of the lock pit. 
 
 1 Annual Report, Chief of Engineers, U.S.A., 1894, p. 1292. 
 
214 CEMENT AND CONCRETE 
 
 The total cost of the plant, exclusive of the boilers, is stated 
 to have been about $8,000, and the average output about two 
 hundred cubic yards of concrete per day of eight. hours. The 
 cost of mixing, depositing and ramming 8,710 cubic yards of 
 concrete in the construction of lock walls was at the rate of 
 $0.884 per cubic yard. 
 
 338. Portland, Maine, Defenses. In the construction of the 
 defenses at Portland, Maine, 1 a five-foot cubical mixer was 
 used. Sand and stone were delivered, by bucket conveyors, 
 in bins directly over the mixer. " Immediately under these 
 bins were two measuring hoppers for stone and sand, respec- 
 tively, and an additional hopper for cement. From these meas- 
 uring hoppers the charge was dumped into the mixer and 
 thence, when mixed, into a car immediately under it. This car 
 delivered the mixed batch by means of a hoisting engine and 
 an inclined track to the site of the battery under a fifty-five 
 foot derrick, which placed it in the work at the point required. 
 Two barrels of cement, sixteen cubic feet of sand, and thirty- 
 two cubic feet of stone constituted a batch. * *-* The usual 
 number of men engaged in the operation of mixing and placing 
 was as follows: Two master laborers, three steam engineers, 
 two stokers and twenty-five laborers." It is said that 200 
 barrels of cement, or 100 batches, could be mixed and placed in 
 a day of eight hours. This would make the labor cost of this 
 portion of the work 50 or 60 cents per cubic yard. The cost 
 stated, however, varies greatly according to the amount of detail 
 in construction, and the lowest cost given for " labor of mixing 
 and placing" is $1.15 per cubic yard. 
 
 339. San Francisco Defenses. A cubical mixer used in the 
 construction of the defenses at San Francisco 2 mixed 250 cubic 
 yards per day with seven men, engineer, fireman, and five men 
 to feed and dump mixer, at a labor cost of $14.67 per day, or 
 about six cents per cubic yard, exclusive of cost of transporta- 
 tion and ramming. The materials and concrete were handled 
 on cars run almost entirely by gravity. 
 
 340. Buffalo Breakwater. In the construction of the Buf- 
 
 1 Report of Charles P. Williams to Maj. Solomon W. Roessler, Corps of 
 Engrs., U. S. A., in charge. Report Chief of Engineers, 1900, p. 745. 
 
 2 Maj. Charles E. L. B. Davis, Corps of Engineers, U. S. A., Report Chief of 
 Engrs., 1900, p. 980. 
 
CONCRETE PLANTS 
 
 215 
 
 falo Breakwater/ the mixing plant, consisting of a cubical 
 mixer with necessary engines and boilers and two derricks, was 
 mounted on a dismantled lake schooner which could be placed 
 beside the section of the breakwater under construction. The 
 broken stone was delivered in a canal boat which could be tied 
 up alongside the schooner, and outside of the canal boat lay the 
 material scow. The latter was made from an old dump scow, 
 the decked pockets serving as bins for cement, sand and gravel. 
 Into a steel bucket on the scow were loaded, by wheelbarrows, 
 the following materials: 
 
 5.4 cu. ft. (H bbls.) cement. 
 10.8 cu. ft. sand. 
 
 5.4 cu. ft. gravel. 
 21.0 cu. ft. total. 
 
 Into a similar bucket on the canal boat 21.6 cubic feet of 
 broken stone were shoveled. As these buckets were filled, they 
 were hoisted by one of the derricks and dumped into the cubical 
 mixer. The latter discharged the mixed concrete into a skip 
 and a derrick deposited the concrete in place. The cost of labor 
 per cubic yard of concrete is as follows: 
 
 ITEMS. 
 
 No. 
 MEN. 
 
 COST 
 
 PER 
 
 Horn. 
 
 Cu. YDS. 
 
 PKB 
 
 HOUR. 
 
 COST o* 
 LABOK 
 
 PER 
 
 Cu. YD. 
 
 Loading material into buckets from scows 
 Mixing, including engine men and derrick 
 men . 
 
 18 
 11 
 
 $S.17J 
 2.85 
 
 18.2 
 182 
 
 .$0.174 
 l'?Q 
 
 Placing, includin op foreman . ... 
 
 13 
 
 2.05 
 
 182 
 
 146 
 
 
 
 
 
 
 Total labor 
 
 42 
 
 88 17 
 
 182 
 
 *0 449 
 
 
 
 
 
 
 The above does not include cost of fuel, nor of transporting 
 materials from the storehouses or yards to the site of the work. 
 
 341. Quebec Bridge. The plant used in the construction 
 of the Quebec Cantilever Bridge 3 consists of a No. 5 rotary 
 stone crusher, with a maximum capacity of thirty cubic yards 
 per hour, discharging into a bucket conveyor which delivered 
 the crushed stone in a small storage bin directly over the con- 
 crete mixer. The latter was of the cubical form, five feet on a 
 side, with a capacity of two cubic yards of concrete per batch. 
 
 1 Emile Low, U. S. Asst. Engr., Engineering News, Oct. 8, 1903. 
 1 Engineering News, Jan. 29, 1903. 
 
216 CEMENT AND CONCRETE 
 
 The cement warehouse and the sand supply were near the 
 mixer. Cement and sand were hoisted to the top of the ma- 
 chine in boxes, with bottoms inclined at forty-five degrees, 
 each holding a batch, and dumped into the charging hopper of 
 the mixer as required. The mixer was elevated sufficiently to 
 permit dumping the concrete directly in a skip on a car, the 
 latter being run to the work. The skip was handled by guy 
 derricks. This plant made the remarkable record of two hun- 
 dred eighty-five batches in ten hours, and on one occasion 
 turned out one hundred fifty batches in five hours, or, if all 
 were two-yard batches, at the rate of sixty yards per hour. 
 
 342. Galveston. For the construction of the Galveston sea 
 wall two concrete mixing and handling machines were designed, 1 
 each consisting of a double-deck car, on eight wheels, with two 
 revolving derricks, one on either side for handling materials 
 and concrete, respectively. The materials are delivered on 
 tracks beside the mixer car track which is parallel to the sea 
 wall. One derrick hoists the loaded skips from the material 
 cars and deposits them on the upper deck of the mixer car, 
 whence they are delivered in measured quantities to the Smith 
 Rotary Mixer located on the lower deck. When mixed, the 
 concrete is dumped into a skip, which is handled by the second 
 derrick and dumped into the forms. 
 
 343. For work having similar requirements to that just 
 described, namely, for retaining walls on track elevation, Chicago 
 & Western Indiana R. R. at Chicago, the problem was met in 
 a somewhat different manner. 2 An ordinary flat car was double 
 decked and the space between decks inclosed to protect the 
 machinery, including the Drake Concrete Mixer. Cars contain- 
 ing cement, sand and stone were coupled in the rear of the mixer 
 car. These material cars were fitted with removable wheeling 
 platforms, making a complete runway along the sides of the 
 cars. The materials were delivered at the mixer car in wheel- 
 barrows and dumped into measuring boxes, and thence fed to 
 the mixer. The concrete was delivered on a belt conveyor 
 mounted on a boom with turntable permitting nearly half of a 
 revolution. The outer end of the conveyor could be raised or 
 lowered as desired, and the concrete was thus deposited where 
 
 1 Engineering News, Jan. 15, 1903. 
 3 Ibid., Feb. 28, 1901. 
 
COST OF CONCRETE 
 
 217 
 
 needed in the work. To permit the mixer train to move along 
 the track, the two ends of a cable were made fast to anchorages 
 placed about a thousand feet apart, one in front of, and the 
 other behind, the train. As this cable had about eight turns 
 around a winding drum on the mixer car, the train could be 
 propelled forward or backward at will. 
 
 A somewhat similar form has been used for street work, 
 where the mixer and electric motor are mounted on a truck 
 with a swinging conveyor for the delivery of concrete anywhere 
 between the curbs. A pair of wheels in the rear serve to carry 
 an inclined runway for wheelbarrows by which the materials 
 are delivered to the mixer. 
 
 344. The data for the following items concerning the cost of 
 mixing concrete for culverts on railroad work are taken from 
 an article in Engineering News. 1 
 
 "The plant is located on a hillside with the crusher bins 
 above the loading floor or platform that extends over the top 
 of the mixer, so that crushed stone can be drawn directly from 
 the chutes of the bins and wheeled to the mixer. The sand is 
 hauled up an incline in one-horse carts and dumped on the? 
 floor, and is also wheeled in barrows to the mixer." The capa- 
 city of the cubical mixer used was seven-eighths cubic yard. The 
 cost of mixing and placing was as follows: 
 
 ITEMS. 
 
 COST 
 
 PER 
 
 DAY. 
 
 COST 
 
 PKIl 
 
 Cu.Yi>. 
 
 One foreman assumed at 82 50 per day 
 
 $2.50 
 
 
 Three men supplying mixer at SI 50 per day 
 
 4.50 
 
 
 One engineman assumed at S2 00 per day 
 
 2.00 
 
 
 Fuel and supplies assumed at . .... 
 
 2.00 
 
 
 
 
 
 Cost of mixing 40 cu. yds. 
 
 SI 1.00 
 
 SO 275 
 
 Two men loading wheelbarrows at SI 50 
 
 $3.00 
 
 
 Four men wheeling wheelbarrows at 8.1. 50 
 
 6.00 
 
 
 
 
 
 Cost of wheeling 40 cu. yds. 100 feet 
 
 $9.00 
 
 0225 
 
 Four men ramming at $1.50 
 Four men wheeling in and bedding large stone in concrete at 
 $1.50 . . 
 
 86.00 
 600 
 
 0.150 
 150 
 
 
 
 
 Total cost mixing and placing 
 
 
 80 800 
 
 
 
 
 1 Location and Construction of the Ohio Residency, Pittsburg, Carnegie 
 & Western R.R., Engineering News, May 21, 1903. 
 
218 CEMENT AND CONCRETE 
 
 It is not explained why six men are required to load and 
 wheel forty cubic yards one hundred feet in ten hours, but it 
 may be that these men assisted in other operations. 
 
 Another contractor on the same work used a different form 
 of mixer with much lower loading platform and handled the 
 mixed concrete with skips and derrick. The cost is estimated 
 as follows: 
 
 1 man feeding mixer $1.50 
 
 1 engineman assumed at 2.50 
 
 1 derrick man assumed at . 2.50 
 
 2 tagmen swinging boom and dumping 3.00 
 
 6 barrowmen supplying mixer 9.00 
 
 2 men tamping 3.00 
 
 Fuel, supplies, etc. 1.50 
 
 Cost of mixing and placing 50 cu. yds. . . $23.00 
 Cost per cu. yd., 46 cents. 
 
 ART. 44. COST OF CONCRETE 
 345. QUANTITIES OF INGREDIENTS IN A CUBIC YARD. As 
 
 has already been indicated, the rational method of proportion- 
 ing concrete is to use just sufficient mortar to fill the voids in 
 the stone, or possibly a very small excess to allow for imperfect 
 mixing; and in ordinary practice this rule should not be de- 
 parted from unless it be for some special reason. When so pro- 
 portioned, a cubic yard of concrete will contain approximately 
 a cubic yard of stone, depending on the method of measure- 
 ment. If we know the percentage of voids in the broken stone 
 or gravel, and consequently the percentage of mortar which 
 should be found in a cubic yard of the finished concrete, we 
 may readily obtain the approximate cost per cubic yard of the 
 latter for a given quality of mortar and given unit prices. 
 
 Thus, suppose we have stone in which the voids are such 
 that the mortar will amount to forty per cent, of the finished 
 concrete, and we wish to have the mortar composed of three 
 volumes of loose sand to one volume packed natural cement, 
 unit prices being as follows: 
 
 Cement, $1.25 per barrel of 300 pounds net, 3.75 cubic feet. 
 
 Sand, $1.00 per cubic yard. 
 
 Stone, $1.75 per cubic yard. 
 
 As in 290, we find the ingredients in one cubic yard of 
 
COST OF CONCRETE 219 
 
 mortar to cost $3.33. Since forty per cent, of the concrete is 
 to be composed of mortar, the mortar in one cubic yard of 
 concrete will cost forty per cent, of $3.33, or $1.33, and one 
 yard of stone at $1.75 will make the total cost of the materials 
 in the concrete $3.08 per cubic yard. 
 
 The diagram herewith may be used to get the approximate 
 cost of the concrete after having obtained the cost of the mortar 
 as before. Thus, if we enter the diagram with the cost of mor- 
 tar $3.33, and follow it to the diagonal line marked forty per 
 cent., we find this is on the ordinate $2.33, the cost of the in- 
 gredients in one cubic yard of concrete when the stone costs 
 one dollar per cubic yard. Hence, $2.33 plus $0.75 equals 
 $3.08, the approximate cost of the materials in a cubic yard of 
 the concrete as desired. 
 
 346. The usual method, however, of stating proportions in 
 concrete is to give the volumes of sand and stone to one volume 
 of cement. Thus, one of cement, three of sand and six of stone 
 would usually mean one volume of packed cement, three vol- 
 umes of loose sand and six volumes of loose broken stone. To 
 arrive at the cost of concrete when proportions are thus ar- 
 bitrarily stated, involves a greater amount of work. From the 
 tables already given (Art. 36), we can determine the amount of 
 mortar which a given quantity of dry ingredients will make, 
 and the consequent cost of the mortar per cubic yard. Then a 
 knowledge of the voids in the broken stone will permit of a 
 close estimate of the amount of concrete made, whence we can 
 determine the cost of the latter. 
 
 For example, suppose it is desired to determine the cost of 
 the materials in a cubic yard of natural cement concrete under 
 the following conditions: 
 
 1 bbl. cement containing 280 pounds net, at $1.00 per bbl. 
 
 3 bbls. sand weighing 100 pounds per cu. ft., at $.75 per cu. yd. 
 
 6 bbls. loose broken stone, having 45 per cent, voids, at $1.25 per cu. yd. 
 
 1 bbl. cement = 3.75 cu. ft. = .139 cu. yd., cost $1.000 
 
 3 bbls. sand = 11.25 cu. ft. = .417 cu. yd., cost .313 
 
 6 bbls. stone = 22.50 cu. ft. = .833 cu. yd., cost 1.041 
 
 Total cost $2.354 
 
 From Table 61, 286, we find that it requires 2.03 barrels of 
 cement to make one cubic yard of one-to-three mortar, when 
 
220 
 
 CEMENT AND CONCRETE 
 
 CONCRETE MAKING 
 Cost of Concrete, Dollars per Cu. Yd. 
 Stone Assumed to Cost $1.00 per Cu. Yd. 
 
 
 
EXAMPLES OF COST 221 
 
 proportions are stated as above; then one barrel of cement 
 would make - - = .493 cu. yd. As forty-five per cent, of the 
 
 Z.Oo 
 stone is voids, the amount of solid stone in six barrels would 
 
 be or" x -55 = .458 cu. yd. Then the mortar plus solid 
 
 stone would be .493 + .458 = .951 cu. yd. It has been found 
 by experiment that the amount of concrete will exceed the sum 
 of the mortar and solid stone by from two to five per cent.; 
 hence we may assume in this case that the amount of concrete 
 made with the above materials would be .95 -f- .03 = .98 cu. 
 yd., and 2.354 -*- .98 = $2.40, the cost of materials in one cubic 
 yard of finished concrete. To obtain the actual cost of con- 
 crete in place, the cost of mixing and deposition must be added 
 (see Arts. 41 and 43). When the volume of mortar used is not 
 greater than the voids in the loose stone, then the amount of 
 rammed concrete made may be less than the volume of loose 
 broken stone. 
 
 347. EXAMPLES OF ACTUAL COST OF CONCRETE. The fol- 
 lowing data are given concerning the cost of concrete on several 
 works where sufficient details have been published to be of 
 value. 
 
 Defenses Staten Island. 1 Cubical box mixer ; proportions by vol- 
 ume, 1 cement, 3 sand, 5 broken stone; 5,609 cu. yds. of concrete. 
 
 ITEMS. 
 
 COST PER Cu. YD. 
 CONCRETE IN 
 PLACE. 
 
 Cement, Portland, at $1.98 per bbl. . 
 
 $2.546 
 
 Broken trap rock . . 
 
 1 041 
 
 Sand drawn from beach . 
 
 0.225 
 
 Receiving and storing materials 
 
 0.149 
 
 Mixing, placing and ramming 
 
 0.879 
 
 Forms, lumber and labor . 
 
 0.477 
 
 Superintendence and miscellaneous .... 
 
 0.190 
 
 Total cost per cu. yd 
 
 .$5.507 
 
 
 
 It is stated that hand mixing for a portion of the concrete 
 used in another emplacement cost fifty-six cents more per 
 cubic yard than machine mixing. 
 
 1 Major M. B. Adams in charge. Report Chief of Engineers, U.S.A., 
 1900, p. 837. 
 
222 CEMENT AND CONCRETE 
 
 348. Defenses Tampa Bay, Florida. 1 Cockburn-Barrow 
 mixer, with cableway for placing concrete. Shell concrete 
 made up of 1 cement, 3^ sand and 5^ shell. 
 
 1.31 bbls. cement, at $2.42 per bbl $3.17 
 
 .71 cu. yd. sand, at 21 cents per cu. yd .15 
 
 1.08 cu. yd. shell, at 50 cents per cu. yd 5.4 
 
 $3.86 
 Labor mixing $0.28 
 
 Labor placing and tamping .33 
 
 Labor on forms . .155 .765 
 
 Total cost per cubic yard $4.625 
 
 The above does not appear to include costs of running ma- 
 chinery, fuel, repairs, and depreciation of plant. 
 
 At the same battery 2 in the following year the cost of broken 
 stone and shell concrete was as follows: 
 
 .9 bbl. cemsnt at $2.46 (including $0.59 per bbl. 
 
 storage) $2.214 
 
 .28 cu. yd. shell, at $0.45 per cu. yd .128 
 
 .47 " sand, at 0.12 " " 056 
 
 .80 ." stone, at 2.95 " " 2.360 
 
 Total materials $4.758 
 
 Mixing and placing $0.623 
 
 Forms 370 
 
 Total .993 
 
 Total cost per cubic yard $5.751 
 
 349. Defenses San Francisco, Cal. 3 Cubical mixer; ma- 
 terials drawn from bins into measuring cars, hoisted by elevator 
 and dumped into hopper of mixer. Mixer given twelve to four- 
 teen turns and concrete dumped into cars, pushed by hand out 
 on trestles, and dumped in place. Average capacity plant, 280 
 cubic yards per day of eight hours. Itemized cost of 8,328 
 cubic yards of concrete in place was as follows: 
 
 1 Report Lieut. Robert P. Johnson, Corps of Engineers, U. S. A., Report 
 Chief of Engineers, 1899, p. 906. 
 
 2 Report Lieut. Frank C. Boggs, Corps of Engineers, U. S. A., Report 
 Chief of Engineers, 1900, p. 931. 
 
 3 Maj. Charles E. L. B. Davis, Corps of Engineers, in charge. Report 
 Chief of Engineers, 1900, p. 980 f ' 
 
EXAMPLES OF COST 
 
 223 
 
 .758 bbl. Portland cement, at $3.03 per bbl . . . $2.298 
 
 .887 cu. yd. rock, at $1.80 per cu. yd 1.597 
 
 .41 cu. yd. sand, at 0.73 per cu. yd .299 
 
 Water .010 
 
 Cost of materials $4.210 
 
 Concrete plant, erection per cu. yd. concrete . . $0.269 
 Concrete plant, running expenses per cu. yd. . . .022 
 Concrete plant, taking down .020 
 
 Cost for plant, exclusive of purchase 
 
 price .311 
 
 Forms materials $0.272 
 
 Forms labor in erecting .346 
 
 Forms labor taking down .079 
 
 Cost forms .697 
 
 Labor mixing, placing and ramming .626 
 
 Total cost per cubic yard $5.844 
 
 350. In the building of a concrete dam for the enlargement 
 of the head of the Louisville and Portland Canal, 1 comparison 
 of cost of hand and machine mixing is given by Asst. Engr. J. 
 H. Casey. 
 
 1.63 bbls. natural cemont, at $0.635 per bbl. . . $1.034 
 
 2 volumes sand, .47 cu. yd., at .87 per cu. yd. . . .408 
 
 5 volumes broken stone, .89, cu. yd. at $.84 . . .756 
 
 Cost of testing cement .081 
 
 Forms, material for .107 
 
 Forms, labor making and setting up .168 
 
 Cost materials and forms $2.554 
 
 Hand mixed concrete: 
 
 . Cost of mixing $1.917 
 
 Cost of placing and tamping .791 
 
 Cost of mixing and placing $2.708 
 
 Total cost hand mixed concrete per cu. 
 
 yd. in place $5.262 
 
 Machine mixed concrete: 
 
 Charging and running mixer $0.864 
 
 Placing and tamping .585 
 
 Cost mixing and placing $1.449 
 
 Total cost machine mixed concrete per 
 
 cu. yd. in place $4.003 
 
 Difference in favor machine mixing $1.26 
 
 1 Capt. George A. Zinn, Corps of Engineers, U. S. A., in charge. 
 Chief of Engineers, 1900, p. 3467. 
 
 Report 
 
224 
 
 CEMENT AND CONCRETE 
 
 Since the above concrete was placed in large masses, the 
 costs of labor are considered high, and it is probable the work 
 was done with exceptional care. 
 
 351. In the construction of the lock at the Cascades Canal 1 
 the concrete plant was so arranged that the materials did not 
 have to be elevated, but much of the work of transportation 
 was done by gravity. The mixing of about eighteen hundred 
 yards by hand permits a comparison to be made with machine 
 mixing by which method about seven thousand eight hundred 
 yards were made. The costs were as follows: 
 
 ITEMS. 
 
 COST PEH Cu. YD. OF CONCRETE. 
 
 HAND MIXED AND 
 PLACED BY DERRICK. 
 
 MACHINE MIXED AND 
 PLACED BY CHUTE. 
 
 AMOUNTS. 
 
 TOTAL. 
 
 AMOUNTS. 
 
 TOTAL. 
 
 .805 bbl. Portland cement at 
 $4.08 . . 
 
 $3.20 
 .47 
 .60 
 
 .54 
 '.15 
 .22 
 
 $4.90 
 
 .37 
 1.09 
 .79 
 
 $3.29 
 .47 
 .60 
 
 .54 
 !l5 
 
 .22 
 
 '.39* 
 
 .04 
 
 141* 
 .05 
 
 $4.90 
 
 .37 
 .43 
 .46 
 
 .456 cu. yd. sand at $1.04 . . 
 .579 cu. yd. gravel at $1.04 . . 
 .317 cu. yd. broken stone at 
 $1.70 
 Cost materials in concrete . 
 Timbering . 
 
 Testing cement and general 
 repairs . 
 
 Forms and tests . 
 
 Mixing, labor 
 
 1.07 
 .02 
 
 " repairs and fuel . . . 
 Total cost mixing 
 
 Placing, labor 
 
 .60 
 .19 
 
 " fuel, tramways, etc. 
 Total cost placing 
 
 Total cost concrete per cu. yd. 
 
 . 
 
 $7.15 
 
 $6.16 
 
 352. In the construction of the retaining walls for the 
 Chicago Drainage Canal, 2 a special plant was designed for the 
 work on account of the large quantities of concrete required, 
 and this, combined with the low cost of materials and the char- 
 acter of the work, resulted in a very low cost concrete. On 
 
 1 Maj. Thomas H. Handbury, Corps of Engineers, U. S .A., in charge. 
 Report Lieut. Edward Burr, Report Chief of Engineers, 1891, Vol. v. Ab- 
 stracted, Engineering News, June 2, 1892. 
 
 2 " Construction of Retaining Walls for the Sanitary District of Chicago," 
 by Mr. James W. Beardsley, and discussion by Mr. Charles L. Harrison. 
 Jour. W. Soc. Engrs., Dec., 1898. 
 
EXAMPLES OF COST 225 
 
 Section 14 the stone was selected from the spoil banks along the 
 canal and could usually be obtained within one hundred feet. 
 This stone, which was delivered to the crusher by wheelbarrows, 
 required some sledging to reduce it to crusher size. An Austin 
 jaw crusher was mounted on a flat car with the Sooysmith 
 mixer. "The cement, sand and stone were raised from their 
 respective bins by means of belt conveyors running at the 
 same rate of speed, but carrying buckets spaced proportional to 
 the required ingredients." "The cost of a second hand plant 
 used on this section was estimated at $9,600, including two 
 crushers and two mixers at $1,500 for each machine. Common 
 labor cost $1.50 per day; firemen, enginemen, and carpenters 
 from $2.00 to $3.00 per day. The itemized cost is as follows: 
 
 ITEMS. 
 
 COST, CENTS 
 PEK CIT.YD. 
 
 General, including superintendent, blacksmith, water boys, etc. 
 Quarrying, i. e., delivering stone to crusher 
 
 Crushing 
 
 Transportation, delivering sand and cement to mixer by teams 
 
 Forms, exclusive of lumber 
 
 Mixing 
 
 Placing and tamping 
 
 7.8 
 30.3 
 
 7.3 
 14.2 
 15.0 
 12.1 
 10.8 
 
 Total 
 
 Cost of plant (no salvage allowance) 
 Cost of cement and sand .... 
 Total cost concrete per cubic yard 
 
 97.5 
 
 40.7 
 
 1(53.3 
 
 33.015 
 
 The amount of concrete used on this section was 23,568 cu. yds. 
 353. On Section 15 of the same work the conditions were 
 somewhat different. The stone had to be quarried within about 
 a thousand feet of the crusher. The stone, after being broken 
 to crusher size, was delivered on the tipping platform of the 
 No. 7 Gates crusher in cars drawn by a cable hoist. "The 
 average output of the crusher for a day of ten hours was about 
 210 cubic yards." The materials were transported to the 
 mixer in four and one-half yard dump cars drawn by a light 
 locomotive. The mixer was of the spiral screw type and de- 
 posited the materials on a rubber belt conveyor. The mixer 
 and operating machinery were mounted on a car which pro- 
 pelled itself by means of rope and winch. The plant for this 
 section was new and estimated to cost $25,420, including $12,000 
 for one crusher. 
 
226 CEMENT AND CONCRETE 
 
 The detailed cost is as follows: 
 
 ITEMS. 
 
 COST 
 
 PER Cu. YD. OF 
 
 CONCRETE, 
 
 CENTS. 
 
 General, including superintendent, blacksmith, teams, etc. 
 Quarrying (exclusive of 8.3 cents for explosives) 
 
 Crushing 
 
 Transportation, delivering cement, sand and stone on a 
 
 platform beside the mixer 
 
 Forms, exclusive of timber 
 
 Mixing, including shoveling materials from platform to 
 
 mixer 
 
 Placing and tamping 
 
 Total 
 
 Cost of plant (no salvage allowance) 
 Powder for quarrying .... 
 Cement and sand . 
 
 8.1 
 14.2 
 
 25.0 
 11.6 
 
 99.1 
 
 50.7 
 
 8.3 
 
 158.6 
 
 Total cost concrete per cu. yd. 
 
 $3.227 
 
 The amount of concrete used on this section was 44,811 
 cubic yards. 
 
CHAPTER XV 
 
 THE TENSILE AND ADHESIVE STRENGTH OF CEMENT 
 MORTARS AND THE EFFECT OF VARIATIONS 
 IN TREATMENT . 
 
 ART. 45. THE TENSILE STRENGTH OF MORTARS OF VARIOUS 
 COMPOSITIONS AND AGES 
 
 354. THE PROPORTION OF SAND. The rate of change in 
 the strength of mortars as the proportion of sand is increased 
 varies greatly for different cements. The fineness and chemical 
 composition of the cement, and the quality of the sand, are the 
 most important factors influencing this rate of change upon 
 which the question of the relative economies of different mor- 
 tars is so largely dependent. 
 
 Table 67 gives the results of tests with two brands of Port- 
 land cement mixed with from two to ten parts of river sand, 
 the age of briquets being six months and two years. It is of 
 interest to notice that the strengths of the mixtures are ap- 
 proximately in the inverse ratio of the number of parts of sand 
 used. Thus the strength with six parts sand is approximately 
 two-sixths of the strength with two parts, while with ten parts 
 sand, the strength is nearly two- tenths of that with mortar 
 containing two parts. 
 
 TABLE 67 
 
 Rate of Decrease in Strength with Addition of Sand 
 PORTLAND CEMENT; RIVER SAND, " POINT AUX PINS " 
 
 
 TENSILE STRENGTH, LBS. PEB SQ. IN. 
 
 
 PARTS SAND 
 
 
 PROPORTIONATE 
 
 
 
 TO 1 
 
 CEMENT BY 
 
 6 MONTHS. 
 
 2 YEARS. 
 
 STRENGTH, 
 Two YEARS, IF 
 
 
 H 
 
 R 
 
 H 
 
 R 
 
 Mean. 
 
 1 TO 2=100. 
 
 2 
 
 512 
 
 504 
 
 634 
 
 548 
 
 541 
 
 100 
 
 3 
 
 390 
 
 335 
 
 363 
 
 355 
 
 359 
 
 66 
 
 4.09 
 
 295 
 
 261 
 
 296 
 
 288 
 
 292 
 
 54 
 
 6 
 
 175 
 
 144 
 
 191 
 
 174 
 
 182 
 
 35 
 
 8 
 
 113 
 
 96 
 
 132 
 
 132 
 
 132 
 
 24 
 
 10 
 
 (54 
 
 74 
 
 104 
 
 116 
 
 110 
 
 20 
 
 227 
 
-228 
 
 CEMENT AND CONCRETE 
 
 355. In Table 68 similar results are given for two samples 
 of Portland cement and two kinds of sand, neat cement speci- 
 mens being included in the comparison. The one-to-one mor- 
 tars give a higher strength than neat cement, and even the 
 mortar containing two parts of the limestone screenings is 
 stronger than the neat specimens. From the one-to-one mor- 
 tars the strengths decrease rapidly as more sand is added, until 
 five parts sand are used, but the strengths then decrease less 
 rapidly as larger additions of sand are made. 
 
 TABLE 68 
 Rate of Decrease in Strength with Addition of Sand 
 
 PORTLAND CEMENT, BRAND R; SAND, CRUSHED QUARTZ AND LIMESTONE 
 
 SCREENINGS 
 
 
 TENSILE STRENGTH, POUNDS 
 TEH SQ. IN. 
 
 PROPORTIONATE 
 STRENGTH IK STRENGTH 
 1 TO 1 MOUTAR= 100. 
 
 PARTS SAND 
 
 
 
 
 
 TO 1 
 
 Sample i 
 
 Sample 
 
 
 
 CEMENT BY 
 
 Cement H H, 
 
 Cement II, 
 
 
 
 WEIGHT. 
 
 Crushed Quartz 
 
 Limestone 
 
 Crushed 
 
 Limestone 
 
 
 Sand, 20-30. 
 Age Briquets, 
 G Months. 
 
 Screenings, 20-30. 
 Age Briquets, 
 6 Months. 
 
 Quartz. 
 
 Screenings. 
 
 
 
 689 
 
 686 
 
 82 
 
 78 
 
 1 
 
 840 
 
 881 
 
 100 
 
 100 
 
 2 
 
 521 
 
 703 
 
 62 
 
 80 
 
 3 
 
 368 
 
 508 
 
 44 
 
 58 
 
 4 
 
 236 
 
 335 
 
 28 
 
 38 
 
 5 
 
 203 
 
 267 
 
 24 
 
 30 
 
 6 
 
 156 
 
 178 
 
 19 
 
 20 
 
 8 
 
 104 
 
 138 
 
 12 
 
 15 
 
 10 
 
 78 
 
 98 
 
 9 
 
 11 
 
 356. In Table 69 two samples of natural cement are treated 
 in a similar manner, from one to eight parts river sand being 
 used in the mortars. With Sample II the strength is dimin- 
 ished rapidly until five parts sand have been added, but with 
 further additions of sand, the strength is decreased more slowly. 
 Sample 18 S gives quite a different curve, as the one-to-two 
 mortar is stronger, and the one-to-three mortar is but little 
 weaker than the one-to-one. With four parts sand the mortar 
 shows a marked falling off in strength, but further additions of 
 sand diminish the strength more slowly. 
 
 357. INCREASE IN TENSILE STRENGTH WITH TIME. In 
 Table 70 are given the results obtained in tests of tensile strength 
 
COMPOSITION AND AGE 
 
 229 
 
 TABLE 69 
 
 Rate of Decrease in Strength with Addition of Sand. Natural 
 Cement, Brand Gn; River Sand, "Point aux Pins" 
 
 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH. 
 
 SAND TO 1 
 CEMENT 
 BY WKKJHT. 
 
 AGE. 
 
 6 MONTHS. 
 
 2 YEARS. 
 
 Proportionate 
 
 
 Sample 
 Cement 
 
 II. 
 
 188. 
 
 18 S. 
 
 Years if 1 to 2 
 r^lOO. 
 
 
 
 
 380 
 
 
 
 
 1 
 
 
 297 
 
 308 
 
 280 
 
 86 
 
 2 
 
 
 200 
 
 314 
 
 324 
 
 too 
 
 3 
 
 
 183 
 
 280 
 
 294 
 
 91 
 
 4 
 
 
 128 
 
 193 
 
 187 
 
 58 
 
 5 
 
 
 81 
 
 101 
 
 165 
 
 51 
 
 
 
 
 69 
 
 142 
 
 172 
 
 53 
 
 7 
 
 
 56 
 
 119 
 
 156 
 
 48 
 
 8 
 
 
 53 
 
 101 
 
 114 
 
 35 
 
 with twelve samples of Portland cement, illustrating the rates 
 of increase in strength from seven days to three years. It is 
 seen that rich mortars gain strength rapidly, neat and one-to- 
 one mortars showing usually eighty to ninety per cent, of their 
 ultimate strength in twenty-eight days. Mortars containing 
 not more than four parts sand to one cement give practically 
 their ultimate strength at six months. It is also of interest to 
 notice that the variations in strength among the several sam- 
 ples are not very great. The lowest strength at the end of 
 two to three years is seventy-five to eighty per cent, of the 
 highest. 
 
 358. In the case of natural cements, results for ten brands 
 of which are given in Table 71, only fifty to seventy per cent, 
 of the ultimate strength is gained in the first twenty-eight 
 days; with mortars containing three parts sand to one cement 
 the average result at twenty-eight days is less than forty per 
 cent, of the strength at two years. Most of the samples gain 
 some strength after six months, but two samples fail at two 
 years which had given a fair result at six months. The varia- 
 tions in strength among the several samples are very much 
 greater than with Portland cements; even omitting the two 
 samples that failed, the strength of the highest is two or three 
 times the strength of the weakest sample at two years. 
 
230 
 
 CEMENT AND CONCRETE 
 
 Strength 
 
 OF SAND TO O 
 BY WEIGHT 
 
 ONE PART SAND TO ON 
 CEMENT BY WEIGHT. 
 
 FINENESS: 
 PER CENT 
 PASSING. 
 
 
 
 soui 9 
 
 soui 9 
 
 stop 82 
 
 soui 9 
 
 COCOCOCOCOCOCOCO 
 
 ^<N, I 
 COCOCO 
 
 CO GO 
 (M o 
 (N CO 
 
 TfH OU CO TC O O rH O O CO 
 
 01 OS GO 00 GO t^- t^- tO ^O l ^ CO 
 
 cococococococococococo 
 
 O -^ 
 GO !> 
 
 -^ co 
 
 CO^^t~^'lt>. 
 OCOOOGOGOCO 
 T^TtlCOCOCOCO 
 
 ^- 
 
 COCOCOCO 
 
 rHt-(NO5iO' ' t^ O rji 
 OC.Oi-HrHOOCO<M^ t- 
 
 -^ 
 (M 
 
 l- t^ CD t^ t^ l^ SO t~- 
 
 i i ?O O CO 
 
 -H^UtJ-HH 
 ?O C5 O 
 
 OCM?Dt^<MC<li-iCOOCOiOTH 
 'C O TJ< r- O 
 
 OOSi-HTOCO^-H/li l^COCO 
 
 lOCDCi^OOOT-ifMT^^Or-i 
 
 00 r- t l>- Ir- r- I- 
 
 - 
 
 CO 
 
 -HH 
 Oi 
 
 OOO5<MCOQOtOi iC5COt~CO CO 
 t^l-lr-lCO<M(7^l^i 1 t^ Tj< CO r- 1 t^ 
 tO t^* CO 
 
 O5COC<1CO' i i i 
 
 , T* .t-OO^rHOOO 
 
 cd ^oicddoicod 
 
 oxenbg; "uj 
 S900' saiOH 
 
 t- .o . co t>- o o o o o 
 
 8O5 ^^ CO ^O CO Oi "^ C^ 
 " 00 * 00 00 OO 00 OO 00 O^ 
 
 pirejg 
 
COMPOSITION AND AGE 
 
 231 
 
 o 
 
 THREE PARTS SAND T 
 ONE CEMENT. 
 
 SAND 
 ENT. 
 
 Two PA 
 ONE 
 
 E PART SAND 
 CEMENT. 
 
 suuaX z 
 
 soiu 9 
 
 sotu g 
 
 stop 85 
 
 soui 9 
 
 stop 8S 
 
 stop 
 
 soui 9 
 
 'SOU! g 
 
 stop 82 
 
 stop 
 
 soui 9 
 
 stop 
 
 stop 
 
 top 
 
 S 
 
 t^iCCiCMOOOOOOl-- CO r- t^ O 
 I-H ;c Ci i~ 00 CO l^ -f ?< CM iO Tfi CO 1-1 
 i i CM CO W rH CN CO CO CO CM CO CM T-H 
 
 'MClCOT-HCOOCOCiO''* OCOOQO 
 7-4 QO rn "N tS O '" '- C O <M QO 'N t 
 T-Ht-i(jqCO(M(MC^i3<l(M<N 
 
 <Mt^c: 
 
 co ci a 
 
 OCO-^^CCM 
 CO n -0 C 
 
 iO -*Ci'!tit^ 
 t- CT. Ci CO 
 
 o i i to .-> ;o t^ C5 o 
 
 O^ O O 
 
 W V W VJV 
 
 S^ CC CC T 
 
 rH >M CO <N CO 
 
 r t>- 
 
 7^ (N 
 
 3 
 
 CD Oi O CO CO 3D CC CO CO 
 
 CC CO CO *O O 1-1 ^ I-H 
 CO t~~- Oi ^O i CO CO i> 
 
 CO^CiOCOCiOiOCO (M CD T<l OS 
 OOl-QO'^C-NQO'-'.-T CCOiQC^: 
 
 CO CO CO C -^ t^ QO *1 OS -efi T< l^ U5 >O 
 
 't^c:ot--H'^cot-o o-^coco 
 
 1- CO OS 
 
 O 
 
 
232 
 
 CEMENT AND CONCRETE 
 
 359. Table 72 shows in some detail the rate of increase in 
 strength of a sample of natural cement when the specimens 
 are maintained in air and in water. This table has several 
 points of interest. When hardened in water, the cement gained 
 steadily in strength up to six months, when it began to fall 
 off, and at two years this cement failed, as is shown in Table 
 71. The neat cement specimens hardened in air are very 
 irregular, as usual. These specimens showed high strength at 
 three months, suffered a marked falling off at six months and 
 one year, but showed a remarkable strength, equal to neat 
 Portland cement, at two years. The strength developed at 
 one year by specimens of this sample containing two parts 
 sand to one cement and hardened in air, is also equal to that 
 shown by similar mortars of Portland cement. 
 
 TABLE 72 
 
 Rate of Increase in Strength in Water and Air 
 
 AGE OF BRIQUETS. 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH. 
 
 NEAT CEMENT. 
 
 MORTAR Two PARTS SAND 
 BY WEIGHT TO ONE 
 CEMENT. 
 
 Water. 
 
 Air. 
 
 Water. 
 
 Air. 
 
 1 day 
 7 days .... 
 14 days .... 
 28 days .... 
 3 raos. .... 
 6 mos. .... 
 1 year .... 
 2 years 
 
 81 
 192 
 232 
 305 
 390 
 437 
 432 
 395 
 
 152 
 254 
 315 
 473 
 551 
 372 
 314 
 731 
 
 ' 135 ' 
 
 ' 232 ' 
 367 
 409 
 249 
 
 ' 142 ' 
 
 271 
 459 
 475 
 
 537 
 
 
 All cement, Brand Hn, Sample 26 S, which fails in water after two years, 
 see Table 71. 
 
 ART. 46. CONSISTENCY OF MORTAR AND AERATION OF CEMENT 
 
 360. EFFECT OF CONSISTENCY OF MORTAR ON TENSILE 
 STRENGTH. The results in Table 73 are from briquets of 
 Portland cement with two parts " Standard" crushed quartz. 
 The consistency of the mortars varied from a " trifle dry/' in 
 which water rose to the surface only after continued tamping, 
 to a wet mortar which would just hold its shape when placed 
 in a heap on the slab. Half of the briquets were immersed, 
 
CONSISTENCY 
 
 233 
 
 while the remainder were stored in the air of the laboratory. 
 The air hardened specimens gave higher results in all cases 
 than those hardened in water. The highest strength was given 
 in general by the dryest mortar, but the differences in strength 
 decrease as the age of the specimen increases. 
 
 TABLE 73 
 
 Effect of Consistency on the Strength of Portland Cement Mortar 
 Hardened in Water and Air 
 
 AGK OK BRIQUETS. 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH. 
 
 Consistency of Mortar. 
 
 Briquets Hardened in Fresh 
 Water. 
 
 Briquets Hardened in Air 
 of Laboratory. 
 
 a 
 
 b 
 
 c 
 
 (I 
 
 e 
 
 a 
 
 b 
 
 c 
 
 d 
 
 e 
 
 7 days . . . 
 
 340 
 
 310 
 
 220 
 
 101 
 
 158 
 
 407 
 
 341 
 
 263 
 
 230 
 
 202 
 
 28 days . . . 
 
 883 
 
 378 
 
 314 
 
 291 
 
 240 
 
 500 
 
 463 
 
 345 
 
 393 
 
 302 
 
 3 months . . 
 
 515 
 
 535 
 
 514 
 
 429 
 
 411 
 
 065 
 
 503 
 
 638 
 
 507 
 
 451 
 
 Cement: Brand R, Sample 18 II, with two parts "Standard" sand. 
 Consistency: a, trifle dry; b, O.K.; c, moist; d, very moist; e, would just 
 hold shape. 
 
 361. Tables 74 and 75 give similar results for Portland and 
 natural cement mortars, respectively, all specimens having 
 hardened in water for three months. The amount of water 
 used in gaging had a wide range, giving mortars of all consis- 
 tencies from very dry to very moist. The richness of the mor- 
 tar was also varied, from neat cement to five parts sand. A 
 comparison of the results in these two tables indicates that the 
 highest strength is usually given by mortars a trifle dryer than 
 that considered right for briquets; that an excess of water is 
 less deleterious to rich mortars than to lean ones, and to Port- 
 land cement than to natural cement, 
 
 362. Conclusions. Although all of these tests indicate the 
 superiority of dry mortars, in considering the effect of consis- 
 tency from a practical standpoint, one must not fail to consider 
 the difference between the conditions existing in the actual use 
 of mortars and in laboratory tests. When mortar is used in 
 
234 
 
 CEMENT AND CONCRETE 
 
 TABLE 74 
 
 Variations in Consistency of Mortar 
 EFFECT ON STRENGTH OF PORTLAND MORTAR AT THREE MONTHS 
 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH, FOR 
 
 PARTS SAND TO 
 
 CONSISTENCY NUMBER. 
 
 1 CEMENT 
 
 
 BY WEIGHT. 
 
 
 
 
 
 
 
 
 
 
 
 1 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 8 
 
 9 
 
 
 
 608 
 
 635 
 
 763 
 
 744 
 
 708 
 
 707 
 
 729 
 
 085 
 
 
 1 
 
 513 
 
 543 
 
 618 
 
 588 
 
 594 
 
 613 
 
 566 
 
 506 
 
 538 
 
 2 
 
 
 
 429 
 
 
 447 
 
 398 
 
 393 
 
 382 
 
 
 3 
 
 
 289 
 
 
 322 
 
 329 
 
 310 
 
 
 279 
 
 
 5 
 
 . . . 
 
 208 
 
 . . . 
 
 230 
 
 201 
 
 189 
 
 
 167 
 
 
 f 1 Very dry ; little or no moisture appeared on 
 Consistency surface of briquets. 
 
 Significance of numbers : J 5 About proper consistency for briquets. 
 Increasing per cent, water | 9 _ Very moist; mortar would barely hold shape 
 used for higher numbers. ^ and shrank in molds in hardening. 
 
 TABLE 75 
 
 Variations in Consistency of Mortar 
 EFFECT ON STRENGTH OF NATURAL CEMENT MORTAR AT THREE MONTHS 
 
 PARTS SAND TO 
 1 CEMENT 
 BY WEIGHT. 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH FOR 
 CONSISTENCY NUMBER. 
 
 1 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 8 
 
 9 
 
 
 1 
 
 2 
 3 
 5 
 
 . . . 
 
 239 
 
 372 
 312 
 255 
 217 
 150 
 
 373 
 314 
 
 283 
 208 
 155 
 
 '277' 
 206 
 125 
 
 305 
 
 286 
 258 
 183 
 101 
 
 281 
 242 
 
 268 
 241 
 204 
 139 
 74 
 
 263 
 207 
 176 
 
 . . . 
 
 
 Consistency Significance of numbers: 
 
 1 Very dry; little or no moisture appeared on surface briquets. 
 5 About proper consistency for briquets. 
 
 9 Very moist ; mortar would barely hold shape and shrank in 
 hardening. 
 
 masonry, the stones or bricks, even though they be dipped, 
 or sprayed with a hose before setting, are very likely to press 
 out or absorb considerable moisture from the mortar. To 
 realize this one has only to raise a heavy stone just after it has 
 been bedded; and the greater ease of setting either stone or 
 
AERATION OF CEMENT 
 
 231 
 
 brick, and obtaining a full mortar bed, with a rather wet mor- 
 tar, is appreciated by all masons. In the laying of concrete, 
 the difficulties of obtaining a compact mass with a dry mortar are 
 also not to be overlooked, but this point is discussed elsewhere. 
 
 363. Effect of Aeration on the Tensile Strength of Cement. - 
 Portland cements that are not perfect in composition and 
 burning, and that therefore contain free lime, may sometimes 
 be rendered sound by exposing them to air, and such exposure 
 was at one time considered almost essential in Portland cement 
 manufacture. 
 
 Fresh Portland cements that are slightly defective may have 
 their properties quite radically changed by such treatment; 
 their rate of setting becoming first more rapid, and then, by 
 further aeration, slower, and their tendency to expand over- 
 come or ameliorated. Portland cements that are perfectly 
 sound suffer some loss in specific gravity by the absorption of 
 carbonic acid and water from the atmosphere, but moderate 
 aeration has no radical effect upon their strength, and Port- 
 lands deteriorate but very slowly by storage, provided the 
 cement is kept dry and does not cake in the package. 
 
 Natural cements, however, usually suffer by aeration, and 
 this is illustrated by tests on several samples of one brand 
 given in Tables 76 and 77. Of the four samples in Table 76, 
 
 TABLE 76 
 Effect of Aeration on Four Samples of Same Brand Natural Cement 
 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH. 
 
 NUMBER 
 
 
 WEEKS 
 
 
 
 CEMENT 
 
 Age of Briquets, 6 Months to 7 Months. 
 
 Age Briquets, 2 Years. 
 
 AERATED. 
 
 
 
 
 Sample QQ 
 
 ss 
 
 NN 
 
 OO 
 
 NN 
 
 OO 
 
 
 
 242 
 
 183 
 
 343 
 
 340 
 
 316 
 
 306 
 
 2 
 
 237 
 
 2(59 
 
 357 
 
 500 
 
 368 
 
 432 
 
 5 
 
 250 
 
 403 
 
 
 
 
 
 7 
 
 268 
 
 358 
 
 . 
 
 
 
 
 10 
 
 . 
 
 . 
 
 225 
 
 212 
 
 246 
 
 284 
 
 11 
 
 313 
 
 279 
 
 
 
 
 
 13 
 
 . . . 
 
 . . . 
 
 213 
 
 218 
 
 200 
 
 258 
 
 Cement: Brand Gn; Sand, two parts crushed quartz to one cement. 
 All briquets of one sample were made by one molder and same percentage 
 water used. 
 
236 
 
 CEMENT AND CONCRETE 
 
 NN and OO showed an improvement by two weeks' exposure 
 to air, spread out in a thin layer, but longer exposure resulted 
 in a serious loss of strength. Of the other two samples, SS was 
 greatly improved by five weeks' aeration, but longer exposure 
 was detrimental, while sample QQ showed a continuous im- 
 provement up to the limit of eleven weeks' exposure to air. 
 
 In Table 77 the effect of aeration on five samples of the same 
 brand is shown. One of these samples was overburned and 
 was rendered practically worthless by fourteen weeks' exposure 
 to air. Nearly all of the samples in this table were seriously 
 affected by six weeks' aeration. 
 
 TABLE 77 
 
 Natural Cement. Effect of Aeration 
 
 
 
 
 TENSILE STRENGTH, POUNDS 
 
 
 
 b 
 
 CEMENT. 
 
 PARTS 
 
 
 PER SQUARE INCH, CEMENT 
 
 
 
 ^ 
 
 
 SAND 
 
 AGE OF 
 
 AERATED. 
 
 
 
 Ss 
 
 
 TO 1 
 
 BRI- 
 
 
 
 
 u~gW 
 
 
 CE- 
 MENT. 
 
 QUETS. 
 
 
 a 
 
 b 
 
 SSfig 
 
 02 
 
 a 
 
 Brand. 
 
 Sam- 
 ple. 
 
 4 to 5 
 
 days. 
 
 11 to 12 
 days. 
 
 45 to 51 
 days. 
 
 99 
 days. 
 
 Gn 
 
 84 
 
 2 
 
 6 ino. 
 
 414 
 
 321 
 
 208 
 
 216 
 
 80.5 
 
 54 
 
 3.01 
 
 " 
 
 83 
 
 ' 
 
 (t 
 
 463 
 
 392 
 
 211 
 
 235 
 
 85.9 
 
 41 
 
 3.11 
 
 t( 
 
 82 
 
 ' 
 
 " 
 
 44o 
 
 3oO 
 
 217 
 
 266 
 
 8-3.6 
 
 34 
 
 3.09 
 
 u 
 
 U' 
 
 ' 
 
 it 
 
 383 
 
 354 
 
 273 
 
 274 
 
 87.8 
 
 23 
 
 2.95 
 
 " 
 
 0' 
 
 ' 
 
 " 
 
 203 
 
 293 
 
 277 
 
 52 
 
 89.7 
 
 97 
 
 3.14 
 
 Fineness expressed as per cent, passing holes .0046 inch square. 
 
 Time setting fresh cement, time to bear T a .j inch | Ib. wire. 
 
 ART. 47. REGAGING CEMENT MORTAR 
 
 364. The Effect of Thorough Gaging. The value of thor- 
 ough gaging is a point frequently overlooked in the preparation 
 of mortars and concretes. Table 78 gives a few of the results 
 obtained in experiments to determine the effect of thorough 
 work in mixing. The tests are made with two brands of natural 
 and one of Portland , with two parts sand to cne cement by 
 weight. The two minutes' mixing with hoe and box method 
 gave a more thorough gaging than could have been accom- 
 plished in the same time with a trowel, and represented 
 about the amount of work put on mortars for testing. We are 
 not, therefore, comparing well mixed and poorly mixed mortars, 
 but rather well gaged and better gaged. The effect of the 
 additional work is shown in all cases; to double the time spent 
 
REGAGING 
 
 237 
 
 in gaging, increases the strength of the resulting mortar about 
 five per cent., while to quadruple the time adds twenty-six 
 per cent, to the strength. 
 
 TABLE 78 
 Effect of Thorough Gaging 
 
 
 CEMENT. 
 
 SAND, Two PARTS 
 TO ONE CEMENT. 
 
 TENSILE STRENGTH, LBS. 
 FEU So,. IN., FOR MORTAR 
 
 REF. 
 
 
 
 * ' ' 
 
 
 Kind. 
 
 Brand. 
 
 Kind of Sand. 
 
 2 Min. 
 
 4 Min. 
 
 8 Min. 
 
 1 
 
 Natural 
 
 Gn 
 
 j Pt. aux Pins J 
 I Pass #10 Sieve ) 
 
 352 
 
 350 
 
 482 
 
 2 
 
 it 
 
 it 
 
 Standard 
 
 418 
 
 451) 
 
 572 
 
 3 
 
 " 
 
 An 
 
 j Pt. aux Pins J 
 I Pass #10 Sieve j 
 
 368 
 
 376 
 
 421 
 
 4 
 
 Portland 
 
 R 
 
 j Pt. aux Pins ) 
 \ Pass #10 Sieve j 
 
 625 
 
 554 
 
 010 
 
 Mean 
 
 416 
 
 430 
 
 523 
 
 Prop 
 
 ortional .... 
 
 100 
 
 105 
 
 126 
 
 
 365. REGAGING. When more mortar is mixed at one 
 time than is required for immediate use, there is always a 
 temptation to retemper the mass and use it, even though it 
 may have been standing for some time. The practice is 
 usually prohibited by specifications and strenuously opposed 
 by engineers. The tests recorded in Tables 79 to 83 were 
 made to determine the effect of regaging on the resulting 
 strength of the mortar. 
 
 366. The results obtained with two brands of Portland ce- 
 ment are given in Table 79. The first result in each line of the 
 table is the strength attained by the mortar when treated as 
 usual. The severity of the treatment of the mortar as regards 
 regaging is shown by the letters heading the columns and the 
 corresponding foot notes. The first general statement to be 
 made concerning the results in this table is that in no case is 
 the effect of regaging Portland mortars containing sand shown 
 to be seriously deleterious to the tensile strength. Neat cement 
 mortar is not improved by regaging, and if allowed to stand 
 more than one hour, and then made into briquets without any 
 further addition of water, the strength is considerably decreased. 
 If water is added and the mortar frequently regaged, however, 
 
238 
 
 CEMENT AND CONCRETE 
 
 s s 
 
 i 
 a 
 
 S 
 
 <N rt 
 Tf rji 
 O iC 
 
 x o 
 
 t r ( 
 <N CO 
 
 00 GO 
 O5 GO 
 
 O ' O 
 
 0100 
 
 Par 
 to 
 Cem 
 
 |S -5 
 
 o o * 
 
 t^ t^ >O >-( -^ . -H 
 l^ i <N 00 00 O 
 ^ O iC 1C O t 
 
 r-H .71 
 
 1C ^ 
 
 s >;s >,2- >,- s 2 
 
 ?OT-lCOr-l?O (M COi-ICO 
 
 (M(M(M(NOOOOi-Hr-l 
 
 -1 
 
 - W 
 
 O t-l 
 
 11.1 
 
 Jji 
 
 Illlil 
 
 5 1 
 
 ""3 O3 O & O & 
 
 C/ -^ .3 -f^ "+* 
 
 ^1, &!* 
 ^||'|_'*'* 
 
 - 
 
 d 
 
 03 o 
 
 0) 0) ( 
 t373 
 
 ^5^3 
 
 flj F ^ ^C r* ^3 
 
 .-^' s ^l 
 
 ) 4) rr, *. 02 02 "^ bC 
 
 rK8a-tfgS l f8 
 
 
 I I I i I I I I I I I 
 
 r 
 
 -1-3 
 
 e 
 1 
 
REGAGINO 
 
 239 
 
 even neat cement mortar does not suffer a great decrease in 
 strength by three to six hours standing. Rich mortars, con- 
 taining one part sand, are not seriously affected by standing 
 three hours if regaged frequently. Poorer mortars, with two 
 to four parts sand, show an actual increase in strength as the 
 effect of such severe treatment as standing five hours, if re- 
 tempered with more water once an hour. These two brands 
 were slow setting Portlands, beginning to set in forty minutes 
 to two hours. The increase in strength of the regaged mortars 
 is doubtless due, at least in part, to the more thorough gaging 
 which they received. 
 
 Table 80 gives similar results of briquets one year old made 
 at another time with two parts river sand. The fact that dur- 
 ing the delay between the making and use of the mortar it 
 should be frequently retempered with water to make up for 
 the loss by evaporation, is plainly shown. 
 
 TABLE 80 
 Regaging Portland Cement Mortar 
 
 TKNHIMS STKKNGTH, POUNDS PKR SQUARK INCH, FOR VARYING TRKATMKNT. 
 
 a 
 
 c 
 
 d 
 
 e 
 
 / 
 
 h 
 
 i 
 
 j 
 
 579 
 
 565 
 
 5(39 
 
 570 
 
 568 
 
 
 . . . 
 
 . . 
 
 554 
 
 579 
 
 . . . 
 
 . . . 
 
 . . . 
 
 627 
 
 624 
 
 560 
 
 Cement: Portland, Brand R, Sample 42 M. Sand: 2 parts " Point aux 
 Pins " passing No. 10 sieve. Age of briquets. 1 year. 
 Treatment: a Molded as soon as gaged. 
 
 c Mortar let stand 1 hour, regaged and briquets made. 
 
 d Mortar let stand 3 hours, regaged each hour. 
 
 h Mortar let stand 3 hours, regaged each hour and water 
 
 added to restore original consistency. 
 e Mortar let stand 5 hours, regaged each hour. 
 i Mortar let stand 5 hours, regaged each hour and water 
 
 added to restore original consistency. 
 
 / Mortar let stand 5 hours, regaged and briquets made. 
 / Mortar let stand 5 hours, regaged and briquets made ; 
 water added to restore original consistency. 
 
 367. Similar tests with natural cements are shown in Table 
 81, and it appears that cements of this class, especially if mixed 
 neat, will not stand the same severe treatments without injury. 
 
240 
 
 CEMENT AND CONCRETE 
 
 
 H 
 
 - 
 
 ' <N O 0^ 
 | ' CN W <>4 . ] 
 
 
 s r 
 
 - 
 
 'oo GO b- ' 
 
 Sq CO <M 
 
 
 *g 
 
 , 
 
 ' l~~ O t SO CO O ^ 
 
 * 00 GO 00 l~ C' OS 
 . G<l CN <M CC (N -^ i i 
 
 
 ^ w 
 
 II 
 
 s 
 
 ' <N 
 31 CO rH 
 ( 99 94 CO 
 
 
 gas 
 
 rfg 
 
 . 
 
 i : :::::: 
 
 
 
 
 
 
 C * 
 
 
 
 
 II 
 
 " 
 
 10 
 
 (>1 CO 
 
 
 w* 
 
 
 ^HO 
 
 
 J ^ 
 
 
 <7<J CO 
 
 M 
 
 oc H 
 
 
 
 rt 
 
 fc - 
 
 
 
 -P 
 
 WO 
 
 
 r-( CO \ 
 
 
 H 
 
 
 ^^ 
 
 1 
 
 g 
 
 a 
 
 iil^iiiii 
 
 s 
 
 0) 
 
 O 
 
 1 
 | 
 
 1 
 
 BKIQUETS. 
 
 05 05 03 05 05 
 
 sS 
 
 w> 
 
 a 
 
 rt 
 W) 
 
 
 Parts to 
 1 Cement, 
 
 M .eo. MM 
 
 
 
 Q 
 
 X 
 
 oc 
 
 d 
 
 s 
 
 O 
 
 1 "S 
 
 8= i i r|s = = 
 
 Qj ^ 
 ^ 
 
 OH" 
 
 
 H 
 
 'a 
 
 aBp6 
 
 
 9 
 
 A 
 
 c^^^, rtc^^^ 
 
 
 
 3 
 
 
 
 ja 
 
 a 
 
 "' 
 
 <U 03 o3 o3 03 o3l s3 
 
 2-e-e-e-e t: -B 
 
 c o o o o c o 
 
 ?5^ fe, 
 
 I I 
 
 Q -O W "?3 
 
 I S 
 
 "1 
 
REGAGING 
 
 241 
 
 s 
 
 Regagillg 
 
 
 .^ 
 
 to 
 
 
 a 
 
 
 
 
 ^ 
 
 | 
 
 g 
 
 * 
 
 co 
 
 
 a 
 
 o 
 fc . 
 
 .- 
 
 ' -M * 00 
 
 "2 "28 ' 
 
 . ^i . CO 
 
 
 INCH 
 
 IT MEN 
 
 -5: 
 
 O * l~- 
 
 co -^ 
 
 
 II 
 
 5V, 
 
 * 
 
 is ii i 
 
 
 PER fr 
 FEREN 
 
 s 
 
 i! : : i 
 
 
 'OUNDS 
 
 <({ DIF 
 
 
 
 3j " o* " 
 
 
 IS 
 
 - 
 
 i ii ii 
 
 
 g3 
 
 a 
 
 5 
 
 
 
 1 il ii 
 
 
 .NSILE 
 
 >Q 
 
 aiis; 
 
 
 Ui 
 
 S 
 
 co co eo co ^ 
 
 
 
 Parts to 1 
 Cement. 
 
 <M 5^ <M ?< M 
 
 ? 
 
 a' 
 
 QQ 
 
 d 
 
 a 
 
 P.P. pass No. 10 
 
 t( 
 
 ii 
 Ii 
 
 u Standard " 
 
 i 
 
 H 
 M 
 
 Sample. 
 
 lO T "^ Ut) 
 
 
 1 
 
 Brand. 
 
 a- d- a 
 
 O " "^ " O 
 
 
 a: 
 
 ia 
 
 -.,- 
 
 
 03 oj 
 
 e -e 
 
 o o 
 
 ^ s 
 
 Illll 
 
 1 1 1 1 1 1 1 1 1 1 1 
 
 e-0 0-T3 < '*^Cft^'^-'^ 
 
 1 
 
242 
 
 CEMENT AND CONCRETE 
 
 Neat cement mortars of these two brands appeared more plastic 
 when they were retempered with more water after standing one 
 hour (column /), but if allowed to stand three hours (column 
 i) t they had then become quite hard set. Mortars containing 
 two parts sand that had stood sixty to ninety minutes with 
 intermediate retempering, showed a slight' increase in tensile 
 strength, but more severe treatment was deleterious. 
 
 In Table 82 the mortars all contain two parts sand to one 
 of cement by weight. The only cases of any serious results of 
 retempering are for mortars standing four hours and regaged, 
 at intervals of one-half hour or one hour, with no water added 
 to th3 original mortar. Briquets made from mortar that had 
 been gaged every half hour, and was molded two hours after 
 first mixed, showed a somewhat higher strength than briquets 
 made of fresh mortar. 
 
 Table 83 shows that the behavior of regaged natural cement 
 mortars, as shown in the preceding tables, is not an eccentricity 
 of one or two brands. Mortars containing two parts sand do 
 not appear to suffer in tensile strength by being allowed to 
 stand two hours if regaged hourly. 
 
 TABLE 83 
 
 Effect of Regaging on Tensile Strength, Five Brands Natural 
 
 Cement 
 
 o 
 
 
 
 
 
 
 TENSILE STRENGTH, POUNDS PER 
 
 H ^ 
 
 AGE 
 BRIQUETS. 
 
 TIME 
 ELAPSED 
 
 BETWEEN 
 
 FIRST GAG- 
 
 L NtTMBE 
 LGINGS. 
 
 INTERVAL 
 BETWEEN 
 
 SUC- 
 CESSIVE 
 
 SQUARE INCH. 
 
 BRANDS. 
 
 I S 
 
 
 ING AND 
 MOLDING. 
 
 H^ 
 
 GAGINGS. 
 
 
 
 
 
 
 
 
 
 
 H 
 
 
 En 
 
 An 
 
 Dn 
 
 Kn 
 
 Hn 
 
 04 
 
 
 
 
 
 
 
 
 
 
 2 
 
 28 days 
 
 
 1 
 
 
 58 
 
 171 
 
 231 
 
 178 
 
 174 
 
 2 
 
 it 
 
 2 hours 
 
 3 
 
 1 hour 
 
 109 
 
 168 
 
 310 
 
 178 
 
 190 
 
 2 
 
 6 months 
 
 
 1 
 
 
 228 
 
 3*8 
 
 306 
 
 361 
 
 273 
 
 2 
 
 " 
 
 2 hours 
 
 3 
 
 1 hour 
 
 284 
 
 382 
 
 307 
 
 416 
 
 347 
 
 4 
 
 28 days 
 
 
 1 
 
 
 39 
 
 49 
 
 149 
 
 23 
 
 58 
 
 4 
 
 tt 
 
 2 hours 
 
 3 
 
 1 hour 
 
 33 
 
 70 
 
 146 
 
 
 56 
 
 4 
 
 6 months 
 
 
 1 
 
 
 104 
 
 146 
 
 188 
 
 '216 
 
 129 
 
 4 
 
 " 
 
 2 hours 
 
 3 
 
 1 hour 
 
 97 
 
 147 
 
 184 
 
 227 
 
 137 
 
 NOTES: Sand, Point aux Pins, passing No. 10 sieve. 
 In general, each result mean of five briquets. 
 All briquets made by one molder and stored in one tank. 
 All mortars appeared about same consistency when molded. 
 No water added in regaging except Brand Kn, 1 to 2 mortar, 
 standing two hours. 
 
MIXING CEMENTS 243 
 
 368. Conclusions. The conclusions to be drawn from these 
 tests appear to be as follows: The cohesive strength of mortars 
 of neat cement is appreciably diminished if they are allowed to 
 stand a considerable length of time after gaging before they 
 are used. Sand mortars, especially of Portland cement, usually 
 develop a higher tensile strength under moderate treatment of 
 this kind; and if regaged frequently, with sufficient water added 
 to keep them plastic, mortars of slow setting cements may be 
 used several hours after made without serious detriment to the 
 tensile strength. Portland cements withstand severe treatment 
 better than natural cements. 
 
 The effect of regaging on the adhesive strength is shown in 
 Table 117, 405. These tests were quite severe and pointed to 
 the conclusion that the adhesive strength is diminished by stand- 
 ing and regaging, rich mortars and natural cement mortars 
 being most affected. 
 
 The effect of regaging on a given sample should be investi- 
 gated before it is permitted to any great extent, or in the most 
 careful work. Regaged mortars are said not to give good re- 
 sults in sea water, and it may be expected that quick setting 
 cement will be injured by regaging. 
 
 ART. 48. MIXTURE OF CEMENT WITH LIME, ETC. 
 
 369. Mixture of Portland and Natural Cements. For cer- 
 tain uses mortar is sometimes made from a mixture of Portland 
 and natural cement, with the idea of retaining some of the 
 properties of the Portland without involving the expense of 
 using a clear Portland mortar. Several tests have been made 
 to determine the rate of hardening and the ultimate strength 
 of such mixtures. 
 
 The mortars used in the tests given in Table 84 contained two 
 parts sand to one of cement, and the cement was composed of 
 one-eighth, one-quarter and one-half Portland to seven-eighths, 
 three-quarters, and one-half natural. Mortars made with Port- 
 land alone and with natural alone are included for comparison. 
 It is seen that the mortars containing some Portland harden 
 more rapidly than the natural cement mortar, so that the in- 
 creased strength developed at short periods is more than pro- 
 portional to the per cent, of Portland used. The results ob- 
 
244 
 
 CEMENT AND CONCRETE 
 
 tained at two and three years, however, indicate that mortars 
 containing only a small proportion of Portland, as one-eighth 
 or one-quarter, do not give a higher ultimate strength than is 
 obtained with clear natural cement mortar. 
 
 TABLE 84 
 
 Tensile Strength of Mortars Made with Mixture of Portland and 
 
 Natural 
 
 w 
 
 
 
 o 
 g 
 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH. 
 
 ti 
 
 BRIQUETS. 
 
 Per I Portland 
 
 100 
 
 50 
 
 25 
 
 12.5 
 
 
 
 a 
 
 
 Cent. ( Natural 
 
 00 
 
 50 
 
 75 
 
 87.5 
 
 100 
 
 i 
 
 7 days 
 
 
 291 
 
 205 
 
 108 
 
 75 
 
 24 
 
 2 
 
 28 days 
 
 
 357 
 
 264 
 
 219 
 
 190 
 
 123 
 
 3 
 
 months 
 
 
 550 
 
 425 
 
 378 
 
 300 
 
 322 
 
 4 
 
 1 year 
 
 . 
 
 574 
 
 441 
 
 360 
 
 336 
 
 291 
 
 5 
 
 2 years 
 
 
 543 
 
 449 
 
 375 
 
 343 
 
 393 
 
 
 
 3 years 
 
 
 592 
 
 501 
 
 428 
 
 370 
 
 429 
 
 NOTES. Portland cement, Brand R, Sample 42 M. 
 Natural cement, Brand Gn, Sample 54 R. 
 Sand, two parts of " Point aux Pins," pass No. 10 sieve, to one 
 
 part cement by weight. 
 
 All briquets made by one molder and immersed in one tank. 
 Each result, mean of ten briquets. 
 
 370. In Table 85 four kmds or mixtures of cement are 
 used, Portland, natural, an " Improved cement" or a cement 
 
 TABLE 85 
 
 Comparisons of Portland, Natural, and " Improved " Cements 
 
 
 c% 
 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH. 
 
 
 H < ^ S 
 
 AGE OF 
 
 
 REF. 
 
 2||S| 
 
 WHEN 
 
 BROKEN. 
 
 Portland, 
 Brand U. 
 
 " Improved," 
 Brand Nn. 
 
 Portland,20%, 
 Natural, 80%. 
 
 Natural, 
 Brand, Mn. 
 
 1 
 
 None 
 
 7 days 
 
 547 
 
 206 
 
 250 
 
 199 
 
 2 
 
 it 
 
 28 " 
 
 586 
 
 293 
 
 341 
 
 270 
 
 3 
 
 One 
 
 7 " 
 
 458 
 
 169 
 
 200 
 
 165 
 
 4 
 
 u 
 
 28 " 
 
 569 
 
 253 
 
 331 
 
 234 
 
 5 
 
 1C 
 
 7 months 
 
 702 
 
 550 
 
 578 
 
 517 
 
 6 
 
 " 
 
 2 years 
 
 577 
 
 563 
 
 534 
 
 497 
 
 7 
 
 Two 
 
 2 " 
 
 522 
 
 510 
 
 573 
 
 529 
 
 8 
 
 Three 
 
 7 days 
 
 176 
 
 52 
 
 80 
 
 49 
 
 9 
 
 it 
 
 28 " 
 
 272 
 
 122 
 
 143 
 
 114 
 
 10 
 
 u 
 
 6 months 
 
 389 
 
 301 
 
 282 
 
 255 
 
 11 
 
 (f 
 
 2 years 
 
 371 
 
 346 
 
 356 
 
 342 
 
 12 
 
 Mean 
 
 2 years 
 
 490 
 
 473 
 
 488 
 
 456 
 
LIME WITH CEMENT 1M5 
 
 sold as a mixed cement, and a sample made by mixing twenty 
 per cent, of the Portland with eighty per cent, of the natural. 
 The first point noticed is that the " Improved" cement does 
 not exhibit the early hardening properties due to the Portland 
 cement in its composition (if any), as strongly as the sample 
 containing twenty per cent. Portland. In only two tests did 
 the "Improved" cement give a higher strength than the clear 
 natural. The results of the two-year tests are of interest as 
 showing how nearly the same ultimate strength is shown by 
 the four samples. The sample of natural cement is of excep- 
 tional quality. 
 
 371. Conclusions. It appears from these tests on the effect 
 of mixing Portland and natural cements that, in general, the 
 full strength of both cements is developed in the mixture; that 
 in the early stages of hardening, the mixture sometimes ex- 
 hibits more nearly the properties of the Portland, gaining 
 strength quite rapidly, but that the ultimate strength of mix- 
 tures containing small amounts of Portland are sometimes as 
 low as mortars made with natural cement alone. It cannot be 
 stated that all samples of Portland and natural cement will 
 give as good results in combination as those obtained in the 
 above tests, and any extended use of such mixtures should be 
 based on full tests of mixtures of the brands that are to be used 
 in combination. 
 
 372. Free Lime in Cement. The presence of free lime in 
 cement is known to be a serious defect. Table 86 gives the 
 results obtained by adding ground quicklime to Portland cement 
 in one-to-two mortars. It appears that eight per cent, quick- 
 lime reduces the strength at six months about twenty-five per 
 cent., and smaller amounts of lime produce approximately 
 proportional decrements. The seven-day results, both hot 
 and cold, show greater proportional effects. The free lime 
 occurring in cements as a result of defects of manufacture is 
 likely to be much more dangerous in character than the lime 
 used in these tests. 
 
 373. THE USE OF SLAKED LIME WITH CEMENT. A small 
 
 quantity of Portland cement is frequently added to lime mortar 
 to hasten the hardening and improve the strength. The ad- 
 dition of a small amount of slaked lime to Portland cement 
 mortar is also practiced. This not only cheapens the mortar 
 
246 
 
 CEMENT AND CONCRETE! 
 
 TABLE 86 
 Mixture of Ground Quicklime with Portland Cement 
 
 BRIQUETS STORED 
 IN WATER. 
 
 AGE OF 
 BRIQUETS. 
 
 TENSILE STRENGTH OF MORTARS IN 
 POUNDS PER SQUARE INCH. 
 
 Lime as Per Cent, of Total Lime and Cement. 
 
 
 
 2 
 
 4 
 
 6 
 
 8 
 
 Hot 80 C. . . ' . . 
 Hot 80 C 
 Ordinary tank 
 Ordinary tank 
 
 3 days 
 
 7 days 
 7 days 
 6 months 
 
 269 
 367 
 
 348 
 604 
 
 223 
 
 297 
 321 
 545 
 
 207 
 266 
 273 
 489 
 
 194 
 223 
 241 
 495 
 
 159 
 191 
 220 
 454 
 
 NOTES. Cement: Portland, Brand R, Sample 83 T. 
 
 Lime: Quicklime ground to pass No. 100 sieve (holes .0065 
 
 in. sq.). 
 Sand: Standard crushed quartz, 600 grams, to 300 grams of 
 
 cement plus lime. 
 Per cent, of lime given replaced the same weight of cement; 
 
 thus: for "4 per cent, lime" the mortar contained 288 grams 
 
 cement, 12 grams lime and 600 grams sand. 
 All briquets made by one molder; each result, mean of five 
 
 briquets. 
 
 but renders it much more plastic, or less " brash," in mason's 
 parlance. It is very difficult to lay bricks in a full mortar bed 
 with Portland cement mortar containing two or three parts 
 sand to one cement, and to use a richer mortar is usually too 
 expensive. The work is very much facilitated by mixing a 
 little slaked lime paste or powder with the mortar. 
 
 374. The tensile strength of such mixtures is shown by the 
 tests in Tables 87 to 89. In the mortars of Table 87 a sample 
 of Portland cement is mixed with slaked lime in two forms, 
 paste and powder. When the briquets are hardened in open 
 air the addition of ten to twenty per cent, of CaO in the form 
 of lime paste decreases the strength about twenty-five per cent. ; 
 seven per cent, of lime in the form of slaked, dry powder has, 
 however, no deleterious effect, and even twenty-eight per cent, 
 gives no serious decrease in strength. For water-hardened 
 specimens the addition of twenty to thirty per cent, of lime in 
 the form of paste appears to increase the strength twenty per 
 cent, and no deleterious effect is shown by the addition of 
 forty per cent. Also for water-hardened specimens, seven to 
 
LIME WITH CEMENT 
 
 247 
 
 twenty-eight per cent, of CaO in the form of slaked powder in- 
 creases the strength nearly twenty per cent. It thus appears 
 that the addition of lime gives better results in mortars that are 
 to harden in water, and that for air-hardened mortars lime 
 powder should be used in preference to lime paste. Similar 
 tests of seven-day briquets showed the lime paste to retard 
 the hardening of the mortar. 
 
 TABLE 87 
 
 Slaked Lime in Portland Cement Mortars 
 
 
 
 
 TKXSILK STRENGTH, 
 
 
 
 PROPORTIONS. 
 
 Poi'XUS PER 
 
 SQUARE INCH., SAMPLE 
 
 
 
 
 STORED IN 
 
 RKF. 
 
 Li MR IN FOIIM 
 OF 
 
 
 
 
 
 
 
 
 
 
 
 CaO in Lime 
 
 
 
 
 
 
 Cement, 
 Grains. 
 
 Paste or 
 Powder, 
 
 Saml, 
 Grams. 
 
 Open Air. 
 
 Water 
 
 Laboratory. 
 
 
 
 
 Grains. 
 
 
 
 
 1 
 
 Paste 
 
 200 
 
 
 
 600 
 
 404 
 
 382 
 
 2 
 
 
 200 
 
 20 
 
 000 
 
 308 
 
 426 
 
 3 
 
 
 200 
 
 40 
 
 600 
 
 292 
 
 450 
 
 4 
 
 
 200 
 
 60 
 
 000 
 
 224 
 
 462 
 
 5 
 
 
 200 
 
 80 
 
 600 
 
 219 
 
 384 
 
 
 
 Powder 
 
 200 
 
 
 
 000 
 
 382 
 
 371 
 
 7 
 
 
 200 
 
 14.3 
 
 600 
 
 385 
 
 443 
 
 8 
 
 
 200 
 
 28.6 
 
 000 
 
 316 
 
 451 
 
 9' 
 
 
 200 
 
 42.8 
 
 600 
 
 338 
 
 431 
 
 10 
 
 
 200 
 
 57.1 
 
 600 
 
 325 
 
 440 
 
 Cement: Portland, Brand R. Sand: Crushed Quartz, 20-30, or "Standard." 
 Age of briquets, 6 months. 
 
 375. In Table 88 only lime paste is used, but both Portland 
 and natural cement are tested, and the specimens are hardened 
 in dry air and damp sand. In the first column of results 
 are given the strengths attained by Portland cement mortar 
 containing three parts sand to one of cement without lime. 
 In the second column, ten per cent. CaO in form of paste is 
 added to the cement. In the third, fourth and fifth columns, 
 respectively, ten, twenty-five and fifty per cent, of the cement 
 is replaced by CaO. 
 
 It appears that ten per cent, of the cement in a one-to- 
 three Portland mortar may be replaced by lime made into paste 
 without diminishing the strength, if the mortar hardens in 
 damp sand. Even in dry air exposure, it is only at one year 
 
248 
 
 CEMENT AND CONCRETE 
 
 that the lime shows any deleterious effect. To replace twenty- 
 five per cent, or more of the cement with lime, however, dimin- 
 ishes the strength of the mortar in a marked degree. 
 
 In the case of natural cement, replacing ten per cent, of the 
 cement with lime is decidedly beneficial, and even twenty- 
 five per cent, lime gives enhanced strength, except for speci- 
 mens hardened in dry air. 
 
 Table 89 gives similar results for one-to-four mortars and 
 different percentages of lime, the briquets being hardened in 
 dry air and damp sand. 
 
 TABLE 88 
 
 Use of Lime Paste in Cement Mortars Containing Three Parts 
 
 Sand 
 
 
 
 
 
 TENSILE STRENGTH, LBS. PER SQ. IN. 
 
 
 
 
 
 Cement, gm. 
 
 200 
 
 200 
 
 180 
 
 150 
 
 100 
 
 FERENCE. 
 
 CEMENT. 
 
 
 
 BRIQUETS 
 STOKED. 
 
 AGE 
 
 BIQUETS. 
 
 
 
 GO 
 
 
 
 300 
 
 Lime Paste, " 
 
 
 
 60 
 
 150 
 
 CaO in Lime 
 Paste, gin. 
 
 
 
 20 
 
 20 
 
 50 
 
 100 
 
 a 
 
 
 
 PQ 
 
 Amt. CaO ex-") 
 
 
 
 
 
 
 
 
 
 
 pressed as % [ 
 of Cement f 
 
 
 
 9 
 
 10 
 
 25 
 
 50 
 
 
 
 
 
 nln T imp 1 
 
 
 
 
 
 
 
 Kind. 
 
 Brand 
 
 
 
 plus .Ljllliu.j 
 
 
 
 
 600 
 
 
 Sand, gm. 
 
 600 
 
 600 
 
 600 
 
 600 
 
 i 
 
 Port. 
 
 X 
 
 Dry aii- 
 
 28 da. 
 
 
 201 
 
 242 
 
 238 
 
 168 
 
 57 
 
 2 
 
 
 
 Damp sand 
 
 (i 
 
 
 294 
 
 330 
 
 309 
 
 238 
 
 95 
 
 8 
 
 
 
 Dry air 
 
 3 uio. 
 
 
 236 
 
 205 
 
 264 
 
 171 
 
 70 
 
 4 
 
 
 
 Damp sand 
 
 u 
 
 
 350 
 
 410 
 
 398 
 
 309 
 
 125 
 
 6 
 
 
 
 Dry air 
 
 l yr. 
 
 
 384 
 
 377 
 
 317 
 
 215 
 
 98 
 
 6 
 
 
 
 Damp sand 
 
 (( 
 
 
 430 
 
 445 
 
 442 
 
 332 
 
 171 
 
 7 
 
 Nat. 
 
 An 
 
 Dry air 
 
 3 mo. 
 
 
 310 
 
 338 
 
 359 
 
 251 
 
 69 
 
 8 
 
 
 
 Damp sand 
 
 u 
 
 
 267 
 
 344 
 
 327 
 
 318 
 
 93 
 
 9 
 
 
 
 Water 
 
 (( 
 
 
 222 
 
 301 
 
 319 
 
 293 
 
 79 
 
 In all of the above tests the mortars containing much lime 
 paste were not only more plastic, but somewhat wetter than 
 the corresponding mortars of cement and sand alone, on ac- 
 count of the water contained in the paste. 
 
 376. The conclusion to be drawn from these tests appears 
 to be that the addition of a small amount, ten to twenty per 
 cent., of slaked lime to cement mortars containing as much as 
 three parts sand, not only renders them more plastic, but 
 actually increases the tensile strength, especially if the mortars 
 are kept damp during the hardening. It also appears that for 
 
PLASTER PARIS WITH CEMENT 
 
 249 
 
 TABLE 89 
 
 Use of Lime Paste in Cement Mortars Containing Four Parts 
 Sand to One Cement 
 
 COMPOSITION OF MORTAR. 
 
 TENSILE STRENGTH OF MORTAR, 
 POUNDS PER SQUARE INCH. 
 
 Cement. 
 
 Lime 
 P't^te 
 
 Lime 
 in 
 
 Sand, 
 
 Stored in Damp 
 Sand. 
 
 Stored in Dry 
 Air. 
 
 
 
 
 Paste, 
 
 Grams. 
 
 Fresh 
 
 Old 
 
 Fresh 
 
 Old 
 
 Kind. 
 
 Grams. 
 
 
 Grams. 
 
 
 Lime 
 
 Lime 
 
 Lime 
 
 Lime 
 
 
 
 
 
 
 Paste. 
 
 Paste. 
 
 Paste. 
 
 Paste. 
 
 Portland, \ 
 
 240 
 
 00 
 
 00 
 
 960 
 
 170 
 
 180 
 
 254 
 
 244 
 
 Brand X, 1 
 
 240 
 
 80 
 
 27 
 
 900 
 
 212 
 
 200 
 
 280 
 
 250 
 
 Sample 1 
 
 200 
 
 120 
 
 40 
 
 900 
 
 198 
 
 212 
 
 227 
 
 237 
 
 41 S I 
 
 180 
 
 180 
 
 00 
 
 960 
 
 204 
 
 194 
 
 232 
 
 184 
 
 Natural, 
 BrandAn, 4 
 Sample L (^ 
 
 240 
 240 
 200 
 180 
 
 00 
 
 80 
 120 
 
 180 
 
 00 
 27 
 40 
 00 
 
 960 
 900 
 900 
 960 
 
 150 
 160 
 160 
 140 
 
 133 
 
 154 
 173 
 166 
 
 127 
 162 
 131 
 124 
 
 142 
 150 
 170 
 154 
 
 NOTE. All briquets three months old when broken. 
 
 mortars exposed to the open air the lime should be in the form 
 of slaked powder rather than paste. It may be added, that in 
 all cases care should be taken that the lime is thoroughly slaked 
 before use, and all lumps should be removed by straining or 
 sifting. Further results on this subject are given in connection 
 with the tests on adhesion of cement mortar to brick (Art. 5). 
 
 377. EFFECT OF PLASTER OF PARIS ON THE COHESIVE 
 STRENGTH OF MORTARS. 
 
 The use of plaster of Paris, or calcium sulphate, in the man- 
 ufacture of cement to regulate the time of setting, has already 
 been mentioned. The amount of such additions at the factory 
 are usually small, the German Cement Makers' Association limit- 
 ing it to two per cent. 
 
 Tests on three brands of Portland cement, showing the effect 
 of small additions of plaster Paris, are given in Table 90. All 
 of these mortars hardened in water. It is not known whether 
 any of the cements had received additions of plaster Paris be- 
 fore leaving the factory. It is probable that brands R and X 
 had been so treated, since they are German cements, but it is 
 not probable that the other brands of Portland had received 
 any addition of plaster. 
 
 It appears that with these brands the addition of from one 
 
250 
 
 CEMENT AND CONCRETE 
 
 to three per cent, of plaster Paris hastens the hardening and 
 increases the strength of the mortar at ages of six months to 
 two years. Six: per cent, plaster sensibly retards the hardening, 
 but, in all cases except one, Brand S, neat, six months, the 
 mortars containing six per cent, plaster, gave higher results 
 on long time tests than did the corresponding mortars to which 
 no plaster had been added. 
 
 TABLE 90 
 
 Plaster of Paris in Portland Cement Mortars, Hardening in Water 
 
 
 ft* 
 
 O H . 
 
 
 
 TENSILE STRENGTH, POUNDS PER SQ. 
 
 1 
 
 fig 
 
 jj 
 
 TEMPERA- 
 TURE WATER 
 
 IN WHICH 
 
 m 
 
 IN., WITH PER CENT. OF CEMENT 
 KEPLACED BY PLASTER 
 OF PARIS. 
 
 w 
 S 
 
 Sg 
 
 "H 
 
 BRIQUETS 
 
 ^D S 
 
 
 d 
 
 s* 
 
 ^ 
 
 STORED. 
 
 M 
 
 
 
 
 
 
 
 r> H^ 
 
 ^ ^ 
 
 
 PH 
 
 
 
 l 
 
 2 
 
 3 
 
 6 
 
 
 
 GO 
 
 
 PP 
 
 
 
 
 
 
 1 
 
 S 
 
 
 
 60to65Fahr. 
 
 7 da. 
 
 487 
 
 626 
 
 600 
 
 519 
 
 380 
 
 2 
 
 
 
 
 
 6 mos. 
 
 743 
 
 746 
 
 754 
 
 742 
 
 660 
 
 3 
 
 
 2 
 
 
 7 da. 
 
 323 
 
 388 
 
 360 
 
 289 
 
 182 
 
 4 
 
 
 2 
 
 
 6 mos. 
 
 492 
 
 530 
 
 547 
 
 607 
 
 663 
 
 5 
 
 
 2 
 
 
 lyr. 
 
 487 
 
 515 
 
 610 
 
 588 
 
 647 
 
 6 
 
 
 2 
 
 
 2yrs. 
 
 533 
 
 586 
 
 612 
 
 659 
 
 684 
 
 7 
 
 R 
 
 
 
 
 7 da. 
 
 562 
 
 608 
 
 726 
 
 709 
 
 432 
 
 8 
 
 
 
 
 
 6 mos. 
 
 745 
 
 751 
 
 799 
 
 804 
 
 795 
 
 9 
 
 
 2 
 
 
 7 da. 
 
 288 
 
 347 
 
 372 
 
 352 
 
 165 
 
 10 
 
 
 2 
 
 
 6 mos. 
 
 532 
 
 538 
 
 624 
 
 638 
 
 642 
 
 11 
 
 
 2 
 
 
 lyr. 
 
 591 
 
 595 
 
 643 
 
 645 
 
 666 
 
 12 
 
 
 2 
 
 
 2yrs. 
 
 590 
 
 623 
 
 680 
 
 673 
 
 666 
 
 13 
 
 X 
 
 
 
 
 7 da. 
 
 351 
 
 368 
 
 405 
 
 450 
 
 204 
 
 14 
 
 
 
 
 
 6 mos. 
 
 560 
 
 606 
 
 580 
 
 645 
 
 797 
 
 15 
 
 
 2 
 
 
 7 da. 
 
 227 
 
 258 
 
 261 
 
 282 
 
 96 
 
 16 
 
 
 2 
 
 
 6 mos. 
 
 494 
 
 54(5 
 
 591 
 
 574 
 
 563 
 
 17 
 
 
 2 
 
 
 lyr. 
 
 572 
 
 580 
 
 586 
 
 583 
 
 652 
 
 18 
 
 
 2 
 
 
 2 yrs. 
 
 592 
 
 575 
 
 592 
 
 592 
 
 667 
 
 19 
 
 S 
 
 2 
 
 176 Fahr. 
 
 5 da. 
 
 296 
 
 307 
 
 362 
 
 391 
 
 422 
 
 20 
 
 R 
 
 2 
 
 140 " 
 
 5 da. 
 
 403 
 
 440 
 
 416 
 
 495 
 
 442 
 
 21 
 
 X 
 
 2 
 
 140 " 
 
 5 da. 
 
 361 
 
 334 
 
 390 
 
 452 
 
 474 
 
 NOTES. Sand, Point aux Pins (river sand) passing No. 10 sieve, except 
 for hot tests, where standard sand was used. Cement and 
 plaster of Paris passed through No. 50 sieve before using. 
 Plaster Paris had no apparent effect on consistency mor- 
 tar at first, but after making first three briquets of batch 
 of five, the mortar containing plaster Paris dried out 
 somewhat. 
 Each result, mean of five briquets. 
 
 Similar tests of natural cement mortars hardening in water 
 are given in Table 91. One of the brands is not much affected 
 
PLASTER PARIS WITH CEMENT 
 
 L>51 
 
 TABLE 91 
 Plaster of Paris in Natural Cement Mortars, Hardening in Water 
 
 
 
 
 
 
 TENSILE STRENGTH, POUNDS PER 
 
 
 CEMENT, 
 
 SAND, 
 
 TEMPER- 
 
 AGE OF 
 
 SQUARE INCH, WITH PEII CENT. 
 OF CEMENT REPLACED BY- 
 
 REF. 
 
 NATURAL 
 BRAND. 
 
 PARTS TO 
 ONE 
 CEMENT. 
 
 ATURE 
 WATER 
 WHERE 
 STORED. 
 
 BRIQUETS 
 WHEN- 
 BROKEN. 
 
 PLASTER OF PARIS. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 1 
 
 2 
 
 3 
 
 6 
 
 
 
 
 Degrees F. 
 
 
 
 
 
 
 
 1 
 
 An 
 
 
 
 00-65 
 
 7 da. 
 
 23:} 
 
 225 
 
 213 
 
 235 
 
 a 
 
 2 
 
 
 
 
 
 mo. 
 
 422 
 
 449 
 
 438 
 
 441 
 
 324 
 
 3 
 
 
 2 
 
 
 7 da. 
 
 111 
 
 109 
 
 97 
 
 144 
 
 a 
 
 4 
 
 
 2 
 
 
 6 mo. 
 
 418 
 
 416 
 
 435 
 
 409 
 
 133c 
 
 5 
 
 
 2 
 
 
 1 yr. 
 
 415 
 
 451 
 
 430 
 
 454 
 
 
 
 
 
 2 
 
 
 2 yrs. 
 
 478 
 
 476 
 
 489 
 
 514 
 
 
 7 
 
 Gi 
 
 
 
 
 7 da. 
 
 146 
 
 156 
 
 115c 
 
 a 
 
 a 
 
 8 
 
 
 
 
 
 6 mo. 
 
 383 
 
 3986 
 
 323 
 
 312e 
 
 234/ 
 
 9 
 
 
 2 
 
 
 7 da. 
 
 62 
 
 80 
 
 94 
 
 a 
 
 a 
 
 10 
 
 
 2 
 
 
 mo. 
 
 374 
 
 312 
 
 355 
 
 86/ 
 
 161/ 
 
 11 
 
 
 2 
 
 
 l yr. 
 
 448 
 
 395 
 
 408 
 
 131/ 
 
 107/ 
 
 12 
 
 
 2 
 
 
 2 yrs. 
 
 456 
 
 437 
 
 397 
 
 172/ 
 
 a 
 
 13 
 
 An 
 
 2 
 
 140 
 
 5 da. 
 
 310 
 
 365 
 
 405 
 
 402 
 
 203 
 
 14 
 
 Gii 
 
 2 
 
 4 . 
 
 . i 
 
 359 
 
 351 
 
 189 
 
 138 
 
 100 
 
 NOTE. Sand, Point aux Pins (river sand) passing No. 10 sieve, ex- 
 cept for hot tests, where standard sand was used. 
 a Found badly swelled and nearly disintegrated after a few 
 
 days in tank. 
 
 6 Surface cracks, 1 inch section swelled to 1 ^ inches. 
 c Surface cracks, 1 inch section swelled to 1 T ] 2 inches. Had 
 
 nearly disintegrated after 2 days. 
 d Surface cracks. 
 e Badly cracked on surface. 
 / Badly cracked on surface, and 1 inch section swelled to 
 
 about ItV inches. 
 
 by additions of one to three per cent., but the other brand is 
 practically ruined by the addition of more than one or two per 
 cent., and both brands are rendered quite unsound by six 
 per cent, plaster. 
 
 378. The briquets reported in the preceding tables were 
 hardened in water, as usual. Table 92 gives some of the results 
 obtained by adding plaster Paris to mortars that are hardened 
 in dry air. The effects on the two samples of the same brand 
 of Portland, one quick setting and one slow setting, are quite 
 different. The strength of the quick setting sample is increased, 
 two per cent, giving the best results, while that of the slow 
 
252 
 
 CEMENT AND CONCRETE 
 
 setting sample is diminished by the addition of plaster. Both 
 brands of natural cement appear to be notably improved by 
 the plaster, the best result being given by three per cent. Such 
 an addition to one brand results in a remarkable increase in 
 strength of 250 per cent. 
 
 TABLE 92 
 
 Plaster of Paris in Cement Mortars, Hardening in Dry Air. Effect 
 on Different Samples, Portland and Natural 
 
 
 
 
 TENSILE STRENGTH POUNDS PER 
 
 
 CEMENT. 
 
 | 
 
 SQUARE INCH, WITH 
 PER CENT. OF CEMENT REPLACED BY 
 
 REF. 
 
 
 || 
 
 PLASTER OF PARIS. 
 
 
 Kind. 
 
 Brand. 
 
 Sample. 
 
 & 
 
 
 
 l 
 
 2 
 
 3 
 
 6 
 
 1 
 
 Port. 
 
 R 
 
 26 R 
 
 6 mo. 
 
 443 
 
 443 
 
 560 
 
 529 
 
 493 
 
 2 
 
 it 
 
 R 
 
 23 R 
 
 u 
 
 559 
 
 483 
 
 419 
 
 436 
 
 337 
 
 3 
 
 Nat. 
 
 An 
 
 L 
 
 (( 
 
 162 
 
 220 
 
 282 
 
 286 
 
 272 
 
 4 
 
 " 
 
 In 
 
 28 S 
 
 u 
 
 76 
 
 110 
 
 151 
 
 269 
 
 240 
 
 NOTES. Sample 26 R, Portland, quick setting, bears ^ inch wire in 18 
 
 minutes. 
 Sample 23 R, Portland, slow setting, bears ^ inch wire in 244 
 
 minutes. 
 
 Sand, two parts Point aux Pins (river sand) to one cement. 
 All briquets stored in air of laboratory until broken. 
 Each result, mean of five briquets. 
 
 For the effect of plaster of Paris on the adhesive strength of 
 mortar, see 407. 
 
 379. Conclusions. It is evident from the above tests that 
 the addition of small amounts of plaster Paris affects different 
 samples of cement in quite different ways, and it is necessary 
 to bear this in mind in the application of general conclusions 
 to special cases. The indications are that the addition to 
 cement of from one to three per cent, of plaster of Paris or 
 sulphate of lime generally hastens the hardening and will not 
 usually result in decreased strength; that some natural cements, 
 however, are sensibly injured by more than one per cent., 
 especially if used neat. The presence of as much as six per 
 cent, plaster of Paris retards the hardening (although hastening 
 the initial set) and is quite apt to ruin either Portland or natural 
 cements. The addition of plaster Paris usually gives better 
 results in air hardened than in water hardened specimens. 
 
CLAY WITH CEMENT 253 
 
 ART. 49. MIXTURES OF CLAY AND OTHER MATERIALS WITH 
 
 CEMENT 
 
 380. EFFECT OF CLAY ON CEMENT MORTAR AND CONCRETE. 
 
 Clay may occur in cement mortar or concrete due to the 
 use of sand or aggregate that is not clean. As the plasticity of 
 cement mortar is increased by the presence of clay, small 
 amounts are sometimes added to produce this effect, and clay 
 is also sometimes used to render mortar stiff enough to with- 
 stand immediate immersion in water. In the case of concrete, 
 the presence of a certain percentage of clay renders it easier to 
 compact the mass by tamping, though if too much clay is pres- 
 ent, the mass becomes sticky. 
 
 A number of tests have been made to determine the behavior 
 of such mixtures of clay and cement. In all of these tests the 
 clay was first dried, pulverized and sifted, and then a weighed 
 quantity equal to a given per cent, of the weight of the cement 
 was added to the latter. In the writer's first tests of this kind 
 small percentages of clay were used, less than ten per cent., but 
 it was found that with lean mortars much larger percentages must 
 be used to determine the point where clay began to be injurious. 
 
 381. Table 93 shows the effect of clay on the time of setting 
 and soundness of neat cement. The effect of small percentages 
 of clay on the time of setting of Portland cement is not very 
 marked, but with natural cement even ten per cent, of clay 
 retards the setting in a marked degree. As to the effect on 
 soundness, Portland cement pats disintegrate with more than 
 twenty-five per cent, of clay added, while the natural cement 
 is affected if more than ten per cent, of clay is present. 
 
 382. Table 94 shows the tensile strength of neat cement 
 mortars to which clay to the amount of 10 to 100 per cent, of 
 the cement has been added. Some of the Portland briquets 
 were immersed as soon as molded, while others were left the 
 customary twenty-four hours in moist air before immersion. 
 
 It is seen that to mix clay with neat Portland cement results 
 in a decided decrease in strength, the results obtained with 
 twenty-five per cent, clay being only about sixty or seventy 
 per cent, of the strength of the mortar without clay. With 
 natural cement the presence of clay seriously retards the hard- 
 ening and results in decreased strength, though it does not 
 
254 
 
 CEMENT AND CONCRETE 
 
 TABLE 93 
 
 Effect of Pulverized Clay on the Time of Setting and Soundness 
 
 of Cement 
 
 
 
 
 TIME TO BEAR j^ INCH WIRE IN MINUTES, 
 
 
 
 
 AND THE CONDITION 
 
 o 
 
 CEMENT. 
 
 CLAY. 
 
 OF PATS AFTER FIVE MONTHS. 
 
 g 
 
 
 
 Clay as Per Cent, of Cement. 
 
 w 
 
 
 
 
 
 Kind. 
 
 Brand. 
 
 Sam- 
 ple. 
 
 Kind. 
 
 
 
 10 
 
 25 
 
 50 
 
 100 
 
 1 
 
 Portland 
 
 X 
 
 41S 
 
 Red 
 
 285 
 
 318 
 
 328 
 
 328 
 
 450 
 
 1 
 
 u 
 
 I 
 
 ' 
 
 t ; 
 
 Good 
 
 Fair 
 
 Good 
 
 Bad a 
 
 Bad a 
 
 2 
 
 i I 
 
 c 
 
 
 
 Blue 
 
 288 
 
 286 
 
 300 
 
 305 
 
 306 
 
 2 
 
 t c 
 
 ' 
 
 c 
 
 " 
 
 Good 
 
 Fair 
 
 Good 
 
 Bad 
 
 Bad a 
 
 3 
 
 Natural 
 
 Gn 
 
 KK 
 
 Red 
 
 69 
 
 123 
 
 195 
 
 345 
 
 445 
 
 3 
 
 " 
 
 ' 
 
 i 
 
 " 
 
 Fair 
 
 Fair 
 
 Bad 
 
 Bad 
 
 Bad a 
 
 4 
 
 " 
 
 
 
 t 
 
 Blue 
 
 98 
 
 173 
 
 215 
 
 350 
 
 415 
 
 4 
 
 " 
 
 i 
 
 c 
 
 H 
 
 Poor - 
 
 Poor 
 
 Bad 
 
 Bad 
 
 Bad a 
 
 NOTE. Results marked a, pats cracked badly in air and were not im- 
 mersed. 
 
 TABLE 94 
 
 Effect of Clay on Tensile Strength; Neat Cement Paste 
 
 
 
 
 * 
 
 
 TENSILE STRENGTH, LBS. PER 
 
 
 
 
 gll 
 
 e 
 
 SQUARE INCH. 
 
 REF. 
 
 CEMENT. 
 
 KIND 
 
 OF 
 
 CLAY. 
 
 Sig 
 
 H H 
 
 Clay Expressed as Per Cent, of 
 Cement. 
 
 
 
 
 * ^ (B 
 
 pq 
 
 
 
 Kind. 
 
 Brand. 
 
 Sample. 
 
 
 |p 
 
 
 
 
 10 
 
 25 
 
 50 
 
 100 
 
 1 
 
 Port. 
 
 X 
 
 41S 
 
 Red 
 
 24 
 
 3 mo. 
 
 658 
 
 535 
 
 474 
 
 336 
 
 253 
 
 2 
 
 " 
 
 ' 
 
 u 
 
 t c 
 
 
 
 " 
 
 660 
 
 .587 
 
 476 
 
 318 
 
 255 
 
 3 
 
 Nat. 
 
 An 
 
 D 
 
 " 
 
 24 
 
 28 da. 
 
 389 
 
 280 
 
 138 
 
 60 
 
 22 
 
 4 
 
 " 
 
 An 
 
 1) 
 
 " 
 
 24 
 
 3 mo. 
 
 376 
 
 365 
 
 323 
 
 219 
 
 176 
 
 have as deleterious an effect as it does with Portland. The mix- 
 ing of clay with neat cement is of course very severe treatment. 
 
 In Table 95 the mortars contain equal parts cement and sand, 
 and the clay is from 50 per cent, to 200 per cent, of the weight 
 of cement. It appears from this table that clay in as large 
 amounts as 50 per cent, of the cement is injurious to one-to- 
 one mortars of either Portland or natural cement. 
 
 383. The mortars in Table 96 are all of Portland, and con- 
 tain three parts sand to one cement. Smaller percentages of 
 
CLAY WITH CEMENT 
 
 TABLE 95 
 
 Effect of Large Amounts of Clay in Mortars Containing Equal 
 Parts Cement and Sand 
 
 
 
 
 
 
 TENSILE STRENGTH, LBS. 
 
 H 
 
 CEMENT. 
 
 
 HOURS 
 
 
 PER SQUARE INCH. 
 
 U 
 
 
 
 KIND 
 
 ELAPSED 
 
 i 
 
 
 
 ;FERE 
 
 
 OF 
 
 CLAY. 
 
 BETWEEN 
 
 MOLDING 
 AND IM- 
 
 w w 
 
 &i 
 
 a 
 
 CLAY EXPRESSED AS PER 
 CENT. OF CEMENT. 
 
 
 
 
 MM 
 
 
 
 
 
 MERSING. 
 
 PQ 
 
 
 
 Kind. 
 
 Brand. 
 
 Sample. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 50 
 
 100 
 
 150 
 
 200 
 
 1 
 
 Port. 
 
 X 
 
 41 S 
 
 Ked 
 
 24 
 
 3 mos. 
 
 747 
 
 512 
 
 337 
 
 239 
 
 193 
 
 2 
 
 t i 
 
 (4 
 
 u 
 
 
 
 
 41 
 
 720 
 
 549 
 
 321 
 
 242 
 
 189 
 
 3 
 
 Nat. 
 
 Gn 
 
 KK 
 
 
 24 
 
 3 mos. 
 
 454 
 
 241 
 
 212 
 
 183 
 
 140 
 
 4 
 
 u 
 
 (i 
 
 " 
 
 
 
 
 " 
 
 413 
 
 231 
 
 206 
 
 157 
 
 128 
 
 5 
 
 u 
 
 An 
 
 1) 
 
 
 24 
 
 " 
 
 442 
 
 259 
 
 194 
 
 167 
 
 152 
 
 (i 
 
 u 
 
 u 
 
 u 
 
 
 
 
 " 
 
 440 
 
 274 
 
 184 
 
 141 
 
 125 
 
 7 
 
 ti 
 
 " 
 
 u 
 
 
 24 
 
 6 mos. 
 
 488 
 
 335 
 
 268 
 
 217 
 
 184 
 
 clay are used, namely, 10 to 40 per cent. The mortars harden- 
 ing in water show a decided improvement due to the presence 
 of clay, but the briquets hardening in the open air indicate that 
 
 TABLE 96 
 
 Effect of Clay in Portland Cement Mortar Containing Three Parts 
 Sand to One Cement 
 
 REFERENCE. 
 
 BRIQUETS STORXD. 
 
 AGE OF 
 BRIQUETS. 
 
 TENSILE STRENGTH, LBS. PER SQUARE 
 INCH. 
 
 CLAY ADDED AS PER CENT. OF CEMENT. 
 
 
 
 10 
 
 '20 
 
 40 
 
 1 
 
 Tank, Laboratory 
 
 6 months.' 
 
 385 
 
 435 
 
 489 
 
 533 
 
 2 
 
 it u 
 
 2 years. 
 
 375 
 
 412 
 
 478 
 
 593 
 
 8 
 
 Open Air 
 
 6 months. 
 
 381 
 
 403 
 
 394 
 
 418 
 
 4 
 
 4 i U 
 
 2 years. 
 
 660 
 
 624 
 
 631 
 
 570 
 
 NOTES. Cement, Portland, Brand R, Sample 83 T. 
 
 Sand, three parts crushed quartz f " to one cement by weight. 
 Clay, red clay dried, pulverized, and passed through No. 100 
 
 sieve. 
 
 Clay added to mortar, amount cement and sand remaining 
 constant. 
 
256 
 
 CEMENT AND CONCRETE 
 
 at two years the mortar without clay is stronger. It may be 
 noted in passing that these results, obtained at two years, with 
 one-to-three mortars hardened in open air, are very high. 
 
 The effect of clay on mortars containing four parts sand to 
 one cement is shown in Table 97. In this case the addition of 
 clay equal to the weight of the cement almost invariably re- 
 sults in increasing the strength of the mortar. Briquets im- 
 mersed as soon as made were especially benefited by the pres- 
 ence of clay, except in one case, the red clay did not appear to 
 increase the ability of the natural cement Gn to withstand 
 early immersion. The red clay appears to give better results 
 than the blue with Portland, while the reverse is true with at 
 least one brand of natural. Whether this difference is a chemi- 
 cal or physical one is not known; the red clay is a good pud- 
 dling clay, while the blue clay is not, but appears to contain 
 some very fine sand. 
 
 TABLE 97 
 
 Effect of Large Amounts of Clay in Cement Mortars Containing 
 Four Parts Sand to One Cement 
 
 
 
 
 
 
 TENSILE STRENGTH, LBS. 
 
 W 
 
 CEMENT. 
 
 
 HOURS 
 
 05 
 
 PER SQUARE INCH. 
 
 
 
 
 KIND 
 
 ELAPSED 
 
 E 
 
 
 
 W 
 K 
 
 
 OF 
 
 CLAY. 
 
 BETWEEN 
 
 MOLDING 
 AND IM- 
 
 o ^ ^ 
 
 S 
 
 CLAY AS PER CEKT. OF 
 
 CEMENT. 
 
 
 
 
 rt 
 
 
 
 
 
 MERSING. 
 
 PQ 
 
 
 
 Kind. 
 
 Brand. 
 
 Sample. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 50 
 
 100 
 
 150 
 
 200 
 
 1 
 
 Port. 
 
 X 
 
 41 S 
 
 Red 
 
 24 
 
 3 mos. 
 
 271 
 
 348 
 
 305 
 
 239 
 
 193 
 
 2 
 
 
 
 
 Blue 
 
 24 
 
 
 227 
 
 304 
 
 250 
 
 179 
 
 145 
 
 3 
 
 
 
 
 Red 
 
 00 
 
 
 156a 
 
 320 
 
 324 
 
 200 
 
 192 
 
 4 
 
 
 
 
 Blue 
 
 00 
 
 
 149a 
 
 270 
 
 215 
 
 148 
 
 114 
 
 5 
 
 Nat. 
 
 Gn 
 
 KK 
 
 Red 
 
 24 
 
 3 mos. 
 
 138 
 
 155 
 
 146 
 
 164 
 
 133 
 
 6 
 
 
 
 
 Blue 
 
 24 
 
 
 118 
 
 167 
 
 200 
 
 167 
 
 134 
 
 7 
 
 
 
 
 Red 
 
 00 
 
 
 83 
 
 87 
 
 39 
 
 86 
 
 72 
 
 8 
 
 
 
 
 Blue 
 
 00 
 
 
 49 
 
 127 
 
 147 
 
 136 
 
 106 
 
 9 
 
 
 
 
 Red 
 
 24 
 
 2>rs. 
 
 194 
 
 348 
 
 306 
 
 256 
 
 190 
 
 10 
 
 
 An 
 
 D 
 
 Red 
 
 24 
 
 3 mos. 
 
 138 
 
 218 
 
 174 
 
 174 
 
 190 
 
 NOTES. Sand, crushed quartz f, ("Standard"), four parts to one 
 
 cement by weight. 
 
 Clay, dried, pulverized and passed through sieve before using. 
 All briquets immersed in tank in laboratory as usual. 
 Each result, mean of five briquets. 
 Results marked "a," briquets disintegrated some on face from 
 
 early immersion. 
 
CLAY WITH CEMENT 
 
 257 
 
 384. Table 98 gives the results of tests by other experimenters, 
 showing the effect of clay on one-to-three mortars of Portland 
 and natural cement. 1 The amount of clay used in these tests 
 appears to be stated as percentage of the total ingredients in- 
 stead of as a percentage of the cement as in the preceding 
 tables. The mortars were mixed quite dry for these experi- 
 ments. The Portland cement mortar seems to be improved by 
 the addition of clay to the amount of twelve per cent, of the 
 mortar. The hardening of natural cement mortar is some- 
 what slower with twelve per cent, clay than with three to six 
 per cent., but at the age of twelve weeks the mortars containing 
 clay were all stronger than that without clay. 
 
 TABLE 98 
 Effect of Clay on the Tensile Strength of One-to-Three Mortars 
 
 
 
 
 TENSILE STRENGTH, POUNDS 
 
 CEMENT. 
 
 PARTS 
 SAND TO 
 
 AOE OF 
 
 BRIQUETS 
 
 PER SQUARE INCH. CLAY EXPRESSED AS 
 PER CENT. OF MORTAR. 
 
 
 CEMENT. 
 
 BROKEN. 
 
 
 
 3 
 
 6 
 
 9 
 
 12 
 
 Portland 
 
 8 
 
 2 weeks 
 
 202 
 
 267 
 
 280 
 
 318 
 
 333 
 
 tl 
 
 3 
 
 4 weeks 
 
 862 
 
 301 
 
 334 
 
 381 
 
 353 
 
 (( 
 
 3 
 
 12 weeks 
 
 451 
 
 506 
 
 521 
 
 522 
 
 5*7 
 
 Natural 
 
 2 
 
 1 week 
 
 68 
 
 117 
 
 101 
 
 99 
 
 65 
 
 tt 
 
 2 
 
 4 weeks 
 
 152 
 
 199 
 
 219 
 
 170 
 
 146 
 
 *' 
 
 2 
 
 12 weeks 
 
 170 
 
 214 
 
 252 
 
 230 
 
 211 
 
 NOTE. Tests by Messrs. J. J. Richey and B. H. Prater. 
 
 385. Conclusions. Always keeping in mind the limitations 
 to be observed in drawing general conclusions from experi- 
 ments having a limited range, it may be said that the indications 
 are as follows: Neat cement and rich mortars are injured by the 
 addition of clay, the rate of hardening and the ultimate strength 
 being diminished. Lean mortars containing three to four parts 
 sand to one cement are usually improved by the addition of 
 clay to the amount of 40 to 100 per cent, of the cement, or 
 10 to 25 per cent, of the combined weight of cement and sand, 
 and the ability of such mortars to withstand early immersion 
 may be greatly enhanced by such additions. It is evident 
 from the above tests that the expense which should be incurred 
 in washing sand to remove a small percentage of clay is limited, 
 
 1 Messrs. J. J. Richey and B. H. Prater, Technograph, 1902-3. 
 
258 
 
 CEMENT AND CONCRETE 
 
 and for certain uses there is no question that mortar may be 
 improved by the addition of clay. 
 
 (For the effect of clay on the compressive strength of con- 
 crete, see Art. 55.) 
 
 386. Powdered Limestone, Brick, etc. Various foreign sub- 
 stances are sometimes used with cement, either in lieu of sand, 
 or to make the mortar more plastic. Such foreign ingredients 
 may also occur in mortar as impurities in the sand used. Pow- 
 dered limestone, slaked lime, powdered brick and clay are some 
 of the materials experimented with in this connection. A few 
 tests of the effects of such mixtures on the setting time of ce- 
 
 TABLE 99 
 
 Foreign Substances in Cement Mortar 
 
 
 CEMENT. 
 
 *% 
 
 
 TENSILE STRENGTH, POUNDS PER 
 SQUARE INCH. 
 
 w 
 
 
 
 
 &% 
 
 AGE OF 
 
 
 
 
 
 
 fe 
 
 u 
 
 
 
 
 
 *> 
 
 fc s 
 
 BRIQUETS 
 WHEN 
 
 Composition of Mortar. 
 
 a 
 K 
 
 H 
 
 Kind. 
 
 Brand. 
 
 Sam- 
 pie. 
 
 
 
 BROKEN. 
 
 
 
 
 
 
 
 
 tf 
 
 
 
 
 fcg 
 
 
 a 
 
 b 
 
 c 
 
 d 
 
 e 
 
 / 
 
 1 
 
 Port. 
 
 R 
 
 JJ 
 
 None 
 
 8 months 
 
 705 
 
 
 674 
 
 583 
 
 615 
 
 667 
 
 2 
 
 
 
 ' 
 
 3.75 
 
 5 days, H 
 
 152 
 
 217 
 
 164 
 
 175 
 
 240 
 
 198 
 
 3 
 
 
 
 ' 
 
 3.75 
 
 3 months 
 
 259 
 
 367 
 
 297 
 
 284 
 
 311 
 
 304 
 
 4 
 
 
 
 ' 
 
 3.75 
 
 1 year 
 
 309 
 
 365 
 
 367 
 
 333 
 
 438 
 
 438 
 
 5 
 
 Nat. 
 
 An 
 
 G 
 
 None 
 
 3 months 
 
 286 
 
 
 203 
 
 307 
 
 154 
 
 203 
 
 6 
 
 
 
 i 
 
 4 
 
 5 days, h 
 
 86 
 
 . . . 
 
 105 
 
 94 
 
 132 
 
 164 
 
 7 
 
 
 
 t 
 
 4 
 
 3 months 
 
 185 
 
 
 214 
 
 157 
 
 239 
 
 215 
 
 8 
 
 
 
 i 
 
 4 
 
 1 year 
 
 210 
 
 :. . 
 
 234 
 
 238 
 
 263 
 
 264 
 
 NOTES. Sand, "Standard." Materials added to mortar were first pul- 
 verized and passed through No. 80 sieve, holes .007 inch 
 square. 
 
 ( 5-day results, H = immersed in hot water, 80 C. 
 I 5-day results, h = immersed in hot water, 60 C. 
 Composition of mortars: 
 
 a No foreign substance. 
 
 b No foreign substance, but additional amount cement added, 
 
 making mortar 1 to 3 instead of Ito 3.75. 
 
 c Kelleys Isd. Limestone, equal to 25 per cent, weight of ce- 
 ment added to mortar. 
 d Slaked lime powder, equal to 25 per cent, weight of cement 
 
 added to mortar. 
 e Red clay, equal to 25 per cent, weight of cement added to 
 
 mortar. 
 
 / Red brick, equal to 25 per cent, weight of cement added 
 to mortar. 
 
FOREIGN SUBSTANCES WITH CEMENT 
 
 259 
 
 ment indicated that the rate of setting of Portland cement was 
 not appreciably affected by the addition of twenty-five per 
 cent, of any of these substances, but the setting time of natural 
 cement appeared to be sensibly hastened by such additions. 
 None of these materials had any appreciable effect on the 
 soundness of either Portland or natural. 
 
 Table 99 shows the effect on the tensile strength of mortar 
 of adding twenty-five per cent, of each of the four substances 
 mentioned. It appears that the strength of neat cement mor- 
 tar, either Portland or natural, is usually diminished by the 
 presence of such materials, but in almost every case mortars 
 containing about four parts sand to one cement are improved 
 by the addition of the substances in question to an amount 
 equal to twenty-five per cent, of the cement. Pulverized clay 
 and brick give the best results, the increased strength amounting 
 to from twenty to forty per cent. 
 
 387. Sawdust. Where a very light and porous mortar is 
 desired for use in floors and similar purposes, the incorporation 
 of sawdust in the mortar is suggested by a similar use in clay 
 building materials. The results in Table 100 show that the 
 use of sufficient sawdust to materially diminish the weight 
 practically ruins the cohesion of the mortar, even ten per cent, 
 of sawdust materially diminishing the strength. 
 
 TABLE 100 
 Sawdust in Cement Mortar 
 
 REFERENCE. 
 
 CEMENT. 
 
 PARTS SAND TO 
 ONE CEMENT. 
 
 BRIQUETS 
 STORED. 
 
 AGE OF 
 BRIQUETS WHEN 
 BROKEN. 
 
 TKNSILK STRENGTH POUNDS PER 
 SQUARE INCH. 
 
 Kind. 
 Port. 
 
 u 
 u 
 (1 
 
 Nat. 
 it 
 
 Brand. 
 
 Sawdust as Per Cent, of Cement. 
 
 
 
 10 
 
 20 
 
 25 
 
 50 
 
 100 
 
 1 
 2 
 
 ;) 
 4 
 5 
 6 
 
 X 
 
 (1 
 (( 
 (1 
 
 An 
 it 
 
 
 
 2 
 2 
 
 2 
 
 Tank 
 Dry air 
 Tank 
 Dry air 
 Tank 
 Tank 
 
 lyr. 
 
 It 
 (( 
 14 
 
 [( 
 
 11 
 
 799 
 074 
 502 
 452 
 433 
 313 
 
 409 
 492 
 
 
 169 
 103 
 
 44 
 28 
 32 
 14 
 
 38 
 58 
 
 31 
 a 
 32 
 
 a 
 b 
 
 20 
 
 253 
 
 129 
 108 
 
 104 
 
 NOTES. Sand, crushed quartz, f $. Sawdust from white pine, passed 
 
 through sieve with one-quarter inch meshes, 
 a Briquets broken in applying initial strain. 
 b Briquets disintegrated in tank. 
 
260 
 
 CEMENT AND CONCRETE 
 
 388. Use of Ground Terra Cotta as Sand. A light weight 
 mortar may also be made by using as sand or aggregate, ma- 
 terials of burned clay, such as brick or terra cotta. The tests 
 in Table 101 were made to determine the value of ground terra 
 cotta for use in place of sand, and it appears that this material 
 gives excellent results. The strength given with one of the 
 brands of natural cement is especially high. 
 
 TABLE 101 
 Use of Ground Terra Cotta as Sand in Cement Mortar 
 
 a 
 
 
 
 TENSILE STRENGTH, LBS. PER SQ. IN. 
 
 <J 
 
 
 
 
 w 
 
 
 AGE OF 
 
 Parts Ground Terra Cotta to One Cement. 
 
 K 
 h 
 
 
 BRIQUETS. 
 
 by Weight. 
 
 
 Kind. 
 
 Brand. 
 
 
 1 
 
 2 
 
 3 
 
 4 
 
 6 
 
 1 
 
 Portland 
 
 X 
 
 3 months 
 
 523 
 
 406 
 
 332 
 
 257 
 
 174 
 
 2 
 
 u 
 
 " 
 
 1 year 
 
 604 
 
 518 
 
 429 
 
 337 
 
 266 
 
 3 
 
 Natural 
 
 An 
 
 3 months 
 
 284 
 
 338 
 
 346 
 
 347 
 
 224 
 
 4 
 
 ' 
 
 (C 
 
 1 year 
 
 262 
 
 360 
 
 351 
 
 361 
 
 186 
 
 5 
 
 u 
 
 En 
 
 3 months 
 
 291 
 
 303 
 
 184 
 
 136 
 
 
 6 
 
 u 
 
 1 1 
 
 1 year 
 
 340 
 
 434 
 
 284 
 
 161 
 
 
 NOTES. Terra Cotta tile, of medium burn, ground and passed through 
 No. 20 sieve, and used in place of sand. 
 
 ART. 50. THE USE OF CEMENT MORTARS IN FREEZING 
 WEATHER 
 
 389. It is frequently desirable to use cement in freezing 
 weather, but to ensure good work under these circumstances it 
 is necessary to take certain precautions. If mortar is frozen 
 immediately after mixing, setting cannot take place until it 
 has again thawed. In the practical use of cement it is always 
 gaged with a larger quantity of water than is required for the 
 chemical combination, and if this excess water is frozen after 
 the setting is somewhat progressed, the consequent expansion 
 may be sufficient to disrupt the partially set mortar. By warm- 
 ing the materials or by lowering the freezing point of the water 
 by the addition of salt, glycerine, or some other substance hav- 
 ing this effect, it is sought to prevent the mortar freezing until 
 the work is protected by another layer of mortar, or otherwise, 
 and thus to avoid the expansion. Salt is generally used much 
 
EXPOSURE TO FROST 261 
 
 too sparingly to prevent freezing. The freezing point of water 
 is lowered about one and a half degrees Fahr. for each per cent, 
 of common salt added; thus a twenty per cent, solution would 
 freeze at about two degrees Fahr. 
 
 390. The following tests are selected as showing typical re- 
 sults of a large number of experiments made under the author's 
 direction to determine the effect of exposing cement mortars to 
 frost, and to indicate what treatment will alleviate the delete- 
 rious effects of low temperature. In making tests with small 
 specimens, it is difficult to approach the conditions existing in 
 the actual use of mortars in freezing weather. A small mass 
 of mortar exposed to the air on all sides sets more quickly than 
 the interior of a large mass; and on the other hand, the effect 
 of frost on a small specimen must be more severe and more 
 quickly apparent. Many of the results are more or less contra- 
 dictory, and the conclusions that have been drawn are such as 
 appear to be indicated by the majority of the tests. The 
 treatment of the briquets, and the conditions existing, are given 
 in some detail, that the limits of applicability of such conclu- 
 sions may be seen. 
 
 391. Exposure to Frost of Mortars Already Set. In the 
 tests recorded in Tables 102 to 104 the briquets were allowed to 
 remain one or two days in the laboratory. It is evident that 
 these results are of but limited practical importance, since it 
 is seldom that mortars which are made in winter can be allowed 
 to set in a warm place before exposure; they are given, how- 
 ever, for what they are worth. Tables 102 and 103 give the 
 results obtained with Portland cement briquets exposed to a 
 severe temperature twenty-four to forty-eight hours after made. 
 The most important deduction, and the one most clearly indi- 
 cated by these tables, is that Portland cement mortar made with 
 fresh water may be subjected to very low temperatures twenty- 
 four to forty-eight hours after molded, without seriously de- 
 creasing the tensile strength given at six months to two years. 
 It also appears that solutions containing as much as fifteen 
 per cent, salt are deleterious, and smaller percentages are not 
 advantageous under these conditions. 
 
 Table 104, giving the results of similar tests with natural 
 cement mortar, indicates that this brand gives good results if 
 allowed to set in warm air before exposure to frost. Solutions 
 
262 
 
 CEMENT AND CONCRETE 
 
 TABLE 102 
 
 Exposure of Portland Cement Mortars to Low Temperatures after 
 Twenty-four Hours in Laboratory 
 
 
 
 AGE 
 
 TENSILE STRENGTH, POUNDS PER SQ. IN. 
 
 SAND, KIND. 
 
 DATE 
 
 WHEN 
 
 
 
 MADE. 
 
 BROKEN. 
 
 a 
 
 b 
 
 c 
 
 d 
 
 e 
 
 / 
 
 g 
 
 Standard . . . 
 
 1-15 
 
 6 mo. 
 
 772 
 
 960 
 
 816 
 
 
 
 524 
 
 507 
 
 Standard . . . 
 
 1-15 
 
 21 mo. 
 
 796 
 
 882 
 
 766 
 
 
 
 443 
 
 642 
 
 Pt. aux Pins, ) 
 
 1-18 
 
 6 mo. 
 
 651 
 
 630 
 
 
 769 
 
 463 
 
 443 
 
 
 pass, sieve f 10 } 
 
 1-18 
 
 21 mo. 
 
 760 
 
 780 
 
 
 
 711 
 
 543 
 
 447 
 
 
 NOTES. Cement, Portland, Brand R. One part sand to one cement. 
 
 Briquets made in laboratory, temp., 64 to 66 Fahr. ; materials 
 
 about 65 Fahr. 
 
 Temperature, open air, Jan. 16 to Jan. 19, 4 to 15 Fahr. 
 Treatment of briquets: 
 
 a Fresh water, briquets stored in water in laboratory. 
 b Fresh water, briquets stored in open air after 24 hours. 
 c Fresh water, briquets alternated, two days in open air 
 and then two days in air laboratory, for fifty-two 
 days, then left in open air. 
 
 d Water 5 per cent, salt; briquets stored in open air. 
 e Water 15 per cent salt; briquets stored in open air. 
 / Water 25 per cent, salt; briquets stored in open air. 
 g Water 25 per cent, salt; briquets stored in water in lab. 
 
 TABLE 103 
 
 Exposure of Portland Cement Mortars to Low Temperatures, 
 Twenty-four to Forty-eight Hours after Made 
 
 SAND, 
 KIND. 
 
 DATE 
 BRIQUET 
 
 MADE. 
 
 AGE 
 WHEN 
 BROKEN. 
 
 TENSILE STRENGTH, POUNDS PER SQ. IN. 
 
 a 
 
 6 
 
 c 
 
 d 
 
 e 
 
 / 
 
 g 
 
 h 
 
 i 
 
 j 
 
 Std. . 
 Std. . 
 
 p. p. 
 p. p. 
 
 1-16 
 1-16 
 1-19 
 1-19 
 
 6 mo. 
 21 mo. 
 6 mo. 
 
 415 
 
 602 
 
 372 
 372 
 
 401 
 438 
 
 262 
 
 384 
 
 202 
 326 
 
 
 
 
 
 
 
 
 
 
 
 381 
 638 
 
 394 
 430 
 
 360 
 418 
 
 371 
 375 
 
 233 
 344 
 
 21 mo. 
 
 
 
 
 
 
 
 
 
 
 
 
 NOTES. Cement, Portland, Brand R. Three parts sand to one cement. 
 
 Briquets made in laboratory. Temp . : Air and materials, 64 to 
 67 Fahr. Open air, Jan. 10 to 20, -15 to +18 Fahr. 
 
 Treatment of briquets: a, b, c, d and e mixed with water con- 
 taining 0, 10, 15, 20 and 25 per cent, salt, respectively; 
 a to d, inclusive, air laboratory 24 hours, water laboratory 
 16 hours, air laboratory 12 hours, then in open air. 
 
 /, g, h, i and /, mixed with water containing 0, 10, 15, 20 and 
 25 per cent, salt, respectively. 
 
 e to /, inclusive, put in open air after about 24 hours in 
 air of laboratory. 
 
EXPOSURE TO FROST 
 
 263 
 
 TABLE 104 
 
 Exposure of Natural Cement Mortars to Low Temperatures, 
 Twenty-four Hours after Made 
 
 PARTS 
 
 SAND TO 
 ONE 
 
 DATE MADE. 
 
 AGE WHEN 
 BROKEN. 
 
 TEN 
 
 SILE STRENGTH, POUNDS PER SQ. IN. 
 
 
 
 
 
 CEMENT. 
 
 
 
 a 
 
 6 | c 
 
 d 
 
 e 
 
 / 
 
 2 
 
 1-20 
 
 H ino. 
 
 297 
 
 404 
 
 319 
 
 
 297 
 
 170 
 
 2 
 
 1-20 
 
 lyr. 
 
 305 
 
 390 
 
 343 
 
 
 273 
 
 217 
 
 4 
 
 1-20 
 
 6 mo. 
 
 222 
 
 318 
 
 319 
 
 344 
 
 
 114 
 
 4 
 
 1-20 
 
 lyr. 
 
 223 
 
 259 
 
 339 
 
 205 
 
 . . . 
 
 150 
 
 NOTES. Cement, Natural, Brand On. Sand, " Pt. aux. Pins" (river 
 sand). 
 
 Temp, materials and air of laboratory where briquets were 
 molded, 65 to 68 Fahr. Temp, open air Jan. 21 to 23, 
 - 1 to + 29. 
 
 Treatment of briquets: a, briquets stored in water in labora- 
 tory, b to /, inclusive, briquets stored in open air after 
 twenty-four hours in air of laboratory. 
 
 a and b, fresh water used for gaging mortar. 
 
 c, d, e and /, water used in gaging had 5, 10, 15 and 25 per 
 cent, salt, respectively. 
 
 containing more than ten per cent, salt are deleterious for such 
 treatment. Briquets of another brand of natural cement, a 
 one-to-one mortar of which gave about 450 pounds tensile 
 strength at one year, failed entirely when placed, one hour 
 after made, in open air for three days, and then immersed in a 
 tank in the laboratory. A 7.4 per cent, solution of salt used 
 for gaging assisted very materially in preserving the mortar 
 under the same severe treatment, although this amount of salt 
 was not sufficient to lower the freezing point of the water below 
 the temperature to which the briquets were subjected. 
 
 392. Effect of Salt in Mortars Hardened in Water and Air. - 
 In the tests recorded in Table 105 the materials used were at "a 
 temperature of forty degrees Fahr., and the briquets were 
 molded in an open warehouse where the temperature was usually 
 below twenty-three degrees Fahr., though for a few of the 
 tests the temperature of the air at time of molding was as high 
 as twenty-seven degrees. The temperature of the mortar 
 when briquets were finished was usually but little above thirty- 
 two degrees Fahr. The briquets were left in a warehouse for 
 three days, when part of them were immersed in cold water 
 
2G4 
 
 CEMENT AND CONCRETE 
 
 Water Used 
 g. 
 
 er Cent. Salt 
 G 
 
 . SO .O .00 .CD .O .O . <M . 
 
 Ol-HCil 1-^T-lTtH 
 
 rt* 't- '-* '^ '"* '^ -^ ' 
 
 . OS .00 .O . >O . t- .CO 
 
 aoco 
 TJH 't- 
 
 . >O . t- .CO .OS . 
 
 iocoi^co 
 'O 'O ' -^ 'CD ' 
 
 i-O . 
 (M 
 O ' 
 
 . G<J . t- .O 
 
 t-QOT-H 
 ' O ' ** 'CO 
 
 'CD ' I>- 
 
 ut> 
 
 CO . O . 
 i^ GO 
 
 O 'O ' 
 
 . CO . CO . CO 
 'CD 'iO 'iO 
 
 .CO . <M .00 .OS 
 >-O O C<l TH 
 
 ' TjH ' O 'O 'TH 
 
 lt in Water Use 
 Gaging. 
 
 CO . Tt* . i 1 
 C<li lOS 
 
 ' 
 
 .CO . <N . rH .CO 
 
 OO'OOCO 
 
 'CO * ' Tj< ' rtl 
 
 N3soug N 
 
 aoy 
 
 0<> - CO - ^o - HN - iT- 3^- r4N - 
 
 H " rH "* 
 
 CSCO ^ ^ ^CO^ " "CO" ^ ^ rH^ 
 
 'd 
 
 ^t^cot^ooocoooocoooco 
 cococococo-^co^cocococo 
 <s i i i i i i i i i i i i 
 
 !7<I'N'N(?IC<ICD(MCD-^C<I-^G<I 
 
 cocococococococococococo 
 .ost--osr-oooooooooot^cot- 
 
 .^ T Tt T T r T T T'T"T T r T i Hr i H 'T 
 
 COCOCOCDOOt^OOt^OS OS 
 
 S^rHC^rHr IrHrHi IT I rH 
 
 ^-,-.-*^p^^^.^,Ox.^^0H^^ , 05 
 ^ ^3 ^^***,,^N^ ^w > ^ j /^i ^ . ^ 
 
 i 
 
 <o 
 
 aav K 
 
 ATURE 
 FAH 
 
 -ug 
 
 3ON3H3JI3'}! 
 
EXPOSURE TO FROST 
 
 265 
 
 (under ice), and the remainder stored in open air on a shelf 
 covered by a rough board roof, but with front left open to the 
 weather. All mortars contained two parts river sand to one of 
 cement by weight. The water used in gaging varied from 
 Iresh to a twenty-five per cent, solution. 
 
 The results indicate that Portland mortars made in low 
 temperatures, to be immersed in cold water, are improved by 
 fifteen to twenty per cent, salt in the water of gaging, but that 
 more than five per cent, salt is deleterious for mortars exposed 
 to the air only. The very high results given by the air-hardened 
 specimens are worthy of notice. 
 
 A similar series of tests of natural cement gave results from 
 which no definite general conclusions could be drawn. The 
 effect of freezing and of the use of salt varied greatly for dif- 
 ferent samples. For any given sample the treatment, as re- 
 gards the use of salt, giving good results in open air, was usually 
 the reverse of that giving good results in cold water. The 
 conclusions indicated for rich mortars were sometimes the re- 
 verse of those shown by lean mortars. 
 
 393. The results obtained with five brands of natural ce- 
 
 TABLE 106 
 Effect of Low Temperatures on Five Brands of Natural Cement 
 
 
 
 
 
 M * 
 
 g 
 
 a' 
 
 K U 
 
 
 H 
 
 O 
 
 w 
 
 Q 
 
 BJ 
 
 g 
 
 1 | 
 
 03 
 
 U 
 
 w 
 
 U "if. 
 
 MEAN TENSILE STRENGTH, 
 
 H 
 U 
 
 U 
 H 
 
 K 
 
 
 
 03 
 
 E 
 
 <* 
 
 Ps 
 
 Jl 
 
 U 
 
 g 
 
 2&I 
 
 BRAND. 
 
 
 
 
 
 
 tf 
 
 q 
 
 
 PH 
 
 u sc 
 H 
 
 K M 
 U 
 
 1 
 
 S 1 
 
 Gn. 
 
 An. 
 
 Kn. 
 
 Hn. 
 
 Jn. 
 
 
 a 
 
 & 
 
 c 
 
 rf 
 
 e 
 
 / 
 
 g 
 
 h 
 
 i 
 
 J 
 
 k 
 
 
 
 
 Mo. Da. 
 
 
 
 Deg. 
 
 
 
 Days 
 
 
 
 
 
 
 1 
 
 2 20 
 
 N 
 
 1 
 
 9-11 
 
 18 
 
 Canal 
 
 
 234 
 
 322 
 
 285 
 
 423 
 
 284 
 
 2 
 
 2 22 
 
 N 
 
 1 
 
 16-19 
 
 
 
 u 
 
 . 
 
 201 
 
 327 
 
 326 
 
 302 
 
 211) 
 
 3 
 
 2 20 
 
 S 
 
 1 
 
 9-11 
 
 18 
 
 Open air 
 
 o' 
 
 344 
 
 416 
 
 412 
 
 321 
 
 292 
 
 4 
 
 2 22 
 
 S 
 
 1 
 
 16-19 
 
 
 
 " 
 
 
 
 367 
 
 305 
 
 480 
 
 360 
 
 311 
 
 5 
 
 2 20 
 
 S 
 
 1 
 
 9-11 
 
 18 
 
 ti 
 
 7 
 
 274 
 
 306 
 
 413 
 
 244 
 
 304 
 
 6 
 
 2 22 
 
 S 
 
 1 
 
 16-19 
 
 
 
 u 
 
 7 
 
 292 
 
 338 
 
 426 
 
 311 
 
 304 
 
 7 
 
 2 21 
 
 N 
 
 2 
 
 7-14 
 
 19 
 
 Canal 
 
 . 
 
 161 
 
 318 
 
 329 
 
 348 
 
 238 
 
 8 
 
 2 23 
 
 N 
 
 2 
 
 9-9 
 
 
 
 it 
 
 
 160 
 
 217 
 
 355 
 
 2-38 
 
 186 
 
 9 
 
 2 21 
 
 S 
 
 2 
 
 9-14 
 
 19 
 
 Open air 
 
 'o 
 
 288 
 
 289 
 
 382 
 
 282 
 
 319 
 
 10 
 
 2 23 
 
 S 
 
 2 
 
 9-9 
 
 
 
 " 
 
 
 
 338 
 
 275 
 
 422 
 
 423 
 
 367 
 
 11 
 
 2 21 
 
 S 
 
 2 
 
 9-14 
 
 19 
 
 u 
 
 3 
 
 268 
 
 271 
 
 340 
 
 240 
 
 295 
 
 12 
 
 2 23 
 
 S 
 
 2 
 
 9-9 
 
 
 
 l( 
 
 7 
 
 317 
 
 333 
 
 414 
 
 345 
 
 356 
 
 NOTE. All briquets broken when six and a half months old. 
 
266 
 
 CEMENT AND CONCRETE 
 
 ment are given in Table 106. The briquets were made in a 
 temperature of nine to nineteen degrees Fahr. Half of the 
 briquets were made with fresh water, and half with water con- 
 taining enough salt to lower its freezing point below that of the 
 
 TABLE 107 
 
 Portland Cement Mortar in Low Temperatures 
 Effect of Heating Materials 
 
 H 
 o 
 
 !fi 
 
 H PS 
 K 
 D < ^ 
 &( W Q 
 
 l| 
 
 
 M 
 
 Pi 
 Qp 
 
 QUETS 
 )KEN. 
 
 TENSILE STRENGTH, POUNDS 
 PER SQUARE INCH. 
 
 55 
 
 s 
 
 P5 
 K 
 
 P 
 
 S fe a 3 
 
 g3 
 
 S w M o 
 
 
 WHERE 
 STORED. 
 
 & 
 
 8 
 
 (3 K 
 
 P3 
 ft, Z 
 
 Cold 
 Ma- 
 
 Warm 
 Ma- 
 
 Cold 
 Ma- 
 
 Warm 
 Ma- 
 
 H 
 
 H H 
 
 P* 
 
 g PS 5 
 
 O^j 1 ^ 
 
 ^ 
 
 
 o 
 
 o a 
 
 terials, 
 
 terials, 
 
 terials, 
 
 terials, 
 
 ri 
 
 
 
 fH^ 
 
 1 Q 
 
 W 
 
 OH 
 
 
 PQ 
 
 M X 
 
 !* 
 
 40. 
 
 110. 
 
 40. 
 
 110. 
 
 
 
 
 
 
 
 Mo. 
 
 
 
 
 
 1 
 
 1 
 
 1-5 
 
 23 
 
 Canal 
 
 Wet 
 
 6 
 
 . 
 
 
 582 
 
 598 
 
 2 
 
 1 
 
 8-9 
 
 o 
 
 
 t 
 
 (( 
 
 590 
 
 593 
 
 
 
 3 
 
 1 
 
 1-5 
 
 23 
 
 
 
 181 
 
 
 
 711 
 
 734 
 
 4 
 
 1 
 
 8-9 
 
 
 
 
 
 a 
 
 770 
 
 737 
 
 
 
 5 
 
 2 
 
 14-16 
 
 14 
 
 
 
 51 
 
 
 
 542 
 
 550 
 
 6 
 
 2 
 
 23-24 
 
 
 
 
 t 
 
 u 
 
 460' 
 
 476 
 
 
 
 7 
 
 2 
 
 14-16 
 
 14 
 
 
 i 
 
 181 
 
 
 
 549 
 
 597 
 
 8 
 
 2 
 
 23-24 
 
 
 
 
 t 
 
 
 
 467' 
 
 541 
 
 
 
 Means 
 
 
 
 
 
 
 
 70 
 
 587 
 
 596 
 
 620 
 
 9 
 
 1 
 
 4-6 
 
 23 
 
 Open air 
 
 Dry 
 
 6i 
 
 
 
 469 
 
 450 
 
 10 
 
 1 
 
 9-10 
 
 
 
 
 " 
 
 
 711 
 
 724' 
 
 
 
 11 
 
 1 
 
 4-6 
 
 23 
 
 
 Wet 
 
 
 
 
 487 
 
 470 
 
 12 
 
 1 
 
 9-10 
 
 
 
 
 (t 
 
 
 628' 
 
 614 
 
 
 
 13 
 
 2 
 
 15-18 
 
 14 
 
 
 Dry 
 
 
 
 
 507 
 
 '542' 
 
 14 
 
 2 
 
 24-25 
 
 
 
 
 u 
 
 
 673 
 
 657 
 
 
 
 15 
 
 2 
 
 15-18 
 
 14 
 
 
 Wet 
 
 
 
 
 422 
 
 453 
 
 16 
 
 2 
 
 24-25 
 
 
 
 
 " 
 
 
 543 
 
 495 
 
 
 
 Means 
 
 
 
 
 
 
 
 639 
 
 622 
 
 471 
 
 479 
 
 Grand 
 
 Means 
 
 
 
 
 
 
 605 
 
 605 
 
 534 
 
 549 
 
 
 
 
 
 
 NOTES. Cement, Portland. Sand, " Point aux Pins." 
 
 When warm materials used, the temperature mortar after briquets 
 finished, 63 to 71 Fahr. 
 
 When cold materials used, the temperature mortar after briquets 
 finished, 32 to 39 Fahr. 
 
 When salt water used for mixing, water was 23 per cent, salt for 
 1 to 1 mortars and 14 per cent, salt for 1 to 2 mortars. 
 
 Briquets stored in canal were left in cold air three days before 
 immersion. 
 
 Part of briquets stored in open air were immersed in tank in labo- 
 ratory one week just before breaking, while others were broken 
 dry as indicated. 
 
 In general, each result is mean of five briquets. 
 
EXPOSURE TO FROST 
 
 267 
 
 air where the briquets were made. The results are chiefly of 
 interest as showing the strength that may be attained by natural 
 cement mortars under these severe conditions. 
 
 Higher results are usually given by the air-hardened speci- 
 mens than by those immersed in cold water, though this de- 
 pends somewhat upon the brand. Salt is usually beneficial if 
 the briquets are immersed, and detrimental for open air ex- 
 posure. 
 
 394. Effect of Heating the Materials. The tests in Table 
 107 were made to determine the effect of heating the materials 
 when working in low temperatures, and thus delaying for a 
 time the freezing of the mortars. The details of the tests are 
 fully given in the table. The conclusion indicated is that the 
 ingredients may be used cold or warm indifferently. A gain of 
 only four per cent, is indicated for warm materials in mortars 
 mixed with salt water and hardened in cold fresh water. In 
 practical work, however, the use of warm materials may so delay 
 the freezing as to permit thorough tamping before the mortar 
 freezes. Table 108 gives similar results with one brand of natural 
 cement, from which it appears that warm materials have a slight 
 advantage for either cold water or cold air hardening. 
 
 TABLE 108 
 
 Natural Cement Mortars in Freezing Weather 
 Effect of Heating Materials 
 
 
 2^ 
 
 
 * 
 
 TENSILE STRENGTH, POUNDS PER SQ. IN. 
 
 a 
 
 a* 
 
 
 Is 
 
 
 fc 
 H 
 H 
 
 Sa 
 
 QQB 
 
 TURE AlR 
 
 WHERE 
 
 ~o 
 
 fflS 
 
 STORED IN CANAL. 
 
 STORED IN OPEN AIR. 
 
 
 00 
 
 BRIQUETS 
 
 fc Z 
 
 
 
 3 
 
 H H 
 
 MOLDED 
 
 o* 
 
 
 
 
 
 
 
 S| 
 
 
 BW 
 
 Materials. 
 
 Materials. 
 
 Materials. 
 
 Materials. 
 
 
 ftn^ 
 
 
 $* 
 
 32 F. 
 
 100 F. 
 
 32 F. 
 
 100 F. 
 
 
 
 Deg. Fahr. 
 
 
 
 
 
 
 1 
 
 3 
 
 15 to 16 
 
 6 mos. 
 
 140 
 
 151 
 
 311 
 
 372 
 
 2 
 
 3 
 
 15 to 19 
 
 9 " 
 
 175 
 
 203 
 
 . . 
 
 . . 
 
 3 
 
 2 
 
 22 to 24 
 
 9 " 
 
 167 
 
 204 
 
 355 
 
 361 
 
 NOTES. Cement, Brand Gn, Natural. All mortars made with fresh water. 
 Briquets made with warm materials were frozen in from 15 to 24 
 
 minutes after made. 
 Each result, mean of ten briquets. 
 
268 
 
 CEMENT AND CONCRETE 
 
 395. Consistency of Mortars to Withstand Frost. Since the 
 injury due to frost is caused by the expansion of the water 
 used in gaging, it would be expected that mortars mixed wet 
 would suffer most. This conclusion is confirmed by the tests 
 in Table 109. The superiority of dry mortars is especially 
 shown in mortars that harden in the air. The treatment to 
 which these briquets were subjected was very severe, yet the 
 results are excellent. 
 
 TABLE 109 
 
 Consistency of Mortars as Affecting Ability to Withstand Low 
 
 Temperatures 
 
 AGE OF 
 BKIQUETS 
 WHEN 
 BROKEN. 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH. 
 
 STORED IN CANAL. 
 
 STORED IN OPEN AIR. 
 
 a 
 
 b 
 
 c 
 
 d 
 
 e 
 
 / 
 
 !l 
 
 h 
 
 6 mos. 
 
 414 
 
 414 
 
 372 
 
 501 
 
 601 
 
 571 
 
 521 
 
 674 
 
 mos. 
 
 474 
 
 4(58 
 
 431 
 
 527 
 
 727 
 
 022 
 
 525 
 
 563 
 
 NOTES. Cement, Portland, Brand R;sand, " Point aux Pins," passing holes 
 .08 inch sq. Two parts sand to one cement by weight. Each 
 "* result, mean of five briquets. 
 Temperature, air where briquets were molded, 13 to 14 Fahr.; 
 
 materials used, 40 Fahr. 
 
 Temperature mortar when molding completed 32 to 36 Fahr. 
 Briquets made with fresh water had frozen after 30 minutes. 
 Treatment briquets: a to d, stored in canal (under ice). 
 
 e to hj stored in open air, January, North- 
 ern Michigan. 
 
 Water used: a and e, 10.4 per cent, fresh water. 
 b and /, 11.9 per cent, fresh water. 
 c and g, 13.3 per cent, fresh water. 
 d and h, 11.9 per cent, water containing 15 per 
 cent. salt. 
 
 396. Fineness of Sand and Effect of Frost. The briquets 
 reported in Table 110 were made from mortar containing one 
 and two parts limestone screenings to one cement, the screen- 
 ings varying from coarse to fine. In general, the results follow 
 the rule applicable to mortars used in ordinary temperatures, 
 namely, that the coarse sands give the best results; but it 
 appears that the briquets made with fresh water and exposed 
 
EXPOSURE TO FROST 
 
 260 
 
 o ^ 
 
 o 75 
 .1? & 
 
 a a 
 
 wj 
 
 0) 
 
 Q 
 
 s a H w 
 
 2 OCQ < 
 
 2 2 
 
 _z jr ^: 
 OS .00 .OS .00 
 
 a S ? c c 
 ^ ^ S ^ op ^ 
 
 2 
 p 
 
 5 
 
 |> 
 
 Jl 
 
 II 
 
 OKIOVQ 
 Nl aaSQ H3XV^ NI 
 
 xivg -XN33 uaj 
 
 UHVJ -o 
 aavj^ bxa 
 
 3H3HM 
 
 aHaxvaadwaj, 
 
 xNiawaj 
 
 XHOiaAi 
 
 AS XN3W3Q 3NQ Oi 
 80NIN33HOg 
 
 anoxeawi^ sxavj 
 
 o : ^ ^ c o2 
 
 GO ^ O -* -M 
 
 ~ I J """^ rH J . ^^ 5^1-H 
 
 " "* (M" co" o?o 
 
 H H >s 
 
 .5 ^ c 
 
 O> 0) O 
 
 o I -2 S 
 
 ^ 2^ 
 
 '8 
 
 S!!! 
 
270 CEMENT AND CONCRETE 
 
 in open air reverse this rule, either the finest sand, fg, or the 
 f g giving the best result. 
 
 397. CONCLUSIONS. The following conclusions concerning 
 the use of cement mortars in freezing weather appear to be indi- 
 cated by the foregoing tests: 
 
 1st, Mortars should not be mixed wet for use in low tem- 
 peratures. 
 
 2d, Portland cement mortars made in cold weather usually 
 develop a good tensile strength, especially when exposed to the 
 open air. 
 
 3d, Portland cement mortars for open air exposure may be 
 benefited by the use of from three to seven per cent, salt in 
 the water used in gaging, and from ten to twenty per cent, 
 salt in the gaging water may prove beneficial for mortars hard- 
 ening in cold water. 
 
 4th, Warming the materials for Portland cement mortar 
 appears to have but little effect on its frost resisting qualities. 
 
 5th, Coarse sand usually gives the best results in Portland 
 mortars made in cold weather, but fresh water briquets ex- 
 posed in open air appear to give better results with fine sand. 
 
 6th, Some natural cements give fairly good results in freez- 
 ing weather, while others are practically destroyed by severe 
 exposure. The effect of variations in treatment on different 
 brands of natural cement is so varied that no general conclu- 
 sions can be drawn from the above tests, but the indications 
 are that salt water for gaging is beneficial if the mortar hardens 
 in cold water, but detrimental for mortars exposed to the open 
 air. 
 
 ART. 51. THE ADHESION OF CEMENTS 
 
 398. THE ADHESION BETWEEN PORTLAND AND NATURAL 
 CEMENTS. The question sometimes arises as to whether Port- 
 land cement will adhere to natural cement already set, and 
 whether fresh natural and Portland cement mortars may be 
 used together, as in the case of a Portland facing mortar used 
 with natural cement concrete. Tests bearing on these points 
 are given in Tables 111, 112 and 113. 
 
 In the tests in Table 111 fresh Portland cement mortar was 
 applied to natural cement mortar that had set seven days. 
 Natural cement briquets, made neat and with one to four 
 parts sand, as seen in the headings of the columns of the table, 
 
ADHESION PORTLAND AND NATURAL 
 
 271 
 
 were broken at the age of seven days. The fresh Portland 
 mortar was applied to the half briquets on the same day that 
 the latter were broken, by placing the half briquet in one end 
 of the mold, and filling the other half of the mold with fresh 
 Portland mortar of the composition shown in the second column. 
 
 TABLE 111 
 
 The Adhesion of Portland Cement to Hardened Natural Cement 
 
 Mortar 
 
 REF. 
 
 PARTS SAND TO 
 ONE PART 
 PORTLAND 
 CKMENT IN 
 FRESH MORTAR. 
 
 ADHESION OF PORTLAND MORTAR TO HALF BRIQUETS OF 
 HARDENED NATURAL CEMENT CONTAINING PARTS SAND. 
 
 
 
 1 
 
 2 
 
 3 
 
 4 
 
 1 
 2 
 3 
 
 
 1 
 
 2 
 
 246 
 185 
 63 
 
 2-35 
 210 
 75 
 
 197 
 186 
 00 
 
 11)4 
 152 
 84 
 
 161 
 120 
 85 
 
 NOTES. Briquets of natural cement, containing parts sand indicated at top 
 of columns, were broken at seven days. The half briquets 
 were then placed in one end of briquet mold and the other 
 end of mold was filled with fresh Portland mortar. 
 
 Fresh mortar made of Portland cement, Brand R. 
 
 Sand, " Point aux Pins," passing No. 10 sieve. 
 
 In general, each result is mean of ten bnquets. 
 
 Nearly all briquets broke at juncture of Portland and natural 
 mortars. 
 
 It is seen that the neat Portland gave the highest results in 
 adheskm, the one-to-one mortar giving a comparatively low ad- 
 hesive strength. It is also seen that the neat and one-to-one- 
 mortars adhered best to the richer natural cement briquets, 
 but the one-to-two Portland gave the greatest adhesi ve strength 
 with the poorer natural cement mortars. All of the tests gave 
 very irregular results. 
 
 399. To make the briquets the results of which are recorded 
 in Table 112, a plate was placed in the center of the mold, one- 
 half of the mold was filled with fresh natural cement mortar, 
 the plate was then withdrawn and the other half of the mold 
 filled with fresh Portland mortar. Briquets in line 1 were 
 made with Portland cement alone, while those in line 2 con- 
 tained orfly natural cement, these briquets being made for pur- 
 poses of comparison. The briquets containing both Portland 
 
272 
 
 CEMENT AND CONCRETE 
 
 and natural were made neat and with from one to three parts 
 sand. By noting the number of briquets that broke at the 
 juncture between Portland and natural, it was found that, in 
 general, the adhesion of rich Portland mortar to rich natural 
 cement mortar is greater than the strength of the natural, but 
 that with the poorer mortars the adhesion is less than the 
 strength of the natural. 
 
 TABLE 112 
 Adhesion bet-ween Fresh Mortars of Portland and Natural Cement 
 
 
 PARTS SAND TO ONE OF 
 
 ADHESIVE OR COHESIVE STRENGTH, POUNDS 
 
 TJ ,T1 
 
 CEMENT. 
 
 PER SQUARE INCH. 
 
 K.EF. 
 
 IN Po RTLAND 
 
 MORTAR. 
 
 IN NATURAL 
 MORTAR. 
 
 28 days. 
 
 3 months. 
 
 6 months. 
 
 1 year. 
 
 1 
 
 2 
 
 
 278 
 
 372 
 
 410 
 
 464 
 
 2 
 
 
 2 
 
 164 
 
 243 
 
 268 
 
 308 
 
 3 
 
 V ' 
 
 
 
 318 
 
 358 
 
 323 
 
 380 
 
 4 
 
 1 
 
 1 
 
 252 
 
 326 
 
 376 
 
 383 
 
 5 
 
 
 
 1 
 
 229 
 
 331 
 
 356 
 
 357 
 
 6 
 
 2 
 
 1 
 
 226 
 
 196 
 
 339 
 
 298 
 
 7 
 
 2 
 
 2 
 
 128 
 
 235 
 
 265 
 
 285 
 
 8 
 
 1 
 
 2 
 
 145 
 
 213 
 
 259 
 
 271 
 
 9 
 
 1 
 
 3 
 
 103 
 
 160 
 
 185 
 
 206 
 
 10 
 
 2 
 
 3 
 
 95 
 
 176 
 
 206 
 
 197 
 
 11 
 
 3 
 
 3 
 
 63 
 
 162 
 
 197 
 
 193 
 
 NOTES. Portland Cement, Brand R. 
 Natural Cement, Brand An. 
 Sand, " Point aux Pins," passing No. 10 sieve. 
 Both mortars mixed fresh and filled in opposite ends of mold. 
 
 400. In Table 113 the natural cement mortar contained 
 three parts sand to one of cement, while the richness of the 
 Portland mortars varied from neat to four parts sand. Four 
 combinations of different brands were used. Brand R, Port- 
 land, and brand An, natural, appear to give the best results 
 together. It is also seen from this table that the adhesion of 
 the rich Portland mortar is greater than the cohesive strength 
 of the natural cement, but when the Portland mortar contains 
 three or four parts sand to one cement, the adhesion is less 
 than the strength of the natural cement mortar. 
 
 401. THE ADHESION TO STONE AND OTHER MATERIALS. 
 Since cement mortars are usually employed to bind <>ther ma- 
 terials together, it follows that the adhesive strength is of the 
 
ADHESION TO VARIOUS MATERIALS 
 
 273 
 
 greatest importance. On account of the difficulty of making 
 tests of adhesive strength, however, the data concerning it are 
 very meager. Two methods have been employed by the au- 
 thor in making such tests. One method, used for brick, is 
 to cement two bricks together in a cruciform shape. The other 
 method consists in placing small blocks of the substance to be 
 used in the center of a briquet mold, and filling the ends of 
 the mold with the desired mortar. 
 
 TABLE 113 
 Adhesion between Fresh Mortars of Portland and Natural Cement 
 
 p 
 a 
 
 CEMENT 
 
 \R. 
 EMENT. 
 
 NATURAL CEMENT 
 MORTAR. 
 
 QUET8. 
 
 ADHESIVE STRENGTH POUNDS PER 
 SQUARE INCH. 
 
 
 
 a HO 
 
 
 M 
 
 Parts Sand to One Cement in Portland 
 
 
 
 
 
 Parts 
 
 
 
 Mortar. 
 
 
 
 
 
 
 
 tf 
 
 ti K 
 
 Cement. 
 
 One 
 
 
 
 
 
 
 
 
 
 (2 ' 
 
 
 Cement. 
 
 
 
 
 1 
 
 2 
 
 3 
 
 4 
 
 1 
 
 R 
 
 An 
 
 3 
 
 3 mo. 
 
 162 X 
 
 173 X 
 
 177 X 
 
 162 J 
 
 127 J 
 
 2' 
 
 R 
 
 Gn 
 
 3 
 
 * 
 
 147 X 
 
 175 X 
 
 167 X 
 
 156 
 
 151 
 
 3 
 
 A 
 
 En 
 
 3 
 
 
 
 l.->6 X 
 
 157 X 
 
 143 X 
 
 136 X 
 
 134 
 
 4 
 
 G 
 
 Bn 
 
 3 
 
 ' 
 
 88 X 
 
 106 X 
 
 115 X 
 
 94 
 
 91 J 
 
 5 
 
 R 
 
 An 
 
 3 
 
 lyr. 
 
 214 
 
 206 X 
 
 206 X 
 
 201 X 
 
 183 J 
 
 6 
 
 R 
 
 Gn 
 
 3 
 
 t 
 
 167 N 
 
 165 X 
 
 180 X 
 
 175 X 
 
 169 
 
 7 
 
 A 
 
 En 
 
 3 
 
 
 
 157 N 
 
 158 X 
 
 173 X 
 
 166 X 
 
 160 
 
 8 
 
 G 
 
 Bn 
 
 3 
 
 4 
 
 .108 J 
 
 127 J 
 
 126,1 
 
 130 J 
 
 114 J 
 
 NOTES: In general, each result is mean of ten briquets. 
 
 Results marked N, briquets broke through the natural cement. 
 Results marked J, briquets broke at juncture of Portland and 
 natural. 
 
 The small blocks were made one inch square and about 
 one-fourth inch thick, two opposite edges of each piece being 
 very slightly hollowed to fit, approximately, the side of the 
 mold. These blocks being placed transversely in the center of 
 the mold, and the ends of the latter filled with the mortar to 
 be tested, formed two joints between the mortar and the block. 
 
 402. Table 114 shows the adhesion of a rich Portland ce- 
 ment mortar to various materials. The mortar adheres most 
 strongly to brick, the adhesion exceeding the strength of the 
 brick itself. A very high result is also obtained with terra 
 cotta, and the adhesion to Kelleys Island limestone is high. 
 The latter is a dolomitic limestone of the corniferous group, 
 which is soft enough to be worked quite easily. The adhesion 
 
274 
 
 CEMENT AND CONCRETE 
 
 to Drummond Island limestone, which is a much harder stone 
 belonging to the Niagara group, is considerably less, and the 
 adhesion to the Potsdam sandstone is very low, A higher re- 
 sult than would be expected is obtained with ground plate 
 glass, but the hammered bar iron gives the lowest result of any 
 of the substances tried. 
 
 TABLE 114 
 Adhesion of Portland Cement Mortar to Various Materials 
 
 w 
 
 o 
 
 
 a 
 
 
 o 
 
 ADHESION, POUNDS PER SQUARE INCH, 
 
 X 
 
 K 
 
 KIND 
 
 oj H 
 
 AGE 
 
 OF 
 
 || 
 
 TO MATERIALS. 
 
 m 
 
 OF 
 
 SAND 
 
 < H S 
 
 SPECI- 
 
 
 
 h 
 
 
 P^ S.S 
 
 MENS. 
 
 
 
 
 
 
 
 
 
 ti 
 
 
 ^ 
 
 
 er 
 
 a 
 
 b 
 
 c 
 
 d 
 
 e 
 
 f 
 
 
 
 1 
 
 Cr. Qtz. 20-30 
 
 1 
 
 28 days 
 
 742 
 
 91 
 
 78 
 
 211 
 
 100 
 
 241 
 
 223 
 
 290 
 
 2 
 
 u u 
 
 1 
 
 6 inos 
 
 775 
 
 103 
 
 122 
 
 201 
 
 252 
 
 284 
 
 310 
 
 395 
 
 NOTES: Cement, Portland, Brand R. 
 
 Adhesion Blocks, 1 in. X 1 in. x in. inserted in center mold. 
 Materials: a Hammered bar iron. 
 
 b Potsdam sandstone, c'eavage surface. 
 
 c Drummond Id. limestone, cleavage surface. 
 
 d Ground plate glass. 
 
 e Kelleys Id. limestone, sawn surface. 
 
 / Soft terra cotta, filed surface. 
 
 g Soft red building brick, sawn surface. 
 
 403. The Adhesion of Neat and Sand Mortars. Table 115 
 shows the cohesive and adhesive strengths of different mortars, 
 the adhesion blocks being all of the same material, Kelleys 
 Island limestone. The Portland mortar giving the highest ad- 
 hesive strength at six months is that containing one-half part 
 sand to one part cement, though the greatest cohesive strength 
 is given by the one-to-one mortar. With natural cement the 
 one-to-one mortar gives the highest strength, both in adhesion 
 and cohesion. The ratio of the adhesive strength to the co- 
 hesive strength is greater for natural than for Portland. It 
 also appears that between twenty-eight days and six months 
 the adhesive strength increases more than the cohesive strength. 
 
 404. Effect of Consistency on Adhesion. Table 116 gives 
 the results of tests to show the relative effects of the consis- 
 tency of the mortar on the adhesive and cohesive strength, It 
 
EFFECT OF CONSISTENCY 
 
 275 
 
 TABLE 115 
 Adhesion of Mortars Containing Different Amounts of Sand 
 
 
 
 
 
 COHESIVE OR ADHESIVE STRENGTH, LBS. 
 
 I 
 
 CEMENT. 
 
 AGE 
 
 COHESION 
 
 PER SQUARE INCH, OF MORTARS WITH 
 SAND, PARTS BY WEIGHT. 
 
 pg 
 
 2 
 
 
 SPECI- 
 
 OK 
 
 ADHESION. 
 
 
 
 
 
 
 a 
 A 
 
 Kind. 
 
 Brand. 
 
 
 
 None. 
 
 Half Part 
 Sand. 
 
 One Part. 
 
 Two 
 Parts. 
 
 1 
 
 Port. 
 
 R 
 
 28 days 
 
 Cohesion 
 
 686 
 
 710 
 
 747 
 
 467 
 
 2 
 
 u 
 
 " 
 
 
 
 Adhesion 
 
 270 
 
 233 
 
 221 
 
 169 
 
 3 
 
 u 
 
 4* 
 
 6 H10S. 
 
 Cohesion 
 
 (581 
 
 787 
 
 816 
 
 551 
 
 4 
 
 u 
 
 ' 
 
 It 
 
 Adhesion 
 
 335 
 
 346 
 
 287 
 
 209 
 
 5 
 
 Nat, 
 
 An 
 
 28 days 
 
 Cohesion 
 
 183 
 
 198 
 
 218 
 
 186 
 
 6 
 
 u 
 
 " 
 
 u 
 
 Adhesion 
 
 94 
 
 104 
 
 116 
 
 66 
 
 7 
 
 " 
 
 ' 
 
 6 inos. 
 
 Cohesion 
 
 203 
 
 334 
 
 383 
 
 376 
 
 8 
 
 " 
 
 ' ' 
 
 u 
 
 Adhesion 
 
 228 
 
 222 
 
 233 
 
 171 
 
 NOTES : Sand, crushed quartz, 20 to 30. 
 
 Adhesion blocks, 1 in. X 1 in. x ^ in., Kelleys Id. limestone, sawn 
 surface, saturated before used. 
 
 is seen that the effect of consistency on the adhesive strength 
 is less than on the cohesive strength, but that the best results 
 in adhesion are given by a mortar that is considerably more 
 moist than that which gives the highest strength in cohesion. 
 The practical bearing of this point on the use of mortars is evi- 
 dent. 
 
 TABLE 116 
 Adhesion of Mortars. Varying Consistency 
 
 
 
 
 
 COHESIVE OR ADHESIVE STRENGTH, LBS. 
 
 a 
 
 g 
 
 u 
 
 CEMENT. 
 
 AGE 
 
 OP 
 
 COHESION 
 
 PER SQUARE INCH, MORTAR OF 
 CONSISTENCY: 
 
 K 
 
 a 
 
 h 
 
 
 SPECI- 
 
 OR 
 
 ADHESION. 
 
 
 
 
 
 
 3 
 
 
 
 MENS. 
 
 
 Trifle 
 
 Trifle 
 
 Quite 
 
 Very 
 
 PH 
 
 Kind. 
 
 Brand. 
 
 
 
 Dry. 
 
 Moist. 
 
 Moist. 
 
 Moist. 
 
 1 
 
 Port. 
 
 R 
 
 28 days 
 
 Cohesion 
 
 541 
 
 502 
 
 443 
 
 372 
 
 2 
 
 
 i 
 
 t 
 
 Adhesion 
 
 148 
 
 1(50 
 
 14--) 
 
 136 
 
 3 
 
 
 
 
 6 inos. 
 
 Cohesion 
 
 697 
 
 660 
 
 616 
 
 539 
 
 4 
 
 
 ' 
 
 ti 
 
 Adhesion 
 
 191 
 
 209 
 
 228 
 
 192 
 
 5 
 
 Nat. 
 
 An 
 
 28 days 
 
 Cohesion 
 
 239 
 
 212 
 
 151 
 
 112 
 
 6 
 
 
 ' 
 
 K 
 
 Adhesion 
 
 96 
 
 96 
 
 87 
 
 70 
 
 7 
 
 
 i 
 
 6 inos. 
 
 Cohesion 
 
 397 
 
 385 
 
 314 
 
 285 
 
 8 
 
 
 i 
 
 it 
 
 Adhesion 
 
 146 
 
 165 
 
 164 
 
 126 
 
 NOTES: Sand," Point aux Pins," pass No. 10 sieve, one part to one cement 
 
 by weight. 
 
 Adhesion blocks, 1 in. X 1 in. X i in., Kelleys Id. limestone, 
 surfaces filed smooth, saturated with water before used. 
 
27G 
 
 CEMENT AND CONCRETE 
 
 405. Effect of Regaging on Adhesive Strength. The tests 
 given in Table 117 were designed to show the effect of regaging 
 on the adhesion of cement mortar to stone. A comparison is 
 made between mortars used fresh and those that were allowed 
 to stand three hours and gaged once an hour. There are but 
 few tests from which to draw conclusions and the treatment is 
 very severe, but it appears that while the regaging to which 
 these mortars were subjected usually resulted in a slight in- 
 crease in cohesive strength, the adhesive strength was consid- 
 erably impaired. The decrease in adhesive strength was greater 
 for natural cement than for Portland, and greater for rich than 
 for poor mortars. The effect of regaging on the cohesive strength 
 is treated in Art. 47. 
 
 TABLE 117 
 Effect of Regaging on Adhesive Strength 
 
 CEMENT. 
 
 ADHESION OR 
 COHESION. 
 
 ADHESION OR COHESION, LBS. PER SQ. IN. 
 
 ONE PART SAND TO ONE 
 CEMENT. 
 
 THREE PARTS SAND TO 
 ONE CEMENT. 
 
 Fresh. 
 
 Regaged. 
 
 Fresh. 
 
 Regaged. 
 
 Portland, Brand X 
 
 Adhesion 
 
 178 
 
 141 
 
 62 
 
 41 
 
 44 4; u 
 
 44 
 
 202 
 
 170 
 
 51) 
 
 61 
 
 44 44 (4 
 
 Cohesion 
 
 718 
 
 764 
 
 327 
 
 343 
 
 Natural, " An 
 
 U 44 44 
 
 Adhesion 
 
 44 
 
 142 
 180 
 
 90 
 120 
 
 17 
 31 
 
 28' 
 
 (4 it 44 
 
 Cohesion 
 
 352 
 
 361 
 
 235 
 
 227 
 
 NOTES: Sand, crushed quartz, |. Each result, mean of two to five speci- 
 mens, broken at age of six months. 
 
 In adhesive tests, pieces Kellcys Id. limestone, 1 in. X 1 in. X i in., 
 placed in center mold and two ends mold filled with mortar. 
 
 Results in columns headed "Fresh" from mortar treated as usual. 
 
 Results in columns headed ''Regaged" mortar allowed to stand 
 three hours before use, mortar being regaged each hour. 
 
 406. Character of Surface of Stone. In the tests recorded 
 in Table 118 all of the adhesion blocks were of Kelleys Island 
 limestone, but part of them were finished with smooth filed 
 surfaces, while the others were grooved with a coarse rasp. In 
 the twenty-eight-day tests there is but little difference in the 
 adhesion to the different surfaces, but at six months the adhe- 
 sion to the smooth surfaces appears to be slightly greater, ex- 
 cept in the case of one-to-two natural cement mortar. 
 
EFFECT OF PLASTER PARIS 
 
 277 
 
 TABLE 118 
 Adhesion of Mortars. Effect of Character of Surface of Stone 
 
 COHESION OB ADHESION 
 
 AND 
 
 CHAHACTKK OF SURFACE. 
 
 AGE OF 
 SPECIMENS. 
 
 ADHESION OR COHESION, 
 LBS. PER SQ. IN. 
 
 Portland Brand R. 
 
 Natural Brand D. 
 
 PARTS SAND TO ONE CEMENT. 
 
 1 
 
 2 
 
 1 
 
 2 
 
 Cohesion 
 
 28 days 
 u 
 ti 
 
 6 inos. 
 
 it 
 
 539 
 151 
 152 
 714 
 238 
 223 
 
 377 
 85 
 115 
 503 
 176 
 154 
 
 343 
 138 
 129 
 387 
 141 
 115 
 
 289 
 113 
 J>8 
 304 
 68 
 96 
 
 Adhesion, smooth surface . . 
 " grooved surface 
 Cohesion 
 
 Adhesion, smooth surface . . 
 " grooved surface 
 
 407. The Effect of Plaster of Paris on the Adhesion of Mortar 
 to Stone. The results in Table 119 show the effect on the 
 adhesive strength of adding small percentages of plaster of 
 Paris to cement mortars of Portland and natural cement. The 
 Portland cement used was a quick setting sample, neat cement 
 pats of which began to set in eighteen minutes. The effect of 
 plaster of Paris on the cohesive strength of mortars from these 
 samples hardened in dry air, is shown in Table 92, 378. It is 
 seen that the addition of from one to three per cent, plaster 
 
 TABLE 119 
 
 Effect of Plaster of Paris on the Adhesive Strength of Cement 
 
 Mortars 
 
 
 
 
 
 ADHESIVE STRENGTH, LBS. PEK 
 
 KEF. 
 
 CEMENT. 
 
 PARTS P.P. 
 SAND TO 
 ONE 
 
 AGE OF 
 SPECIMENS. 
 
 SQ. IN., OF MORTARS IN WHICH 
 PER CENT. OF CEMENT REPLACED 
 BY PLASTER OF PARIS. 
 
 
 Kind. 
 
 Brand. 
 
 Sample. 
 
 CEMENT. 
 
 
 
 
 1 
 
 2 
 
 3 
 
 6 
 
 1 
 
 Port. 
 
 R 
 
 26 R 
 
 
 
 1 year 
 
 263 
 
 311 
 
 376 
 
 291 
 
 81) 
 
 2 
 
 <' 
 
 U 
 
 " 
 
 2 
 
 K 
 
 130 
 
 107 
 
 144 
 
 157 
 
 34 
 
 3 
 
 Nat. 
 
 An 
 
 L 
 
 
 
 n 
 
 88 
 
 97 
 
 87 
 
 133 
 
 a 
 
 4 
 
 u 
 
 An 
 
 it 
 
 2 
 
 u 
 
 64 
 
 74 
 
 89 
 
 82 
 
 93 
 
 NOTES: Adhesion pieces between two halves of briquet were of Kelleys Id. 
 
 limestone, sawn surfaces, saturated with water before used. 
 Cement and plaster Paris passed through No. 50 sieve. 
 All briquets stored in tank in laboratory. 
 Each result, mean of four to ten briquets. 
 
 a Found badly cracked and separated from limestone prisms after 
 three days. 
 
278 CEMENT AND CONCRETE 
 
 has no deleterious effect on the adhesive strength of these 
 samples at one year. Six per cent, plaster, however, ruins the 
 Portland and the neat natural cement. 
 
 408. THE ADHESION OF CEMENT MORTAR TO BRICK. 
 Tests of the adhesion of cement mortar to brick were made by 
 cementing pairs of brick in a cruciform shape, with a one-fourth 
 inch joint of mortar. The brick were placed together flatwise, 
 with the bed down, so that in the case of stock brick, one 
 stock mark, or depression in one side, was filled with mortar. 
 The mortar was made more moist than was ordinarily used for 
 briquets, but not so moist as would be used in brickwork. The 
 top brick of each pair was slightly tapped to place with the 
 handle of a pointing trowel, and the excess mortar cut away. 
 About forty-eight hours after cemented, the pairs of brick were 
 packed in damp sand in a large box prepared for the purpose, 
 and the sand was kept in a moist condition by a thorough daily 
 sprinkling. For pulling the bricks apart, a special clip was de- 
 vised to equalize the pull on the two ends of each brick, and a 
 simple lever machine was used to measure the force required. 
 
 409. Tensile tests were made of briquets from mortars 
 similar to those used in the adhesive tests and stored in damp 
 sand, and the results are used for comparison with the adhesive 
 tests. The cohesive strength given by the briquets is not 
 strictly comparable with the adhesive strength shown in the 
 tests with brick, because of the great difference in the area of 
 the breaking sections in the two cases. It has been well estab- 
 lished in tensile tests of cohesion that briquets of large cross- 
 section break at a lower strength than those of small section. 
 It is quite possible also that even with the special clip devised, 
 cross-strains were more likely to occur in the adhesive tests 
 than in the briquet tests. An opportunity was furnished of 
 comparing the tensile strength of neat natural cement mortar 
 under the two conditions, for in one case six joints broke di- 
 rectly through the mortar, the adhesion being greater than the 
 cohesion. It was found that the strength per square inch 
 given by the briquets was at least six times that given by the 
 large joint. This difference should be kept in mind in making 
 comparisons in the tables between the cohesion and adhesion 
 as given. It should also be noted that some of the highest 
 results of adhesive strength represent in reality the strength 
 
ADHESION TO BRICK 
 
 279 
 
 of the brick rather than the adhesive strength of the mortar, as 
 chips were pulled from the brick, leaving the mortar joint 
 undisturbed. The brick were of a rather poor quality, but 
 selected with a view to obtaining those of a uniform degree of 
 burning. 
 
 TABLE 120 
 
 Adhesion of Cement Mortar to Brick. Variations in Richness 
 
 of Mortar 
 
 
 
 
 TENSILE STKENGTH, POUNDS PER 
 
 
 
 
 SQUARE INCH, OF MORTARS CONTAINING 
 
 OKMENT. 
 
 AGE OP 
 
 ADHESION 
 
 OR 
 
 PAHTS SAND TO ONE CEMENT. 
 
 
 Mo UTAH. 
 
 
 
 
 
 COHESION. 
 
 None. 
 
 I 
 
 1 
 
 2 
 
 3 
 
 Portland, X, 41 S 
 
 28 days 
 
 Cohesion 
 
 632 
 
 596 
 
 589 
 
 409 
 
 270 
 
 U I ( 
 
 t i 
 
 Adhesion 
 
 48 
 
 42 
 
 24 
 
 20 
 
 11 
 
 u t 
 
 3 months 
 
 Cohesion 
 
 676 
 
 728 
 
 694 
 
 423 
 
 325 
 
 a i i 
 
 ( 
 
 Adhesion 
 
 64 
 
 52 
 
 41 
 
 24 
 
 12 
 
 t; i ( 
 
 6 months 
 
 Cohesion 
 
 723 
 
 764 
 
 679 
 
 52 1 
 
 374 
 
 U I i 
 
 U 
 
 Adhesion 
 
 60 
 
 66 
 
 39 
 
 20 
 
 14 
 
 Natural, Gn, KK 
 
 3 months 
 
 Cohesion 
 
 180 
 
 240 
 
 317 
 
 279 
 
 181 
 
 U U It 
 
 tt 
 
 Adhesion 
 
 46 
 
 62 
 
 42 
 
 28 
 
 15 
 
 U U U 
 
 6 months 
 
 Cohesion 
 
 276 
 
 444 
 
 388 
 
 331 
 
 236 
 
 4 ' " " 
 
 it 
 
 Adhesion 
 
 44 
 
 52 
 
 50 
 
 38 
 
 18 
 
 NOTES: Bricks were cemented together in pairs in cruciform shape and 
 
 kept in damp sand until time of test. Briquets for cohesion 
 
 tests stored in same manner. 
 Each result in cohesion, mean of five briquets. 
 Each result in adhesion is in general mean of six results, three 
 
 with common die cut brick and three with sand molded stock 
 
 brick. 
 When adhesion exceeded 50 pounds per square inch, bricks were 
 
 about as likely to break as the joint between brick and mortar. 
 
 410. Adhesion of Neat and Sand Mortars of Portland and 
 Natural. Some of the results of these tests are given in Table 
 120. The most noteworthy point developed is that for mor- 
 tars containing more than one-half part of sand to one part 
 cement, the adhesion of the natural cement is greater than 
 that of the Portland with the same proportion of sand, al- 
 though the Portland mortar was much the stronger in cohesion. 
 The mortars giving the highest adhesive strength are those 
 containing not more than one-half part sand to one part cement. 
 
 The addition of sand lowers the adhesive strength more 
 rapidly than it does the cohesive strength. This point would 
 
280 
 
 CEMENT AND CONCRETE 
 
 be shown still more clearly if the true adhesive strength of the 
 richest mortars was obtained, as we may be certain that the 
 adhesion of these mortars would be shown to be considerably 
 greater if the brick were strong enough to allow this strength 
 to be developed. With natural cement mortars containing not 
 more than two parts sand to one cement, the adhesion is one- 
 sixth to one-ninth the cohesion, and with Portland mortars 
 containing not more than one part sand, the adhesion is about 
 one-fifteenth the cohesion. (But see 409 in this connection.) 
 
 411. Effect of Lime Paste on Adhesive Strength of Cement 
 Mortars. A number of tests were made to determine the 
 effect, on the adhesive and cohesive strength of mortars, of 
 mixing lime paste with the cement. Tables 121 and 122 give 
 the results of a few preliminary tests on this subject. 
 
 For the tests recorded in Table 121 the mortars were al- 
 lowed to harden in dry air. From the cohesive tests it is seen 
 that lime in form of paste to the amount of ten per cent, of 
 
 TABLE 121 
 
 Adhesion of Cement Mortar to Brick. Effect of Lime Paste in 
 Mortar Hardened in Dry Air 
 
 CEMENT. 
 
 AGE 
 
 OF 
 
 MORTAR. 
 
 COHESION 
 
 OB 
 
 ADHESION. 
 
 TENSILE STRENGTH, POUNDS PER 
 SQUARE INCH. 
 
 Composition of Mortar. 
 
 A 
 
 B 
 
 C 
 
 D 
 
 E 
 
 Portland, X, 41S 
 
 It it it 
 
 3 months 
 4 " 
 
 Cohesion 
 Adhesion 
 
 97 
 18 
 
 99 
 29 
 
 101 
 20 
 
 46 
 
 22 
 
 59 
 13 
 
 it It U 
 
 6 " 
 
 a 
 
 
 24 
 
 20 
 
 19 
 
 11 
 
 Natural, Gn, LL 
 
 3 " 
 
 Cohesion 
 
 18 
 
 38 
 
 21 
 
 22 
 
 68 
 
 U U it 
 
 4 " 
 
 Adhesion 
 
 39 
 
 32 
 
 30 
 
 28 
 
 11 
 
 U it it 
 
 6 " 
 
 U 
 
 26 
 
 31 
 
 25 
 
 27 
 
 
 NOTES: Brick, sand molded stock. 
 
 All briquets and brick stored in dry air. 
 
 Composition of mortars: A B C D E 
 
 Grams P. P. river sand, 480 480 480 480 480 
 
 Grams cement, 120 120 90 60 
 
 Grams lime paste, 40 30 60 120 
 
 Grams lime contained in lime paste, 14 10 20 41 
 Lime in paste expressed as per cent, 
 of cement plus lime, 
 
 10 10 25 100 
 
 Consistency about same as mason's mortar. 
 
EFFECT OF LIME PASTE 
 
 281 
 
 the cement had little effect on one-to-four Portland mortars, 
 but that a larger amount of lime was very deleterious for dry 
 air exposure. The sample of natural cement used did not harden 
 well in dry air, and the highest result is given by the lime mor- 
 tar without cement. It appears that the adhesive strength of 
 the Portland mortar was slightly increased by the addition of a 
 small amount of lime paste, but the adhesive strength of natural 
 was not greatly affected. The adhesive strength of the natural 
 cement is, in general, higher than the Portland. The nat- 
 ural cement appeared to harden better in the joints than in the 
 briquets, and we have, as a peculiar result, the adhesive strength 
 exceeding the cohesion. This illustrates a statement already 
 made, that to store briquets in dry air does not approach very 
 nearly the ordinary conditions of use. 
 
 412. In Table 122 are given a few tests of mixtures of Port- 
 
 TABLE 122 
 
 Adhesion of Cement Mortar to Brick. Effect of Lime Paste in 
 Portland Cement 
 
 
 
 TENSILE STRENGTH, POUNDS PER SQUARE INCH. 
 
 MORTAR HARDENED IN 
 
 COHESION 
 
 OR 
 
 COMPOSITION OF MORTAR. 
 
 
 ADHESION. 
 
 
 
 
 A 
 
 B 
 
 C 
 
 D 
 
 E 
 
 Tank in Laboratory 
 
 Coh'n. 
 
 177 
 
 203 
 
 183 
 
 158 
 
 82 
 
 Dry air, 
 
 (i 
 
 167 
 
 180 
 
 167 
 
 150 
 
 81 
 
 Damp sand, " 
 
 
 
 173 
 
 198 
 
 171 
 
 154 
 
 88 
 
 Dry air, 
 
 Adh'ii. 
 
 15 
 
 36 
 
 40 
 
 33 
 
 26 
 
 Damp sand, " 
 
 it 
 
 15 
 
 33 
 
 35 
 
 32 
 
 27 
 
 NOTES: Bricks cemented together in pairs in cruciform shape. 
 
 Age of all mortars when tested, three months. 
 
 Cement, Portland, Brand R, Sample 14 R. Sand, " Point aux Pins." 
 
 Lime paste slaked about six months before use. 
 
 Each result in cohesion, mean of five to ten briquets. 
 
 Each result in adhesion, mean of eight to sixteen pairs of bricks. 
 
 Half of pairs were hard burned brick and half soft burned. 
 
 Composition of mortars: A B C D E 
 
 Grams P. P. (river) sand, 960 960 960 960 960 
 
 Grams cement, 240 240 200 180 120 
 
 Grams lime paste, 80 120 180 360 
 
 Grams lime contained in paste, 27 40 60 120 
 
 Lime as per cent, lime plus cement, 10 16.7 25 50 
 
282 CEMENT AND CONCRETE 
 
 land cement and lime paste, the mortars being hardened in dry 
 air and in damp sand. Cohesive tests are also given of briquets 
 hardened in damp sand, water and dry air. It appears that 
 the addition of ten per cent, of lime in the form of paste to 
 mortars of this sample of Portland increases the tensile strength, 
 the effect being least when the mortars harden in dry air. The 
 substitution of lime for one-sixth of the cement in a one-to-four 
 mortar has little effect on the tensile strength. Larger propor- 
 tions of lime result in decreased strength, and if one-half of the 
 cement is replaced by lime, the resulting strength is only about 
 one-half that given by the cement mortar without lime. The 
 results of the adhesive tests show that if half of the cement in 
 the mortar is replaced by an equal weight of lime in the form 
 of paste, the resulting strength is increased by nearly 100 
 per cent., and that if smaller amounts of lime are used, the 
 adhesive strength is increased by about 150 per cent, over 
 that given by the cement mortar without lime. 
 
 413. The results of a more complete set of tests on this sub- 
 ject 'are given in Table 123. The mortars used included one 
 made with four parts sand to one of cement by weight; one in 
 which about ten per cent, of lime by weight, which had pre- 
 viously been made into lime paste, was added to the mortar; a 
 third in which lime, in the form of paste, was substituted for 
 one-sixth of the weight of the cement used in the first mortar; 
 a fourth, in which lime was substituted for one-fourth of the 
 cement; and finally, a mortar composed of lime paste and sand 
 only. 
 
 The adhesive strengths of the mortars are given in the 
 table. The difference in the adhesion of Portland cement 
 mortar to hard brick and to soft brick is not clearly brought 
 out. Neither is the strength of air-hardened specimens much 
 different from that of the mortars stored in damp sand. The 
 use of lime paste with Portland cement in the amounts tried 
 here more than doubles the adhesive strength of the mortar. 
 
 The first point to notice in the case of natural cement is 
 that the adhesive strength of this mortar without lime is nearly 
 double the adhesive strength of Portland mortar without lime. 
 The adhesive strength of mortars hardened in damp sand is 
 somewhat greater than the strength of similar mortars hard- 
 ened in dry air. The addition of a small amount of lime paste 
 
EFFECT OF LIME PASTE 
 
 283 
 
 PIO 
 
 jpug P.ITJH 
 
 ill 
 
 
 
 
 = 
 
 'QNVg 
 
 8KQ 'axsvj awrj 
 
 CO t-_ 
 
 01 CS M O 
 
 C5 T-H I- O 
 
 O T-H 00 ^ - 
 CO ^CCOO 
 
 ISXO 
 
 cb ci co --H 
 
 T-H 00-^r^ 
 
 CO G<IO<M 
 
 t^ ^-' o " 
 
 CO rf-rtCC 
 
 O O'C'N 
 
 O5 O CO O 
 ^H ^CC^ 
 
 OCOX 
 
 't 
 
 CO 
 
 i-' o' r? ss **' 
 
 rH CO "^ CO CO 
 
 CO t-i 
 O O -t C<l 
 
 O5 CO CO 
 
 t- -^ 55 t- 
 
 10 co ^M' o 
 
 i-t CO-^Tjl 
 
 CO CO O 
 
 O CC O O 
 
 T-I (M U3 CO O5 
 UD O 'N O O5 
 
 U5 CC CO CO C5 CO t- 
 ^' ^H' ^.' ,_(' i.o CO CO 00 
 i-i CO-^-^ CO COCO5<I 
 
 t- CO CO CO CO Oi CO 55 
 
 T-i ^ ^' ^' o co Tt< co 
 
 CN TfCOCO CO T^'tCO 
 
 !>. O CO T*< CO CO i-H I- 
 
 Tt< N CO O 
 CO TjICOCO 
 
 S 33'S; 
 
 S28 $'3' 
 
 CO t^CCt^ t^ CO-^CO 
 r ' co' O cb t ' ' 
 
 CO 
 
 SO O O O O C: O 
 00 (N CO O CO 5^ CO 
 
 O OOO O 
 
 ^ Tfoco TJ< 
 
 (M GNlCV|i-H (M 
 
 30N3H3.I3H 
 
 I-H CMCO'* iO COt-CO 
 
 73 cc! 
 
 11 
 
 !"S 
 
284 CEMENT AND CONCRETE 
 
 increases the adhesive strength somewhat, and when as much 
 as twenty-five per cent, of the cement is replaced by lime in 
 the form of paste, the adhesive strength of the natural cement 
 mortar is not usually diminished. The effect of lime paste, 
 however, on the adhesive strength is not nearly so great as it 
 is in the case of Portland mortars. 
 
 The following conclusions may be briefly stated: The ratio 
 of adhesive to cohesive strength is much greater with natural 
 cement than with Portland. If a high adhesive strength is 
 desired, Portland cement should not be mixed with more than 
 two parts sand unless lime paste is added to the mortar, as 
 the use of lime paste materially increases the adhesive strength 
 of lean mortars. Tests of cohesion of similar mortars contain- 
 ing lime paste are given in Art. 48. 
 
 414. THE ADHESION OF CEMENT TO RODS OF STEEL AND 
 
 IRON. The tests recorded in Tables 124 and 125 were made 
 to determine the adhesion of cement mortar to iron rods, or 
 the strength of a bolt anchorage secured with cement mortar, 
 and the style of rod and kind of mortar which would give the 
 best results. The bars were made in an ordinary concrete 
 mold, ten inches by ten inches by four and one-half feet. The 
 rods or bolts were placed in a row along the center of the box, 
 being spaced about nine inches apart, and the mortar was 
 rammed about them. After being allowed to set in a warm 
 room for twenty-eight days, the rods were pulled by means of 
 two hydraulic jacks, a special grip being used to grasp the free 
 end of the rod, and an hydraulic weighing machine serving to 
 measure the pull required to start it. The supports against 
 which the hydraulic jacks were braced bore at points on the 
 concrete bar about three or four inches from the center of the 
 rod which was being tested. 
 
 415. The rods given in Table 124 were imbedded in mortar 
 composed of one part of Portland cement to two parts lime- 
 stone screenings. The rods were cut from bar iron and were 
 perfectly plain, without nuts or fox wedges. The results in- 
 dicate that the force required is proportional to the area of 
 contact. Comparing the different styles and sizes of plain rods, 
 no difference in favor of one style or size can be determined; 
 the apparent higher resistance per square inch offered by one- 
 inch rods would probably disappear in a large number of tests. 
 
ADHESION TO STEEL RODS 
 
 285 
 
 TABLE 124 
 
 Resistance to Pulling of Iron Rods of Various Forms Imbedded 
 
 in Mortar 
 
 
 
 
 
 
 
 
 POUNDS PULL. 
 
 
 
 
 
 PERIM- 
 
 DEPTH 
 
 
 REF. 
 
 gg 
 o 
 SK 
 p 
 
 MORTAR, 
 BAR No. 
 
 DESCRIPTION OF Ron. 
 
 ETER 
 OF ROD, 
 INCHES. 
 
 IM- 
 BEDDED, 
 INCHES. 
 
 Per In. 
 Depth 
 Im- 
 
 PerSq. 
 In. Area 
 in Con- 
 
 
 Z 
 
 
 
 
 
 bedded. 
 
 tact. 
 
 1 
 
 3 
 
 2, 6, 7 
 
 Plain, i ' diameter 
 
 1.57 
 
 8 to 10 
 
 700 
 
 447 
 
 2 
 
 3 
 
 i* 
 
 V 
 
 3.14 
 
 " 
 
 1750 
 
 556 
 
 3 
 
 3 
 
 i < 
 
 H' 
 
 3.03 
 
 u 
 
 2060 
 
 524 
 
 4 
 
 3 
 
 (i 
 
 ^ square 
 
 2.00 
 
 11 
 
 1085 
 
 543 
 
 ;"> 
 
 4 
 
 2,5,6,7 
 
 r " 
 
 4.00 
 
 u 
 
 22oO 
 
 5(52 
 
 <l 
 
 3 
 
 2, 6, 7 
 
 iy 
 
 5.00 
 
 it 
 
 2170 
 
 434 
 
 7 
 
 3 
 
 4, 5 
 
 I Twisted 1" square, j 
 I 1 turn in 8" length ( 
 
 4.3 1 
 
 o 
 
 25D5 
 
 608 
 
 8 
 
 3 
 
 t{ 
 
 ( Twisted 1" square, \ 
 1 2 turns in 8" length J 
 
 4.3 1 
 
 9 
 
 2215 
 
 516 
 
 1) 
 
 3 
 
 11 
 
 { Twisted 1" square, ( 
 j 3 turns in 8" length j 
 
 4.3 1 
 
 0-9.5 
 
 2405 
 
 561 
 
 NOTES: Cement, Portland, Brand R. 
 
 Sand, limestone screenings passing f inch slits, two parts by 
 
 weight to one cement. 
 Mortar one month old when tension was applied to rods. 
 
 The rods given in lines seven to nine were made by twisting 
 a piece of one inch square bar iron. The twisted portion was 
 eight inches in length. Comparing the plain one inch square 
 bolts with the twisted bolts, it appears that the former offered 
 a resistance of 2,245 pounds per inch depth while the latter 
 gave 2,405 pounds, an increase of less than eight per cent. 
 
 416. In the tests recorded in Table 125, the ordinary river 
 sand used in construction was employed. The mortar was 
 made neat and with two and four parts sand to one of cement. 
 The depth the rods were imbedded varied from two inches to 
 ten inches. The one-to-two mortar gave nearly as good results 
 as neat cement, but the one-to-four mortar gave much lower 
 results. The resistance seems to vary directly as the area of 
 contact without reference to the depth imbedded, except as 
 
 1 In computing adhesion, or shear, or pounds pull per square inch of area 
 in contact, perimeter considered circumference of a circle of diameter equal 
 to the distance between opposite edges of rod after twisting. A core of mor- 
 tar of this diameter, was torn from bar in pulling. 
 
 To perceive effect twisting, compare pounds pull per inch depth imbedded. 
 
286 
 
 CEMENT AND CONCRETE 
 
 this enters in obtaining the said area. The results obtained in 
 this table do not compare favorably with those obtained in 
 Table 124, where limestone screenings were used. 
 
 TABLE 125 
 
 Resistance to Pulling of Iron Rods Imbedded in Mortar. Variations 
 in Depth Imbedded and in Richness of Mortar 
 
 PARTS 
 
 SAND TO 
 ONE 
 CEMENT. 
 
 ADHESION, POUNDS PER SQUARE INCH OF SURFACE IN CONTACT FOR DIFFERENT 
 DEPTHS IMBEDDED. 
 
 Depths ) 
 Imbed- 
 ded, In ) 
 
 1.9-2.2 
 
 3*2 
 
 4 
 
 4.5^1.8 
 
 5.8-6 
 
 7.8-8 
 
 8.8 
 
 9.6-10 
 
 No. 
 Re- 
 sults. 
 
 Mean. 
 
 
 2 
 4 
 
 
 340 
 
 272 
 
 74 
 
 294 
 
 346 
 270 
 119 
 
 262 
 
 313 
 255 
 117 
 
 247* 
 
 100 
 
 228 
 
 340 
 275 
 142 
 
 5 
 15 
 10 
 
 313 
 264 
 111 
 
 
 
 
 NOTES : Cement, Portland, Brand R. 
 
 Sand, "Point aux Pins," river sand. 
 Mortar one month old when rods pulled. 
 Rods, round, 1 inch diameter. 
 
 417. Tables 126 and 127 are from similar tests made by 
 Messrs. Peabody and Emerson. 1 The rods in Table 126 were 
 of various shapes and included some having rivets through 
 them. The ^ inch by 1 inch bars gave lower adhesion per 
 square inch than the square and round rods. When two rods 
 are twisted together and imbedded in a small specimen, the 
 tendency is to split the specimen. The bars containing rivets 
 broke before the adhesion was overcome, although the depth 
 imbedded Was but six inches. 
 
 In Table 127 neat cement paste and concretes of several 
 compositions are tried. These results are of interest as show- 
 ing that concretes show as great adhesion to steel rods as do 
 mortars. The very low result obtained with neat cement in 
 this table is not explained and is in opposition to the results 
 in Table 125. 
 
 Engineering News, March 10, 1904. 
 
ADHESION TO STEEL RODS 
 
 287 
 
 TABLE 126 
 
 Adhesion of Mortar to Steel Rods of Various Shapes, Imbedded 
 about Six Inches 
 
 No. 
 
 OF 
 
 TESTS. 
 
 DESCRIPTION OF ROD. 
 
 PERIMETER 
 OF ROD, 
 INCHES. 
 
 POUNDS PULL. 
 
 Per Inch 
 Depth 
 Imbedded. 
 
 Per Square 
 Inch Area 
 in Contact. 
 
 4 
 3 
 4 
 4 
 4 
 
 4 
 
 Plain, \" square. 
 Plain, \" square. 
 Plain, \" round. 
 Twisted, \" square. 
 V by 1 ". 
 
 Two \" rods twisted 
 together. 
 
 1.0 
 2.0 
 1.57 
 
 2.5 
 
 369 
 864 
 804 
 1259 
 744 
 
 369 
 432 
 512 
 
 293 
 
 Three specimens split. 
 One rod broke at 8,000 Ibs., or when 
 adhesion was 1,250 Ibs. per inch 
 depth. 
 
 4 
 
 \" by 1" with I" rivets 
 through. 
 
 One specimen split. 
 Three rods broke at lirst rivet with 
 9,800 to 10,500 pounds, or when 
 adhesion was 1,500 to 1,660 Ibs. 
 per inch depth. 
 
 NOTES: Tests by Messrs. George A. Peabody and Samuel W. Emerson. 
 
 Mortar composed of one part cement (Portland) to three parts sand. 
 Specimens approximately 6 inch cubes. One rod imbedded in 
 
 each, 6 to 6 inches. 
 Rods pulled forty and eighty days after mortar was made. 
 
 TABLE 127 
 
 Adhesion of Mortars and Concretes of Various Compositions to 
 One Inch Square Steel Rods Imbedded about Ten Inches 
 
 
 COMPOSITION OF MORTAR OR 
 
 POUNDS PULL. 
 
 No. 
 
 CONCRETE. 
 
 
 OF 
 
 
 
 
 TESTS. 
 
 
 Per Inch 
 Depth 
 Imbedded. 
 
 Per Square 
 Inch Area 
 in Contact. 
 
 Cement. 
 
 Sand. 
 
 Stone. 
 
 Gravel. 
 
 4 
 
 
 
 
 
 
 
 
 1112 
 
 278 
 
 4 
 
 
 3 
 
 
 
 
 
 1644 
 
 411 
 
 4 
 
 
 3 
 
 6 
 
 
 
 1912 
 
 478 
 
 4 
 
 
 3 
 
 
 
 6 
 
 2062 
 
 516 
 
 4 
 
 
 2 
 
 4 
 
 
 
 2348 
 
 587 
 
 4 
 
 1 
 
 2 
 
 
 
 4 
 
 2187 
 
 547 
 
 NOTE: Tests by Messrs. George A. Peabody and Samuel W. Emerson. 
 
CHAPTER XVI 
 
 THE COMPRESS1VE STRENGTH AND MODULUS OF 
 ELASTICITY OF MORTAR AND CONCRETE 
 
 ART. 52. COMPRESSIVE STRENGTH OF MORTAR 
 
 418. The compressive strength of cement mortar is from 
 five to ten times the tensile strength. As the result obtained 
 in tests of either compression or tension depends upon the shape 
 and size of the specimen, no definite value can be assigned to 
 the ratio of compression to tension. Comparative tests have 
 indicated in a general way that the cements giving the best 
 results in tension show also the highest compressive strength; 
 but with variations in treatment, different kinds and brands of 
 cement do not give the same variations in the ratios of the two 
 kinds of strength. 
 
 Mortar is not usually employed alone in large masses. It 
 more frequently forms the binding medium between fragments 
 of other substances, such as brick and stone. The dependence 
 of the strength of masonry upon the strength of the mortar 
 increases with the roughness of the stone or brick, and the 
 thickness of the bed joints. In fine ashlar masonry this depend- 
 ence is comparatively small, in brickwork it is important, and 
 in concrete any increase in the strength of the mortar increases 
 the strength of the concrete in nearly the same ratio. 
 
 Piers of brickwork may give a crushing resistance either 
 greater or less than the strength of cubes made from mortar 
 of the same composition as that used in building the piers. 
 Thin beds of mortar between strong materials resist high com- 
 pressive stresses, while in walls or piers built with weak blocks, 
 the mortar is destroyed by the cracking of the blocks at a lower 
 stress than the mortar would withstand in a cube pressed 
 between steel plates. Since in brick and stone masonry the 
 mortar forms but a small part of the structure, it is not econom- 
 ical to use a poor quality of mortar with good brick and stone. 
 
 288 
 
CEMENT MORTAR 
 
 289 
 
 419. Ratio of Compressive to Tensile Strength. M. E. 
 
 Candlot has made many experiments showing the effect of 
 certain variations in the preparation of mortars upon the com- 
 pressive and tensile strength. A few of the results of one 
 series are presented in Table 128. The reduction from the 
 metric system has been made, and a column added giving 
 approximately the number of parts of sand to one of cement by 
 weight, the accurate proportions appearing in the form of 
 weight of cement to one cubic yard of sand. These results in- 
 dicate that the ratio of the strength in compression to that in 
 tension increases with the age of the mortar and also with its 
 richness. 
 
 TABLE 128 
 
 Resistance of Cement Mortars to Tension and Compression, with 
 
 Varying Proportions of Normal Sand 
 
 SPECIMENS HARDENED IN FKESH WATER 
 
 [From Ciments et Chaux Ifydrauliqucs, par M. E. Candlot.] 
 
 -S S 8* 
 
 h 
 
 RESISTANCE IN POUNDS PER SQUARE INCH IN TENSION AND 
 
 
 
 0^j< H 
 
 * Q J 
 
 COMPRESSION. 
 
 * u 
 
 ! g .** 
 
 J 
 
 
 |H 
 
 
 
 
 
 
 Q H w^* 
 
 s s - 
 
 7 days. 
 
 28 days. 
 
 1 year. 
 
 2 years. 
 
 3 years. 
 
 O n w 
 
 O ^^ W 
 
 g So 
 
 
 
 
 
 
 2 ^ E 
 
 KOC *-* & 
 
 [^O ^ 
 
 
 
 
 
 
 
 
 
 
 
 H fcH 
 
 l| 
 
 Pi 
 
 T. 
 
 C. 
 
 T. 
 
 C. 
 
 T. 
 
 C. 
 
 T. 
 
 C. 
 
 T. 
 
 C. 
 
 5 
 
 10.8 
 
 250 
 
 27 
 
 266 
 
 38 
 
 408 
 
 70 
 
 507 
 
 74 
 
 572 
 
 108 
 
 738 
 
 6.8 
 
 6.4 
 
 420 
 
 128 
 
 643 
 
 143 
 
 1164 
 
 212 
 
 1730 
 
 209 
 
 1630 
 
 219 
 
 1775 
 
 8.1 
 
 4.6 
 
 590 
 
 139 
 
 1040 
 
 234 
 
 1940 
 
 337 
 
 2980 
 
 284 
 
 2930 
 
 341 
 
 3080 
 
 9.0 
 
 3.5 
 
 7(50 
 
 233 
 
 1520 
 
 393 
 
 3080 
 
 435 
 
 4020 
 
 400 
 
 4400 
 
 462 
 
 4590 
 
 9.9 
 
 2.9 
 
 930 
 
 251 
 
 2110 
 
 462 
 
 3690 
 
 490 
 
 5580 
 
 490 
 
 5680 
 
 557 
 
 6060 
 
 10.9 
 
 2.6 
 
 1100 
 
 341) 
 
 2630 
 
 551 
 
 5020 
 
 594 
 
 5820 
 
 557 
 
 6060 
 
 616 
 
 6480 
 
 10.5 
 
 2.0 
 
 1350 
 
 368 
 
 3360 
 
 550 
 
 5020 
 
 713 
 
 7750 
 
 805 
 
 7860 
 
 784 
 
 8710 
 
 11.1 
 
 1.6 
 
 1690 
 
 4 13 
 
 3310 
 
 561 
 
 5070 
 
 767 
 
 7670 
 
 907 
 
 8800 
 
 815 
 
 9180 
 
 11.3 
 
 From a study of the results of nearly three thousand tests 
 made by Professor Tetmajer, the late Professor J. B. Johnson 
 concluded that for mortars containing three parts sand to one 
 cement the ratio of the compressive strength to the tensile 
 strength is equal to 8.64 -f- 1.8 log. A, where A is the age of 
 the mortar in months. It is shown above that the ratio in- 
 creases with increasing proportions of sand. 
 
 420. Table 129 gives some results obtained at the Water- 
 town Arsenal in tests of cement mortar cubes. 1 The mortars 
 
 Prepared by Mr. George W. Rafter for the State Engineer of New York. 
 
290 
 
 CEMENT AND CONCRETE 
 
 TABLE 129 
 Compressive Strength of Cement Mortar. Portland and Natural 
 
 TESTS OF 12 INCH CUBES, TWENTY MONTHS OLD, MADE AT WATERTOWN 
 ARSENAL FOR STATE ENGINEER OF NEW YORK 
 
 METHOD OF STORAGE OF 
 CUBES. 
 
 CEMENT. 
 
 CONSISTENCY 
 
 OF 
 
 MORTAR. 
 
 CRUSHING STRENGTH, LBS. PER 
 SQUARE INCH, FOR MORTARS 
 CONTAINING PARTS SAND TO ONE 
 CEMENT BY VOLUME: 
 
 Kind. 
 
 Brand. 
 
 1 
 
 2 
 
 3 
 
 4 
 
 Mean 
 
 
 
 
 f Dry 
 
 3479 
 
 2200 
 
 1154 
 
 
 2278 
 
 Water 3 to 4 mo., 
 then buried in sand. 
 
 Nat. 
 
 Buffalo 
 
 1 Plastic 
 1 Excess 
 
 '2795 
 2161 
 
 1783 
 1698 
 
 1000 
 776 
 
 
 1859 
 1545 
 
 
 
 
 I Mean 
 
 2812 
 
 1894 
 
 977 
 
 
 1894 
 
 Covered with burlap; 
 
 
 
 f Dry 
 
 3347 
 
 2000 
 
 961 
 
 
 2103 
 
 kept wet for several 
 weeks, then exposed 
 
 Nat. 
 
 Buffalo 
 
 I Plastic 
 1 Excess 
 
 2476 
 2070 
 
 1294 " 
 1358 
 
 692 
 738 
 
 
 1487 
 1389 
 
 to weather. . . 
 
 
 
 [_ Mean 
 
 2631 
 
 1551 
 
 797 
 
 . . . 
 
 1660 
 
 
 
 
 {Dry 
 
 2844 
 
 2051 
 
 987 
 
 
 1961 
 
 In cool cellar 
 
 Nat. 
 
 Buffalo 
 
 Plastic 
 
 2514 
 
 1256 
 
 883 
 
 ! ! 
 
 1551 
 
 
 
 
 Excess 
 
 2159 
 
 1386 
 
 678 
 
 
 1408 
 
 
 
 
 Mean 
 
 2504 
 
 1564 
 
 849 
 
 . 
 
 1640 
 
 Fully exposed to 
 weather 
 
 Nat. 
 
 Buffalo 
 
 C Dry 
 ! Plastic 
 
 3272 
 
 2667 
 
 1879 
 1356 
 
 1054 
 822 
 
 
 2068 
 1615 
 
 
 
 
 1 Excess 
 
 1996 
 
 1311 
 
 669 
 
 
 1325 
 
 
 
 
 I Mean 
 
 2645 
 
 1513 
 
 848 
 
 
 1669 
 
 
 
 
 f Dry 
 
 3236 
 
 2032 2 
 
 1039 
 
 
 2102 
 
 Means 
 
 
 
 ^ Plastic 
 
 2613 
 
 1421 
 
 849 
 
 . 
 
 1628 
 
 
 
 
 [ Excess 
 
 2097 
 
 1438 
 
 715 
 
 . . . 
 
 1417 
 
 Grand mean . . 
 
 
 
 
 2649 
 
 1630 
 
 868 
 
 
 1716 
 
 Water 3 to 4 mo., 
 -then buried in sand 
 
 Port. 
 
 Empire 
 
 f Dry 
 
 \ Plastic 
 
 . . . 
 
 3897 
 3642 
 
 2494 
 
 2168 
 
 1782 
 1717 
 
 . . . 
 
 Covered with burlap; 
 
 
 
 
 
 
 
 
 
 kept wet for several 
 
 
 
 
 
 
 
 
 
 weeks, then exposed 
 to weather. 
 
 Port. 
 
 Empire 
 
 f Dry 
 | Plastic 
 
 . 
 
 3880 
 3672 
 
 2492 
 2168 
 
 1489 
 1726 
 
 
 In cool cellar . 
 Fully exposed to 
 weather . . . 
 
 Port. 
 Port. 
 
 Empire 
 Empire 
 
 Plastic 
 Dry 
 Plastic 
 
 . . . 
 
 3397 
 3313 
 4059 
 
 3589 
 
 2132 
 2164 
 
 2450 
 2270 
 
 1614 
 1679 
 1715 
 1465 
 
 . . . 
 
 
 
 
 Dry 
 
 
 3808 
 
 2392 
 
 1650 
 
 
 Means 
 
 
 
 Plastic 
 
 
 3554 
 
 2193 
 
 1647 
 
 
 
 contained one, two and three volumes of sand to one of natural 
 cement, and two to four parts sand to one volume of Portland. 
 
 1 Result interpolated. 
 
 * 2,043 omitting interpolated result. 
 
CONCRETE 291 
 
 The proportions of water used were such as to give mortars of 
 different consistency, "dry," like damp earth, "plastic," of the 
 consistency usually employed by masons, and "excess," quak- 
 ing like liver with slight tamping. The specimens were twelve 
 inch cubes and four methods of storage were used, as indicated. 
 
 Comparing the results with similar tests of tensile strength, 
 it appears that the strength in compression decreases more 
 rapidly as sand is added than does the tensile strength. The 
 same conclusion was drawn from Table 128. 
 
 The strength of the Portland mortar with four parts sand 
 is about equal to the strength of the natural with two parts. 
 The dry mortar gives the highest strength with natural cement, 
 but with Portland the "dry" and "plastic" give about the 
 same result. 
 
 Concerning the consistency, it has already been pointed out 
 that the conditions of the actual employment of mortar are 
 , such as to favor, in general, the use of a wetter mixture than 
 that which gives the best results in laboratory tests of mortars. 
 As to storage, the specimens kept in water for three or four 
 months after made, give the highest results with natural ce- 
 ment. There seems to be no choice between the other three 
 methods of storage. 
 
 ART. 53. COMPRESSIVE STRENGTH OF CONCRETE WITH VARIOUS 
 PROPORTIONS OF INGREDIENTS 
 
 421. With the increasing use of concrete in arch bridges, 
 in foundation piers and in columns of buildings, and especially 
 in connection with steel in beams, etc., the compressive strength 
 of the material becomes of the greatest importance. Moreover, 
 the composition of concrete may vary so much, the range of 
 available aggregates is so wide, and the methods of manipula- 
 tion are so diverse, that many tests must be studied before 
 one can judge of the probable strength of a given mixture. 
 
 For any very extended work, it may be found economical 
 to make a series of tests using the materials available, and 
 combining them as nearly as possible in the manner proposed 
 in actual construction. This practice has been followed in 
 several important works, and the data thus accumulated have 
 added much to our hitherto somewhat vague notions of the 
 probable strength of different mixtures under varying condi- 
 
292 
 
 CEMENT AND CONCRETE 
 
 tions of use. It is possible here to abstract but a few of the 
 more reliable and complete tests of this kind, selecting those 
 which indicate the value of certain special kinds of aggregate 
 or the effect of certain variations in manipulation. 
 
 422. In connection with the design of the Boston Elevated 
 R. R. 7 Mr. George A. Kimball, Chief Engineer, prepared a series 
 of concrete cubes of mixtures usually employed in practice, 
 and with the materials available for the work in hand, and these 
 cubes were tested at the Watertown Arsenal in 1899. A por 
 tion of the results of these experiments are given in Table 130, 
 where the details concerning character of the materials and 
 the preparation of the specimens are shown. As each result 
 in the table is the mean of at least twenty specimens, the ir- 
 regularities frequently appearing in compressive tests have 
 been largely eliminated, and the results are worthy of much 
 confidence. 
 
 TABLE 130 
 
 Compressive Strength of Concrete 
 TESTS OF 12 INCH CONCRETE CUBES FOR BOSTON ELEVATED RAILROAD. 
 
 COMPOSITION OF CONCRETE BY 
 VOLUME. 
 
 CRUSHING STRENGTH, POUNDS PER SQUARE INCH, 
 AT AGE, 
 
 Cement. 
 
 Sand. . 
 
 Stone. 
 
 7 days. 
 
 1 month. 
 
 3 mouths. 
 
 6 months. 
 
 1 
 1 
 1 
 
 2 
 
 3 
 
 
 
 4 
 6 
 12 
 
 1525 
 1232 
 583 
 
 2440 
 2063 
 1042 
 
 2944 
 2432 
 1006 
 
 3904 
 29(>9 
 1313 
 
 NOTES: 
 
 Materials: Cement, mean results with four brands Portland, two Ameri- 
 can, two German. 
 
 Sand, coarse, clean, sharp, voids 33 per cent, loose. 
 Broken stone, conglomerate passing 2^ inch ring, voids 49| 
 
 per cent, loose. 
 
 Mixing : Sand and cement turned twice, mortar and stone turned twice. 
 Storage : Cubes removed from molds three or four days after made and 
 
 buried in wet ground until about a week before testing. 
 Each result, mean of twenty or more tests. 
 
 Tests made at Watertown Arsenal, for George A. Kimball, Chief Engineer, 
 Boston Elevated R.R. "Tests of Metals," 1899. 
 
 At the time of making these tests some cubes were crushed 
 with a die having a smaller area than the face of the cube. 
 
CONCRETE 293 
 
 With a die 8 by 8^ inches on one compression face, the area of 
 the die being thus about .46 of the area of the cube face, the 
 strength per square inch under the die was about twenty-five 
 per cent, higher than when the entire face of the cube was 
 pressed. This is in line with the behavior of all brittle sub- 
 stances under compression, as shown by Professor Bauschinger 
 in testing sandstone specimens. 
 
 423. Tables 131 and 132 give a summary of a part of a very 
 valuable series of tests of concrete cubes prepared by Mr. George 
 W. Rafter and tested at the Watertown Arsenal for the State 
 Engineer of New York. 1 
 
 The results summarized in Table 131 are those obtained with 
 four brands of Portland cement made in the State of New 
 York, namely, Wayland, Genessee, Empire and Ironclad. Tests 
 were also made with a sand-cement, and with one brand of 
 natural, but these results are not included in the table. The 
 aggregate was sandstone of the Portage group, broken by hand 
 to pass a two inch ring. 
 
 The mortars used in making the cubes were of three degrees 
 of consistency: (a) In the dry est blocks the mortar was only a 
 little more moist than damp earth, and much ramming was 
 required to flush water to the surface. (6) In another set the 
 mortar was about the consistency of ordinary mason's mortar, 
 (c) In the third set, the mortar was wet enough to quake like 
 liver under moderate ramming. 
 
 424. The mortar was composed of one volume of loose ce- 
 ment to two, three or four volumes of loose sand. Other pro- 
 portions were also employed, but in this table only those re- 
 sults are included in which the series of tests was complete as 
 to variations in consistency and storage. 
 
 The voids in the stone were about forty-three per cent, 
 when the measure was slightly shaken, and thirty-seven and 
 a half per cent, when rammed without mortar. The amount 
 of mortar used was made either thirty-three per cent, or forty 
 per cent, of the volume of the loose stone. 
 
 Four methods of storage were used as follows: 1st, blocks 
 immersed in water as soon as they were removed from the 
 molds, and after three or four months they were buried in sand; 
 
 1 Report of State Engineer of New York, 1897. 
 
294 
 
 CEMENT AND CONCRETE 
 
 TABLE 131 
 
 Compressive Strength of Concrete 
 
 MEAN RESULTS WITH FOUR BRANDS PORTLAND CEMENT, ILLUSTRATING EFFECTS 
 
 OF PROPORTIONS, CONSISTENCY, AND METHODS OF STORAGE. TESTS OF 
 
 CONCRETE CUBES, ABOUT TWENTY MONTHS OLD, MADE FOR 
 
 STATE ENGINEER OF NEW YORK 
 
 
 
 <* >.o 
 
 05 O5 
 
 CM CM "* CO 
 
 ^2212 
 CM CM CM CM 
 
 t- 
 OS 
 
 CM O CM t* *^H CO OS 
 
 CD T-H O CM CO T-H CD 
 
 O O T-H T-H O O O 
 
 CM CM CM CM CM CM CM 
 
 T-H CM tO 
 
 co o: TJH 
 
 CO T-H r-H 
 
 CM CM CN 
 
 OS 00 ^ 
 
 ,. CO CM OS 
 
 OS OS !> 
 
 (M OS >O OS 
 
 r-H t- (M CO 
 CO CO CO 
 
 OS 5 
 
 2 
 
 t^ -<ti i-O <M 
 CO CO COt- 
 
 ^ Tj< CM t- 
 
 T-H O CO T-H 
 
 CO CD CO 
 
 CM * 00 
 
 CO tO T-H 
 
 CO O CO CM 
 t- CO O t~- 
 
 O OS OS OS 
 
 OtOCOO 
 tO -* CM Tfl 
 (M CM CM CM 
 
 alll 
 
 tO CO CO CM 
 CO T-H T-H t"~- 
 
 
 CM rJH T-H. O5 
 CO Cb T-H t>- 
 
 CO T-H <N 
 CO CO CO CO 
 
 t^ T-H CO O5 
 tO O5 -* O5 
 iO O CD O 
 <M (?1 <N <N 
 
 co os oo 
 
 id CM 00 
 O O Ttl 
 CO CM CM 
 
 to CO OS t^- 
 
 Tt* to 
 CM CM 
 
 tO 
 CM 
 
 ^ 
 
 T-H 
 
 OS 
 
 ^ |c^ 
 t- O 
 CM ICO 
 
 tO to 
 GC t^- 
 
 CM CM 
 
 H a 
 z ^ 
 w < 
 
 
 " 
 
 
 a s 
 
 ^ "S 
 
 s a 
 
CONCRETE 295 
 
 2d, blocks covered with burlap and wet frequently for several 
 weeks, after which they were exposed to the weather; 3d, kept 
 in a cool cellar from the time of fabrication until shipped for 
 testing, and 4th, fully exposed to the weather throughout. 
 
 425. In Table 131 each result is the mean of four cubes, one 
 of each brand. The mean results are so arranged as to show 
 the effects of variations in the amount, the richness, and the 
 consistency of the mortar, and of the different methods of storage. 
 
 Taking up first the question of consistency, it appears from 
 column "/" that the use of plastic mortar, marked " mason's," 
 gave from 92 to 97 per cent, of the strength given by the dry 
 mortar of about the consistency of "moist earth;" and that 
 the "quaking" concrete gave from 89 to 95 per cent, of the 
 strength of that marked "moist earth." From the three lines 
 at the bottom of the table it is seen that in the poor concrete, 
 one-to-four mortar, the wettest mortar gave nearly as good 
 results as the dryest, while in the rich concrete, one-to-two 
 mortar, the strength of the wet was but 89 per cent, of the dry. 
 The explanation of this may be found in the fact that in the 
 poor concrete the mortar was "brash," and the concrete did 
 not ram well with a dry mortar, while the rich mortar was 
 "fuller" and more plastic, so that the excess of water was not 
 needed to make a compact mass. 
 
 426. Turning to the question of the amount of mortar, it is 
 plainly shown that the concrete containing forty per cent, is 
 but little better than that containing thirty-three per cent. 
 This is in line with what has been said elsewhere, that an excess 
 of mortar, as well as a deficiency, may be an actual detriment 
 to the strength of the concrete. In this case the thirty-three 
 per cent, mortar was not quite sufficient to fill the voids in 
 the stone, and forty per cent, was a very slight excess. 
 
 Some interesting conclusions are indicated by the results in 
 the line marked "ratios," near the bottom of the table. The 
 ratios of the strength of the concrete containing thirty-three 
 per cent, mortar to the strength of that containing forty per 
 cent, are 91.6 per cent., 98.5 per cent, and 102.6 per cent., re- 
 spectively, for one-to-two, one-to-three, and one-to-four mor- 
 tars. That is, with a rich mortar forty per cent, may be used 
 to advantage, but if the mortar is of poor quality, the strength 
 of the concrete is not increased by an excess of rnortar. 
 
CEMENT AND CONCRETE 
 
 Finally, as to the strength developed under different con- 
 ditions of storage, column " k " shows that for these cements the 
 highest strengths are attained by immersing the concrete in 
 water. In comparison, the strength developed by the concrete 
 covered with wet burlap is 84 per cent. ; in cool cellar, 82 per cent. ; 
 and in the open air fully exposed to the weather, 81 per cent. 
 
 427. The results given in Table 132 are the mean crushing 
 strengths obtained in the same series of tests as described above, 
 so arranged as to bring out the effect of the richness of the 
 mortar. Although several brands were tested, only the results 
 obtained with a single brand of Portland, namely, Milieu's 
 "Wayland," are included here, since the series was not com- 
 pleted with other brands. From similar tests with concretes 
 containing one-to-two and one-to-three mortars only, it was 
 found that three other brands of Portland gave from 91 to 102 
 per cent, of the strength obtained with the Wayland, and a 
 brand of sand-cement gave 66 per cent. 
 
 TABLE 132 
 
 Compressive Strength of Concrete. Effect of Richness of Mortar 
 MEAN RESULTS, FOUR METHODS or STORAGE 
 
 
 
 MORTAR, PROPORTIONS CEMENT TO SAND. 
 
 VOLUME MOR- 
 
 
 
 TAR \8 
 
 CONSISTENCY 
 
 
 
 
 
 
 PER CENT. OF 
 
 OF 
 
 1-1 
 
 1-2 1-3 
 
 1-4 
 
 1-5 
 
 VOLUME OF 
 
 CONCRETE. 
 
 
 
 
 
 
 AGGREGATE. 
 
 
 
 
 
 Crushing Strength, Lbs. per Sq. In. 
 
 ( 
 
 Moist Earth . 
 
 4267 
 
 2888 
 
 2056 
 
 1810 
 
 1537 
 
 33 \ 
 
 Mason's . . . 
 
 4072 
 
 2777 
 
 2207 
 
 1600 
 
 1568 
 
 I 
 
 Quaking . . . 
 
 3764 
 
 2847 
 
 1723 
 
 1767 
 
 1441 
 
 ( 
 
 Moist Earth . 
 
 3966 
 
 3404 
 
 2179 
 
 1671 
 
 1559 
 
 40 I 
 
 Mason's . . . 
 
 4123 
 
 2960 
 
 2027 
 
 1750 
 
 146o 
 
 I 
 
 Quaking . . . 
 
 3256 
 
 3168 
 
 2016 
 
 1670 
 
 1400 
 
 
 Mean 
 
 3908 
 
 3007 
 
 2035 
 
 1711 
 
 1495 
 
 
 
 
 
 
 
 
 Proportional . 
 
 100 
 
 77 
 
 52 
 
 44 
 
 38 
 
 NOTES: One brand Portland cement. 
 
 Aggregate, Portage sandstone, broken to pass two-inch ring. 
 Age of cubes about twenty months. 
 Each result, mean of four cubes. 
 
CONCRETE 
 
 297 
 
 Each result in the table is the mean of four cubes, each 
 stored in a different manner. Tests with four brands (Table 
 131) where the concretes were made with one-to-two, one-to- 
 three and one-to-four mortars, indicated that the percentages of 
 the mean strength developed in the several methods of storage 
 were as follows: If stored in water, the cubes developed 115 per 
 cent, of the mean result; covered with burlap kept wet, the 
 cubes developed 97 per cent. ; stored in a cool cellar, 95 per 
 cent.; and fully exposed to weather, 93 per cent, of the mean 
 strength. 
 
 The mean results given at the bottom of the table represent 
 each a mean of twenty-four cubes made with two different 
 amounts of mortar, three degrees of consistency, and four 
 methods of storage. By applying the percentages given above 
 the probable corresponding result for any set of conditions 
 may be obtained. The last line of the table shows the propor- 
 tions that the strength of the concretes made with poorer mor- 
 tars, bear to the strength obtained with one-to-one mortar. 
 
 428. Table 133 gives the results of a series of tests made by 
 J. W. Sussex at the University of Illinois. 1 The materials used 
 were ''Chicago AA Portland cement, sand containing a small 
 
 TABLE 133 
 
 Compresaive Strength of Concrete. Relative Strength of Dry, 
 Medium, and Wet Mixtures 
 
 
 
 TENSILE STRENGTH, POUNDS PER 
 
 
 
 
 SQUARE INCH, AT AGE OP 
 
 PROPOR- 
 
 CONSISTENCY. 
 
 TAMPING. 
 
 
 TIONAL VALUK 
 
 
 
 
 
 
 AT 3 Mo*. 
 
 
 
 7 days. 
 
 1 month. 
 
 3 months. 
 
 
 Dry . . 
 
 Light 
 
 1200 
 
 1750 
 
 2500 
 
 82 
 
 Medium . . 
 
 u 
 
 2290 
 
 221)0 
 
 2150 
 
 71 
 
 Wet . . 
 
 tt 
 
 1040 
 
 2230 
 
 3040 
 
 100 
 
 Dry . . 
 
 Hard 
 
 1340 
 
 1960 
 
 2000 
 
 86 
 
 Medium . . 
 
 u 
 
 1330 
 
 2565 
 
 2580 
 
 85 
 
 NOTES: Concrete composition: 
 
 Cement, Portland, one volume. 
 
 Sand, containing some fine gravel, three volumes. 
 
 Six volumes broken limestone passing one-inch mesh. 
 Specimens, six-inch cubes. 
 Results by J. W. Sussex, Univ. of 111. 
 
 Technograph, 1902-03. 
 
298 CEMENT AND CONCRETE 
 
 percentage of fine gravel, and crushed limestone which would 
 pass through a sieve with one-inch mesh." The proportions 
 were three parts sand and six of broken stone to one volume of 
 loose cement. The cubes were six inches on a side. The 
 treatment during storage is not stated. The consistency of 
 the concrete was as follows: "Dry," water 6.0 per cent., as 
 moist as damp earth, no free water flushed to surface in ram- 
 ming. "Medium/' 7.8 per cent, water; water flushed to surface 
 and concrete quaked only after being well rammed. "Wet/' 
 water 9.4 per cent., concrete quaked in handling and could be 
 tamped but lightly. 
 
 Each result in the table is the mean of three cubes. The 
 concrete was tamped in layers about one inch thick with a 
 rammer weighing 11^ pounds and dropped six inches. Ten 
 blows of the rammer constituted "light" tamping and twenty 
 blows "hard" tamping. The results show that the "medium" 
 concrete gains its strength more rapidly than the "wet," but 
 that at one month the "wet" concrete has a higher strength 
 than the dry, and that at three months the wet surpasses in 
 strength both the dry and the medium. 
 
 ART. 54. CONCRETES WITH VARIOUS KINDS AND SIZES OF 
 AGGREGATES 
 
 429. It has already been stated that the character of the 
 aggregates is second only to the quality of the mortar in its 
 effect on the strength of concrete. The materials available for 
 aggregate in different localities are so varied that only a general 
 idea of their relative values may be obtained from a limited 
 number of tests. 
 
 The results given in Table 134 are from tests made at the 
 Watertown Arsenal/ and show the compressive strengths of 
 concretes made with broken trap and gravel of different sizes. 
 The concretes are all very rich, and the strengths correspond- 
 ingly high, although the oldest specimens have hardened less 
 than three months. The results are somewhat irregular, and 
 the conclusion to be drawn concerning the best size for the 
 aggregate is not very clearly brought out. The one-inch trap 
 gives uniformly good results, as do the mixtures of two or 
 
 1 "Tests of Metals/' 1898. 
 
 
CONCRETE 
 
 299 
 
 more sizes. The trap rock gives a higher result than the gravel, 
 the mortar being sufficient to fill the voids in the trap, and in 
 excess for the gravel. 
 
 TABLE 134 
 
 Compressive Strengths of Rich Concretes at Different Ages 
 TESTS OF TWELVE INCH CTBES 
 
 
 WT. PKK 
 
 
 
 CU.FT.OF 
 
 C'oMPKKMHivn STRENGTH, POUNDS IKR 
 
 
 ('ONTKKTK 
 
 SQUARE INCH, AT AGE, DAYS, 
 
 CHARACTER OF AGGREGATE. 
 
 WHKN 
 
 
 
 Mo. OLD, 
 IN LBS. 
 
 7-8 
 
 19-23 
 
 29-34 
 
 61-76 
 
 Trap \" 
 
 148 6 
 
 1391 
 
 2220 
 
 2800 
 
 5021 
 
 y 
 
 148 5 
 
 1900 
 
 2760 
 
 3200 
 
 
 
 150.8 
 
 3800 
 
 4254 
 
 4917 
 
 5272 
 
 \\" 
 
 150 2 
 
 3180 
 
 4000 
 
 4662 
 
 2583 
 
 2J" 
 
 160.2 
 
 2400 
 
 4143 
 
 4140 
 
 4523 
 
 i"-i, 2r-2. . . . 
 
 158.4 
 
 2800 
 
 3786 
 
 4340 
 
 4544 1 
 
 *"-!, 1"-1, 2}"-l 
 
 150.8 
 
 2800 
 
 4156 
 
 4800 
 
 5542 
 
 Mean results, trap rock alone 
 
 
 2553 
 
 3619 
 
 4110 
 
 4581 
 
 Pebbles I" . 
 
 148.2 
 
 1208 
 
 2600 
 
 2002 
 
 3870 
 
 " H" 
 
 161.0 
 
 2276 
 
 3186 
 
 3817 
 
 4018 
 
 " f'-l, lJ"-2 . . . 
 
 150.3 
 
 1004 
 
 3023 
 
 3800 
 
 345M) 
 
 " t"-i, '-!, ij"-i 
 
 147.8 
 
 1480 
 
 2676 
 
 3000 
 
 3800 
 
 Mean, pebbles alone . 
 
 
 1764 
 
 2871 
 
 3402 
 
 3794 
 
 NOTES: Tests made at Watertown Arsenal, "Tests of Metals," 1898. 
 
 All concretes composed of one cubic foot of Alpha Portland cement, 
 weight 06i to 106 Ibs. per cu. ft., one cu. ft. of bank sand, 
 weight 93^ to 104 Ibs. per cu. ft., and 3 cu. ft. of aggregate, 
 weighing from 93 to 105 Ibs. per cu. ft. 
 
 The size of aggregate indicated gives the larger of the two screens 
 used in separating it into different sizes; thus, " f inch" 
 means passing f inch mesh and retained on \ inch mesh. 
 
 The compressive strength of twelve inch cubes of one-to-one mor- 
 tar alone was 3,833 Ibs. per sq. in. at seven days, and 4,800 
 Ibs. per sq. in. at seventy-five days. 
 
 430. In 1896-97 Mr. A. W. Dow 2 prepared a number of 
 twelve-inch cubes of concrete for the Engineer Commissioner 
 
 1 Not fractured. 
 
 2 Report Operations, Engineer Department, District of Columbia, 1897. 
 Also Baker's "Masonry Construction," p. 112 r. 
 
300 
 
 CEMENT AND CONCRETE 
 
 of the District of Columbia. These cubes are of interest as 
 showing the strength of natural cement concrete as well as 
 Portland, and the results are abstracted in Table 135. 
 
 TABLE 135 
 
 Compressive Strength of Concrete 
 
 TESTS OF TWELVE-INCH CUBES FOR THE ENGINEER COMMISSIONER OF THE 
 DISTRICT OF COLUMBIA 
 
 REF. 
 
 COMPOSITION OF CONCRETES. 
 
 PER 
 CENT. 
 
 VOIDS IN 
 AGGRE- 
 GATE. 
 
 CRUSHING 
 STRENGTH, LBS. 
 PER SQVAKE 
 INCH AT ONE 
 YEAR. 
 
 Cement. 
 
 Sand. 
 
 Broken Stone. 
 
 Gravel. 
 
 Port- 
 land. 
 
 Natural. 
 
 Coarse. 
 
 Average. 
 
 Average. 
 
 Small. 
 
 1 
 2 
 3 
 4 
 5 
 6 
 
 1 
 1 
 1 
 1 
 1 
 1 
 
 2 
 2 
 2 
 2 
 2 
 2 
 
 6 
 
 
 
 
 45.3 
 45.3 
 30.5 
 29.3 
 35.5 
 36.7 
 
 1850 
 3060 
 2700 
 2820 
 2750 
 2840 
 
 829 
 915 
 800 
 763 
 841 
 915 
 
 6 
 6 1 
 
 ' 3' ' 
 4 
 
 
 
 
 
 
 
 3 
 2 
 
 NOTES: Materials: 
 
 Cement, Portland, "Atlas" (American), 104 Ibs. per cu. ft.; Natural, 
 
 "Round Top," 70 Ibs. per cu. ft. 
 Sand, 15 per cent, retained on No. 8 mesh, 75 per cent, between 8 and 
 
 40 mesh, 10 per cent, passing 40 mesh. Sand was used damp, and 
 
 weighed in that condition 90 Ibs. per cu. ft. 
 
 Stone, Bluestone, "Average," 93 percent, between ^ inch and 2 inches. 
 "Coarse," 89 per cent, between 1| inches and 2$ inches. 
 Gravel, "Average," 90 per cent, between inch and 1 inches. 
 
 "Small," 90 per cent, between inch and f inch. 
 Granolithic, 92 per cent, between -^ inch and inch. 
 Mixing, thorough by experienced man. 
 
 Tamping, light, in 4 inch layers, just sufficient to bring mortar to surface. 
 Storage, cubes thoroughly wet twice a day. 
 Age of specimens when broken, one year. 
 
 The concretes all contained two parts sand and six parts 
 aggregate to one cement, but the character of the aggregate 
 varied as shown. The natural cement concrete gave from one- 
 quarter to one-third the strength of the Portland concrete. 
 The best result seems to be given by the average size broken 
 stone, which was in reality a mixture of various sizes, ninety- 
 
 Mixture of one part granolithic size to one of concrete stone. 
 
CONCRETE 
 
 301 
 
 three per cent, of it being retained on a one-third inch mesh 
 and passing a two-inch mesh. The mortar was probably in- 
 sufficient to fill the voids in the stone for the first three cubes 
 in the table, and under these conditions the gravel, with its 
 smaller percentage of voids, makes a good showing. This illus- 
 trates what we have already said, that the relative value of 
 broken stone and gravel for aggregate depends upon the pro- 
 portion of mortar used. 
 
 TABLE 136 
 
 Compressive Strength of Concrete. Portland Cement 
 TESTS OF SIX-INCH CUBES OF VARIOUS MIXTURES 
 
 UKFKHENCE. 
 
 PARTS BY VOLUME TO 
 ONE CEMENT. 
 
 CRUSHING STRENGTH, POUNDS PER SQUARE 
 INCH, AT AGE OF, 
 
 Sand. 
 
 Gravel. 
 
 Broken 
 Stone. 
 
 7 days. 
 
 30 days. 
 
 90 days. 
 
 1 
 
 
 
 
 
 
 
 3412 
 
 5318 
 
 6140 
 
 2 
 
 
 
 3 1 
 
 
 
 1077 
 
 1908 
 
 2517 
 
 3 
 4 
 
 1 
 2 
 
 2 
 2 
 
 i* 
 
 1430 
 420 
 
 2215 
 2117* 
 
 2903 
 1324 
 
 5 
 
 2 
 
 3 
 
 4 
 
 640 
 
 1199 
 
 1290 
 
 6 
 
 7 
 
 l\ 
 
 5 
 
 
 
 5 
 
 566 
 739 
 
 1385 
 2033 
 
 1609 
 1783 
 
 8 
 
 2 
 
 2 
 
 21 
 
 792 
 
 1482 
 
 2014 
 
 9 
 
 3 
 
 o 
 
 o 
 
 767 
 
 1345 
 
 1409 
 
 10 
 
 3 
 
 3 
 
 4 
 
 714 
 
 ' 1028 
 
 1818 
 
 Means, Actual 
 Means, Per Cent 
 
 
 1056 
 46 
 
 2003 
 88 
 
 2284 
 100 
 
 
 NOTE : Results of Messrs. Ketchum and Honens. 
 
 431. The results in Table 136 were obtained by Messrs. R. 
 B. Ketchum and F. W. Honens at the laboratory of the Uni- 
 versity of Illinois, 3 and illustrate the rate of gain in strength 
 of several mixtures. The cement used was Baylor's Portland, 
 fine and of good quality. The sand and gravel were composed 
 principally of silica, with 10 to 30 per cent, of limestone. About 
 60 per cent, of the sand passed a " number thirty" sieve. The 
 unscreened gravel had about 42 per cent, caught on a " num- 
 ber five" sieve and eighteen per cent, of it passed a "number 
 
 1 Unscreened. 
 
 2 Result irregular. 
 
 3 Technograph, 1897-98. 
 
302 CEMENT AND CONCRETE 
 
 thirty." Except in one mixture, however, the gravel and 
 broken stone were screened, and only that portion passing a 
 two-inch ring and retained on a " number five" sieve was used. 
 The stone was a magnesian limestone. 
 
 The concrete was mixed dry, so that considerable tamping 
 was required to bring water to the surface. The cubes were 
 first kept under a damp cloth for one day, immersed six days, 
 and then stored in air in a room until broken. In crushing; 
 "the direction of the force applied was parallel to the tamped 
 surface." 
 
 432. Each result in the table is the mean of six specimens. 
 Comparing number 2 with number 9 indicates that the strength 
 obtained with one part cement to three parts unscreened gravel 
 is much higher than with mortar of one part cement to three 
 parts sand. Comparing 9 and 10 indicates that seven parts 
 gravel and stone may be mixed with one-to-three mortar and 
 give higher strength than the mortar alone. A comparison of 
 6, 7, and 8 shows that in case there is sufficient mortar to fill 
 the voids in the aggregate, angular fragments give a somewhat 
 higher strength than rounded ones, but that a mixture of broken 
 stone and gravel is better than either alone. One of the most 
 important points brought out by the tests is that the strength 
 at seven days is 46 per cent., and at thirty days is 88 per cent., 
 of the strength attained at three months. 
 
 ART. 55. CINDER CONCRETE, ETC. 
 
 433. For such purposes as floors for buildings, cinders are 
 used in concrete to a considerable extent on account of their 
 light weight. Cinder concrete weighs only from two-thirds to 
 three-fourths as much as broken stone or gravel concrete. 
 The strength, however, is correspondingly less, and whether for 
 a given strength a floor may be made lighter by the use of 
 cinders will depend upon the conditions of use and the charac- 
 ter of the reinforcement. 
 
 Table 137 gives the results of the tests of eight-inch cylinders, 
 fifteen inches high, made by Mr. George Hill. 1 In these cylin- 
 ders, cinders, broken stone, and gravel were used as aggregates. 
 The character of the materials is shown in the foot-note of the 
 
 Trans. Am. Soc. C. E., Vol. xxxix, p. 632, 
 
CINDER CONCRETE 
 
 303 
 
 table. As the specimens were but one month old when tested, 
 the results are low, but since in the construction of floor arches 
 the centers are usually removed in less than one month, the 
 strength developed in a short time has a special interest. 
 
 TABLE 137 
 
 Compressive Strength of Concrete about One Month Old 
 TESTS OF CYLINDERS, EIGHT INCHES DIAMETER, FIFTEEN INCHES HIGH 
 
 
 PROPORTIONS BY VOLUMK. 
 
 AGE, 
 
 COMPRESSIVE STRENGTH, 
 LBS. PER SQ IN. 
 
 AGGREGATE. 
 
 Cement. 
 
 Sand. 
 
 Aggregate. 
 
 Days. 
 
 American 
 Portland 
 Cement. 
 
 Slag 
 Cement. 
 
 Cinders. 
 
 
 .3 
 
 6 
 
 33 
 
 246 
 
 
 
 
 
 3 
 
 6 
 
 18 
 
 292 
 
 
 
 
 
 2 
 
 5 
 
 33 
 
 305 
 
 
 it 
 
 
 2 
 
 5 
 
 33 
 
 4(J4 
 
 
 it 
 
 
 2 
 
 5 
 
 32 
 
 490 
 
 
 K 
 
 
 2.4 
 
 6 
 
 32 
 
 590 
 
 
 i( 
 
 
 1.7 
 
 4.2 
 
 30 
 
 
 342 
 
 (t 
 
 
 1.8 
 
 4 
 
 30 
 
 
 330 
 
 (I 
 
 
 1.8 
 
 4 
 
 31 
 
 
 766 
 
 (( 
 
 
 1.8 
 
 4 
 
 31 
 
 
 765 
 
 Stone. 
 
 
 3 
 
 6 
 
 30 
 
 398 
 
 
 " 
 
 
 2.4 
 
 4.1 
 
 30 
 
 503 
 
 
 u 
 
 
 2.4 
 
 4 
 
 33 
 
 
 645 
 
 u 
 
 
 2.4 
 
 4 
 
 30 
 
 
 730 
 
 Gravel. 
 
 
 3 
 
 6 
 
 30 
 
 oi7 
 
 618 
 
 
 
 1 
 
 2.4 
 
 4.8 
 
 30 
 
 
 650 
 
 u 
 
 1 
 
 2 
 
 7 
 
 25 
 
 880 
 
 
 (1 
 
 1 
 
 1.8 
 
 6.5 
 
 31 
 
 
 730 
 
 Stone and gravel, ) 
 graded . . . . \ 
 
 1 
 
 2 
 
 10 
 
 30 
 
 625 
 
 ... 
 
 NOTES: 
 
 Cement, American Portland, tensile strength 624 Ibs. per sq. in., neat, seven 
 days. 
 
 Slag cement, a little less than 400 Ibs. per sq. in., neat, seven days. 
 Sand, clean, sharp, bank sand of mixed sizes, from moderately fine up to 
 
 some pebbles size of bean. 
 Cinders, ordinary steam, dust to inch size. 
 Stone, broken trap, nearly uniform size passing l\ inch ring. 
 Gravel, clean, washed, \ in. to 1^ in. 
 Abstract of tests by Mr. George Hill, M. Am. Soc. C. E., Vol. xxxix, p. 632. 
 
 It is evident that cinder concrete should not be loaded very 
 heavily within a month after made. The gravel gives a better 
 result than broken stone. 
 
304 
 
 CEMENT AND CONCRETE 
 
 434. In Table 138 are given the results of some tests of 
 twelve-inch cubes of cinder concrete made at the Watertown 
 Arsenal for the Eastern Expanded Metal Companies. Steam 
 cinders were used, practically as they came from the furnace, 
 only the larger clinkers being broken. Two proportions were 
 used and the specimens were broken at one month and three 
 months. It is seen that the one-one-three mixture is about 
 twice as strong as the one-two-five with all brands. The varia- 
 tions between the several brands are also very great. 
 
 TABLE 138 
 
 Crushing Strength of Cinder Concrete. Portland Cement 
 TESTS OF TWELVE-INCH CUBES AT WATERTOWN ARSENAL 
 
 BRAND 
 
 OF 
 
 CEMENT. 
 
 STRENGTH, POUNDS PER SQUARE INCH. 
 
 Mixture A, 1-1-3. 
 
 Mixture B, 1-2-5. 
 
 Age of Specimens. 
 
 Age of Specimens. 
 
 1 month. 
 
 3 months. 
 
 1 month. 
 
 3 months. 
 
 A 
 B 
 C 
 D 
 
 2329 
 1602 
 1438 
 1032 
 
 2834 
 2414 
 1890 
 1393 
 
 940 
 696 
 744 
 471 
 
 1600 
 1223 
 880 
 685 
 
 NOTES: 
 
 Concretes mixed rather dry, 10 to 12J pounds of water per cubic foot of 
 
 concrete. 
 
 Mixture "A," one part cement, one part sand, three parts cinders. 
 Mixture "B," one part cement, two parts sand, five parts cinders. 
 Weight of concrete, 104 to 116 pounds per cubic foot. 
 Tests for Eastern Expanded Metal Companies. Data from "Tests of 
 Metals/' 1898. 
 
 435. Table 139 gives the results of other tests in the same 
 series, using a single brand of cement and five mixtures, the 
 richest containing three parts cinders and one part sand to one 
 volume cement, and the poorest six parts cinders and three 
 parts sand to one cement. The weight per cubic foot of the 
 several concretes is also given. 
 
 Tests of cinder concrete prisms made by the late Prof. J. B. 
 Johnson at Washington University * indicated that the mixture 
 
 Materials of Construction," p. 628. 
 
CONCRETE WITH CLAY 
 
 305 
 
 containing one part sand and three parts cinders to one volume 
 cement gave the highest strength, or about twelve hundred 
 pounds per square inch, at one month. The same mixture gave 
 the highest values for the ratios of strength to cost, and of 
 strength to weight per cubic foot. 
 
 TABLE 139 
 
 Crushing Strength of Cinder Concrete. Various Proportions with 
 
 Germaiiia Portland Cement 
 TESTS OF TWELVE-INCH CUBES AT WATERTOWN ARSENAL 
 
 PROPORTIONS IN CONCRETE. 
 
 WEIGHT PER 
 Cu. FT. AT 98 
 TO 102 
 DAYS, POUNDS. 
 
 CRUSHING STRENGTH, 
 POUNDS PER SQUARE INCH, 
 AT AGE, 
 
 Cement. 
 
 Sand. 
 
 Cinders. 
 
 29 to 39 days. 
 
 98 to 102 days. 
 
 1 
 1 
 1 
 1 
 1 
 
 1 
 2 
 2 
 2 
 3 
 
 3 
 3 
 4 
 5 
 6 
 
 110.4 
 112.8 
 107.9 
 105.3 
 103.5 
 
 1406 
 1008 
 904 
 760 
 529 
 
 2001 
 1634 
 1325 
 1084 
 788 
 
 NOTE: Tests for Eastern Expanded Metal Companies, "Tests of Metals," 
 1898. 
 
 436. Clay in Concrete. The effect of clay on the tensile 
 strength of mortars has already been shown (Art. 49). Aggre- 
 gates available for concrete frequently contain a certain amount 
 of clay, and the question arises whether such aggregate must 
 be washed, or whether certain small percentages may be per- 
 mitted in the concrete, using, perhaps, a trifle richer mortar. 
 The results in Table 140 were made to determine the. effect of 
 clay on the crushing strength of concrete. 1 
 
 The test specimens were six-inch cubes, and were broken 
 when one week to twelve weeks old in an Olsen machine. The 
 proportions were two parts sand and six parts gravel by weight 
 to one of Portland cement, or two parts sand and four parts 
 gravel by weight to one of natural cement. The clay is ap- 
 parently expressed as the per cent, of total aggregates. It is 
 seen that while six or twelve per cent, clay retards the harden- 
 ing of both Portland and natural cement concrete, the strength 
 of the Portland concrete after four weeks is increased by six per 
 
 1 Tests by Messrs. J. J Richey arid B. H. Prater, Technograph, 1902-03. 
 
306 
 
 CEMENT AND CONCRETE 
 
 cent, clay, while at the same age the strength of the natural 
 cement concrete is not greatly affected. The ramming of con- 
 crete is facilitated by the presence of a small amount of clay, 
 but larger amounts may render the mass sticky and difficult 
 to ram. 
 
 TABLE 140 
 
 Effect of Clay on Crushing Strength of Concrete 
 SIX-INCH CUBES 
 
 CEMENT. 
 
 PROPORTIONS BY 
 WEIGHT, No. PARTS 
 TO ONE CEMENT. 
 
 AGE OF CUBES 
 
 WHEN 
 
 CRUSHING STRENGTH, POUNDS PER SQ. 
 IN.; CLAY AS PER CENT. OF CONCRETE, 
 
 
 Sand. 
 
 Gravel. 
 
 BROKEN. 
 
 
 
 6 
 
 12 
 
 Port. 
 
 2 
 
 6 
 
 1 week 
 
 1030 
 
 1001 
 
 692 
 
 ti 
 
 2 
 
 6 
 
 4 weeks 
 
 1398 
 
 1525 
 
 1287 
 
 u 
 
 2 
 
 6 
 
 12 " 
 
 2110 
 
 2760 
 
 1865 
 
 Nat. 
 
 2 
 
 4 
 
 1 week 
 
 208 
 
 131 
 
 81 
 
 u 
 
 2 
 
 4 
 
 4 weeks 
 
 428 
 
 364 
 
 283 
 
 u 
 
 2 
 
 4 
 
 12 " 
 
 786 
 
 722 
 
 480 
 
 ART. 56. THE MODULUS OF ELASTICITY OF CEMENT MORTAR 
 
 AND CONCRETE 
 
 437. With the increasing use of concrete and steel in com- 
 bination, the modulus of elasticity of cement mortar and con- 
 crete assumes a new importance, since the ratio of the stresses 
 in the two materials depends upon the relative moduli of elas- 
 ticity. Some of the earlier determinations of the modulus of 
 mortar gave very high values. This may have been due to the 
 use of richer mixtures, and the exercise of greater care in the 
 manipulation, than are employed in actual construction, and 
 also to the fact that the determinations were based upon the 
 deformations resulting from the application of very limited 
 loads. 
 
 It is now considered that the ratio of stress to strain is not 
 constant, even for moderate loads, but that the modulus of 
 elasticity decreases with increasing stress, and this fact is brought 
 out in the following tables. The tests cited bring out a wide 
 range of values for concretes and mortars made from a variety 
 of sand and aggregate and of various compositions and ages. 
 
 438. Modulus of Elasticity of Natural and Portland Cement 
 Mortars. Table 141 gives the modulus of elasticity of mortars 
 as determined by tests of twelve-inch cubes at the Watertown 
 
MODULUS OF ELASTICITY 
 
 307 
 
 
 T 
 
 
 8 
 
 00 "N 
 
 
 
 g 
 
 100-600 
 
 .... iO 
 
 u 
 S 
 p 
 
 
 er Square ! 
 
 1 
 
 : : : : : 
 
 
 
 2 
 
 Pounds i> 
 
 100-1000 
 
 1 : : : 1 : 
 
 vxaoj^ N 
 
 
 J 
 ^ 
 
 I 
 
 IPS i ; 
 
 
 z 
 
 CO 
 
 g 
 
 
 between ] 
 
 j 
 
 ;;:n 
 
 u 
 
 s 
 
 2 
 
 "s 
 3 
 = 
 
 j 
 
 11*1 11 
 
 o 
 
 
 
 
 
 
 
 3RTIONS 
 
 
 'S 
 
 100-600 
 
 nsss 3 
 
 ~t -N ?q or o c: 
 
 1-1 i-l i-l i- OC r-t 
 
 1 
 
 
 s of Elasti 
 
 j 
 
 Ci O O? I-H 
 
 oo *N or 00 
 
 rc c QO o 
 
 
 2 
 
 s 
 1 
 
 100-1000 
 
 <N O O Ci ' ' 
 1- C O 
 ^ft 00 00 I> 
 
 
 
 
 i 
 
 t^- o cc t'* 
 
 1-1 iM (?} (M . . 
 
 
 CON- 
 SISTENCY 
 
 OF 
 
 MORTAR. 
 
 
 O 32 O 
 
 
 s 
 
 
 Brand. 
 
 !- J- 
 
 
 1 
 
 
 d 
 
 a 
 
 HM 
 
 K f^ 
 
 
 
 0, 
 
 I 
 
 I 
 
308 
 
 CEMENT AND CONCRETE 
 
 Arsenal. 1 These specimens were a portion of those prepared 
 by Mr. Rafter, the compressiva strength being given in Table 
 129. As each value is the result of but one determination, the 
 results are not as regular as might be desired. In general the 
 strength and the modulus decrease together as the amount of 
 water used in mixing is increased. The modulus also decreases 
 with the strength as the proportion of sand increases. 
 
 439. Modulus of Concretes One Month to Six Months Old. - 
 In the compressive tests of twelve-inch concrete cubes made 
 for Mr. George A. Kimball and abstracted in Table 130, many 
 of the specimens were also gaged for compression under load to 
 determine the modulus of elasticity, and a part of the results 
 are presented in Table 142. 
 
 TABLE 142 
 Modulus of Elasticity of Concrete 
 
 TESTS MADE ON TWELVE-INCH CUBES OF PORTLAND CEMENT CONCRETE AT 
 WATERTOWN ARSENAL FOR BOSTON ELEVATED RAILROAD 
 
 AGE 
 OP CUBES 
 WHEN 
 CRUSHED. 
 
 CONCRETE 1-2-4. 
 
 CONCRETE 1-3-6. 
 
 CONCRETE 1-6-12. 
 
 Modulus of Elasticity in Thousands, between Loads, 
 in Pounds per Square Inch, of 
 
 100-600 
 
 100-1000 
 
 1000-200:) 
 
 100-600 
 
 100-1000 
 
 1000-2000 
 
 100-600 
 
 100-1000 
 
 7 days 
 1 month 
 3 months 
 6 months 
 
 2592C 
 2662c 
 3670 
 3646 
 
 2053c 
 2444c 
 3170 
 3567 
 
 1351a 
 
 1462c 
 2157 
 2581 
 
 1869c 
 2438 
 2976 
 3608 
 
 15296 
 2135 
 2656 
 3503 
 
 
 
 
 1210a 
 1805 
 1868 
 
 1376 
 1642 
 1820 
 
 1363 
 1522 
 
 are means of five or more tests of one brand, 
 are means of five or more tests on each of 
 
 are means of five or more tests on each of 
 
 NOTES: Results marked "a ! 
 Results marked "b 
 
 two brands. 
 Results marked "c 
 
 three brands. 
 Results not marked are means of five or more tests on each of 
 
 four brands, two American, two German. 
 For compressive strengths of similar cubes, see Table 130. 
 
 It is seen that the modulus increases with the age and rich- 
 ness of the specimens, and decreases as the load increases. For 
 one-two-four concrete the modulus at one month, for loads 
 between a hundred and a thousand pounds, is about two and 
 
 1 "Tests of Metals," 1899. 
 
r 
 
 MODULUS OF ELASTICITY 309 
 
 one-half million, and for six months, three and a half million. 
 The corresponding values for the one-three-six concrete are two 
 million and three and one-half million. When the ultimate 
 strength is approached, the modulus of elasticity decreases 
 rapidly, and between loads of one thousand and two thousand 
 pounds per square inch, the richest concrete gives only about 
 one and one-half and two and one-half million at one month 
 and six months, respectively. 
 
 440. Modulus of Concrete Dependent on Richness of Mortar. 
 - The results in Table 143 are abstracted from the extensive 
 
 tests made at the Watertown Arsenal for the State Engineer 
 of New York. Although several brands were testetl, the results 
 in the table are from one brand only, namely, "Waylaml" 
 Portland. These cubes were all stored in the same manner, 
 namely, in water three to four months, and then buried in damp 
 sand until broken at the age of twenty months. The mean 
 ultimate strengths of similar cubes stored according to four 
 methods are given in Table 132. 
 
 Since in all of these mixtures the quantity of mortar was a 
 given percentage, either thirty-three or forty, of the volume of 
 aggregate, the effect of the richness of the mortar may be studied. 
 While the proportional strengths of the concretes made with 
 mortars containing from one to five parts sand are 100, 77, 52, 
 .44, and 38, the corresponding proportional moduli of elasticity 
 are 100, 92, 77, 60, and 55, the modulus decreasing less rapidly 
 than the strength, with the addition of sand. 
 
 441. Gravel and Trap Aggregates. Table 144 gives the re- 
 sults of the determinations of the modulus of elasticity of con- 
 crete specimens made and tested at the Watertown Arsenal, 1 
 the strength of which was given in Table 134. As these are all 
 rich concretes, the moduli and the strengths are high. The 
 values of the modulus for the gravel concretes are about 70 
 per cent, of those for the trap, but the strengths of the gravel 
 concretes are in general about 80 per cent, of those obtained 
 with concretes having trap aggregate. In a general way, how- 
 ever, the modulus and strength vary together. 
 
 442. Modulus of Cinder Concrete. The modulus of elas- 
 ticity of cinder concrete prepared for the Eastern Expanded 
 
 'Tests of Metals," 1898. 
 
310 
 
 CEMENT AND CONCRETE 
 
 CO CO O CD "* CO 
 t~- 00 Ci t- rH T* 
 l^ CO 'O CO Oi 'CO 
 
 CO * ' CO 
 >* to -^ 
 O ' i i ' O 
 
 T-H (N O Oi O CO 
 t- CO O OO O ii 
 Tfi O (M CO fN CO 
 
 ^ ^ 
 
 O t- CO I-H 5<l 
 O CO i 1 l-~ C 
 
 (M CO CO -^ i 
 
 iO t- (M 
 
 TfH CD O 
 
 O CO 
 
 
 CO OS ' I-H <M 
 CO CM i I CD 
 - 
 
 CO O 
 O O 
 
 O CO CD 
 t~ ^ >-O 
 
 GO T-H Oi 
 
 CO QO CO CO CO O 
 CO t^ GO CO t^ O 
 O t^ O O CM O 
 
 CO r t>- t^ Oi i-i 
 
 ^ co o 
 
 TH t- QO 
 CM rH T-l 
 
 CO l>- t^ rH 
 rH rH CM ^ 
 
 PROPORTION 
 CEMENT TO 
 SAND IN MORTAR. 
 
 nsistency 
 Concrete. 
 
 axvoaHooy awmoA 
 
 -XN33 H3J 
 
 BV uvxaoj\[ 
 
MODULUS OF ELASTICITY 
 
 311 
 
 TABLE 144 
 
 Modulus of Elasticity of Rich Concretes with Gravel and 
 Trap as Aggregates 
 
 TESTS OF TWELVE-INCH CUBES AT WATERTOWN ARSEXAL. 1-1-3, ALPHA 
 
 CEMENT 
 
 CHARACTER OF AGGREGATE. 
 
 MODULUS OF ELASTICITY IN THOUSANDS, BE- 
 TWEEN LOADS OF 100 AND 1,000 
 POUNDS PER SQUARE INCH, AT AGE, DAYS, 
 
 7-8 
 
 19-23 
 
 29-34 
 
 61-76 
 
 Trap \ 
 
 1875 
 3214 
 
 4091 
 4500 
 3214 
 5000 
 3401 
 
 2500 
 2808 
 6429 
 5025 
 5<J25 
 4500 
 4500 
 
 3750 
 5025 
 5(525 
 5000 
 4500 
 7500 
 5025 
 
 3750 
 
 ' 5025 ' 
 4091 
 7">00 
 5025 
 7500 
 
 
 1 
 
 1 . 
 
 2 
 
 -1, 2y-2 .... 
 -1, 1"-1, 2J"-1 . . 
 
 Mean Results, trap rock aloue 
 
 3022 
 
 4507 
 
 5375 
 
 5082 
 
 Pebbles f " 
 
 1800 
 3750 
 
 2812 
 1800 
 
 3750 
 4091 
 3461 
 3214 
 
 3461 
 3750 
 4091 
 3461 
 
 3214 
 3000 
 4500 
 3214 
 
 " 1$" 
 
 " f"-l, lJ"-2 . . . 
 " I"- 1 * I"- 1 * li"- 1 
 
 Mean Results, pebbles alone 
 
 2540 
 
 3629 
 
 3091 
 
 3482 
 
 NOTES: 
 
 Tests at Watertown Arsenal, "Tests of Metals," 1898. 
 
 For crushing strength of these concretes, see Table 134. 
 
 The modulus of elasticity of twelve-inch mortar cubes, one volume cement 
 to one volume sand, was, for loads between five hundred and one thou- 
 sand pounds per square inch, 3,401,000 at seven days and 5,000,000 at 
 seventy-five days. 
 
 Metal Companies is given in Table 145. The results are seen 
 to be low, as is the crushing strength. The permanent set in 
 five-inch gaged length for a load of six hundred pounds per 
 square inch is also shown in the table. 
 
31L> 
 
 CEMENT AND CONCRETE 
 
 g 
 
 I 
 
 ^ 
 
 5 5 
 
 3 
 
 2 o 
 
 H o | 
 m ^ 
 
 ? g 
 
 H !l 
 
 M ,T 
 
 Ml W 
 
 g 
 
 . o 
 
 -3 5 
 
 1 5 
 
 S 
 
 E 
 
 Big 
 
 s . 
 
 
 
 
 
 
 02 H 
 
 ** u 
 
 
 
 
 
 
 HHH 1 
 
 56 fc 
 
 O O T 1 rH i 1 
 
 ^ 
 
 G<1 00 * OS t CO 
 
 UT) 
 
 
 PERMANEJ* 
 
 FIVE INCH 
 
 AFTER Lo 
 
 600 POUNI 
 SQUARE I 
 
 
 
 
 00000^ 
 000000 
 
 o o o o o o 
 
 o 
 
 CO 
 
 
 
 
 
 
 
 OS 
 
 fc 
 
 
 
 
 
 
 GO 
 rH 
 
 gh 
 
 fc 
 
 i 
 
 
 
 
 
 -fJ 
 
 ^t M 
 
 TJ 
 
 ^1 Ci *O 
 
 
 
 
 O> 
 
 go 
 
 i 
 
 "* rH CO ' 
 
 
 
 
 J^ 
 
 !i 
 
 
 
 
 
 
 *o 
 
 il 
 
 1 
 
 
 
 CO I >C Cl 
 
 
 1 
 
 ii 
 
 i 
 
 1C (M 00 O 00 
 
 * 
 
 oo o co -HH 
 
 (M OS GO OO 
 
 
 
 
 
 3 
 
 <M CO rH rH rH 
 
 C^l 
 
 
 
 
 
 3 Q 
 
 8 
 
 
 
 
 
 J 
 
 O 
 
 1-1 
 
 
 
 
 
 5 
 
 !1 
 
 
 
 
 
 
 
 s 
 
 g* 
 
 1 
 
 IO rH O CO 'N 
 
 1 
 
 rH t- t^ O .t! OO 
 
 1^ 00 00 OS CO O 
 
 HH O O rH L~ O 
 
 00 
 
 co 
 
 3 
 
 S w 
 
 ^ 
 
 CM -* CN rH CM 
 
 (7<l 
 
 rH i 1 rH rH 
 
 rH 
 
 ^3 
 
 w 
 
 S 
 
 
 
 
 
 4> 
 
 H 
 
 . 
 
 
 
 
 
 13 
 
 ^ X 
 
 3 o 
 
 
 
 
 
 T^ 
 
 li 
 
 p<3 
 
 Is 
 
 O 00 O t^ OS 
 
 t- CO O 00 CC 
 (N (M (N rH rH 
 
 CN 
 
 (M 
 
 O (M CO O t- rH 
 
 rH O CO O rH 1 
 
 00 
 rH 
 
 1 
 
 X 
 
 a - 
 
 o 2, 
 
 
 
 
 
 W 
 
 O K 
 
 ^^ 
 
 
 
 x 
 
 
 S 
 
 
 
 
 
 
 
 <u 
 
 
 . 
 
 
 
 
 
 ^_ 
 02 
 
 
 
 
 
 55 
 
 
 
 
 
 
 h* 
 
 
 CO CO CO CO CO 
 
 CO 
 
 ^O tO >-O tO -O *O 
 
 O 
 
 rM 
 
 OD 
 
 j^ 
 
 
 
 
 
 JH 
 
 O 
 
 M 
 
 5 
 
 
 
 
 
 2 
 
 li 
 
 
 
 
 
 
 1 
 
 t 
 
 g 
 
 i 1 i 1 i 1 rH rH 
 
 ^ 
 
 G<l (M <M <M C< <M 
 
 <M 
 
 H 
 
 l 
 
 08 
 
 02 
 
 
 
 
 
 1 
 
 9 
 
 
 
 
 
 
 H 
 
 p 
 | 
 
 "c 
 
 
 
 
 
 1 
 
 
 S 
 
 S 
 
 
 
 
 
 z; 
 
 o 
 
 
 
 
 
 
 
 Is' 
 
 Q 
 
 
 a 
 
 
 p, 
 
 
 gi 
 II 
 
 1 
 
 
 cS 
 
 0) 
 
 ^<^pq WOO 
 
 1 
 
 
CHAPTER XVII 
 
 THE TRANSVERSE STRENGTH AND OTHER PROPERTIES OF 
 MORTAR AND CONCRETE 
 
 ART. 57. TRANSVERSE STRENGTH 
 443. TENSILE, TRANSVERSE AND COMPRESSIVE STRENGTHS 
 
 OF MORTAR COMPARED. The tests given in Tables 146 and 
 147 were designed to compare the strengths of cement mortars 
 in tension, bending and compression, and to show the relative 
 effect on the three kinds of strength of certain variations in 
 manipulation. 
 
 The tensile specimens were briquets of the ordinary form, 
 made in brass molds. The transverse and compressive speci- 
 mens were made in wooden molds, the bars for transverse tests 
 being two by two by eight inches and molded horizontally, 
 while the specimens for compressive tests were two-inch cubes. 
 Specimens of the three forms were made from the same batch 
 of mortar to obviate, as far as possible, variations due to differ- 
 ence in gaging. Two cubes, two briquets and one bar were 
 usually made from one gaging of mortar. 
 
 The briquets were broken in the usual manner on a Riehle 
 cement testing machine. The bars were broken on a home- 
 made lever machine. Two fixed knife edges were placed five 
 and one-third inches apart, and the breaking stress was applied 
 through a third knife edge at mid-span. The lengths of the 
 lever arms of the testing machine were in the ratio of one to 
 twenty-five, and water was allowed to run gently into a vessel 
 at the end of the longer arm. The span of five and one-third 
 inches was chosen because at this length the modulus of rup- 
 ture, for a two inch square specimen, has the same numerical 
 value as the center load applied. 
 
 The cubes were crushed in a crude machine, improvised for 
 the purpose, consisting of two iron plates, two hydraulic jacks, 
 with hydraulic weighing gage and proper framework. The 
 upper plate was fastened to the base of the framework by 
 
 313 
 
314 CEMENT AND CONCRETE 
 
 means of two bolts which worked freely in the lower plate, and 
 the latter was connected to the weighing gage at the top of the 
 framework by . two bolts which worked freely in the upper 
 plate. An hydraulic jack was placed under either end of a 
 yoke, at the middle of which was supported the weighing gage. 
 While the tensile and transverse tests are doubtless good, the 
 compressive tests are lacking in accuracy because of the crude 
 method of crushing. 
 
 444. Table 146 shows the comparative tensile, transverse 
 and compressive strengths of two samples of cement, one of 
 Portland and one of natural, with different proportions of sand. 
 It is seen that the modulus of rupture, or stress on the extreme 
 fiber in transverse tests, computed by the ordinary formula, is 
 considerably greater than the strength obtained in direct tensile 
 tests. The ratio of the transverse to the tensile strength varies 
 from 1.25 to 1.90 for Portland and from 0.95 to 2.19 for natural. 
 
 4 These tests indicate that the ratio of the compressive strength 
 to the tensile strength diminishes with the addition of sand, 
 but the reverse has been found to be true in other series of 
 tests where the facilities for making compressive tests were 
 better. The result obtained here may be attributed to the fact 
 that richer mixtures gave cubes with smoother and more regular 
 faces, and thus less subject to eccentric loading. The com- 
 pressive strength increases between three months and one year 
 much more than the tensile and transverse strengths. Tests on 
 ten brands of Portland and ten brands of natural showed that 
 in general the brands giving the highest strength in tension 
 gave also the highest strength in transverse and compressive 
 tests. 
 
 445. A few results to show the effect of consistency of the 
 mortar on the three kinds of strength are given in Table 147. 
 With Portland cement the highest strength in transverse and 
 compressive tests is given by a wetter mortar than that giving 
 the highest strength in tension, but with natural cement the 
 compressive strength is lowered more than the tensile strength 
 by an excess of water. All oj the specimens were one year 
 old when broken. 
 
 446. TRANSVERSE TESTS OF CONCRETE BARS. The effect 
 
 on the strength of concrete of variations in manipulation and 
 treatment is most satisfactorily investigated by tests of large 
 
MORTAR AND CONCRETE 
 
 315 
 
 an 
 
 a 
 
 S 
 
 H 
 
 8 I 
 
 H tt 
 
 H S 
 H! a 
 
 n S 
 <3 o 
 H O 
 
 T3 
 
 a 
 8 
 
 MEAN STRENGTH, POUNDS PER SQUARE INCH, FOR VARYING RICHNESS OF MORTARS, 
 PARTS SAND TO ONE OF CEMENT BY WEIGHT. 
 
 T3 
 co 
 
 t: 
 
 cS 
 OH 
 O 
 
 S. 
 
 o o ' t~- o 
 
 S I- _CO T* 
 
 Trans. 
 
 ' ^ CS ' O -f 
 M CO I-H 7^1 
 
 ^ 
 
 .r- n .O CO 
 
 00 O -f -* 
 
 1 
 
 CO 
 
 1 
 
 H 
 
 1 
 
 c oo o o n i.o 
 
 C: OO * 7-J ^il- 
 O <M t - ^ ^ ' 
 
 X 
 
 3 
 
 ^r cT ^t oo . o -N . 
 
 <M CO O iO I-H TO 
 
 OD 
 
 1 
 
 'N t^ 00 CS . O 1- . 
 
 1 
 
 00 
 
 1 
 
 O 
 H 
 
 & 
 
 I 
 
 4j< ' CS ' CO i- ^ "N 
 
 I 
 
 t- ' -ti ' oo -N o n 
 
 s 
 
 5< .^h . ^ co co 
 
 d 
 
 1 
 
 
 
 
 
 
 
 t 
 
 III! s ;i ; 
 
 i 
 
 H 
 
 o S ^ oo i5 .cb . 
 
 O CJ 1-1 -H I-H O 
 
 1 
 
 ft o >-7 oo o .00 . 
 
 QC CO O' 'N CS O 
 
 Neat Cement. 
 
 g 
 O 
 O 
 
 CO t^ I- O 2< O 
 c^^^I . ^2^ . 
 
 Trans. 
 
 vO t- O * t O O 
 
 i-H CO "^ CO 1- i-H 
 
 I-H -M CO ' <N -* . 
 
 
 
 H 
 
 1IH : S^i : 
 
 
 
 aoy 
 
 ee es 2 ^ ce 2 ^; 
 'O'Cd>> 'd'd >> 
 
 -- ^g-, 
 
 XN3W3Q 
 
 ^^^^ ^^ z 
 
 h 
 
 Ed 
 ti 
 
 -"* o*-co 
 
 o* 
 
 I 
 
 DO 
 
 -d* 
 
 t! 
 11 
 
316 
 
 CEMENT AND CONCRETE 
 
 sized specimens either in compression or bending. In the prep- 
 aration of "such large specimens the conditions of actual con- 
 struction may be closely reproduced, and the results, although 
 likely to be quite irregular, as the strength of concrete in struc- 
 tures is not uniform throughout, are nevertheless very valu- 
 able On account of the expense connected with such tests, 
 the number of specimens is usually so limited that the natural 
 irregularities in strength mask the true conclusions. 
 
 TABLE 147 
 
 Comparative Tensile, Transverse and Compressive Tests, 
 of Varying Consistency of Mortar 
 
 Effect 
 
 
 
 
 
 TRANSVERSE AND 
 
 
 
 WATER 
 
 MEAN STRENGTH, POUNDS PER 
 
 COMPRESSIVE 
 
 
 
 AS PER 
 
 SQUARE INCH. 
 
 STRENGTH AS PER 
 
 REF. 
 
 CEMENT. 
 
 CENT. OF 
 DRY 
 
 
 CENT. OF TENSILE. 
 
 
 
 INGRE- 
 DIENTS. 
 
 Tensile. 
 
 Trans. 
 
 Comp. 
 
 Trans. 
 
 Comp. 
 
 1 
 
 P 
 
 9 
 
 516 
 
 837 
 
 1731 
 
 162 
 
 335 
 
 2 
 
 P 
 
 12 
 
 533 
 
 987 
 
 2173 
 
 185 
 
 408 
 
 3 
 
 P 
 
 15 
 
 467 
 
 850 
 
 2498 
 
 180 
 
 533 
 
 4 
 
 P 
 
 18 
 
 461 
 
 966 
 
 2823 
 
 209 
 
 612 
 
 5 
 
 P 
 
 21 
 
 430 
 
 1022 
 
 2487 
 
 239 
 
 578 
 
 6 
 
 N 
 
 12 
 
 272 
 
 447 
 
 2270 
 
 164 
 
 835 
 
 7 
 
 N 
 
 14 
 
 325 
 
 516 
 
 2141 
 
 158 
 
 659 
 
 8 
 
 N 
 
 16 
 
 319 
 
 519 
 
 1481 
 
 163 
 
 464 
 
 9 
 
 N 
 
 20 
 
 304 
 
 509 
 
 1512 
 
 167 
 
 497 
 
 10 
 
 N 
 
 24 
 
 315 
 
 462 
 
 1317 
 
 147 
 
 418 
 
 NOTES: Cement, P = Portland, Brand R; N = Natural, Brand In; 
 
 Sand, "Point aux Pins/' pass No. 10 sieve. Age of specimens, 
 one year. Two parts sand to one cement by weight. 
 
 In Tables 148 to 156 are given some of the results obtained 
 in testing over two hundred concrete bars at St. Marys Falls 
 Canal. The molds for making the concrete bars were ten 
 inches square by four and one-half feet long inside. The con- 
 crete was rammed into the mold with a light wooden rammer. 
 The bars were, in general, covered with moist earth soon after 
 completed, to await the time of breaking. To break them they 
 were supported on knife edges placed four feet apart, and the 
 load was applied at mid-span through an iron bolt laid across 
 the bar. In the earlier tests a direct load was imposed by means 
 of a platform which was gradually loaded with one-man stone, 
 but in the later tests the load was applied by means of hydraulic 
 
CONCRETE 
 
 317 
 
 jacks, au hydraulic gage being used to measure the force. In 
 many cases the half bars were again broken at a later date 
 with a twenty-inch span, as shown in the tables. 
 
 447. Variations in Richness of Mortar. In Table 148 sev- 
 eral concretes made with mortars having different proportions 
 of sand are compared, and the results of briquet tests on similar 
 mortars are also given. Although the briquets were not broken 
 at the same age as the bars, the tests on the latter at the differ- 
 ent ages show that they were not gaining strength rapidly, 
 and the results may therefore be compared without serious 
 error. 
 
 TABLE 148 
 Transverse Tests of Concrete. Variations in Richness of Mortar 
 
 
 
 
 hi 
 
 = 2 3 
 
 MODULUS OF RUPTURE. 
 
 No. 
 BARS. 
 
 DATE 
 MADE. 
 
 CEM- 
 ENT. 
 
 *z'z 
 
 !jj*M 
 
 Four Foot Sp.'in. 
 
 Twenty Inch Span. 
 
 
 
 
 ||s 
 
 jipy 
 
 o cr 
 
 
 
 No. 
 Tests. 
 
 Age. 
 
 Mean. 
 
 No. 
 
 Tests. 
 
 Age. 
 
 Mean. 
 
 
 Mo. Da. 
 
 
 
 
 
 Yr Mo. 
 
 
 
 Yr. Mo. 
 
 
 76-77 
 
 11-2 
 
 Port. 
 
 
 
 717 
 
 2 
 
 1 7 
 
 503 
 
 4 
 
 2 9 
 
 600 
 
 78-70 
 
 " 
 
 ' 
 
 1 
 
 700 
 
 2 
 
 
 680 
 
 4 
 
 
 698 
 
 80-81 
 
 u 
 
 
 
 2 
 
 505 
 
 2 
 
 
 538 
 
 4 
 
 
 677 
 
 82-83 
 
 u 
 
 i 
 
 3 
 
 432 
 
 2 
 
 
 480 
 
 3 
 
 
 415 
 
 84-85 
 
 11-3 
 
 
 
 4 
 
 335 
 
 2 
 
 
 370 
 
 4 
 
 
 385 
 
 86-87 
 
 " 
 
 4 
 
 5 
 
 252 
 
 2 
 
 
 284 
 
 4 
 
 
 316 
 
 88-80 
 
 t ( 
 
 ' 
 
 6 
 
 218 
 
 2 
 
 
 262 
 
 4 
 
 
 279 
 
 00-01 
 
 11-4 
 
 Nat. 
 
 1 
 
 483 
 
 2 
 
 
 420 
 
 4 
 
 
 450 
 
 02-03 
 
 
 
 
 
 2 
 
 306 
 
 2 
 
 
 332 
 
 4 
 
 
 387 
 
 04-0--, 
 
 K 
 
 u 
 
 3 
 
 330 
 
 2 
 
 
 240 
 
 4 
 
 
 224 
 
 06-07 
 
 11 
 
 
 4 
 
 237 
 
 2 
 
 
 186 
 
 4 
 
 
 205 
 
 NOTES: 
 
 Portland, Brand R, Sample 82 M. 
 
 Natural, Brand Gn, Sample 83 T. 
 
 Sand, from *'' Point aux Pins" (river sand). 
 
 Stone, Potsdam sandstone, retained on f inch square mesh, and no pieces 
 
 larger than 3 inches in one dimension. 
 Amount mortar used in each case equal to voids in stone measured loose, 
 
 except in case 1-2 natural, when mortar exceeded voids by seven per 
 
 cent. 
 The fracture showed concrete very compact in nearly all cases. 
 
 The results obtained with natural cement show that the 
 tensile strength of the mortar in pounds per square inch was 
 greater than the modulus of rupture obtained for the concrete. 
 
318 
 
 CEMENT AND CONCRETE 
 
 This is also the case with rich mortars of Portland cement, 
 but for Portland mortars containing more than three parts sand 
 to one of cement the concrete gives the higher result. The 
 strength of the concrete with one-to-four mortar is fifty-five per 
 cent, of the strength with one-to-one mortar for Portland, and 
 forty-five per cent, for natural. The decrease in strength due 
 to larger proportions of sand in the mortar is usually greater 
 than the decrease in cost. 
 
 TABLE 149 
 Transverse Tests of Concrete. Variations in Quantity of Mortar 
 
 
 
 
 
 3 H| 
 
 w w 
 
 MODULUS OF RUPTURE. 
 
 No. 
 BAR. 
 
 DATE 
 MADE. 
 
 SfS 
 
 ll! 
 
 Four Foot Span. 
 
 Twenty Inch Span. 
 
 
 
 
 H tfo, ^ 
 
 
 
 
 
 <|^ 
 
 jjjj 
 
 No. 
 Tests. 
 
 Age. 
 
 Mean. 
 
 No. 
 Tests. 
 
 Age. 
 
 Mean. 
 
 
 Mo. Da. 
 
 
 
 
 
 
 
 Yr. Mo. 
 
 
 42 
 
 7 3 
 
 31 
 
 88 
 
 1 
 
 lyr. 
 
 247 
 
 2 
 
 1 10 
 
 363 
 
 37-40 
 
 7 1 
 
 38 
 
 92 
 
 2 
 
 
 284 
 
 3 
 
 n 
 
 447 
 
 38-41 
 
 7 1 
 
 47 
 
 104 
 
 2 
 
 
 
 350 
 
 4 
 
 n 
 
 596 
 
 39-43 
 
 7 1,3 
 
 60 
 
 112 
 
 2 
 
 " 
 
 346 
 
 4 
 
 t 4 
 
 589 
 
 NOTES: 
 
 Cement, Portland. Brand R, Sample 64 T. 
 
 Sand, " Point aux Pins," three parts by weight dry to one cement. 
 Stone, Drummond Island limestone, passing 1 inch slits and retained on 
 f inch slits. 
 
 448. Variations in Amount of Mortar Used. Bars 37 to 
 43, Table 149, were all made with the same kind and quality 
 of stone and the same quality of mortar, three parts sand to one 
 cement by weight, but the amount of mortar varied; thus, in 
 bars 41 and 38 sufficient mortar was used to fill the voids in 
 the stone, while the bars above were deficient in mortar, and 
 those below contained an excess. It is seen that the highest 
 result is given by the bars in which the mortar was just suf- 
 ficient to fill the voids in the stone, though the bars containing 
 an excess of mortar gave practically the same result, while a 
 deficiency of mortar resulted in decreased strength. 
 
 449. Variations in the Amount of Sand for Fixed Quantities 
 of Cement and Stone. In Table 150, bars 68 to 75 were all 
 made with the same kind and quantity of cement and stone, 
 but the amount of sand, and consequently the quantity and 
 
CONCRETE 
 
 319 
 
 quality of the mortar, varied. The highest strength is given by 
 the concrete in which the weight of the sand was three times 
 the weight of the cement; this quantity of sand gave sufficient 
 mortar to fill the voids in the stone. The richer mortars, 
 though stronger, were deficient in quantity, while four parts 
 sand made an oxcess of mortar having a lower strength. 
 
 TABLE 150 
 
 Transverse Tests of Concrete. Variations in Quantity of Sand for 
 Fixed Quantities of Cement and Stone 
 
 
 8 
 
 if. 
 
 
 ^ 
 
 i a . 
 .2 ' 9 a 
 
 w w 
 
 MODULUS OF RUPTURE. 
 
 
 pj 
 
 PQ 
 
 & 
 
 
 
 P 
 
 ,- _ 
 < a 
 
 OQ W 
 
 U Eh O 
 
 -^ < ft. K 
 
 - 1 W S 55 
 
 
 M 
 PC 
 
 Four Foot Span. 
 
 Twenty Inch Span. 
 
 o 
 
 / 
 
 fc o 
 
 H ' 
 
 b*** 5 g w 
 
 2; S^r^ 
 
 
 
 H 
 
 
 w 
 a 
 
 K 
 
 
 
 0^ 
 
 ^ 
 
 |3| 
 
 3^ 
 
 ' O w 
 
 No. 
 
 Tests. 
 
 Age. 
 
 Mean. 
 
 No. 
 Tests. 
 
 Age. 
 
 Mean. 
 
 - 
 
 
 
 
 
 
 
 
 Yr.Mo. 
 
 
 
 Yr.Mo. 
 
 
 
 74-75 
 
 (').", 
 
 65 
 
 1 
 
 16 
 
 95 
 
 2 
 
 1 8 
 
 299 
 
 4 
 
 2 10 
 
 295 
 
 a 
 
 72-73 
 
 65 
 
 130 
 
 2 
 
 24 
 
 101 
 
 2 
 
 tt tt 
 
 335 
 
 4 
 
 tt tt 
 
 303 
 
 6 
 
 70 71 
 
 65 
 
 195 
 
 3 
 
 32 
 
 104 
 
 2 
 
 tt tt 
 
 324 
 
 4 
 
 tt tt 
 
 354 
 
 c 
 
 68-69 
 
 65 
 
 260 
 
 4 
 
 42 
 
 110 
 
 2 
 
 t tt 
 
 322 
 
 4 
 
 tt tt 
 
 321 
 
 d 
 
 NOTES: 
 
 Cement, Portland, Brand R, Sample 768. 
 
 Sand, " Point aux Pins." 
 
 Stone, Potsdam sandstone, screened with f inch mesh, and all pieces larger 
 
 than 3 inches in one dimension rejected. 
 Appearance of fracture: a, very porous; 6, many voids; c, some voids; 
 
 d, few voids. 
 
 450. Consistency of Concrete. -- The bars, the results of 
 which are given in Table 151, were made to show the effect of 
 the consistency of the concrete on the strength obtained. It is 
 seen that the highest strength is given when the consistency 
 is such that a little moisture is shown when ramming is com- 
 pleted; the decrease in strength from an excess of water is much 
 less than that caused by a corresponding deficiency. The re- 
 sults of briquet tests on similar mortar are also given in the 
 table, and it appears that the highest result is given by the 
 mortar containing the least water, which shows the familiar 
 fact that the mortar for concrete should be more moist than 
 that which gives the best results in briquet tests. 
 
 451. Value of Thorough Mixing. Bars 182 to 189, Table 
 152, were made to show the effect of thorough mixing of the 
 
320 
 
 CEMENT AND CONCRETE 
 
 TABLE 151 
 Transverse Tests of Concrete. Variations in Consistency 
 
 
 
 
 
 o - 
 
 MODULUS OF RUPTURE. 
 
 
 & 
 
 
 
 PRO- 
 
 
 
 a c 
 
 
 k 
 
 o 
 
 
 CEM- 
 
 PORTIONS. 
 
 
 Sg w 
 
 
 
 o 
 fc 
 
 W 
 
 BAR. 
 
 ENT, 
 
 
 6 
 
 OH >^ E^H 
 
 4 Foot Span, 
 13 Months 
 
 20 In. Span, 
 2 Years. 
 
 w 
 
 1 
 
 
 Kind. 
 
 
 
 M 
 
 Z Pd pq 
 
 
 
 i 
 
 
 
 
 Cem- 
 
 
 _, 
 
 Pop 
 
 
 
 
 
 o 
 
 *3 
 
 
 
 ent, 
 Lbs. 
 
 Sand, 
 Lbs. 
 
 
 
 II" 
 
 No. 
 Tests. 
 
 Mean. 
 
 No. 
 Tests. 
 
 Mean. 
 
 u 
 
 a 
 H 
 
 138-139 
 
 Port. 
 
 120 
 
 237 
 
 0.61 
 
 7.31 
 
 2 
 
 354 
 
 2 
 
 289 
 
 a 
 
 509 
 
 136-137 
 
 
 120 
 
 237 
 
 0.83 
 
 7.12 
 
 2 
 
 450 
 
 3 
 
 482 
 
 6 
 
 404 
 
 140-141 
 
 
 120 
 
 237 
 
 1.03 
 
 7.00 
 
 2 
 
 450 
 
 4 
 
 442 
 
 c 
 
 415 
 
 142-143 
 
 
 120 
 
 240 
 
 1.16 
 
 7.12 
 
 2 
 
 385 
 
 4 
 
 417 
 
 d 
 
 400 
 
 146-147 
 
 Nat. 
 
 115 
 
 230 
 
 0.83 
 
 7.64 
 
 2 
 
 180 
 
 4 
 
 156 
 
 a 
 
 267 
 
 144-145 
 
 
 115 
 
 230 
 
 1.03 
 
 7.31 
 
 2 
 
 223 
 
 4 
 
 282 
 
 6 
 
 187 
 
 148-149 
 
 
 115 
 
 230 
 
 1.16 
 
 7.12 
 
 2 
 
 234 
 
 4 
 
 256 
 
 c 
 
 145 
 
 150-151 
 
 
 115 
 
 230 
 
 1.35 
 
 7.12 
 
 2 
 
 202 
 
 4 
 
 177 
 
 d 
 
 127 
 
 152-153 
 
 
 115 
 
 230 
 
 1.51 
 
 7.12 
 
 2 
 
 155 
 
 3 
 
 170 
 
 e 
 
 116 
 
 NOTES: Portland cement, Brand R, Sample M. 
 
 Natural cement, Brand Gn, Sample 88 T. 
 
 Sand, " Point aux Pins" (river sand). 
 
 Stone, Potsdam sandstone, 7 cubic feet to each batch. 
 
 Results in last column give tensile strength at one year of briquets 
 
 made from similar mortar. 
 Consistency : a, very dry ; no moisture shown on ramming. 
 
 6, slight moisture appeared at surface after continued 
 ramming. 
 
 c, quaked somewhat. 
 
 d, quaked and water rose to surface in ramming, 
 
 e, too wet to ram. 
 
 TABLE 152 
 Transverse Tests of Concrete Bars. Value of Thorough Mixing 
 
 
 
 MODULUS OF RUPTURE. 
 
 No. BAR. 
 
 MIXING OF CONCRETE. 
 
 Four Foot Span. 
 
 Twenty Inch Span. 
 
 
 
 No. 
 Tests. 
 
 Age. 
 
 Mean. 
 
 No. 
 Tests. 
 
 Age. 
 
 Mean. 
 
 182-186 
 
 Turned once and back 
 
 2 
 
 lyr. 
 
 290 
 
 4 
 
 21^ mo. 
 
 373 
 
 183-187 
 
 " twice " " 
 
 2 
 
 ii. 
 
 294 
 
 4 
 
 u 
 
 353 
 
 184-188 
 
 " 3 times " ." 
 
 2 
 
 u 
 
 306 
 
 4 
 
 u 
 
 444 
 
 185-189 
 
 u 4. tt u u 
 
 2 
 
 it 
 
 328 
 
 4 
 
 u 
 
 474 
 
 NOTES: Cement, Portland, Brand X, 200 Ibs. 
 Sand, " Point aux Pins," 600 Ibs. 
 Stone, Potsdam sandstone, 15 cubic feet. 
 
CONCRETE 
 
 321 
 
 concrete. Comparing the concrete turned once or twice, and 
 back, with that turned three or four times, and back, it is seen 
 that the mean strength of twelve tests with the former is 328 
 pounds per square inch, while the mean strength of the same 
 number of tests with the more thoroughly mixed concrete is 
 388 pounds per square inch, an increase of eighteen per cent. 
 
 TABLE 153 
 Transverse Tests of Concrete. Variation in Size of Aggregate 
 
 
 
 
 STONE. 
 
 s.* 
 
 * X H 
 
 MODULUS OF RUPTURE. 
 LBS. PESg. IN. 
 
 No. 
 BAU. 
 
 EMENT SAMI 
 
 
 AMOUNT Cr 
 icr STONE I 
 CUBIC FEE' 
 
 AMOUNT 
 RAMMED Co 
 CRETE MAD 
 CUBIC FEE 
 
 
 Kind. 
 
 Sizes. 
 
 PER 
 
 CENT. 
 VOIDS 
 
 IN" 
 
 COM- 
 
 One Bar, 4 
 Ft. Span, 
 Age 1 Yr. 
 
 Half Bar, 
 20 In. 
 Span, Age, 
 21 Mo. 
 
 
 
 
 
 
 PACT. 
 
 a. 
 
 
 
 
 202 
 
 XRO 
 
 a * 
 
 V 
 
 45 
 
 3.75 
 
 3 75 
 
 259 
 
 3(57 
 
 199 
 
 t ; 
 
 a 
 
 J V, f F 
 
 43 
 
 3.75 
 
 3.75 
 
 259 
 
 347 
 
 201 
 
 (4 
 
 a 
 
 M 
 
 44 
 
 3.75 
 
 3.75 
 
 216 
 
 269 
 
 200 
 
 (C 
 
 } 
 
 \ each, 
 Vj F, & M 
 
 } 
 
 3.75 
 
 3.75 
 
 245 
 
 292 
 
 11)8 
 
 u 
 
 1 
 
 i each, 
 K, V, F, & M 
 
 
 3.75 
 
 3.75 
 
 288 
 
 390 
 
 ire 
 
 XM8 
 
 d 
 
 V 
 
 32 
 
 3.75 
 
 
 216 
 
 311 
 
 195 
 
 u 
 
 d 
 
 F 
 
 88 
 
 3.75 
 
 3 8fi 
 
 186 
 
 302 
 
 197 
 
 ii 
 
 d 
 
 M 
 
 34 
 
 3.75 
 
 3.75 
 
 131 
 
 208 
 
 194 
 
 " 
 
 '! 
 
 each, 
 V, F, & M 
 
 j 30 
 
 . . . 
 
 . . . 
 
 207 
 
 302 
 
 NOTES: 
 
 All mortar, three parts sand to one part Portland cement by weight. 
 Quantity of mortar about one-third volume of compact stone. 
 Stone: a = Potsdam sandstone; d = gravel. 
 Size : K = T ' ff inch to inch. 
 
 V = i " i " 
 
 F = \ " 1 " 
 
 M = 1 " 2 " 
 
 452. Variations in Size of Stone and Volume of Voids. The 
 bars given in Table 153 were all made with mortar composed 
 of three parts sand to one of Portland cement by weight. The 
 stone for these bars was sorted into different sizes, and these 
 were recombined in the proportions indicated in the table. 
 The sizes are denoted as follows: that passing one-half inch 
 mesh and retained on one-quarter inch mesh, is called V; one- 
 half inch to one inch is called F; one inch to two inches, M; 
 two inches to three inches, C; and coarse sand, one-tenth inch 
 to one-quarter inch, is called K. 
 
322 CEMENT AND CONCRETE 
 
 The first five bars were made with broken sandstone, and it 
 is seen that the coarsest stone, size one inch to two inches, 
 gave the lowest result. The size V, one-quarter inch to one- 
 half inch, although containing no smaller percentage of voids, 
 gave a much higher strength. The highest result was given 
 by the bar made with a mixture of four sizes, the voids in this 
 mixture being only thirty-six per cent. 
 
 The bars containing gravel as aggregate indicate that the 
 strength decreases as the size of stone and volume of voids 
 increase, but a mixture of three sizes gives nearly as good a 
 result as the fine gravel alone. Comparing the results with 
 similar sizes of the two kinds of aggregate, it appears that the 
 broken sandstone gives somewhat better results than gravel, 
 notwithstanding that the proportion of voids in the former 
 exceeds that in the latter. 
 
 453. Sandstone and Bowlder Stone Compared. --The results 
 given in Table 154 are from a series of tests made for the Mich- 
 igan Lake Superior Power Company by Mr. H. von Schon, 
 Chief Engineer, 1 and show the strength of concretes made with 
 two kinds of aggregate available at Sault Ste. Marie. Two 
 samples of Portland cement, one made from marl and one from 
 limestone, a slag cement, and a natural cement, are used in 
 these tests. 
 
 The two samples of Portland cement give nearly the same 
 result, the slag less than half the strength, and the natural 
 quite weak. The ratio of the strength obtained with crushed 
 bowlders to that made with sandstone is about 1.6 with 
 Portland, and the superiority of the former is shown with all 
 cements. 
 
 454. Various Kinds of Aggregate. Table 155 gives the re- 
 sults obtained at St. Marys Falls Canal in using various kinds 
 of stone. In bars 25 to 30, three kinds of stone are compared. 
 The superiority of the Kelleys Island Limestone " shavings " 
 from the stone planers is evident. The shape of the pieces may 
 have had a considerable influence on this result, the planer 
 shavings being flat, or lenticular in form. Bar 34 was made 
 with a hard limestone from Drummond Island, 33 with gravel, 
 and 31 and 32 with gravel and stone mixed in equal propor- 
 
 Tests reported by H. von Schon in Trans. A. S. C. E., Vol. xlii, p. 135. 
 
CONCRETE 
 
 323 
 
 TABLE 154 
 
 Transverse Strength of Concrete with Crushed Sandstone and 
 
 Bowlders 
 
 AGGREGATE. 
 
 MIXTURE No. 
 
 MODULUS OF RUPTURE, POUNDS PER SQUARE INCH. 
 
 Portland. 
 (Marl.) 
 
 Portland. 
 (Rock.) 
 
 Slag. 
 
 Natural. 
 
 Sandstone . . . 
 
 u 
 
 U 
 U 
 
 1 
 2 
 3 
 4 
 5 
 
 328 
 283 
 220 
 178 
 106 
 
 312 
 205 
 178 
 173 
 186 
 
 122 
 161 
 118 
 74 
 131 
 
 43 
 34 
 40 
 
 '35' 
 
 Mean, Sandstone 
 
 223 
 
 223 
 
 121 
 
 38 
 
 
 Bowlder Stone . 
 
 1 i it 
 
 U U 
 (1 il 
 (4 U 
 
 1 
 2 
 3 
 4 
 5 
 
 407 
 
 377 
 332 
 327 
 333 
 
 397 
 395 
 374 
 351 
 325 
 
 145 
 167 
 176 
 146 
 123 
 
 36 
 67 
 55 
 52 
 60 
 
 Mean, Bowlder Stone .... 
 
 355 
 
 368 
 
 151 
 
 54 
 
 Katio of j Bowlder Stone j 
 
 1.59 
 
 1.65 
 
 1.25 
 
 1.42 
 
 Moduli j Sandstone f 
 
 NOTES: Cross breaking tests of 6 in. by 6 in. by 24 in. bars made for 
 Michigan, Lake Superior Power Co. 
 
 Materials: Cement, representative brands of each of four classes. 
 
 Sand, river sand, " Point aux Pins," mostly quartz, 96J Ibs. per cubic foot. 
 Voids, 41.7 per cent. Fineness, 96 per cent, passing No. 20 sieve, 
 
 39 per cent, passing No. 40 sieve. 
 Stone, Sandstone, broken Potsdam 1 to H inch size. 
 
 Bowlder stone, broken gneiss and granite bowlders, 
 
 1 to l\ inch size. 
 
 Proportion in mortar, 1 part cement to 2.4 parts sand by volume. 
 Mixing: Consistency, plastic; cement and sand mixed dry, then wet and 
 
 mixed; mortar added to wet aggregate and concrete mixed by hand. 
 Storage: Bars stored in shed, protected from rain, fully exposed to air. 
 
 Age of specimens when broken, sixty days. 
 Mixture: 1. Mortar 15 per cent, in excess of quantity required to fill voids. 
 
 2. Mortar 10 per cent, in excess of quantity required to fill voids. 
 
 3. Mortar 5 per cent, in excess of quantity required to fill voids. 
 
 4. Mortar just sufficient to fill voids in stone. 
 
 6. Mortar 15 per cent, in excess of amount required to fill voids in 
 stone, but this 15 per cent, excess made with lime instead of 
 cement. 
 
324 
 
 CEMENT AND CONCRETE 
 
 TABLE 155 
 Transverse Tests of Concrete. Value of Different Kinds and Sizes of Aggregate 
 
 MODULUS OF RUPTURE. 
 
 Twenty Inch Span. 
 
 (D 
 
 
 Aggregate: a = Potsdam sandstone; size, f inch to 1 inch. 
 b = Drummond Island limestone; size f inch to 3 inches. 
 c = Kelleys Island limestone; shavings from stone planers; size, inch to 3 inches. 
 d = Gravel. 
 e Broken brick. 
 
 i* o o cr o 
 
 CO CO ^ ^ CO CD 
 
 
 
 
 
 
 '2 a K- 
 
 Ss 2 ::;*>> 
 
 ::::::& ? 
 
 % 
 
 00 
 
 4 
 
 
 
 
 
 Four Foot Span. 
 
 1 
 
 *O<NTtfO"*iOfNCsas<NCO 
 (MOTfCiOiOaOClCOT-HC'-i 
 
 rH T-H <N <M I-H , 1 <M <M CO Tfl 
 
 g 
 
 <3 
 
 1- 
 
 11 i i 
 rH 
 
 lj 
 
 
 
 PROPORTIONS. 
 
 || 
 IJ5 
 
 OS CD O O 
 t--. -.^^^o^^^oO 
 
 (N CO i>- t~ 
 
 P 
 
 ^ 
 
 OS T)H ^H O O 
 
 i 1 1-1 i i <M (M 
 
 1-S 
 
 I J 
 
 00 
 
 ^ 2 - - 3 2! 
 
 aivoaaooy 
 
 '' 
 
 ?StOO!3hOOfO' i e rH|Ci WO 
 
 i<ri 
 
 p^HM 
 
 
 
 9 
 "S 
 
 DO 
 
 - fe 
 
 S 2 2 - GQ2 
 
 K 
 
 T3 
 
 3 
 
 ^' - - t3 - 
 
 1 fi 
 
 1 
 
 ir^ OS 
 
 
 >OCDt-OOCiOT^CO(Mr-i r Y T 7 
 ^(MC^OIiMJOCOCOCOCO^oD 
 O iO 
 
 1 1 T 1 
 
CONCRETE 
 
 325 
 
 tions The gravel and hard limestone gave about the same 
 result, but it is seen that the mixture gave a higher strength. 
 Bars 154 to 159 were made to test the value of broken brick for 
 use in concrete. It is seen that the strength obtained with 
 brick is considerably lower than that obtained with the soft 
 limestone. Had a poorer mortar been used, the brick would 
 doubtless have given a better comparative result, since with 
 the one-to-two mortar, the brick are not strong enough in them- 
 selves to utilize the full adhesive strength of the mortar. 
 
 TABLE 156 
 
 Transverse Tests of Concrete. Use of Screenings with Broken 
 
 Stone 
 
 
 
 SAND AND 
 
 
 
 
 
 STONE. 
 
 STONE 
 TO 80 LBS. 
 
 H . 
 
 tS 
 
 MODULUS OF RUPTURE. 
 
 
 
 CEMENT. 
 
 J^ 
 
 
 
 
 No. 
 
 R 4 R 
 
 
 
 p 
 
 J2r^ 
 
 FOUR FOOT 
 
 TWENTY INCH 
 
 
 
 
 
 
 13 AH. 
 
 c? 
 
 *> 
 
 
 
 
 
 " a* 
 
 a 3 
 
 SPAN. 
 
 SPAN. 
 
 
 1 
 
 ||i 
 
 fj 
 
 4*5 
 
 || 
 
 ^0 
 
 AOE, 11 Mos. 
 
 AGE, 2 YRS. 
 
 
 
 S's 
 
 f 
 
 3? 3 
 
 i 
 
 
 
 No. 
 Tests. 
 
 Mean. 
 
 No. 
 Tests. 
 
 Mean. 
 
 114-115 
 
 a 
 
 49 
 
 240 
 
 7.0 
 
 3.43 
 
 3.05 
 
 2 
 
 233 
 
 4 
 
 237 
 
 124-125 
 
 b 
 
 48.4 
 
 243 
 
 7.0 
 
 3.39 
 
 3.05 
 
 2 
 
 196 
 
 4 
 
 210 
 
 112-113 
 
 c 
 
 44 
 
 240 
 
 7.0 
 
 3.08 
 
 
 2 
 
 194 
 
 3 
 
 236 
 
 1KJ-117 
 
 d 
 
 44 
 
 138 
 
 7.8 
 
 3.43 
 
 2.16 
 
 2 
 
 227 
 
 3 
 
 311 
 
 118-119 
 
 e 
 
 40.5 
 
 243 
 
 7.0 
 
 2.83 
 
 3.10 
 
 2 
 
 201 
 
 4 
 
 219 
 
 120-121 
 
 f 
 
 38.8 
 
 243 
 
 7.0 
 
 2.72 
 
 3.05 
 
 2 
 
 122 
 
 4 
 
 164 
 
 122 
 
 
 
 243 
 
 7.0 
 
 
 3.05 
 
 1 
 
 130 
 
 2 
 
 141 
 
 Screenings replacing 
 
 NOTES: 
 
 Cement: Natural, Brand Gn, Sample 92 T, 80 Ibs. 
 Stone: a = Drummond Island limestone, screened. 
 
 6 = 10 parts screenings to 100 parts stone. 
 
 c = 17 parts screenings to 100 parts stone. 
 
 d = 17 parts screenings to 100 parts stone. 
 equal amount sand. 
 
 e = 50 parts screenings to 100 parts stone. 
 
 / = 100 parts screenings to 100 parts stone. 
 
 g = Screenings only, no broken stone. 
 
 455. Use of Screenings with Broken Stone. Table 156 gives 
 the results of a number of tests made to show the effect of 
 mixing screenings with the broken stone. A smaller amount 
 of mortar is required to fill the voids in a given volume of stone 
 and screenings mixed than is required for the same volume of 
 
326 CEMENT AND CONCRETE 
 
 stone. It is seen that, with natural cement, when the same 
 volume of mortar is used in the two cases, the presence of 
 screenings to the amount of one-third of the total aggregate 
 does not make a material change in the strength of the result- 
 ing concrete, but when the screenings are allowed to take the 
 place of a part of the sand in the mortar, as in bars 116 and 
 117, a much stronger concrete results. Natural cement mixed 
 with sand and screenings alone, bar 122, does not make a strong 
 concrete, but Portland cement with screenings without sand 
 was found to give excellent results. 
 
 456. Deposition in Running Water. A few tests were made 
 to show the effect of depositing concrete in rapidly running 
 water. The molds were placed in the stream and weighted 
 down in twelve inches of water. The concrete for two bars 
 was deposited as soon as mixed, that for two other bars was 
 allowed to stand in the air three hours before deposition, until 
 it should have acquired an initial set, and two bars were made 
 after the mortar had been allowed to stand five hours and 
 twenty minutes before deposition; by this time the mortar had 
 set quite hard. No attempt was made to ram the concrete, 
 which was deposited by lowering it carefully into the water 
 with shovels, the molds being filled as rapidly as possible. A 
 very large amount of the cement was washed out by the current 
 in all cases. After a few months the bars were removed from 
 the stream and covered with earth as usual. The tests at eleven 
 months did not appear to show any advantage in allowing the 
 mortar to stand some time before deposition, but the tests at 
 two years showed a distinct advantage in this treatment. 
 
 457. Use of Concrete in Freezing Weather. Table 157 gives 
 the results obtained with Portland cement concrete made in 
 the open air during cold weather. The conditions as to tem- 
 peratures and the character of the materials are fully given in 
 the table. The experiments are too limited to permit of draw- 
 ing definite conclusions, but the following points are indicated 
 by the results obtained. The use of warm water, 100 to 156 
 Fahr., in freezing weather appears to give somewhat better 
 results than cold water. Salt should not be used unless the 
 temperature is below the freezing point, but in very cold weather 
 the use of enough salt in the water to lower its freezing point 
 below the temperature of the air seems to hasten the harden- 
 
CONCRETE 
 
 327 
 
 TABLE 157 
 
 Transverse Tests of Concrete Bars. Use of Concrete in Low 
 Temperatures 
 
 ji 
 
 izrr::^: '. '-.'.': '^ '.'.'.'.'. 
 
 *a 
 
 |i 
 
 CO ^ C^ CO Ct CO * ^ t- Ci l"- O C* t*- O CO O " H GO 
 
 CO <N ^ >-< -* TO O I-H O r-t t -N CO CO CO X) O O 
 
 - - 
 
 
 
 o 
 
 
 
 p 
 
 
 
 
 "I 
 
 >'Tt < '>ic < it--'.'Tcc?ic>ao*t<cb'Mi (oJ^i^-oo 
 
 (MCMtMCSa^CCCCCC-^CC^CCOO^CCOlCCCOCM 
 
 a 
 
 . -^ . ^ . . . , . . _ 
 
 
 C = WO'-OOCOO03--S5OOO^S C5 ' ; " | iei gJC5?:' a O" H< - | Cl'OOSOC;-OCi<0 
 
 "^ 
 
 '' 
 
 sxsax -OK 
 
 
 -_ . . . . . . 5- 
 
 & " a 
 
 Bg|| 
 
 ', O '. ' O \ TH \ ^ -M 
 
 is 
 
 11 
 
 ^ooTt*^^ cccccc^cc* ' o 
 
 ii 
 
 1*8 
 
 O500(NCCCMO5t > -t~-t~->.'5CC^1 < ^OO(>IOOl^ 
 rl i-t <M(NO<CCCCCCi-ir-(T-(T-i5<ICM 
 
 H 
 
 ^s 
 
 
 
 SS 
 
 S Sco 'oo t^^ ^oocc 
 
 w 
 ffi 
 
 ^3 
 
 o "' o o . . o "' o o o o o " 
 
 5 
 1 
 
 ll's; 
 
 a ----- g uc . . 
 
 5 J 5 w - 
 
 5 
 
 CG^Q 
 
 
 
 * 
 
 O-----O--;-.--Q--- 
 
 2 
 
 15 
 
 S 2 
 
 2 
 
 .. 
 
 
 is 
 
 a ^ 
 
 o^^^^^o^.^^^^^o^^^ 
 
 
 o 
 
 
 a* 
 o 
 
 ^ OS * w~" * ^-* OS-* O5" OS~* OS" 
 
 a 
 
 S'o*- -CC" -(N-O-CC"OrO5" - - 
 -H " i-t .-H rH 
 
 H 
 
 Q 
 
 < 
 
 -t (N - "G<)- ^^H^ <M-(N~.-t-CC- " ^ 
 
 Q 
 
 ^ 
 
 REMARKS: 
 
 a, Concrete frozen after 10 to 20 min.; 6, frozen in 45 min.; c, began to 
 freeze in 15 min.; d, frozen hard following morning after mo'ding; 
 e, concrete still soft 9 A.M. morning after molding; /, bar defective. 
 NOTES: Cement, Portland, Br. R. Sand, "Point aux Pins." Stone, 
 Potsdam sandstone. 
 
328 
 
 CEMENT AND CONCRETE 
 
 ing as well as to increase the ultimate strength. (For tests of 
 mortars in freezing weather, see Art. 50.) 
 
 ART. 58. RESISTANCE TO SHEAR AND ABRASION 
 
 458. Shearing Strength. The shearing strength of mortars 
 and concretes is of importance not only because of its intimate 
 relation to the compressive strength, but because of the shear- 
 ing stresses to which these materials are subjected in structures 
 reinforced with steel. But few tests of shearing strength have 
 been made, however, partly because of the lack of appreciation 
 of their value, and partly because it is difficult to subject a 
 specimen to a purely shearing stress. It is frequently stated 
 that the shearing strength is somewhat in excess of the tensile 
 strength, perhaps as much as twenty per cent. 
 
 Table 158 gives the results of a series of tests made by Prof. 
 Bauschinger in 1878. 1 The values in shear are very closely 
 twenty per cent, in excess of the tensile strengths of similar 
 mortars tested at the same time. 
 
 TABLE 158 
 
 Shearing Strength of Portland Cement Mortar Cubes Hardened 
 
 in Air 
 
 
 
 SHEARING STRENGTH, POUNDS PER 
 
 
 
 KZ . 
 
 SQUARE INCH. 
 
 TENSILE 
 
 CEMENT. 
 
 gg 
 
 Age of Mortar. 
 
 OP SIMILAR 
 MORTARS 
 
 
 s 
 
 
 AT EIGHT 
 
 WEEKS. 
 
 
 
 
 
 
 |DQ 
 
 1 week. 
 
 2 weeks. 
 
 4 weeks. 
 
 8 weeks. 
 
 
 Quick setting Port- ( 
 
 None 
 
 225 
 
 270 
 
 257 
 
 259 
 
 210 
 
 land, mean results < 
 
 3 
 
 108 
 
 128 
 
 154 
 
 196 
 
 169 
 
 of four brands. ( 
 
 5 
 
 67 
 
 94 
 
 112 
 
 168 
 
 139 
 
 Slow setting Portland, ( 
 
 None 
 
 301 
 
 323 
 
 341 
 
 377 
 
 256 
 
 mean results of four ) 
 
 3 
 
 124 
 
 164 
 
 199 
 
 237 
 
 181 
 
 brands. ( 
 
 5 
 
 78 
 
 122 
 
 138 
 
 199 
 
 169 
 
 NOTE: Cement, each result mean of four brands. Sand, medium grain, 
 
 clean. Mortars hardened in dry air. 
 Tests by Professor Bauschinger, 1878. 
 
 459. A distinction should be drawn between the resistance 
 offered by a thin mortar bed to the sliding of one stone or brick 
 on another and to shear of the mortar itself. The former re- 
 
 Quoted by Mr. Emil Knichling in a Report on Cement Mortars. 
 
SHEAR AND ABRASION 329 
 
 sistance involves the adhesion of the mortar to the surface of 
 the brick or stone, and the values for this resistance are usually 
 much less than the shearing strength, and not greatly in excess 
 of the adhesive strength. The one is of importance in the de- 
 termination of the stability of masonry dams, retaining walls, 
 etc., but the latter is the resistance in question in the design 
 of monolithic concrete structures. 
 
 460. Resistance to Abrasion. The resistance of cement 
 mortar to abrasion depends on the quality of the sand as well 
 as the cement. The abraiding surface wears away the cement 
 or pulls the particles of sand out of their beds in the cement 
 matrix. If the adhesion to the sand grains is strong, the sand 
 particles receive the wear and withstand it until nearly worn 
 away. With hard sand particles, therefore, the resistance to 
 abrasion should increase as the proportion of sand increases, 
 until the volume of the cement matrix becomes relatively too 
 small to thoroughly bind the sand grains together. This limit is 
 reached, however, when the mortar contains not more than two 
 parts sand. With soft sand grains, the neat cement will usually 
 give the highest resistance to abrasion, at least in the case of 
 Portland. It has been found that specimens hardened in the air 
 are brittle and wear more rapidly than those hardened in water. 
 
 461. Table 159 gives the results of several tests made to 
 determine the relative wearing qualities of different mortars for 
 such uses as sidewalk construction. The specimens were two- 
 inch cubes, hardened in water and dried for a few hours just 
 before grinding. An emery plate, set horizontally, was used 
 in most of the tests. The results in any given line of the table 
 are comparable, but, owing to changes in the grinding plate 
 and in the methods used, the results in different lines are not 
 all intercomparable. It is seen that when soft sand is used, 
 such as limestone screenings, the greatest resistance to abrasion 
 is offered by the neat cement mortar, and the resistance de- 
 creases constantly as the amount of sand is increased. When 
 hard sand, such as the siliceous river sand, from Point aux 
 Pins ("P.P." in the table) is employed, the greatest resistance 
 is offered by mortars containing about equal parts of sand and 
 cement. A comparison of lines 5 and 10 indicates that rich 
 natural cement mortars lose about twice as much as similar 
 mortars of Portland, but natural cement mortars containing 
 
330 
 
 CEMENT AND CONCRETE 
 
 +> 
 
 I 
 
 4 
 
 a 
 a, | 
 
 GRAMS PE 
 ABRASIO 
 
 MEN 
 CTED 
 
 S 
 
 S 
 
 WEIGHT O 
 OF SURFACE 
 
 CO O OS 
 
 id o T*I 
 
 i-l <M t- CO 
 
 rH O 
 
 i--' co oo 
 
 T-lCOOOO 
 
 !>' i-i CM' co 
 
 DOO 
 
 co os ^' 
 
 CM OS 
 
 i-i o 
 
 co i ( co 10 i i o 
 
 O 1-1 * 
 
 co i i CM' 
 
 CM OS OS 
 
 co o CM' 
 
 r-( r*l CO 
 
 I-H o Tt? 
 
 OOOSCO 
 (M O CM 
 
 OS-*CO OcM^tl 
 
 O O CO CO CO r-i 
 
 * CO "* 
 
 co rH CM' 
 
 CO -^ t^ 3 <* OS 
 
 o o CM' rti o i-i 
 
 rH rH CM i- O rH 
 
 (Mlr-CO 
 (M O CO* 
 
 .SOIXHO.IOHJ; nova: 
 JO 'o 
 
 ! : 
 
 CO 
 
 <M CO 
 
 1 
 
 sqq; 'uauiioadg uo 
 uj 'bg jad aanssajj 
 
 r^GOCOCO CO-^CO 
 
 oo oooo ooo 
 coco ocococo cococo 
 
 uo pasn 
 peqsn.10 
 
 O 
 
 rt. 
 
 O- 
 
 o ^ o 
 
 H 
 
 Mg = 
 
EXPANSION AND CONTRACTION 
 
 331 
 
 more than two parts by weight of sand do not give relatively 
 as good results. 
 
 ART. 59. THE EXPANSION AND CONTRACTION OF CEMENT MOR- 
 TAR, AND THE RESISTANCE OF CONCRETE TO FIRE 
 
 462. Change in Volume during Setting. Cement mixtures 
 shrink somewhat when hardened in air, while specimens stored 
 in water expand a trifle during hardening. Although several 
 experiments have been made on this subject the specimens used 
 have been so small that the results obtained by various author- 
 ities do not agree, and the effect of variations in the character of 
 the mixtures has not been thoroughly investigated. The impor- 
 tance of the question is found in the necessity of providing ex- 
 pansion joints in long walls or sheets, and in the effect of such 
 changes in volume in producing initial stresses in concrete or 
 steel where these materials are used in combination. 
 
 Certain general conclusions are well established and may be 
 stated as follows: 1st. The shrinkage of mortar and concrete 
 hardening in air is considerably greater than the expansion of 
 similar specimens hardening in water; 2d. The amount of 
 change in volume increases with the proportion of cement used 
 in the mixture; 3d. The change in volume is continuous up to 
 one year, but about one half of the change occurs in the first 
 week, and it is very slow after 3 to 6 months. 
 
 The following values of the change in linear dimensions are 
 derived from the results of several experimenters, and show in a 
 general way what changes are to be expected at the end of three 
 months. 1 Variations in the character of the cement and the 
 consistency of the mortar will affect the result. 
 
 COMPOSITION: PARTS SAND 
 TO ONE PORTLAND 
 CEMENT. 
 
 SHRINKAGE OF MORTARS 
 HARDENED IN AIR. 
 
 EXPANSION OF MORTARS 
 HARDENED IN WATER. 
 
 CHANGE IN LINEAR DIMENSIONS, ONE UNIT IN 
 
 Neat cement .... 
 One part sand .... 
 Three parts sand . . . 
 
 300 to 800 
 600 to 1200 
 700 to 1200 
 
 500 to 2000 
 1200 to 3000 
 3000 to 5000 
 
 1 For more detailed results the reader is referred to the following authori- 
 ties: Dr. Tomei, Trans. A. S. C. E., Vol. xxx, p. 16. Mr. John Grant, 
 Proc. Inst. C. E., Vol. Ixii, p. 108. Prof. Bauschinger, Trans. A. S. C. E. 
 Vol. xv, p. 722. 
 
332 CEMENT AND CONCRETE 
 
 463. The Coefficient of Expansion of Cement and Concrete. 
 
 Concerning the coefficient of expansion of cement mortars of 
 various compositions, we know but little. The result obtained 
 by M. Bonniceau, giving the coefficient of neat Portland cement 
 as about .000006 per degree Fahr., is frequently quoted. This 
 is very nearly the value for iron and steel, and has formed a the- 
 oretical basis for combining these materials. In the case of 
 cement mortars and concretes, however, it is highly probable 
 that the coefficient follows quite closely the behavior of the sand 
 and stone used in the mixture, and is much less dependent upon 
 the coefficient of the cement. This was indicated by the results 
 of M. Bonniceau who obtained a value of about .000008 for con- 
 crete. 
 
 464. A number of experiments to determine the coefficient 
 of expansion of cement concretes were carried out under the 
 direction of Prof. Wm. D. Pence by students of Purdue Uni- 
 versity. 1 As a mean of seven tests with one-two-four concrete 
 of Bedford oolitic and Kankakee limestones combined with 
 Portland cements of two well-known brands, the mean result 
 for the coefficient was .0000055, the lowest result being .0000052, 
 and the highest result .0000057. The coefficient of a bar cut 
 from the Kankakee limestone was .0000056, the same result as 
 obtained from the mean of three tests of concrete containing 
 broken stone of this variety. 
 
 The average result of four tests of gravel concrete composed 
 of one part Portland cement, two parts sand and four parts 
 screened gravel, or one part Portland cement to five parts un- 
 screened gravel, gave .0000054 as the coefficient of expansion. 
 
 These values differ from the coefficient of steel enough to 
 indicate that in positions where the range in temperature is 
 great, the resulting stresses in the concrete and steel may be 
 considerable, and worthy of attention. 
 
 465. THE FIRE-RESISTING QUALITIES OF CONCRETE. The 
 
 value of concrete as a material to be used in the construc- 
 tion of the walls and floors of buildings, is largely dependent 
 on its fire-resisting qualities. That its use for such purposes 
 is rapidly extending, is some evidence that these qualities are 
 
 1 Paper read before the Western Society of Engineers, Engineering News, 
 Nov. 21, 1901. 
 
REMHTANCE TO FIRE 333 
 
 as satisfactory as in other classes of materials devoted to the 
 same use. 
 
 Under favorable circumstances, a fire in a building filled with 
 combustible materials may reach a temperature of 2,000 to 
 2,300 Fahr. If a small specimen of cement mortar or concrete 
 is subjected to a temperature approaching this intensity, the 
 cement loses its water of crystallization and becomes friable. 
 If cooled suddenly in water, the specimen cracks and disinte- 
 grates. If cooled gradually, the outer edge of the specimen 
 crumbles away. From such tests on small specimens some very 
 erroneous conclusions have been drawn as to the value of con- 
 crete as a fire-resisting material. Such conclusions have done 
 much to prejudice the public mind against concrete, and to re- 
 tard its introduction in buildings designed to be fireproof. 
 
 466. Conductivity. The great value of concrete as a fire 
 resistant is due to its low conductivity of heat, and while the 
 surface of a mass of concrete exposed to an intense flame for 
 some time is ruined, and may be flaked off by the application 
 of a strong stream of water from a fire hose, the depth to which 
 the heat penetrates is very limited. Steel is said to lose ten 
 per cent, of its strength at about 600 Fahr. and fifty per cent, 
 at about 750 Fahr. The importance of protecting the steel 
 framework of a building, not only from warping and complete 
 destruction due to flames, but from loss of strength from over- 
 heating, is therefore evident. 
 
 Among engineers and architects it is recognized that the 
 term " fireproof construction" is only relative, although the 
 lay mind is apt to give a definite and literal meaning to the 
 term. It is well known that fireproofing tile, whether hard or 
 porous, will fall to pieces if subjected to a temperature above 
 that employed in its manufacture. The practical question then 
 is, what type of construction will withstand long continued 
 intense flame, and subsequent quenching with water, with the 
 least injury to the strength of the structure. The results of 
 fire tests that have been conducted in several places, and no- 
 tably those made by the Department of Buildings of New York 
 City, have shown that floor arches properly constructed of 
 concrete-steel are equal to any style of floor with which they 
 come in competition. 
 
 The low conductivity of concrete is shown by the fact, 
 
334 CEMENT AND CONCRETE 
 
 stated by Mr. Howard Constable in connection with the dis- 
 cussion of fire tests of concrete floor arches, 1 " that in some 
 thirty-five cases where the temperature ranged from 1,500 to 
 2,400 degrees, the time of exposure being from one to six hours, 
 the temperature of the upper flanges of six-inch to ten-inch 
 beams might be approximately place~d at not much above 200 
 degrees." He also says "in one case, where the beam was pro- 
 tected by three inches of concrete, the fire was maintained for 
 five hours, and the temperature went as high as 2,300 degrees, 
 and there was no practical or permanent set produced in the 
 beams." 
 
 467. Behavior in Conflagrations. As to the behavior of 
 concrete-steel arches in an actual fire, a board of experts was 
 appointed by the insurance companies to investigate the causes 
 and extent of damage to the fireproof buildings in the Pitts- 
 burg, Pa., fire of May 3, 1897. This board stated in their 
 report that they believed that in important structures of this 
 class "the fireproofing should be in itself strong and able to 
 resist severe shocks, and should if possible, be able to prevent 
 the expansion of the steel work "; and continued, "There seems 
 to be but one material that is now known that could be utilized 
 to accomplish these results, and that is first-class concrete. 
 The fire-resisting qualities of properly made concrete have been 
 amply proven to be equal, if not better than fire clay tile, as 
 shown by the tests carried on by the Building Department of 
 the City of New York." 
 
 468. In a report on the Baltimore fire, Captain Sewall, 2 
 Corps of Engineers, U. S. A., says concrete "undergoes more or 
 less molecular change in fire; subject to some spalling. Molecu- 
 lar change very slow. Calcined material does not spall off 
 badly, except at exposed square corners. Efficiency on the 
 whole is high. Preferable to commercial hollow tiles for both 
 floor arches or slabs and column and girder coverings. In form 
 of reinforced concrete columns, beams, girders and floor slabs, 
 at least as desirable as steel work protected with the best com- 
 mercial hollow tiles. Stone concrete spalls worse than any 
 
 1 Trans. A. S. C. E., Vol. xxxix, p. 149. 
 
 2 Report to the Chief of Engineers, U. S. A., by Capt. John Stephen 
 Sewail, Corps of Engineers, Published in Engineering News, March 24, 1904. 
 
RESISTANCE TO FIRE 335 
 
 other kind, because the pieces of stone contain air and moisture 
 cavities, and the contents of these rupture the stone when 
 hot. Gravel is stone that has had most of these cavities elimi- 
 nated by splitting through them, during long ages of exposure 
 to the weather. It is therefore better for fire-resisting concrete 
 than stone. Broken bricks, broken slag, ashes and clinker all 
 make good fire-resisting concrete. Cinders containing much 
 partly burned coal are unsafe, because these particles actually 
 burn out and weaken the concrete. Locomotive cinders kill 
 the cement, besides being combustible. On the whole, cinder 
 concrete is safe only when subjected to the most rigid and 
 intelligent supervision; when made properly, of proper ma- 
 terials, however, it is doubtful whether even brickwork is much 
 superior to it in fire-resisting qualities, and nothing is superior 
 to it in lightness, other things being equal." 
 
 469. Aggregate for Fireproof Work. Since air is a poor 
 conductor of heat, the more porous concretes are the better 
 protectors against fire. On this account, as well as because of 
 its lightness, cinder concrete is preferred for fireproofing. Care 
 should be taken that cinders to be used in fireproofing concrete 
 do not contain any appreciable amount of unburned coal; in 
 concrete to be used next to steel members the cinders should 
 also be practically free from iron rust. (See 473.) 
 
 The strength of cinder concrete is much inferior to that 
 made with the ordinary aggregates, and there should be no 
 difficulty in making a porous concrete with the latter. In fact, 
 in many other classes of construction it has been seen that 
 great precautions must be taken to avoid porosity. By the 
 use of insufficient mortar to fill the voids in the stone, voids 
 may be left in the concrete, though a't the expense of dimin- 
 ishing somewhat the strength of the mixture. In adopting such 
 an expedient one should not lose sight of the fact that in order 
 to preserve the imbedded steel from corrosion, it must be fully 
 covered with the mortar. 
 
 470. Broken bricks are excellent for fireproofing concrete. 
 The bricks themselves are fire resistant, porous and light, while 
 the adhesion of cement mortar to bricks is so great that unless 
 a very weak mortar is used, the strength of the concrete is 
 limited only by the strength of the brick employed. 
 
 Sandstones, especially those with . siliceous cementing ma- 
 
336 CEMENT AND CONCRETE 
 
 terial, are also well adapted for this purpose. Limestone, on 
 account of the low temperature at which it is broken up, is 
 not good, though as to just how far a limestone concrete would 
 be disintegrated by the heat of an ordinary building fire has 
 not, so far as the author knows, been fully investigated. It is 
 known, however, that limestone masonry is calcined to a cer- 
 tain depth in a conflagration. 
 
 Granite in large pieces is cracked by only a moderate degree 
 of heat, and spalls badly. Just how much danger there might 
 be of a similar action in concrete aggregates of this material is 
 not known, nor whether small pebbles or fine gravel would 
 have this property in the same degree, though it is believed 
 they would not, and this view has been confirmed by observa- 
 tions of the Baltimore ruins. 
 
 Before adopting a given aggregate for fireproof work, one 
 should satisfy himself by actual test as to the suitability of the 
 materials available, but such tests should be conducted upon 
 concretes containing the proposed aggregates, rather than upon 
 fragments of the materials not incorporated with mortar. 
 
 ART. 60. THE PRESERVATION OF IRON AND STEEL BY MORTAR 
 
 AND CONCRETE 
 
 471. The rusting of steel members in modern buildings and 
 other engineering structures is one of the most serious menaces 
 to their permanence. The introduction of concrete-steel con- 
 struction has given rise to some discussion, especially among 
 those unfamiliar with the properties of concrete, as to the 
 effect of the concrete upon the steel. 
 
 472. Action of Corrosion. The rusting of iron takes place 
 only in the presence of moisture, air and carbon dioxide. In 
 perfectly dry air, or in perfectly pure water, iron does not rust. 
 Under the proper conditions, however, the iron, water and 
 carbonic acid combine to form ferrous carbonate, which at once 
 combines with oxygen from the air to form ferric oxide, the 
 carbonic acid being liberated to act on a fresh portion of the 
 metal. It is seen that only a very small amount of the carbon 
 dioxide is necessary. If, however, the carbon dioxide or other 
 acid filling the same role, is neutralized by the presence of an 
 alkaline substance, the foregoing reactions cannot take place. 
 As cement is strongly alkaline, it thus furnishes an almost per- 
 fect protection against rusting. 
 
t 
 EFFECT ON CORROSION OF METAL 337 
 
 473. Tests of Effect of Concrete on Corrosion of Metal. - 
 
 To determine the cause of occasional rusting of steel surrounded 
 by cinder concrete, and consequently the proper methods of 
 applying cement mortar or concrete to steel, Prof. Chas. L. 
 Norton, engineer in charge of the Insurance Engineering Ex- 
 periment Station at Boston, made tests on several hundred bri- 
 quets in which steel was imbedded in mortars and concretes 
 of various compositions. 1 The briquets were subjected to air, 
 steam and carbon dioxide, others to air and steam, to air and 
 carbon dioxide and to the ordinarily dry air of a room. At the 
 end of three weeks it was found that neat Portland cement had 
 furnished a perfect protection in all cases. The corrosion of 
 the steel in other specimens was always at a point where a void 
 existed in the concrete, or where a badly rusted cinder had lain. 
 In every case where the concrete or mortar had been mixed wet, 
 and the surface of the steel had been thus coated with a thin 
 layer of grout, no rust spots occurred. 
 
 In the first tests made by Professor Norton the specimens 
 were thoroughly cleaned before being imbedded in the concrete, 
 but later tests indicated that in specimens that had begun to 
 corrode before treatment, the rusting was arrested by the coat- 
 ing of cement mortar or concrete. After from one to three 
 months in tanks holding steam and carbon dioxide, specimens 
 which had' been in all stages of corrosion before being im- 
 bedded in the concrete had not suffered any sensible change 
 in weight or size except when the concrete had been poorly 
 applied. 
 
 474. The results of these experiments showed that the steel 
 need not necessarily be freed from rust before being imbedded 
 in the concrete; that the concrete to be applied next the steel 
 should be mixed wet, or that the steel should be first coated 
 with grout by dipping or brushing; and it appeared that the rust- 
 ing sometimes found in cinder concrete is due to the rust in the 
 cinders rather than to the sulphur, and that if proper precau- 
 tions are taken, cinder concrete is nearly as effective as stone 
 concrete in preventing corrosion. Prof. Norton says, "In the 
 matter of paints for steel there is a wide difference of opinion. 
 I cannot believe that any of the paints of which I have 
 
 Report III of Insurance Engineering Exp. Station, Boston, Mass. 
 
338 CEMENT AND CONCRETE 
 
 any knowledge can compare with a wash or painting with 
 cement.'' 
 
 475. Sulphur in Cinders. The conclusions drawn by Booth, 
 Garrett and Blair from a series of tests made for the Roebling 
 Construction Co., were that cinders from anthracite pea coal 
 contained about two-tenths per cent, of sulphur which they 
 considered sufficient to cause corrosion of unprotected iron- 
 work, more or less rapidly, depending on the presence or absence 
 of moisture; but they further concluded that a "full" concrete 
 (one in which the voids in the cinders were entirely filled by 
 mortar of cement and sand) would fully protect the steel. 
 
 In a paper read before the Associated Expanded Metal Com- 
 panies, Prof. S. B. Newberry has this to say concerning cinder 
 concrete: 1 "The fear has sometimes been expressed that cinder 
 concrete would prove injurious to iron, on account of the sulphur 
 contained in the cinders. The amount of this sulphur is, how- 
 ever, extremely small. Not finding any definite figures on this 
 point, I determined the sulphur contained in an average sample 
 of cinders from Pittsburg coal. The coal in its raw state con- 
 tains rather a high percentage of sulphur, about fifteen per 
 cent. The cinders proved to contain only 0.6 per cent, sulphur. 
 This amount is quite insignificant, and even if all oxidized to 
 sulphuric acid, it would at once be taken up and neutralized 
 in concrete by the cement present, and could by no possibility 
 attack the iron." 
 
 476. Precautions. While so far as the corrosion of steel 
 is concerned, the above experiments by Prof. Norton show that 
 the rusting is corrected by the concrete, yet it is quite possible 
 that the adhesion of cement to steel may be impaired by a coating 
 of rust. The cleaning of the steel may be accomplished by 
 first brushing with wire brushes to remove all scales, followed 
 by treatment with hot dilute sulphuric acid, and finally apply- 
 ing an alkaline wash such as hot milk of lime to neutralize all 
 traces of the acid. Oxalic acid may be used in place of the 
 sulphuric, and the application of the milk of lime dispensed with, 
 since the acid oxidizes. The crystals of oxalic acid as purchased 
 commercially should be mixed with about seven parts hot water 
 and the solution applied with a brush or sponge. When the 
 
 Engineering News, Apr. 24, 1902. 
 
f 
 
 PRESERVATION OF STEEL 339 
 
 adhesion of the mortar or concrete to the steel is of any impor- 
 tance, as it is in all concrete-steel construction where the stresses 
 are divided between the steel and concrete, any of the ordinary 
 oil paints will not only be quite unnecessary, but may be a very 
 serious detriment to the construction. 
 
 The experiments quoted indicate the importance of having 
 the steel covered with an unbroken coating of cement or cement 
 mortar. To insure this the steel must either be coated with a 
 layer, preferably of neat Portland, by dipping or brushing, or 
 the mortar placed next the steel must be wet enough to insure 
 intimate contact throughout. It may be added also, that the 
 addition of a small amount of thoroughly slaked lime to Port- 
 land cement mortar or concrete will not only render the mate- 
 rial more alkaline, but will make the mortar more plastic, and 
 thus insure a better coating of the steel. Such small additions 
 have no deleterious effect on the mortar. 
 
 477. Practical instances of the preservation of iron by con- 
 crete are not wanting. The writer has stored in water, briquets 
 with small iron plates imbedded in Portland cement mortar, 
 and at the end of six months the plates were found moist, but 
 entirely free from corrosion except where they projected beyond 
 the mortar. A concrete-steel water main built on the Monier 
 system at Grenoble, France, was taken out and examined after 
 fifteen years service in damp ground. The metal imbedded in 
 the mortar showed no signs of corrosion, and the mortar could 
 only be detached from it by hammering. 
 
 Mr. W. G. Triest * relates that in breaking up cast-iron, con- 
 crete-filled pillars, a wrench was found that had been buried 
 in the concrete for twenty-two years. The wrench had main- 
 tained its metallic surface in the concrete, while a part of it that 
 had been imbedded in coal ashes had corroded badly. 
 
 Similar instances showing the action of concrete on steel 
 and iron might be multiplied, but it is sufficient to state that 
 the preservation of iron or steel properly imbedded in Port- 
 land cement mortar or concrete is now seldom questioned, and 
 the use of cement paint, in* place of the ordinary oil paints, 
 as a steel preservative, has been adopted in many places. 
 
 1 Trans. A. S. C. E., April, 1894. 
 
340 CEMENT AND CONCRETE 
 
 ART. 61. POROSITY AND PERMEABILITY; EFFLORESCENCE; 
 POINTING; USE u; SEA WATER 
 
 478. The porosity and permeability of mortars have been 
 thoroughly investigated by M. Paul Alexandre, who has pub- 
 lished his results in " Recherches Experimentales Sur Les Mortiers 
 Hydr antiques" l The results and conclusion in the following 
 notes on the subject are largely a resume of the systematic 
 investigation made by M. Alexandre. 
 
 The two qualities, porosity and permeability, should not be 
 confused, nor should it be thought that a porous mortar is 
 always very permeable, or that a permeable mortar must of 
 necessity be very porous. Porosity is measured by the amount 
 of water which will be absorbed by a specimen after drying, 
 while permeability is measured by the amount of water which 
 will pass through a specimen in a given time under certain de- 
 fined conditions of thickness, water pressure and area of face. 
 
 479. Porosity. The porosity of mortars is due to, and in 
 fact is measured by, the volume of the voids contained. These 
 voids may be divided into three classes, according to the causes 
 to which they may be attributed, as follows: 1st, apparent 
 voids, due to the mortar not being properly compacted; 2d, 
 latent voids due to the imprisonment of air in the mortar when 
 made; and 3d, voids resulting from the evaporation, during har- 
 dening, of a portion of the water used in gaging. 
 
 480. Apparent voids may occur as the result of using insuf- 
 ficient cement to fill the voids in the sand, or, in the case of 
 concretes, insufficient mortar to fill the voids in the aggregate. 
 They may also be due to improper manipulation as to tamping, 
 or improper mixing, giving an excess of matrix in one place and 
 a deficiency in another. It was found by experiment that 
 mortars made with coarse sand had the largest volume of ap- 
 parent voids. 
 
 It has been shown elsewhere that if dry sand be moistened 
 and agitated, the bulk of the sand is increased. This is caused 
 partially by the imprisonment of air bubbles in the mass, and 
 if a measure of sand so treated is filled with water, the bubbles 
 will rise to the surface on jarring the vessel. Latent voids in 
 mortar are due to a similar action, and hardened mortars con- 
 
 1 Extrait des Annales des Fonts et Chaussees, September, 1890. 
 
POROMTY AND PERMEABILITY 341 
 
 taining such voids refuse to absorb water to replace the air 
 bubbles, at least for a long time. 
 
 481. A portion of the water used in mixing mortar enters 
 into chemical combination with the cement, another portion is 
 absorbed by the sand grains, and a third portion goes to moisten 
 the sand. The quantity absorbed by the grains depends upon 
 the character of the sand, and the amount required to moisten 
 the sand depends upon the superficial area of the grains in a 
 given volume, being greatest for fine sands and least for coarse 
 ones. At least one fourth of the water ordinarily used in 
 mixing neat cement is given off later, if the hardened mor- 
 tar is allowed to remain in dry air. The water required to 
 moisten the sand, and at least a part of that absorbed by 
 the sand grains, also dries out, leaving voids of the third class 
 mentioned. 
 
 The apparent voids may be reduced to a very small per- 
 centage by care in the proportions and preparation of the 
 mortar. The latent voids may amount to six or seven per 
 cent, of the total volume. The evaporation of water may 
 leave from six to eighteen per cent, of voids in the mass. 
 
 482. The conclusions drawn from M. Alexandre's experi- 
 ments are briefly as follows: The porosity varies between wide 
 limits according to the fineness of the sand and the richness of 
 the mortar. It may be as low as thirteen per cent, and may 
 exceed thirty-one per cent. 
 
 With sand of the same degree of fineness, the porosity di- 
 minishes as the proportion of cement in the mortar increases. 
 
 With the same quantity of cement per volume of sand, the 
 porosity increases with the fineness of the sand. This is es- 
 pecially marked in rich mortars, where the increase in porosity 
 may reach 50 to 100 per cent., while in lean mortars the use 
 of a fine sand may not increase the percentage of voids more 
 than 20 per cent. 
 
 The least porous mortars are those rich in cement and made 
 with coarse sand. Mortars made with fine sand are relatively 
 very porous, even when, made rich with cement. 
 
 Mortars gaged dry are more porous than those of ordinary 
 consistency, and mortars gaged wet are also likely to be more 
 porous, unless the manipulation is such as to allow the excess 
 water to rise to the surface of the mortar.. 
 
342 CEMENT AND CONCRETE 
 
 483. Permeability The degree of permeability of mortars 
 is a more important property than the porosity, since not only 
 does it affect the suitability of the mortar for certain uses, 
 but the life of the structure may depend upon the difficulty 
 with which water may percolate the mass. 
 
 The permeability of mortar decreases as the proportion of 
 cement is augmented, and in the case of concretes the per- 
 meability diminishes as the percentage of mortar increases, at 
 least to the point where the latter is in excess of the voids in 
 the stone. 
 
 From experiments made at the Thayer School of Civil En- 
 gineering, Messrs. J. B. Mclntyre and A. L. True found that a 
 five-inch layer of concrete containing from 30 to 45 per cent, 
 of one-to-one Portland cement mortar, and some of the speci- 
 mens containing 40 to 45 per cent, of one-to-two mortar, were 
 impermeable with pressures of 20 to 80 pounds per square 
 inch, maintained for two hours. 
 
 484. Mortars made with fine sand are much less permeable 
 than those made with coarse sand. This difference is so marked 
 that a less permeable mortar is made with one barrel of cement 
 per cubic yard of fine sand, passing a sieve having, say, fifty 
 meshes per inch, than with two barrels of cement per cubic 
 yard of very coarse sand in which the grains are, say, one- 
 tenth inch in diameter. Mortars made with sands composed 
 of a mixture of grains of various sizes are neither very porous 
 nor easily permeated. 
 
 Mortars mixed very dry or very wet have greater permeabil- 
 ity than those of the ordinary consistency, and in the case of 
 concretes, it would probably be found that a deficiency of 
 water would result in a much more permeable mass than the 
 use of what might be considered an excess. 
 
 All of the above conclusions indicate that a mortar may be 
 quite porous, and yet so long as the voids are very minute, the 
 percolation of water through it will be slow. This is especially 
 shown by the fact that mortars of coarse sand, not porous, 
 are more permeable than the porous mortars of fine sand. 
 
 485. When water is permitted to percolate continuously 
 through a mass of mortar, the interstices gradually become filled, 
 and the permeability decreases in marked degree. M. Alex- 
 andre found that a volume of water which passed a certain 
 
WA TER-PROOFING .>43 
 
 mass of mortar in twenty minutes at the beginning of the ex- 
 periment, required five hours to percolate the mass at the end 
 of a month. M. R. Feret has obtained similar results in making 
 extensive experiments l on the subject of permeability, and 
 considers that fine particles of cement or lime are carried along 
 by the water, forming efflorescence at the surface and tending 
 to stop the flow. 
 
 486. The Preparation of Water-Proof Mortar and Concrete. 
 - To enumerate briefly the precautions necessary to attain 
 
 water-tightness in mortars and concretes, it may be said that 
 different brands of cement present different characteristics in 
 this regard. Fine grinding is a prime requisite, and sand 
 cement or silica cement, containing as it does very fine grains of 
 sand intimately mixed with cement particles of extreme fine- 
 ness, is admirably adapted to such uses. 
 
 The Sand should, if possible, be composed of a mixture of 
 grains of various sizes, because such a mixture gives a mortar 
 not only little permeable, but one that is not porous, and that 
 has, besides, a good strength. The amount of cement in the 
 mortar should be in excess of the voids in the sand, not less, 
 in general, than three barrels of cement per cubic yard of sand. 
 
 In concrete the volume of mortar should exceed the volume 
 of voids in the aggregate, and to obtain this result without 
 too great expense, the aggregate should be so selected as to have 
 a minimum of voids. Gravel concrete properly proportioned 
 may be made water-tight somewhat more easily than broken- 
 stone concrete, but a mixture of gravel and broken stone will 
 give good results not only in this regard, but in the matter of 
 strength as well. 
 
 487. To make a compact mortar for use where the facilities 
 for tamping are ordinarily good, the consistency should be 
 neither very wet nor very dry. When the mortar is struck 
 with the back of a shovel, moisture should glisten on the surface, 
 but in a pile the mortar should appear but little moister than 
 fresh earth. This is the consistency which, with a moderate 
 amount of tamping, gives the least volume of mortar with 
 given quantities of dry materials. In places difficult of access, 
 
 1 "La Capacit^ des Mortier Hydrauliques," Annales des Fonts et Chausstes, 
 July. 1892. 
 
344 CEMENT AND CONCRETE 
 
 or in the preparation of concrete, better results will be obtained 
 with a mortar somewhat wetter than the above, since large 
 voids will be less likely to occur in the more plastic mass. In 
 fact, unless the supervision is very close, it is advisable to use 
 a rather wet mixture in preparing concrete where water-tight- 
 ness is desired. 
 
 488. Washes. The application of certain washes to the 
 surfaces of walls intended to be water-proof, and the introduc- 
 tion of foreign materials into the mortar or concrete to make 
 it less permeable, have been practiced to some extent. Alter- 
 nate coatings of soap and alum solutions are applied with a 
 brush, not only to concrete, but to brick and stone masonry 
 surfaces. These penetrate the pores of the masonry, forming 
 insoluble compounds which prevent percolation. Washes of 
 grout, composed of cement, or of cement and slaked lime, are 
 used for a similar purpose. 
 
 "Sylvester's Process for Repelling Moisture from External 
 Walls" consists in applying first a solution of three quarters 
 of a pound of soap to one gallon of water, followed, after twenty- 
 four hours, by the application of a solution containing two 
 ounces of alum per gallon of water. Both solutions are applied 
 with a brush, the soap solution boiling hot, and the alum so- 
 lution at 60 to 70 Fahr. The applications are alternated, 
 with twenty-four hours intervening each time. Experiments at 
 the Croton Reservoir l indicated that four coats of each wash 
 were required to render brickwork impervious to a head of forty 
 feet of water and the cost of the four double applications was 
 about ten cents a square foot. 
 
 In Reservoir Number Two of the Pennsylvania Water Co., 
 two washes of each solution were used on the walls at a cost 
 for materials and labor of twenty-three cents per hundred 
 square feet, and the results were said to be good. 
 
 A modified recipe for such a wash in which but one solution 
 is made is given as follows: 2 A stock solution is prepared of 
 one pound lye, five pounds powdered alum, dissolved in two 
 quarts water. One pint of the stock is used to a pail of water 
 in which ten pounds Portland cement has been well mixed. 
 
 1 Trans. A. S. C. E., Vol. i, p. 203. 
 
 2 J. H. G. Wolf, Engineering News, June 30, 1904. 
 
WATER-PROOFING 345 
 
 489. In a few cases the use of alum and soap solutions in 
 the body of the mortar has been tried with apparently success- 
 ful results. Mr. Edward Cunningham, 1 in making experiments 
 on water-proof concrete vessels, used powdered alum equal to 
 one per cent, of the combined weight of the sand and cement, 
 mixing this with the dry ingredients. To the water used in 
 mixing, one per cent, of yellow soap was added. The results 
 were said to be very satisfactory. In the above proportions, 
 however, the amount of alum is made to depend upon the 
 amount of cement and sand used, while the soap added depends 
 upon the amount of water, whereas the soap should bear a de- 
 finite ratio to the alum. 
 
 In experiments with mortar composed of one part cement 
 to two and one-half parts of bituminous ash, Prof. W. K. Hatt 2 
 found that the alum and soap mixed with the mortar at the 
 time of gaging increased the strength and hardness of the ash 
 mortar about fifty per cent., and diminished the absorption by 
 the same percentage. One half of the water used for gaging 
 was a five per cent, solution of ground alum, the other half 
 being a seven per cent, solution of soap. The alum solution 
 was used first and the gaging completed with the soap solution. 
 
 Mr. W. C. Hawley 8 employed a stock solution of two pounds 
 caustic potash, five pounds powdered alum, and ten quarts 
 water, and used in the finishing coat three quarts of this solu- 
 tion in each batch of mortar containing two bags of cement. 
 The mortar was mixed with two volumes of sand to one of ce- 
 ment and covered forty-eight square feet to a depth of about 
 one-half inch. The extra cost for materials and preparing so- 
 lution was only about nine and a half cents per hundred square 
 feet. With less than two parts sand to one cement, it was 
 found the finishing mortar checked in setting. It was also 
 found that any organic matter in the sand was softened by the 
 potash, and an excess of potash caused checking, although an 
 excess of alum had no deleterious effect. 
 
 490. Use of Lime, etc. The introduction of slaked lime in 
 mortars designed to be water-proof is suggested by the fact 
 
 1 Trans. A. S. C. E., Vol. li, p. 128. 
 8 Trans. A. S. C. E., Vol. li, p. 129. 
 1 Journal New England Water- Works Association, 1904. 
 
346 CEMENT AND CONCRETE 
 
 that the permeability of mortar diminishes if water is allowed 
 to percolate it for some time, the theory being that fine par- 
 ticles of cement and lime are dislodged by the passage of the 
 water to form a deposit at or near the surface, and check the 
 flow. This suggestion, however, needs experimental confirma- 
 tion, since it seems quite possible that the introduction of a 
 substance containing such a large proportion of water as does 
 slaked lime, may increase the percentage of voids in the mor- 
 tar, if not the permeability. 
 
 The use of pulverized clay and pozzolanic materials for a 
 similar purpose has been suggested. It has already been shown 
 that moderate doses of clay have no deleterious effect on the 
 strength of mortars for ordinary exposures. The action of the 
 pozzolanic substances has been found by Dr. Michaelis and 
 M. Feret to be not mechanical alone, but chemical, and the 
 effect on the strength of the resulting mortar depends upon the 
 exposure to which it is subjected, such admixtures being dele- 
 terious for mortars hardened in air. 
 
 491. Efflorescence. The white deposit sometimes formed 
 at the surface of brick and masonry walls is usually due to the 
 filtration of water through the mortar, dissolving out salts of 
 potash, soda, etc., and depositing these salts on the surface by 
 evaporation or by the formation of sodic carbonate. The ab- 
 sorption of water from the atmosphere may also account for 
 this deposit in some degree, especially near the sea. The same 
 term is applied to a more harmful deposit, sulphate of calcium, 
 which may be supplied by the filtrating water or may come 
 from the cement, either from the addition of gypsum or from 
 the fuel used in burning. The crystallization of this salt in the 
 pores of the masonry at the surface may cause disintegration. 
 
 On the other hand efflorescence may be quite harmless, as 
 when it is formed by washing out from the mortar an excess of 
 hydrate of lime. A portion of the latter may then be changed 
 to carbonate of lime near the surface of the wall and actually 
 stop up the pores or voids, and prevent further filtration. 
 
 492. The discoloration of brickwork and fine masonry by 
 efflorescence is sometimes serious. To ameliorate these condi- 
 tions, the use of water-proof mortars, and careful pointing of 
 the work, are precautions to be recommended. General Gill- 
 more, in " Limes, Hydraulic Cements and Mortars/' suggests 
 
EFFLORESCENCE 347 
 
 the use of about ten pounds of animal fat to one hundred pounds 
 of lime and three hundred pounds of cement; the object of the 
 fat being to saponify the alkaline substance, the lime in form 
 of paste serving only as a vehicle for the fat. A more practical 
 method, however, would seem to be the application of soap 
 and alum washes on the surface, or the use of soap and alum 
 in the preparation of the mortar to be used near the face of the 
 wall, and especially for pointing. The remedy to be adopted, 
 however, will depend upon the cause of the efflorescence. 
 
 493. Pointing Mortar. Pointing serves the double purpose 
 of making the joint practically water-tight at the edges, and giv- 
 ing a finish to the face of the wall. If the edge of the joint is 
 not well filled, moisture collects there either from the face or 
 from seepage through the wall. Subsequent freezing or the 
 crystallization of certain salts may spall the stones or loosen 
 them from their bed. 
 
 In laying cut-stone masonry, the joints should be raked out 
 for about two inches back from the face to be pointed. 
 Pointing mortar should be prepared from fine sand and the 
 best Portland cement. The proportion of sand should not 
 exceed two parts by weight to one cement, and in the highest 
 class work, equal parts of cement and sand are sometimes 
 used. No advantage is gained, however, by using a mortar 
 richer in cement than the one last mentioned. The use of fine 
 sand and rich mortars are specified not only because such mor- 
 tars are practically water-tight, but because they take a fine 
 finish. 
 
 494. The tools required for pointing are a bent iron to rake 
 out the joints (though this should be partially done while the 
 mortar is green), a mortar board and small trowel, a calking 
 iron and wooden mallet, a brush for moistening the joint, and 
 one or more beading tools. After raking out the joint it is 
 moistened by the brush, and the mortar, which is mixed quite 
 dry. is filled in with the trowel. When enough mortar is in 
 place to fill half the depth of the joint, it is tamped with the 
 calking iron and mallet much as a ship's seam is calked with 
 oakum. The joint is then filled to the face, and again tamped. 
 The bead is then formed by running the beading iron back 
 and forth over the joint. This beading iron is of steel with the 
 handle parallel to, but some two or three inches out from, the 
 
348 CEMENT AND CONCRETE 
 
 line of the blade forming the bead. The blade is three to five 
 inches long and " hollow ground" or finished with a smooth 
 concave surface. Only such a length of joint is pointed at one 
 operation as may be quickly carried to completion. The wall 
 must be kept moist for some time after the pointing is done, 
 and it should be protected from the direct rays of the sun, as 
 fine cracks are very likely to appear in this rich, finely finished 
 mortar. If possible, pointing should be done in moderate 
 weather and must be entirely suspended in temperatures ap- 
 proaching the freezing point. 
 
 495. Cements in Sea Water. The theory of the action of 
 sea water upon cements is not fully understood. It is known 
 that some cement structures exposed to the worst conditions 
 have given most satisfactory results, while others have failed 
 in greater or less degree. It may be said at once, however, 
 that many of the most eminent and conservative engineers 
 consider that the failures that have occurred in the use of Port- 
 land cement in sea water are due to improper specifications, 
 proportions and manipulation, rather than to any defect in 
 Portland cements as a class. 
 
 496. It is thought the following represents, in the main, the 
 most generally accepted theory of the chemical action. In the 
 setting of cements that are rich .in lime, the whole of the lime 
 is not engaged in stable compounds, and when placed in the 
 sea the sulphate of magnesia of the sea water is able to com- 
 bine with the lime, forming calcic sulphate, the magnesia being 
 precipitated. The discovery of magnesia in decomposed mor- 
 tars led, at first, to the supposition that the cause of failure 
 was the presence of magnesia in the cement when used. If 
 the water level about the structure changes frequently, as is 
 usual, or if the wall is at times subjected to a greater head on 
 one side than on the other as in tide docks, the percolation of 
 water through the wall is stimulated, and the sulphate of lime 
 may then be washed out if the mortar is quite pervious, and 
 more will be formed from a fresh supply of sea water attacking 
 the lime of the cement, until the latter is destroyed. If, how- 
 ever, the sulphate of lime is not washed out, it may crystallize 
 and thus cause swelling of the mortar. 
 
 497. It would appear from the above that for successful 
 use in sea water the hydraulic index of the cement should be 
 
ACTION OF SEA WATER 349 
 
 high; that is, that the lime should be comparatively low in or- 
 der that the lime compounds may be more stable. For this 
 reason it is not impossible that some of our natural cements, 
 which are so much more nearly uniform than the Roman ce- 
 ments of Europe that have been condemned for this reason, 
 may give fairly good results in sea water. The fact that the 
 mortars of natural cement are more permeable than those of 
 Portland, is, however, a serious defect. 
 
 Following a similar reasoning, Dr. Win. Michaelis has ad- 
 vanced the theory that if trass, or other pozzolanas of proper 
 composition, be mixed with Portland cement subsequent to 
 the burning, the hydrate of lime which separates from the ce- 
 ment in hardening will at once combine with the pozzolanas, 
 forming a stable compound. This view, however, has been 
 vigorously opposed by the Society of German Portland Cement 
 Manufacturers, as well as by many engineers, especially of 
 France, and the discussion is not yet at an end. 
 
 M. Candlot l says that, from the experiments of various en- 
 gineers, "we have arrived at this conclusion, that the only 
 remedy to adopt against decomposition is to prevent the sea 
 water from penetrating the mortar. We are led thus to dis- 
 miss the chemical reactions of sea water on mortars and to 
 consider their action from a purely physical standpoint." 
 
 498. To resist the attacks of sea water the mortar should not 
 only be impervious, but also as little porous as possible. The 
 cement should be finely ground and should not contain free 
 lime. The content of magnesia and of sulphuric anhydride 
 should be as low as possible, the latter not exceeding one and 
 five-tenths per cent. The proportion of lime should not be 
 too high, and above all, special pains should be taken with the 
 manufacture to insure proper comminution and mixing of the 
 raw materials, and uniform burning. The addition of sulphate 
 of lime to regulate the setting is believed to be injurious for 
 cements to be used in sea water; even two or three per cent, 
 is said to cause rapid disintegration, and in the specifications 
 for recent extensive works in dock construction, the addition 
 of gypsum or other foreign matter was entirely prohibited. 
 
 "Le Cimcnt," September, 1896, quoted by F. H. Lewis, M. Am. Soc. 
 C. E. Trans. A. S. C. E., Vol. xxxvii, p. 523. 
 
350 CEMENT AND CONCRETE 
 
 Although slag cements have given good results in the sea 
 for a short time, it is considered that they will not, in general, 
 resist the action of sea water for long periods. 
 
 499. Sand or aggregate containing argillaceous or soft cal- 
 careous matter should be avoided for works in the sea. Two 
 instances of failure of sea walls in which shells were used as the 
 aggregate are mentioned by Col. Wm. M. Black, 1 and although 
 the failures are not definitely traced to the calcareous matter in 
 the concrete, the fact that experiments have shown that cal- 
 careous sands do not withstand the action of sea water, makes 
 it probable that this was an important cause of the failure. 
 
 Fine sands that give porous mortars, though not easily per- 
 meable, are to be strictly avoided. Coarse sands giving per- 
 meable, though not porous, mortars are better, but still leave 
 much to be desired as to immunity from decomposition. The 
 best sands are those containing various grades of sizes of par- 
 ticles from coarse to fine, as mortars made with such sands are 
 not only compact, but practically impermeable. 
 
 500. Since the mortar and concrete should be made as com- 
 pact as possible, the precautions mentioned under the head of 
 water-proof mortar and concrete should be taken in the prep- 
 aration of mortars and concretes for use in the sea. That is, 
 the proportion of cement should exceed the voids in the sand 
 and the mortar should exceed the voids in the aggregate. 
 
 M. Alexandre has found that the mortars mixed to the or- 
 dinary consistency are attacked least by sea water. When 
 specimens are merely immersed in the water, those mixed dry 
 suffer the most, but some tests indicate that if mortars are 
 submitted to the filtration of water soon after made, those 
 mixed wet are most easily decomposed. As to whether fresh 
 or salt wafer should be employed in mixing mortars to be used 
 in sea water, although Mr. Eliot C. Clarke, M. Paul Alexandre 
 and many others have investigated this subject, the conclusions 
 are not definite and it is probable that either may be used as 
 convenient. 
 
 Trans. A. S. C. E., Vol. xxx, p. 601. 
 
PART IV 
 
 USE OF MORTAR AND CONCRETE 
 CHAPTER XVIII 
 
 CONCRETE : DEPOSITION 
 
 501. Concrete may be molded into blocks which are allowed 
 to set and then are transported to the structure and laid as 
 blocks of stone. This is the block system of construction. The 
 adaptability of concrete to being built in place, however, is 
 one of its chief merits, and consequently the monolithic method 
 of construction is far more common Since it has been found 
 that expansion and contraction, due to changes in temperature, 
 affect concrete walls as they do any other walls of masonry, 
 it has become customary to mold the concrete in sections, usu- 
 ally alternate sections of equal size and shape being built first, 
 and the omitted sections built in later. This method of con- 
 structing a long wall is also called monolithic, since the blocks 
 are of large size and are built in place. 
 
 502. When concrete is deposited either in air or in water, 
 molds must be provided to keep the mass in the desired shape 
 until it has lost its plasticity and acquired sufficient strength 
 to stand alone. In foundations, the earth at the sides of the 
 excavation may supply the place of a mold, and sometimes 
 the mold forms a part of the permanent structure, as in the 
 case of masonry piers with concrete hearting, and in steel cylin- 
 der piers filled with concrete. 
 
 ART. 62. TIMBER FORMS OR MOLDS 
 
 503. The construction, placing and removal of forms fre- 
 quently represent a considerable percentage, from five to thirty 
 per cent., of the total cost of the concrete, and it is therefore 
 evident that an improper design may result in a considerable 
 waste of money, as well as in marring the appearance of the 
 
 351 
 
352 CEMENT AND CONCRETE 
 
 work. The character of the form will of course depend on the 
 character of the work; in the construction of a large number 
 of small blocks of the same shape, where one mold may be used 
 over and over, the thickness of the pieces should not be stinted, 
 and the ease of knocking down the mold should be carefully 
 considered. When a form can be used but once, the size of 
 pieces should be no larger than necessary to give the requisite 
 stiffness, and the ease of first construction is a main considera- 
 tion. Forms should be left in place forty-eight hours to allow 
 the concrete to set, and in the case of arches and beams a 
 much longer time is necessary, so that the concrete may assume 
 considerable strength before it is called upon to support its 
 weight. 
 
 504. Sheathing. Forms for massive walls of monolithic 
 construction usually have vertical posts, with iron ties across, 
 or braced by battered posts outside. The sheathing planks 
 are then placed horizontal. In a few cases horizontal wales 
 have been placed within the posts and vertical sheathing laid 
 against the wales. 
 
 The strength of the sheathing must be sufficient to stand 
 the pressure transmitted to it through the concrete when the 
 rammer is used close to the face of the mold. The concrete is 
 seldom built up fast enough to bring upon the sheathing a great 
 head of fluid pressure, but the ramming brings a heavy local 
 pressure upon it. If supported at intervals of four feet, two- 
 inch lumber dressed to one and three-quarters inches thick is 
 usually sufficient; for spans of more than 5 feet, 2 } inch lum- 
 ber is required to make a perfect face. Boards seven-eighths 
 inch thick are suitable only when supports are not more than 
 about 2 feet apart. In placing concrete in molds under water 
 there is more danger of bursting the mold by the weight of 
 semi-fluid concrete, and if the work is .to be built up rapidly, 
 this must be guarded against by sufficient bracing. 
 
 505. For exposed faces, the duty to be performed by the 
 lagging includes leaving as smooth a finish as possible on the 
 concrete after the removal of the forms. If green lumber is 
 employed, the boards may shrink before use, leaving openings 
 between the sheathing that will show plainly on the face of the 
 work. A slight tendency of this kind may be checked by 
 keeping the boards well wet with a hose until the concrete is 
 
TIMBER FORMS 353 
 
 placed. On the other hand, thoroughly seasoned lumber will 
 swell when the concrete is placed; to obviate this difficulty the 
 lower edge of each sheathing plank may be beveled on the outer 
 edge; the thin edge on the inside will then crush when the 
 planks swell. 
 
 The use of tongue and grooved lagging has been tried, but 
 is not usually satisfactory, as there is no opportunity to expand, 
 and the planks are particularly hard to place a second time. 
 To give a good face in work under water, however, tongue and 
 groove sheathing will assist in preventing washing of the cement. 
 Yellow pine lumber is found to be excellent for sheathing; on 
 account of the large amount of pitch contained, it absorbs 
 water slowly and holds its shape. For a similar reason, fir 
 timber would be suitable. 
 
 In order that the face of the mold shall be perfectly smooth, 
 it is necessary to size and dress the plank on at least one side 
 and two edges. 
 
 As it is almost impossible to avoid having some line of de- 
 marcation shown in the concrete at the joints of the sheathing 
 planks, care should be taken that the lagging is of uniform 
 width throughout, and laid horizontal so that consecutive sec- 
 tions show the joint continuous. The sheathing may be placed 
 for the entire form before concreting is commenced, or the 
 plank may be raised on the posts as the work advances. The 
 former method will usually give the neater appearance, but is 
 too expensive for high walls. 
 
 506. Lining. The appearance of the finished concrete is 
 much improved, and the labor of preparing the forms probably 
 not increased, since less care may be taken in surfacing, by 
 lining the mold with thin sheet iron. Iron of number twenty 
 gauge (.035 inch thick, 1.42 pounds per square foot) has been 
 used for this purpose, but where the same lining is used several 
 times, a heavier iron is preferable. The joining of one sheet of 
 lining to another may present greater difficulties than the join- 
 ing of planks, but joints will occur less frequently. 
 
 In the construction of the Marquette Breakwater, Mr. 
 Clarence Coleman, Asst. Engineer, used sheet steel one eighth 
 of an inch thick for lining molds for building monolithic blocks. 
 Concerning the use of the steel, Mr. Coleman says: 1 "Very 
 
 Report Chief of Engineers, V. S. A., 1898, p. 2254. 
 
354 CEMENT AND CONCRETE 
 
 smooth surfaces were produced on the slopes of the concrete 
 and the work of the molders was greatly facilitated on account 
 of the comparative ease with which the concrete was compacted 
 under the slope pieces of the molds. The steel effectually pre- 
 vented the aggressive friction of the sharp particles of broken 
 stone on the wooden surfaces of the molds, thus increasing the 
 life of the molds and decreasing the cost of molding the con- 
 crete." 
 
 507. Oiling the Forms. Oiling or greasing the face of the 
 mold, in order that the latter may be removed without detach- 
 ing particles from the concrete face, is usually advisable. Soap, 
 crude oil, linseed oil, bacon fat, are some of the materials that 
 have been used for this purpose; the first mentioned probably 
 gives the best results, and if not applied too freely will have no 
 injurious effect upon either the finish or strength of the work. 
 Applying shellac to the molds improves the appearance of the 
 concrete surface. When the forms are lined with steel, the 
 adhesion of the concrete to the lining is more difficult to over- 
 come. In this case the ordinary oils are not entirely success- 
 ful, but fat salt pork has been found to give satisfactory 
 results. 
 
 508. Joints and Corners. If desired, triangular strips may 
 be nailed to the inside of the forms in such a way as to block 
 off the face to represent stone masonry, and in this way the 
 marks of joints between planks or between strips of lining may 
 be avoided. Square corners should not be allowed on exterior 
 angles, as it is difficult to so tamp the concrete as to make the 
 corner perfect, and they are so likely to be chipped off. Tri- 
 angular strips or moldings should be tacked along the corners 
 of the mold as a fillet to cut off the corner by a plane making 
 equal angles with the adjacent faces. This plane may be from 
 one inch to two inches wide. 
 
 To form water drips on projecting ledges, such as door caps 
 and sills, abutment copings, etc., a small half-round should be 
 nailed to the upper surface of the mold a short distance back 
 from the projecting face. This leaves a ridge at the edge of 
 the under side of projection so that the water must drip from 
 the edge, and not follow back to the main wall face. 
 
 509. POSTS AND BRACES. The sizes of posts and braces 
 must be such as to make a practically unyielding support to 
 
TIMBER FORMS 355 
 
 the sheathing. With one and three-quarters inch lagging, posts 
 may be four feet apart; if five feet four inches apart (three to 
 each sixteen foot length), some yielding of the sheathing may 
 be expected if it is less than two and three quarter inches. 
 If sheathing is four inches thick, the distance between posts 
 may be six or seven feet. 
 
 Fir, yellow pine, and Norway pine are suitable for posts. 
 Three-inch by eight-inch is an ordinary size, and a post of 
 these dimensions should be supported, either by ties or braces, 
 at intervals of four to six feet. Where the posts are four inches 
 by ten inches, supports may be six to eight feet centers, while 
 with six-inch by twelve-inch posts, the distance between cen- 
 ters of supports may be eight to ten feet. Posts should be 
 sized arid dressed on the side which is to receive the sheathing, 
 in order that the alignment may be perfect. 
 
 510. Methods of Bracing. The general plan of the mold 
 may vary according to conditions, the following methods hav- 
 ing been employed on heavy work to support the vertical posts: 
 1st, With outside inclined braces, leaving the interior of the 
 mold unobstructed. 2d, Tie rods across the interior of the 
 mold connecting opposite posts at frequent intervals. 3d, Each 
 post trussed vertically and tied across at top and bottom only. 
 4th, Horizontal trussed wales outside of posts, spaced four to 
 five feet apart in the vertical and tied across at the ends. 
 
 511. Inclined Braces. The sizes of inclined braces depend 
 on their lengths, the inclination to the vertical, and the amount 
 of shoring used. An approximate rule for the size of braces 
 under usual conditions and using ordinary dimension stuff, 
 not boards, is that the number of square inches area of cross- 
 section of brace should equal length of span in feet. If thin 
 planks are employed, they should be in pairs, one on either 
 side of the vertical post, and made to act together by cross- 
 pieces nailed to the two planks. 
 
 The aim should be to make the whole form practically un- 
 yielding under the action of the tampers, as it has been found 
 that this action is usually more severe than the mere pressure 
 of the concrete in a semi-liquid condition. The sizes of pieces 
 cannot, therefore, be accurately computed, but the above sizes 
 are derived from the generafe-result of experience as to what 
 has proved satisfactory. 
 
356 CEMENT AND CONCRETE 
 
 The advantage of the form of construction just described is 
 that the interior of the mold is left entirely unobstructed. On 
 high walls, however, the amount of timber required for braces 
 is excessive, and the braces may be almost as objectionable as 
 tie rods, since the former prevent the laying of tracks along 
 the side of the form. 
 
 512. Tie Rods. When the vertical posts are supported by 
 tie rods across the mold and the wall is thin, it may be possible 
 in removing the mold to withdraw the bolts or rods if they 
 have been thoroughly greased or wrapped with stiff paper be- 
 fore the concrete is placed. If it is designed to leave the rods 
 in the concrete, they should be provided with sleave nuts near 
 the end, which, when unscrewed, will leave the end of the rod 
 within the concrete mass not less than two inches from the face. 
 The hole left by the nut should be carefully filled with mortar 
 after the mold is removed. 
 
 With vertical posts four feet apart, this method of support 
 is objectionable, as it leaves a network of ties within the forms 
 interfering seriously with the operation of a skip and with the 
 ramming. It is not necessary, however, to place all of the tie 
 rods to the top of the mold before beginning the concreting, 
 as it is sufficient to keep one or two rods in place above the 
 plane where tamping is being done, 
 
 A modification of this method is to use wires of large diam- 
 eter with an eye at the end just inside the finished face of 
 the concrete. A short bolt, with hook at one end and threaded 
 at the other, passes through the post, hooks into the eye of the 
 wire, and is tightened by a nut on the threaded end outside 
 the post. After removing the nut, the rod is unhooked and 
 the hole in the face filled with mortar, the wire remaining in 
 the concrete. 
 
 513. Trussed Posts. The third method of support, where 
 the posts are trussed and provided with heavy tie rods at the 
 top, and held at the bottom either by tie rods or some other 
 means, seems to have fewer objections than the methods just 
 described. Less timber will usually be required to build this 
 form than for that where inclined braces are used, and the 
 obstruction to operations will usually be less than with either 
 of the other styles. This mold is also very readily taken down, 
 though the posts are heavier and more difficult to handle. 
 
TIMBER FORMS 357 
 
 To secure the bottoms of the posts, they may be set in the 
 ground, or rest against sills braced to some other portion of 
 the structure, or to piles. A suitable support may also be 
 obtained by dumping a mass of concrete around the bottom 
 of each post and allowing it to set. Forms erected on rock 
 may have the posts rest against blocks bolted to the rock. 
 
 514. Trussed Wales. The fourth method of supporting the 
 posts is particularly applicable where the work is divided into 
 blocks of moderate size in horizontal cross-section, say twenty 
 feet square. In longer lengths the horizontal trussed wales 
 become rather heavy for convenient handling. Within these 
 limits, however, this is an excellent form. In the construction 
 of Lock No. 2 between Minneapolis and St. Paul, 1 a form of 
 this kind was used for blocks about twelve by fifteen feet at the 
 bottom. The sheathing was one and three-quarters inches, 
 lined with No. 20 galvanized iron. Verticals were four by 
 twelve, spaced about two feet centers. The trussed wales were 
 twelve by twelve inch, trussed with one and one-quarter inch 
 rod, the king-post being of twelve by twelve, inch about two 
 and one-half feet long, making depth of truss three and one- 
 half feet. The ends of opposite wales were connected by one 
 and one-quarter inch rod passing outside of the sheathing. 
 Each pair of the longitudinal wales was just above the corre- 
 sponding pair of transverse wales, so that they did not inter- 
 fere at the corners. The mold was twenty-nine feet in 
 height. 
 
 In describing this mold, Mr. Powell says: "One complete 
 form weighs twenty-eight tons; each piece about seven tons. 
 Each piece is moved separately by the cable-way in forty-five 
 to sixty minutes. The operation of removing one complete 
 form requires from three to four hours time. After being 
 moved, a small crew of men occupy nearly a day in plumbing 
 and bolting together the form." "The boxes containing 1.7 
 cubic yards of concrete are landed on top of the form by cable- 
 way and tipped from that position. Although the jar and 
 strain is severe, the forms have shown no ill effects therefrom, 
 remaining tight and secure." 
 
 1 Major Frederic V. Abbot in charge. Mr. A. O. Powell, Asst Engr., 
 Report Chief of Engineers, 1900, p. 2778. 
 
358 CEMENT AND CONCRETE 
 
 ART. 63. DEPOSITION OF CONCRETE IN AIR. 
 
 515. Transporting to Place of Deposition. In. depositing 
 the concrete in place, care must be taken not to undo the work 
 of mixing. If the concrete is allowed to fall freely a distance 
 of several feet or to slide down an inclined plane, the stones 
 will be likely to separate from the mass, and the result will be 
 a layer of broken stone followed by a layer of mortar. If the 
 concrete is deposited in a pile, the stone will roll down the out- 
 side of the cone. This action is especially bad in concrete that 
 is mixed rather dry. The author has seen a pavement founda- 
 tion in which the limits of each wheelbarrow load of concrete 
 could be distinguished, the foundation presenting the appear- 
 ance of the cross-section of a honeycomb, made up of irregular 
 hexagons outlined by broken stone having a deficiency of 
 mortar. In all such cases, if the action cannot be avoided by 
 some other method of dumping, then care must be taken to 
 remix the concrete. 
 
 There is one method by which the concrete may be deposited 
 by gravity without separation of the materials. This consists 
 in allowing the material to slide down a tube, but the tube 
 must be kept continually full, the concrete being allowed to 
 run out at the bottom only as fast as it is filled in at the top. 
 This method is only applicable where the mixing is continuous, 
 as in the case of machine mixers. 
 
 516. Sometimes it will be found possible to mix the con- 
 crete so near to the place of deposition that it may be shoveled 
 directly into place. In mixing by hand this is practicable, as 
 the mixing platforms may usually be easily moved, and this 
 method of deposition is carried out even in street work where 
 the concrete is in thin layers and hence requires much moving 
 of platforms. 
 
 Where a machine mixer is used that is so mounted as to be 
 portable, the concrete may be delivered in place by a belt 
 conveyor. Such an arrangement for the building of walls and 
 for foundations of pavements, has already been described in 
 Chapter XIV. 
 
 The conditions are usually such, however, as to preclude the 
 possibility of mixing the concrete so close to the work that it 
 may be shoveled into place or handled economically on a con- 
 
PLACING IN AIR 359 
 
 veyor of the style mentioned. The next cheapest method is 
 to use a derrick to handle skips or bottom-dump buckets, pro- 
 vided the work is sufficiently concentrated to have one posi- 
 tion of the derrick serve to place a large quantity of con- 
 crete. The skips should hold about a cubic yard, and if a batch 
 mixer is used, the skip should hold a batch, whatever that 
 may be. 
 
 517. If the concrete is mixed on the same level and within 
 less than two hundred feet of the work, wheelbarrows may be 
 used, but for greater distances, carts, or, what is usually cheaper, 
 cars running on a track, should be employed. 
 
 For large masses of concrete a cableway may be employed 
 to advantage, provided there is sufficient use for it to repay 
 the high original cost of plant. The selection to be made from 
 among these common methods is dependent on economy as in 
 handling other material, the only requirements being that the 
 concrete shall be conveyed to place quickly, and that the ma- 
 terials shall not be allowed to separate as a result of any of the 
 manipulation. In laying large quantities of concrete, the dif- 
 ference between success and failure from a financial standpoint 
 may easily rest in the proper transportation of the materials 
 to and from the mixer. 
 
 518. Ramming. The concrete should be deposited in hori- 
 zontal layers about six inches thick, leveled with a shovel and 
 thoroughly rammed. The length of time ramming should be 
 continued and the vigor with which it should be done depend 
 largely on the degree of plasticity of the concrete. If the con- 
 crete is made of such a consistency that when struck a smart 
 blow with the back of a shovel a film of moisture will just show 
 on the surface, it should have vigorous ramming to insure a 
 compact mass. A flushing of water to the surface will then 
 indicate when to cease tamping. 
 
 With a little more water there is less danger of the larger 
 stones " bridging" and leaving large voids in the mass, and 
 less work will be required to flush water to the surface. With 
 such a consistency, cutting the mass with a spade before start- 
 ing, the ramming may assist in expelling air bubbles and pre- 
 venting voids. With still wetter mixtures ramming becomes 
 difficult, as the concrete will soon begin to quake, after which 
 the ramming should not be long continued as the mass is then 
 
360 CEMENT AND CONCRETE 
 
 semi-fluid, and the stones may gradually work themselves to 
 the bottom of the layer, forcing the mortar to the top. 
 
 519. Rammers are frequently made of wood, but those of 
 iron are believed to be better. The weight of a rammer is 
 limited by the capacity for work of the man who wields it. 
 They are usually made to weigh from twenty to forty pounds. 
 If a man lifts and drops a forty-pound rammer with forty square- 
 inch face twenty times a minute, he is doing less good to the 
 concrete than if he dropped a twenty-pound rammer with twenty 
 square-inch face forty times a minute. If the face of the ram- 
 mer exceeds thirty-six square inches, the result is apt to be a 
 mere patting of the surface of the concrete, unless the rammer 
 is so heavy as to require two men to operate it. Iron rammers 
 with face,- say, three by six inches, and weighing twenty to 
 thirty pounds, are believed to be the most efficient. Still thinner 
 rammers than this may be necessary in work involving such 
 detail as for filling in between iron beams, and are desirable 
 for tamping near the face of the mold. 
 
 520. Rubble Concrete. In massive work the embedding of 
 stones of "one-man size," or larger, in the concrete is a practice 
 that has long been in vogue. The objection is sometimes made 
 that this interferes with the homogeneity of the wall and that 
 variations in expansion may result in injury to the work. It 
 is thought, however, that in large masses this danger is more 
 theoretical than real, and the author sees no objection to this 
 form of construction for many purposes if properly carried out, 
 and it is frequently permitted in important works. Thin walls, 
 the arch rings of bridges, shallow foundations, etc., should not 
 of course be built in this way, because the stresses to which such 
 structures are subjected should be met by a uniform resistance, 
 to avoid the effects of eccentric or irregular loading. In such 
 structures as dams, lock walls, breakwaters, retaining walls, 
 and in many cases bridge piers and abutments, the work may 
 be considerably cheapened without sacrificing the fitness of the 
 structure. The stones thus embedded should be perfectly 
 sound and should not lie nearer one to another than six inches, 
 nor should they lie nearer than this to the face of the wall. 
 The concrete should be mixed rather wet, and much care taken 
 that each stone is completely surrounded by a compact mass of 
 concrete. The stones should be settled into the concrete al- 
 
PLACING IN AIR 361 
 
 ready laid far enough to assure their having a full bed. Stones 
 used in this manner are sometimes called " plums." 
 
 521 . Another class of rubble concrete differing from the 
 above more in degree than in kind, is formed by placing large 
 stones in the work, and filling the joints between them with a 
 rath?r wet concrete in which spalls may be rammed if desired. 
 The difficulty of obtaining a compact wall by this method is 
 perhaps a little greater than when smaller stone are used, but 
 in either case if really water-tight work is desired, the inspec- 
 tion must be thorough. 
 
 The saving in cost by the use of rubble concrete depends 
 upon the local conditions, but under ordinary circumstances 
 when broken stone is employed, the cost of crushing the stone 
 and the cost of cement, for a volume of concrete equal to the 
 volume of the stone imbedded, are practically saved. 
 
 522. Joints in Concrete. In the construction of large 
 masses of concrete in place, joints cannot be avoided; that is, 
 it is not possible to make the entire mass monolithic, as force 
 enough could not be employed to carry up the entire struc- 
 ture at once. Even if this were possible, it would not be de- 
 sirable, since the changes in length of the wall due to changes 
 in temperature would probably result in cracks which would 
 be irregular in outline and mar the appearance of the wall, 
 if they had no more serious effect. 
 
 When the concrete is subjected to vertical forces only, as 
 in foundations for buildings, horizontal joints are less objec- 
 tionable than vertical joints. But in the construction of con- 
 crete lock walls, dams, and breakwaters, vertical joints are de- 
 sirable to confine the cracks to predetermined planes. In the 
 building of such structures, therefore, the method has been 
 adopted of dividing the work into sections of such horizontal 
 dimensions as may be thought best, and completing each sec- 
 tion as a monolith. This will sometimes require the contin- 
 uous prosecution of work for twenty-four or forty-eight hours. 
 Whether this method, involving work at night, which is always 
 more expensive and usually less thorough, is justified by the 
 end sought, depends upon the character of the structure. 
 
 523. If this method is not adopted, and a horizontal plane 
 of weakness is a serious defect, special means should be pro- 
 vided for avoiding this plane of weakness. Such provision may 
 
362 CEMENT AND CONCRETE 
 
 be made by iron dowels set in the concrete at the end of the 
 days work and projecting above the surface to be covered by 
 the concrete placed the next day; steps or hollows, or grooves 
 parallel to the length of the wall, may be left to be filled by 
 the next layer. Large stones weighing a hundred pounds or 
 more are frequently imbedded one half their depth in the 
 last layer of a days work to form a bond with the following 
 layer. 
 
 In any case special care should be taken to thoroughly wash 
 and clean the surface of hardened concrete before continuing 
 the work, using preferably wire brooms for this purpose and 
 removing any stones at the surface that appear to be loose. 
 A thin layer of rich cement mortar should then be laid upon it, 
 into which the first layer of fresh concrete is well rammed. 
 
 If the appearance of the finished face is of importance, special 
 care must also be exercised in joining at this point. Before 
 leaving a layer which is to be allowed to harden before contin- 
 uing the work, the line limiting the height of the concrete at 
 the face should be made perfectly horizontal, for a slight crack, 
 or at least a noticeable line, may be expected at this point, and 
 if not straight it will be the more unsightly. 
 
 524. If for any reason a layer of concrete cannot be carried 
 over the whole area of the wall or foundation, it should neve r 
 be allowed to taper off to a wedge, but a plank equal in width 
 to the thickness of the layer should be set on edge, firmly se- 
 cured, and the concrete tamped against it. In the construction 
 of arches, culverts and sewers, such stop planks may well be set 
 normal to the surface of the intrados instead of vertical. In 
 case more than one layer is left incomplete, they should be 
 stepped back, making an offset for each layer of at least one or 
 two feet. The concrete should never be built up on a smooth 
 batter if new concrete is to be joined to it later. 
 
 525. Keeping Concrete Moist. All concrete should be kept 
 moist from the time it is in the wall until it has become well 
 hardened. Surfaces exposed to the air should therefore be 
 sprinkled frequently for at least several days after placing. 
 An excellent practice is to cover the surface with burlaps which 
 may be kept saturated, as this not only furnishes the necessary 
 moisture, but protects the work from the direct rays of the 
 sun. The interior of a large mass will probably take care of 
 
SURFACE FINISH 363 
 
 itself in this regard, but the precaution has sometimes been 
 taken of leaving vertical holes or wells in the mass, which are 
 kept filled with water for some weeks and are then filled with 
 concrete. 
 
 523. FINISH. Some of the precautions that must be taken 
 to secure a good finish to the face of concrete work have al- 
 ready been mentioned in considering the forms and the meth- 
 ods of deposition. These are usually supplemented, however, 
 by certain special means when the appearance is of much im- 
 portance. 
 
 We must say first, that the application of a plaster of ce- 
 ment mortar to a finished and set concrete face will almost 
 never be permanent. It is seldom that it will adhere with suf- 
 ficient strength to prevent scaling due to differences in expan- 
 sion of the materials of different composition and age. If 
 plaster must be used on the face of a wall, it should be applied 
 before the concrete has set, but it is safer to avoid plastering. 
 It is of course advisable to fill with rich mortar any voids that 
 may appear in the face of the work, but such places should be 
 few. 
 
 If the molds are removed while the concrete is still moist, 
 the face may be coated with a thin grout and then immediately 
 scraped off with the edge of a trowel. This results in filling the 
 small voids in the face of the work, but does not leave a coat of 
 plaster on the surface to scale off. 
 
 527. A good finish may be obtained when the molds are 
 smooth if the workmen will force the blade of a spade or shovel 
 between the fresh concrete and the mold, and pull the handle 
 away from the mold. This has the effect of forcing the large 
 stone back from the face and allowing the mortar to flow in. A 
 layer of mortar is thus left next the mold with no marked line 
 of junction between mortar and concrete, as may be the case* in 
 using a mortar facing. A similar effect may be produced by 
 throwing the concrete against the face of the mold with such 
 force that the larger pieces of aggregate rebound. In very 
 finely finished work this may mar the surface of the sheathing, 
 but ordinarily this method is effective. 
 
 528. When a special layer of mortar is used for facing, there 
 is more danger, perhaps, of making the layer too thick than too 
 thin. As to the richness of the mortar, two parts sand by 
 
364 CEMENT AND CONCRETE 
 
 measure to one volume packed cement is usually sufficient, 
 though a more glossy finish may be made if desired, by using 
 equal parts of cement and sand. It is better to avoid too great 
 a variation between the richness of the mortar used for facing 
 and that used in the body of the concrete. 
 
 One of the best ways of applying such a layer is to prepare a 
 sheet of steel of width equal to the thickness of one layer of con- 
 crete, usually six to eight inches, with two handles on the upper 
 edge to facilitate moving it. At the ends of the sheet, on the 
 side next the mold, rivet short pieces of 1J in. by 1J in. or 2 in. 
 by 2 in. angle iron. This sheet of iron with the projecting legs 
 of the angles against the face of the molds, forms, with the 
 latter, a space one and one-half or two inches thick, which 
 is to be entirely filled with the finishing mortar made rather 
 moist and tamped lightly with edge rammers. The concrete is 
 filled in behind the iron, after which the latter is withdrawn by 
 means of the handles, and the whole mass is thoroughly rammed. 
 The end sought is that the finishing mortar shall have some 
 approximately definite thickness, and that the stones of the 
 concrete shall be tamped into the finishing mortar, but not 
 through it, and thus destroy any sharp line of demarcation 
 between mortar and concrete, and ensure a perfect bonding of 
 the two. It is evident that this can only be accomplished by 
 placing the mortar and concrete at the same time. 
 
 529. One other cautionary remark concerning the use of fin- 
 ishing mortar. With the present state of our knowledge con- 
 cerning the rates of expansion of mortars and concretes of dif- 
 ferent composition, it is not considered wise to use too many 
 combinations in the same structure. To illustrate, a pavement 
 or surfacing of a large concrete structure was once built in layers 
 as follows: first, thick natural cement grout was placed on the 
 concrete foundation; second, natural cement concrete; third, 
 Portland cement concrete; fourth, a richer Portland cement 
 concrete; fifth, Portland granolithic; sixth, rich Portland mortar; 
 and seventh, floated with dry Portland cement and sand. We 
 cannot be absolutely sure that this is bad practice, but it would 
 seem that this structure might have served its purpose with 
 fewer varieties of material, and it is usually considered very 
 doubtful whether Portland cement mixtures will always ad- 
 here well to mixtures of natural cement, although the author 
 
SURFACE FINISH 365 
 
 knows of instances where they have been used in juxtaposition 
 apparently with good results. 
 
 530. Granolithic is a facing or surfacing mortar composed 
 of crushed granite and cement. The granite is usually specified 
 to contain no particles larger than f inch to one inch, and about 
 one and one-half to two and one-half parts are used to one vol- 
 ume of cement. This is more frequently used for foot walks 
 and other places where resistance to wear is required, but may 
 also be used to surface walls, to line reservoirs, etc. It will be 
 mentioned again in connection with cement sidewalk construc- 
 tion. 
 
 531. Exposed concrete surfaces frequently present a patchy 
 appearance. This may be the result of lack of care in placing 
 the concrete next the mold, or it may be due to variations in 
 the purity of the sand or in the amount of water used in mixing. 
 On mortar-faced work this lack of uniformity is less noticeable. 
 The use of slag sand, or of a little fine pozzolanic material, may 
 be advantageous, and a small amount of lampblack in the facing 
 mortar also tends towards uniformity in appearance. 
 
 A very pleasing finish may be given by applying to the set 
 concrete a thin wash of cement and plaster of Paris, though the 
 permanence of such a wash may be open to question. The 
 sheathing should be removed as early as it is perfectly safe to 
 do so, and the concrete surface cleaned from any oil or grease 
 that may have come from the mold planks. The wash, which 
 should be very thin, may be applied with a whitewash brush. 
 A mixture of equal parts Portland cement and plaster of Paris 
 gives a very light gray finish, and one part plaster to three parts 
 cement gives a trifle darker shade. 
 
 532. A rubbed finish of excellent appearance may be given 
 by removing the sheathing before the concrete has set very 
 hard, say after twenty-four to forty-eight hours, and rubbing 
 the surface with white brick or with a wooden float. If there 
 are small voids in the surface, it may be covered with a thin 
 grout of equal parts of cement and sand and then rubbed 
 hard with a circular motion. The grout should not leave a 
 scale on the work, the object being only to fill surface imperfec- 
 tions. 
 
 If the mold boards are removed at just the proper time, a 
 good finish may be given by rubbing with a wooden float, with- 
 
366 CEMENT AND CONCRETE 
 
 out the coating of thin grout. A somewhat similar effect is 
 produced by brushing the surface with brooms or stiff brushes. 
 
 533. " Pebble-dash." What is called a pebble-dash finish 
 was used in the construction of a bridge in the National Park at 
 Washington, D. C. 1 Eighteen inches of the concrete next the 
 face was made of one part cement, two parts sand, and five parts 
 of gravel and rounded stone from one and one-half to two inches 
 in their smallest diameter. After the removal of the forms the 
 cement and sand were b rushed from around the face of the gravel 
 next the surface exposed to view. It was found by experiment 
 that the brushing should be done when the concrete was about 
 twenty-four hours old. At twelve hours the gravel was displaced 
 by the brushing, and after thirty-six hours the mortar had be- 
 come so hard as to be removed from the surface of the stones 
 with difficulty. The forms were therefore designed so that 
 sections 'of the lagging could be removed as desired. The cost 
 of the brushing was said to be about sixty cents per square yard. 
 
 A somewhat similar method is employed in giving to con- 
 crete the appearance of cut stone. The materials used in the 
 surfacing mortar are Portland cement and crushed rock, the 
 character of the rock depending upon the color and texture de- 
 sired in the finish. The molds having been removed after the 
 proper time has elapsed, the mortar covering the face of the 
 particles of crushed rock is removed by brushing or by washing 
 the surface with a weak acid solution, followed by clean water, 
 and finally by an alkaline solution to prevent any further action 
 of traces of the acid which might be left on the face. This last 
 method is said to be patented, "the patent covering the obtain- 
 ing of a natural stone finish for concrete by mechanical, chemi- 
 cal or other means." 2 It is hoped that such a blanket patent is 
 somewhat less formidable than it appears. 
 
 If the sheathing planks of the molds can be removed about 
 twenty-four hours after the concrete is placed, the same effect 
 may be produced without the use of acid. By using plenty of 
 water the cement and finer portions of crusher dust in the face 
 
 1 Capt. Lansing H. Beach, Corps of Engrs., U. S. A., in charge. Work 
 described by Mr. W. I. Douglas, Engr. of Bridges, D. C., Engineering News, 
 Jan. 22, 1903. 
 
 2 Engineering News, May 21, 1903, 
 
SURFACE FINISH 367 
 
 may be washed out with a stiff corn broom, leaving the facets 
 of the crushed rock exposed. 
 
 534. Pointing and Bush-hammering. If the molds have 
 been left in place until the concrete is set hard and it is found 
 that the face of the concrete is not what is desired, it may still 
 be improved although it may not be plastered. With this ob- 
 ject the face is sometimes tooled with the stone cutter's point to 
 give the appearance of rough pointed or rock face masonry. 
 Grooves may be cut to block off the work into rectangles of the 
 proper size, then a draft of one to two inches may be left along 
 all of these artificial joints, and within the draft line the rough 
 pointing may be done. 
 
 A cheaper method, however, is to bush-hammer the entire 
 face, and this tends to mask any lack of uniformity in color or 
 smoothness. Hush-hammering may be done by ordinary labor- 
 ers at a small cost, as one man can go over from fifty to one 
 hundred square feet in ten hours, making the cost from 1-J cents 
 to 3^ cents per square foot, with labor $1.75 per day. Where it 
 is decided beforehand to bush-hammer the work, less pains need 
 be taken in dressing the lagging of the forms. 
 
 535. Colors for Concrete Finish. The addition of coloring 
 matter to cement and concrete is not at present widely prac- 
 ticed, and consequently experience has not been sufficient to in- 
 dicate just what colors may be used without detriment to the 
 work. Lampblack has been most commonly employed, giving 
 different shades of gray according to the amount used. In any 
 large work where the use of coloring matter is desirable and 
 there is not time to institute thorough tests, the advice of a 
 cement chemist should be sought. The dry mineral colors, 
 mixed in proportions of two to ten per cent, of the cement, 
 give shades approaching the color used. Bright colors are diffi- 
 cult to obtain and would not be in keeping with a masonry 
 structure except in architecture. 
 
 When mixed with an American Portland cement mortar, 
 containing one part cement to two parts by weight of a yellow 
 river sand, the particles of which are largely quartz, the colors 
 indicated in the following table are obtained. 
 
 With no coloring matter added, the mortar was a light green- 
 ish slate when dry. Ultra marine green, in amounts up to 8 
 per cent, of the cement, had no apparent effect on the color of 
 
368 
 
 CEMENT AND CONCRETE 
 
 this mortar. Variations in character of cement and sand will 
 affect the result obtained in using coloring matter. The colors 
 indicated below are for dry mortars ; when the mortar is wet 
 the shades are usually darker. None of the materials mentioned 
 in the table seems to affect the early hardening of the mortar, 
 though very much larger proportions might prove injurious. 
 With red lead, however, even one per cent, is detrimental, and 
 larger proportions are quite inadmissible. 
 
 COLORED MORTARS. 
 
 Colors Given to Portland Cement Mortars Containing Two Parts 
 River Sand to One Cement. 
 
 
 
 .|o 
 
 DRY 
 
 WEIGHT OF DRY COLORING MATTER TO 100 POUNDS OF CEMENT. 
 
 Ir j 
 
 MATERIAL 
 
 
 ^ E 
 
 USED. 
 
 
 ""S *-" 
 
 
 i Pound. 
 
 1 Pound. 
 
 2 Pounds. 
 
 4 Pounds. 
 
 II 
 
 
 
 
 
 Dark Blue 
 
 
 Lamp Black 
 
 Light Slate . 
 
 Light Gray 
 
 Blue Gray . 
 
 Slate . 
 
 15 
 
 Prussian 
 
 Light Green 
 
 Light Blue 
 
 
 Bright Blue 
 
 
 Blue . . 
 
 Slate . . 
 
 Slate . . . 
 
 Blue Slate . 
 
 Slate . 
 
 50 
 
 Ultra Marine 
 
 
 Light Blue 
 
 
 Bright Blue 
 
 
 Blue 
 
 
 Slate . . . 
 
 Blue Slate 
 
 Slate . 
 
 20 
 
 Yellow 
 
 
 
 
 
 
 Ochre . . 
 
 Light Green . 
 
 
 . 
 
 Light Buff 
 
 3 
 
 Burnt 
 
 Light Pinkish 
 
 
 Dull Laven- 
 
 
 
 Umber 
 
 Slate . . 
 
 Pinkish Slate . 
 
 der Pink 
 
 Chocolate . 
 
 10 
 
 Venetian 
 
 Slate, Pink 
 
 Bright Pinkish 
 
 Light Dull 
 
 
 
 Red . . 
 
 Tinge . . 
 
 Slate . . . 
 
 Pink . . 
 
 Dull Pink 
 
 2 
 
 Chattanooga 
 
 Light Pinkish 
 
 
 Light Terra 
 
 Light Brick 
 
 
 Iron Ore . 
 
 Slate . . 
 
 Dull Pink . . 
 
 Cotta . . 
 
 Red . . 
 
 2 
 
 
 
 
 
 Light Brick 
 
 
 Red Iron Ore 
 
 Pinkish Slate 
 
 Dull Pink . . 
 
 Terra Cotta 
 
 Red . . 
 
 2* 
 
 536. In some cases it may be sufficient to color the surface 
 of the work by painting. Ordinary oil paints are sometimes 
 applied after washing the surface of the wall with very dilute 
 sulphuric acid, one part acid to 100 parts water, but the per- 
 manence of such a finish seems very questionable. 
 
 The method of obtaining a gray finish by painting with a 
 thin grout of cement and plaster of Paris has already been de- 
 scribed ( 531). Similar methods may be used with the dry 
 mineral colors, and, while their permanency cannot be vouched 
 for, it seems a more reasonable procedure than to paint a con- 
 
PLACING UNDER WATER 369 
 
 crete surface with oil paints. One pound red iron ore to ton 
 pounds cement mixed dry, and then made into a very thin grout 
 and applied to a well cleaned concrete surface with a white- 
 wash brush, gives a pleasing brick-red color; and a rich dark 
 red is given by one pound red iron ore to three pounds cement. 
 The earlier this is applied after the concrete has set, the more 
 likely is it to remain permanent. 
 
 ART. 64. PLACING CONCRETE UNDER WATER 
 
 537. In building a concrete structure under water where the 
 site cannot be coffered, it must be expected that the expense 
 of the work will be increased, and the quality of concrete poorer. 
 The methods employed for subaqueous construction are: 1st, 
 the laying of freshly mixed concrete in roughly prepared forms; 
 2d, placing the fresh concrete in bags of burlap or canvas which 
 are deposited while the concrete is still soft; and 3d, molding 
 in air concrete blocks which are placed in the work when well 
 set. 
 
 In the first method some cement will certainly be washed 
 out of the concrete, the extent of this loss depending upon the 
 condition of the water in which the work is done (i.e., its depth 
 and the amount of current and wave action) and the care with 
 which the concrete is lowered to place. Tamping cannot be 
 done with this method, and any movement of the concrete to 
 level it will cause further loss of cement. 
 
 In the second method the loss of cement will be much less, 
 but the adhesion between the different masses will be slight. In 
 the third method there is no loss of cement and the concrete 
 can be well rammed ; but if small blocks are used, there may 
 be difficulty in so placing them under water as to make a solid 
 structure, while if large blocks are used, special hoisting ma- 
 chinery is required to handle them. 
 
 538. DEPOSITING IN PLACE. The first method mentioned 
 above, depositing fresh concrete in place, is usually the cheap- 
 est and most expeditious method, though it is not likely to 
 dve the best results. When concrete is lowered through water, 
 there is a tendency for the cement to separate from the sand 
 and stone. This tendency seems to be exhibited in a more 
 marked degree with some cements than with others. In con- 
 nection with the construction of the concrete foundations of 
 
370 CEMENT AND CONCRETE 
 
 the Charlestown bridge, a test was devised for determining the 
 relative values of the different lots of cement for depositing in 
 water. 1 Concrete was laid, through a small chute, in a cement 
 barrel placed in a hogshead filled with salt water. It was found 
 that while some specimens would retain their form after twenty- 
 four hours when the barrel was removed, others showed but 
 little cohesion after twenty-four to forty-eight hours. In the 
 former, the cement and gravel remained well distributed through- 
 out the mass, but in the latter much of the cement had sepa- 
 rated from the gravel, and settled in the bottom of the barrel, 
 where it remained in an inert state, while the central portion of 
 the concrete, robbed of its cement, had many voids. As a 
 result of this test, some lots of cement were not accepted for 
 use. 
 
 The finest portion of the cement is very liable to separate 
 from the remainder as the concrete passes through the water, 
 and if subjected to the action of waves or a current, much of 
 the cement will be washed away. In exposed situations it is 
 especially necessary to inclose the site of the work with sheet pil- 
 ing or cribs, or a wall constructed by the bag or block method. 
 When the water level outside the form is constantly changing, 
 the flow of water through^ the joints in the sheathing is especi- 
 ally effective in washing out the cement, and in such conditions 
 the sheathing should be made as nearly water-tight as possible. 
 To this end tongue and groove lagging may t>e used, or the face 
 of the mold may be covered with tarred felt, or canvas, tacked 
 in place. 
 
 539. Laitance is the term applied to the whitish spongy 
 material that is washed out of concrete when it is deposited in 
 water. Before settling on the surface of the concrete, which 
 it does slowly, it gives to the water a milky appearance, hence 
 the name. In fresh water this semi-fluid mass is composed of 
 the finest flocculent matter in the cement, containing generally 
 hydrate of lime. It remains in a semi-fluid condition for a 
 long time and acquires very little hardness at the best. In sea 
 water the laitance is more abundant and is made up of silica, 
 lime and magnesia, with carbonic acid and alumina, its exact 
 
 1 Report of Mr. William Jackson, Chief Engineer. Third Annual Report 
 Boston Transit Commission. 
 
PLACING UNDER WATER 371 
 
 composition depending upon the character of the cement. This 
 interferes seriously with the bonding of the layers of concrete, 
 and when it has settled it should be cleaned from the surface 
 before another layer is placed. 
 
 540. The Tremie. A method frequently employed to pre- 
 vent, as much as possible, the loss of cement, is to make use of 
 a large tube of wood or sheet iron, made in sections so that 
 its length is adjustable, and provided with a hopper at the 
 upper end. Such a tube is called a tremie. The hopper is 
 always above water, and the lower end of the tube, which may 
 also terminate in a hopper, rests upon the bottom of the founda- 
 tion. 
 
 The tremie is first filled with concrete, a box placed over the 
 lower end serving to prevent the escape of the concrete while 
 the tube is being lowered until the end rests upon the bottom. 
 The tube is then lifted from the bottom sufficiently to allow 
 the concrete to escape as fast as fresh concrete is added at the 
 top. The surface of the concrete in the tube should be kept 
 continuously above the water surface. The tremie may be 
 held in position by a crane, or it may be so supported as to al- 
 low of two motions at right angles to each other. Such an 
 arrangement was used in building the piers for the Boucicault 
 Bridge, the tube traveling along a platform, which in turn 
 could move on a track at right angles to the first motion. In 
 using a tremie the thickness of a layer may be regulated at will. 
 
 In the construction of the Charlestown Bridge 1 a tube was 
 used fourteen inches in diameter at the bottom, and about 
 eleven inches in diameter at the neck, above which was a hopper 
 to receive the concrete. When the attempt was made to place 
 too thick a layer at one operation, it was found that the charge 
 was likely to be lost, and the best results were believed to be 
 obtained with layers two feet to two and one-half feet thick. 
 Some experiments were made with a plug designed to keep the 
 water from flowing up through the concrete when the tube 
 was being refilled after a loss of the charge. This plug was 
 made with a central core of wood and sides of canvas expanded 
 by steel ribs. It worked fairly well, but its use was not con- 
 tinued. 
 
 1 Third Annual Report, Boston Transit Commission. 
 
372 CEMENT AND CONCRETE 
 
 541. This principle was employed by Mr. Daniel W. Mead 
 in placing concrete in a small shaft in ninety feet of water. 1 
 An eight inch, wrought iron pipe was screwed together in sec- 
 tions, and provided with a hopper at the upper end and a wooden 
 plug at the lower end. After lowering the pipe into the shaft, 
 the pipe was filled with concrete and it was expected that its 
 weight would force out the plug at the bottom when the pipe 
 was raised. On the first attempt, however, the plug failed to 
 drop out, and on raising the pipe the cause was apparent. The 
 plug had evidently leaked, and as the first concrete was dropped 
 into the pipe it had separated, the broken stone being at the 
 bottom, the sand next, and the cement above had so plugged 
 the pipe as to support the weight of the concrete. The second 
 attempt, when a tighter wooden plug was used and a small 
 pipe placed inside the larger one to assist in loosening the plug 
 if necessary, was successful. 
 
 542. The Skip. Since in submerged work the concrete 
 should be deposited in as large masses as possible, the use of a 
 large skip will probably give better results than the tremie. 
 A box form may be used with hinged lids at the top to permit 
 filling, and two hinged doors at the bottom which may be 
 opened from the surface by a tripping rope when the box has 
 reached the place for depositing the concrete. 
 
 A convenient form of skip is made in two halves, each half 
 having a cross-section either of a right angled triangle or a 
 quadrant of a circle. The two boxes are hinged at their upper 
 inside corners and the pieces through which the hinge rod 
 passes are prolonged upward, the lowering cables being at- 
 tached to their ends. Two opening cables are fastened to the 
 outer corners of the boxes. Two sheets of iron may be used 
 as covers to the boxes, being attached to the hinge rod that 
 serves for the two halves of the skip. 
 
 It is seen that the skip will work on the principle of a pair 
 of ice tongs. While being filled with concrete the box is sup- 
 ported by the lowering cables, and the hinged lids are kept up 
 by some simple contrivance. When full, the lids are closed 
 and the skip lowered till it rests on the bottom; the skip being 
 then hoisted slowly by means of the opening cables, the con- 
 
 1 Trans. Assn. of Civil Engineers of Cornell University, 1898. 
 
PLACING UNDER WATER 373 
 
 crete is gently deposited in place. Su?h skips are supplied by 
 the makers of concrete machinery. 
 
 543. In depositing concrete by means of skips it is well to 
 have the latter of large size, holding not less than a cubic yard, 
 and preferably two cubic yards or more. The larger quantity 
 will compact itself better on account of the greater weight, 
 and the surface which is subjected to wash will have a lesser 
 area in proportion to the volume of the mass. The skip should 
 be completely filled with concrete and tightly closed while it is 
 being lowered. It is important also that the skips be lowered 
 slowly, in order that the inclosed air may be replaced by water 
 without commotion. 
 
 544. The Bag. Mr. Wm. Shield ' devised a bag for de- 
 positing concrete under water which is said to work very satis- 
 factorily. The top of the bag is closed, and has a three-quarter 
 inch wrought iron bar fastened across the end with a loop to 
 receive the hook of the lowering line. The mouth of the bag 
 is slightly larger than the upper end, to facilitate the discharge 
 of the concrete. The bag is inverted to be filled, and the mouth 
 is then secured by a turn of a line provided with loops through 
 which a small tapering pin is passed. This pin is attached to 
 a tripping line, and when the bag has reached the place of 
 deposition, a pull of the tripping line releases the pin; when 
 the bag is. gently lifted, the concrete is deposited in place with 
 such slight commotion that but little cement is said to be lost. 
 
 545. Other Methods of Depositing in Situ. For deposition 
 under water the materials for concrete are sometimes mixed 
 dry, but this is not good practice. The mere soaking of water 
 into cement does not form a compact mortar; the moistened 
 materials need to be thoroughly mixed and, if possible, rubbed 
 together in order to obtain perfect adhesion. Then, too, if the 
 dry materials are lowered to place and water is suddenly al- 
 lowed access to the mass, much of the cement will be washed 
 away in the disturbance caused by the sudden inrush of water. 
 
 M. Paul Alexandre 2 found by experimenting on mortars of 
 "dry" (stiff), "wet" and "ordinary consistency," that mortars 
 
 1 "Subaqueous Foundations," London Engineering, 1892. Abstract in 
 Engineering News, Vol. xxviii, p. 379. 
 
 2 " Rec.herches Experimentales sur les Mortiers Hydrauliques," par M. Paul 
 Alexandre, pp. 93-96. 
 
374 CEMENT AND CONCRETE 
 
 mixed "dry" suffered the greatest decrease in strength by im- 
 mersion in running water. Mortars mixed "wet" suffered the 
 least loss, though their resistance was less than those mixed to 
 the ordinary consistency, since when not subject to the current 
 of water, the wet mortars gave much lower results than those 
 of ordinary consistency. 
 
 546. In order to avoid the washing, out of the cement, the 
 concrete is sometimes allowed to partially set before deposition. 
 Mr. Robert W. Kinipple has used this method and advocates 
 its adoption. 1 In employing this method, the concrete, which 
 should be deposited when of the consistency of stiff clay, re- 
 quires careful watching that it does not set so hard as not to 
 reunite after being broken up. Under ordinary supervision, 
 this will probably not prove as successful as some of the other 
 devices, but it may be found valuable under certain circum- 
 stances. The writer made a few experiments with this method 
 on a small scale in swiftly running shallow water. Much of the 
 cement appeared to be washed out by the current, but the 
 results were somewhat better than were obtained when the 
 concrete was deposited fresh. (See 456.) M. Paul Alexandre 
 made some short time experiments on this point, which indi- 
 cated that but little advantage was gained in allowing the 
 cement to partially set before deposition. 
 
 547. DEPOSITING CONCRETE IN BAGS. The second method 
 of depositing concrete under water, namely/by placing the freshly 
 mixed concrete in coarse sacks and immediately lowering them 
 to place, is very convenient under certain conditions. This 
 method is of especial value in leveling a foundation to receive 
 concrete blocks, or to form a base for concrete deposited in situ. 
 Small bags of concrete have been successfully used in filling the 
 spaces between pile heads which were to support an open caisson. 
 In such a case the bags should be lowered to a diver who places 
 and rams them. If the bags be properly leveled and the earth 
 firm, a part of the load is thus transmitted to the material be- 
 tween the pile heads, while if the earth be very unstable, the 
 bag construction compels the piles to act together, giving lateral 
 stiffness to the foundation and tending to prevent over turning. 
 
 1 "Concrete Work under Water," Proc. Inst. C. E., Vol. Ixxxvii. See 
 also "Notes on Concrete," by John Newman, pp. 116 and 117. 
 
PLACING UNDER WATER 375 
 
 548. The bag method was successfully used in replacing 
 with concrete the timber superstructure of the breakwater at 
 Marquette, Mich. 1 The main portion of the breakwater was 
 built of monolithic blocks on the rock-filled timber substruc- 
 ture. After removing a portion of the rubble filling, a bed was 
 made for the monolithic blocks by laying concrete in place two 
 feet thick, extending from one foot below to one foot above 
 low water datum. This method was afterward replaced by 
 the use of concrete in bags, which made it safe to remove a 
 lesser amount of the rock filling of the crib at the center, and 
 thus decreased the expense of the work. The bags were of 
 eight ounce burlap made 6 feet long and 6 feet 8 inches in cir- 
 cumference, and held about one ton of concrete. They were 
 filled while lying on a skip specially constructed, so that when 
 the skip was in place it could be tripped and the bag placed in 
 its exact position in the work. 
 
 549. In connection with this work a practical indication of 
 the character of the concrete deposited in this manner was 
 given by some small bags of concrete that were laid to protect, 
 during the winter storms, a portion of the crib filling. Mr. 
 Coleman says of this, 2 "Only one layer of these sacks, laid 
 slightly overlapping from the lake side of the crib, was used. 
 The sacks were so lightly filled that when laid as described, the 
 average thickness of the concrete covering was not more than 
 six inches. The crib was storm swept many times without 
 displacing a single sack, and when they were removed in the 
 following spring to facilitate the work, they came away, when 
 pulled up with the floating derrick, a dozen or more at a time, 
 so firmly were they cemented together, and in many cases 
 large rubble stones were lifted up along with them, because of 
 the adhesion of the cement to their surfaces." 
 
 550. The Cost of the concrete in bags was as follows: - 
 
 Materials, cement, sand, stone, burlaps, etc .$5.281 
 
 Mixing concrete and filling bags 912 
 
 Transportation 157 
 
 Depositing .408 
 
 Total cost per cubic yard $6.758 
 
 Or, cost in bags, exclusive of materials .... 1.477 
 
 Major Clinton B. Sears, Corps of Engineers, in charge; Mr. Clarence 
 Coleman, Asst. Engineer. 
 
 J Report Chief of Engineers, U. S. A., 1897, p. 2620. 
 
376 CEMENT AND CONCRETE 
 
 The cost of the first plan, placing a two foot layer of con- 
 crete in situ, where different methods of handling were em- 
 ployed, was, for labor: 
 
 Loading scow with materials $0.411 
 
 Mixing concrete 846 
 
 Depositing 524 
 
 Cost in situ, exclusive of materials $1.781 
 
 551. When concrete bags are used in forming a foundation, 
 the lower layers should usually cover a considerably greater 
 area than that required for the top. Especially is this true if 
 building upon insecure earth. This increased area at the bot- 
 tom may be obtained by building the sides on a batter, or by 
 the use of footing courses. If the latter are used, they should 
 be so designed that in any case the projection beyond the course 
 next above is not greater than the thickness of the layer. 
 
 Before filling the concrete into the bags it should be thor- 
 oughly mixed, as for deposition in the ordinary manner. The 
 practice of using dry concrete for this purpose is reprehensible 
 for the same reason as has been given in 545. It has also been 
 found that if the concrete is mixed and filled into the bags in 
 a dry state, a layer of concrete on the outside may cake before 
 the water has had time to reach the interior portion. The 
 bags should be filled about three-fourths full, leaving the mass 
 free to adjust itself to inequalities in the rock, or to the irreg- 
 ular surface of the previously deposited layer. When strength 
 and compactness are desired, the bags should be placed by a 
 diver and gently rammed. In this way the mass may be well 
 bonded by " breaking joints." 
 
 552. Large Masses in Sacks. Very large bags of concrete 
 are sometimes employed, as in the construction of a breakwater 
 at New Haven, England. 1 "The top of the breakAvater has a 
 width of thirty feet, is ten feet above high water, and is sur- 
 mounted by a covered way and parapet running along the outer 
 side, both sides battering one in eight. The breakwater is 
 unsheltered from the force of the Atlantic, is founded on the 
 rough, natural sea bottom, and the foundation course has a 
 
 1 From London Engineering, quoted in Engineering News, Vol. xxvii, 
 p. 551. 
 
PLACING UNDER WATER 377 
 
 width of fifty feet; the lower portion of the structure, from the 
 bottom up to a level of two feet above low water, consists of 
 one-hundred-ton sacks of concrete deposited while plastic. 
 The canvas with which the concrete was enveloped was of 
 jute, weighing about twenty-seven ounces per square yard. 
 The sacks were dropped into place by a specially designed 
 steam hopper barge. The ' sack-blocks' in the finished work 
 became flattened to a thickness of about two feet six inches. 
 With the exception of this sack work the breakwater is built 
 of plastic concrete laid in situ." Similar sack-blocks of one 
 hundred sixty tons have been employed in breakwater con- 
 struction. 
 
 It is evident that this method of depositing concrete in 
 large sacks is peculiarly suited to forming a foundation on a 
 soft bottom, since, if the bags are made to project well beyond 
 the sides of the molded concrete to be deposited above, they 
 act in the double capacity of a mattress to prevent scour, and 
 a foundation for the upper part of the structure. 
 
 553. Other uses for Bags of Concrete. In the construc- 
 tion of the Merchants' Bridge at St. Louis, bags of concrete 
 were used to check the scour which occurred beneath the up- 
 stream cutting edge of one of the caissons while it was being 
 grounded. The bags were thrown into the river at such a dis- 
 tance above the" pier that they settled to the bottom at the 
 point where the scour was taking place. 
 
 Burlap bags were used at St. Marys Falls Canal for laying 
 concrete under water next the face of the form to prevent 
 washing of the cement in building concrete superstructure for 
 canal walls. As the bags were placed by hand they were made 
 to hold only about two cubic feet of concrete. 
 
 554. Paper Sacks. Paper sacks are sometimes employed 
 instead of jute bags. Dr. Martin Murphy 1 describes the meth- 
 ods employed in filling steel cylinders for the substructure of 
 the Avon Bridge as follows: "Bags made of rough brown paper 
 well stiffened with glucose, were employed and slipped into the 
 water over the required place of deposition. Each bag held 
 about one cubic foot of concrete; smaller ones were used be- 
 
 " Bridge Substructure and Foundations in Nova Scotia," by Martin 
 Murphy. Trans. A. S. C. E., Vol. xxix, p. 629. 
 
378 CEMENT AND CONCRETE 
 
 tween dowels. The bags were quickly made up and dropped 
 one after another, so that the one following was deposited 
 before the cement escaped from the former one. The paper 
 was immediately destroyed by submersion, and the cement 
 remained; it could not escape. The bags cost one dollar thirty- 
 five cents per hundred, or thirty-five cents per cubic yard." 
 The success of this method will depend upon the character of 
 the sacks, for in some experiments on a small scale with sacks 
 of stiff manila paper the author found that the bags were not 
 destroyed, and that no adhesion took place between the separate 
 sacks. 
 
 555. THE BLOCK SYSTEM OF CONCRETE CONSTRUCTION. - 
 
 The advantage of the block system of construction lies in the 
 fact that the individual blocks may be made with the greatest 
 care, and as they are allowed to harden thoroughly before being 
 put in place, the loss of cement incident to the other systems 
 is avoided. There is, however, the difficulty of forming a joint 
 between adjacent blocks. The joints are of great importance 
 when small blocks are employed, since the latter may not have 
 sufficient weight to escape being washed out of the work. Large 
 blocks may make a very solid structure by being simply super- 
 imposed, but special hoisting machinery will be required to 
 place such blocks. 
 
 Sometimes a large bed of mortar is laid in coarse sacking 
 and carefully lowered and spread on the block last laid, the 
 next block being placed upon it immediately. A very rich 
 mortar should be used for this purpose. Usually, however, it 
 is not attempted to place mortar in the horizontal joints in 
 concrete block work laid under water, but it is considered that 
 all vertical joints should be filled with rich Portland cement 
 mortar when the work is to be exposed to wave action. If 
 settlement is anticipated, and large blocks are used, no attempt 
 should be made to break joints in the direction of the longer 
 dimension of the work, but the blocks should bond in a direc- 
 tion transverse to the wall. Concrete blocks may be advan- 
 tageously employed to form the faces of a structure built under 
 water or exposed to wave action, the concrete hearting or 
 backing being built in situ. 
 
 556. For convenience in handling, a groove to receive a 
 chain or cable should be left down two sides and across the 
 
PLACING UNDER WATER 379 
 
 bottom of the blocks to enable them to be placed close together 
 and to facilitate the withdrawal of the hoisting chain. These 
 grooves may afterward be filled with concrete; such recesses 
 are sometimes molded for the sole purpose of filling them with 
 fresh concrete when in place, and thus binding the work to- 
 gether. The molds for forming the blocks should be carefully 
 made in order that the finished blocks may have good bearings 
 one upon another. If the corners are rounded, they are less 
 likely to be chipped off in handling or by having an undue 
 strain come upon the corner when in place. 
 
 If any recesses are desired in the blocks, the pieces placed 
 in the mold to form them should be trapezoidal in cross-section 
 with the longer parallel face against the side of the mold. If 
 such filling pieces are made rectangular, difficulty will be ex- 
 perienced in removing them when the concrete has set. The 
 molds should, of course, be so constructed as to be readily 
 taken apart to be used again. The opposite sides may be kept 
 from spreading by rods which pass through the mold, but such 
 rods are an inconvenience in packing the concrete into the 
 mold, and it is therefore better to truss the mold outside. If 
 such tie rods are used, they may be left imbedded in the con- 
 crete, or removed with the mold, as desired. 
 
 557. Cost of Molding Blocks. An illustration of the use of 
 the block method is furnished in the United States breakwater 
 at Marquette. 1 The general plan of this breakwater has already 
 been briefly noted and two methods of laying a two foot layer 
 of subaqueous concrete, as a foundation for monolithic blocks 
 forming the superstructure proper, have been described. A 
 third method was to mold footing blocks, seven feet by five feet 
 in section and two feet high, which were afterward laid flush 
 with the lake side of the substructure cribs and filled in behind 
 with concrete laid in place. The footing blocks thus assured 
 a good quality of concrete beneath the toe of the monolithic 
 block on the lake side where it was most necessary to provide 
 a good foundation, and also served as a protection behind which 
 the remainder of the two foot layer could be placed with greater 
 facility. 
 
 Many of these blocks were built during the winter in a shed 
 
 Report Chief of Engineers, 1897, p. 2624. 
 
380 CEMENT AND CONCRETE 
 
 artificially heated, the materials being thawed out as required. 
 The molds were of six by six inch and four by four inch pine, 
 lined with two by eight inch plank dressed on one side. Strips 
 of trapezoidal cross-section, nailed inside the mold, provided for 
 two parallel grooves on the bottom and two sides of the block 
 to receive hoisting chains. A dovetail at the back of the block 
 was also formed by three wedge-shaped pieces placed against 
 the back face of the mold. The cost per cubic yard of making 
 forty blocks is as follows: 
 
 1.42 bbls. Portland cement, at $2.75 $3.90 
 
 .45 cu. yd. sand, at $0.45 20 
 
 1.0 cu. yd. stone screenings passing f" sieve, at $1.10, 1.10 
 
 Cost material in concrete per cu. yd $5.20 
 
 Superintendence, labor and watchman $2.21 
 
 Fuel 31 
 
 10 per cent, of cost of warehouse and molds 52 
 
 Total cost of making per cu. yd 3.04 
 
 Total cost per cu. yd. of blocks ready to place 
 
 in work $8.24 
 
CHAPTER XIX 
 
 CONCRETE-STEEL 
 
 558. The ratio between the compressive and tensile strengths 
 of steel is nearly unity. The same thing is approximately true 
 of wood and some other materials of construction. In cement 
 and concrete, however, the conditions are quite different, the 
 strength in compression being from five to ten times the strength 
 in tension. Concrete cannot, therefore, be economically used to 
 resist tension, and in structures requiring transverse strength 
 concrete is at a great disadvantage. 
 
 559. The idea of supplementing the tensile strength of con- 
 crete by the use of iron in combination with it, seems to have 
 been suggested independently by a number of men. It is 
 known that combination beams were tested by Mr. R. G. Hat- 
 field as early as 1855. In 1875 Mr. W. E. Ward, 1 M. Am. Soc. 
 Mech. Engrs., constructed a dwelling entirely of "beton," the 
 floors, roofs, etc., being reinforced with light iron beams and 
 rods. These early uses of the combination have some bearing 
 upon the ability of patentees to cover in their blanket patents 
 more than the peculiar form of the steel member which they 
 advocate in their particular system. 
 
 ART. 65. MONIER SYSTEM 
 
 560. A much more picturesque beginning of the concrete- 
 steel industry is furnished in the story, quite true*, that about 
 1876, a French gardener, Jean Monier, used a wire netting as 
 the nucleus about which to construct his pots for flowers and 
 shrubs, and seeing that the practice might be extended, he 
 called to his aid engineers and capitalists who developed the 
 Monier system of construction. 
 
 This system consists of imbedding in the concrete two sets 
 of parallel rods at right angles to each other, the rods of the 
 two sets being tied together with wire at all intersections. 
 
 1 Proc. Am. Soc. Mech. Engrs., VoJ . iv, p. 388. 
 
 381 
 
382 CEMENT AND CONCRETE 
 
 The larger wires run in the direction of the greater tensile stresses 
 and are usually spaced two to four inches apart. The rods at 
 right angles to these main tension members are to assist in dis- 
 tributing the stress to the main members and may be of smaller 
 diameter. 
 
 The iron rods in this system are designed primarily to resist 
 the tension only, and the form of the bars is not such as will 
 stiffen the structure while the concrete is fresh. In an arch, 
 two systems of netting are used, one near the intrados and one 
 near the extrados. 
 
 561. The main advantages which this system has over some 
 of its competitors are the simple shapes required, that is, round 
 rods, which may always be obtained without difficulty, and 
 the fact that these may be so readily put together by ordinary 
 workmen under supervision. This system is especially adapted 
 to vertical walls, whether curved or straight, and found its 
 first extensive use in the construction of tanks and reservoirs. 
 It has been extended, however, to the construction of sewers, 
 floors, roofs, and arch bridges. 
 
 One of the practical disadvantages of the system is that the 
 nets are somewhat difficult to handle and keep in position, and 
 in thin sections it has not been found practical to imbed the 
 nets in concrete containing broken stone of the ordinary size. 
 The use of cement mortar, usually one part cement to three 
 sand, has been found necessary in order to get a perfect con- 
 nection between the wires and the body of the work. This, 
 of course, increases the cost. Another objection has been 
 urged against it, namely, that the transverse rods do not in 
 general have any ^duty to perform, and are simply a waste of 
 material so far as the final strength of the structure is con- 
 cerned. While this may be so in certain forms of construction, 
 it may be met by the statement that these cross-rods may be 
 made as small as desired if they are to act merely as spacers 
 for the main rods. In slabs, walls, etc., however, these cross- 
 rods have a purpose, and in some other systems members are 
 supplied to fulfill this necessary function. 
 
 562. Some very bold arches have been built on the Monier 
 system, including three bridges in Switzerland having 128 foot 
 span, 11 foot rise, and a thickness of but 6| inches at the crown 
 and 10 inches at the abutments. 
 
PATENTED SYSTEMS 383 
 
 A Moriier arch of 32.8 foot span, rise one-tenth of span, 
 width 13.2 feet, in which the mortar at the crown was six inches 
 thick and eight inches at the abutments, was tested in Austria 
 in 1890. It held a fifty-three ton locomotive on half the arch, 
 and finally failed at the abutments under a load of 1,700 pounds 
 per square foot over half the span. 
 
 563. Pipes are now made by this system at yards and trans- 
 ported to the place of use. It has also been used as a substi- 
 tute for iron in cylinders for bridge piers. A novel use of this 
 system consists in making a" pipe covering for piles exposed to 
 marine borers. The pipe, which is long enough to reach from 
 above the water surface to below the bed of the waterway, is 
 slipped over the head of the pile and settled a short distance 
 into the mud or silt of the bottom with the aid of a water jet. 
 A question, however, has been raised as to the action of con- 
 crete and iron in combination in sea water on account of the 
 possible setting up of galvanic action. 
 
 ART. 66. WUNSCH, MELAN, AND THACHER SYSTEMS 
 
 564. Wiinsch System. This system, which was invented 
 in 1884 by Robert Wiinsch of Hungary, consists of two iron 
 members of angle irons and plates imbedded in concrete, the 
 lower member being arched and conforming to the outline of 
 the soffit, while the upper one is horizontal and continuous. 
 The two members are riveted together at the crown, and at the 
 abutment are rigidly connected by a vertical member. The 
 several systems of rib bracing thus constructed are connected 
 laterally at the abutment by channel bars running transverse 
 to the arch and riveted to the bottom of each vertical in the 
 abutment. Assuming that the abutments are stable, it is evi- 
 dent that we have here not simply an arch, but also some ele- 
 ments of the cantilever. The spandrels being built up solid of 
 concrete, there is no definite arch ring, and the quantity of 
 material required, especially in long spans, is likely to be much 
 greater than in other systems. On the other hand, the great 
 depth at the springing permits the use of concrete only moder- 
 ately rich in cement. 
 
 565. A bridge of this type, built at Neuhausel, Hungary, 
 consists of six spans of about 56 feet each, rise 3.7 feet, thick- 
 ness at crown 9.8 inches, and at springing line 54.3 inches. 
 
384 CEMENT AND CONCRETE 
 
 The total width of the arch was 19.7 feet and contained thir- 
 teen systems of arch ribs. Concrete in the abutments below 
 water was made mainly of Roman cement. Above water it 
 was composed of one part Portland to eight or ten parts sand 
 and gravel. Ten to twelve inches of the arch was built of 
 strong Portland concrete rammed in layers at right angles to 
 radial lines of the arch, special care being taken with that part 
 below the bottom arched member. An arch was usually com- 
 pleted in one day, and the centers remained in place thirty to 
 forty days, the greatest settlement on the removal of centers being 
 two-thirds of an inch. This bridge contained 1,346 cubic yards of 
 concrete and 88,180 pounds of iron, and cost, complete, $13,700. 
 
 566. Melan System. This system, invented by an Austrian 
 engineer, Joseph Melan, consists of arched ribs between abut- 
 ments as in bridges, or between beams or girders as in floor 
 construction, the space between the ribs being filled with con- 
 crete. Steel I-beams curved to the proper form are usually 
 employed for the reinforcement, though angle iron flanges with 
 lattice connections have been used in some of the large bridges. 
 The steel members extend into the piers or abutments and are 
 there connected by angles or other shapes, and firmly imbedded 
 in the concrete. 
 
 567. This system as adapted to bridge construction has 
 probably met with greater favor among American engineers 
 than any other form. Perhaps this is because of the stiffness 
 of the form of iron beam used, and because by assuming a 
 rather high fiber stress for steel the reinforcement may be de- 
 signed to withstand the entire bending moment without exces- 
 sive dimensions for the steel members. There is thus a feeling 
 of security in its use that is not felt in the same degree with 
 other systems. The arch dimensions are determined by com- 
 puting the forces and required thickness of arch ring after 
 assuming certain safe working stresses for the steel and con- 
 crete; but if desired, the size of steel members may then be 
 increased slightly where necessary to such dimensions that 
 with unit stresses of, say, one-half the elastic limit, the entire 
 bending moment shall be taken by the steel. Some of the 
 largest bridges built after this system in the United States are 
 the five-span bridge at Topeka, Kan., and the three-span bridge 
 at Paterson, N. J. 
 
PATENTED SYSTEMS 385 
 
 568. Thacher System. A modification of the Melan system 
 is that invented and patented by Mr. Edwin Thacher. Steel 
 bars are used in pairs and imbedded in the concrete near the 
 intrados and extrados of the arch and extending well into the 
 abutments. The bars of each pair may be connected by bolts 
 or stirrups, though Mr. Thacher's original idea seems to have 
 been to have no connection between two bars of a pair ex- 
 cept through the concrete. The bars are provided with pro- 
 jections which may be in the form of rivet heads, lugs, or 
 bolts, to increase the resistance of the bars to slipping in the 
 concrete. 
 
 569. Mr. Thacher has more recently designed a special form 
 of rolled bar having projections that serve the same purpose 
 as the rivet heads mentioned above. Several bridges have 
 been built on this system, one of the most notable of these being 
 the Goat Island bridge at Niagara Falls, one span of which is 
 110 feet in length. 
 
 570. In the construction of arch bridges many of the other 
 systems are simply modifications of the Melan. The shapes of 
 the steel members may have different forms, and the connec- 
 tions between the pairs of bars forming the arch ribs may vary 
 to suit the idea of the inventors. But though these systems 
 lose their identity in long-span arches, their distinctive features 
 are more apparent in the construction of floors, roofs, columns, 
 etc. 
 
 ART. 67. OTHER SYSTEMS OF CONCRETE-STEEL 
 
 571. The Hennebique System. The rods are here arranged 
 in pairs, one above the other, in a vertical plane. In girders, 
 the bar in the tension side is straight, while the other one of 
 the pair is horizontal for a short distance along the center of the 
 span, the ends being inclined upward near the ends of the 
 beam. The two bars are connected by bent straps or U-bars 
 so that the steel reinforcement may be compared to a queen 
 post truss within the concrete. This system has been used in 
 the construction of bridges, both arch and girder, floors, roofs, 
 stairways, etc., but it is in beams and girders that its distin- 
 guishing characteristics are best displayed. 
 
 572. A beautiful arch on this system is the bridge over the 
 river Vienne at Chatellerault, France, consisting of three spans, 
 
386 CEMENT AND CONCRETE 
 
 the central one of which is 164 feet long, with rise of 15 feet, 
 8 inches. Four arch ribs 20 inches deep .support the roadway, 
 25 feet wide, by posts forming a skeleton spandrel. 
 
 573. Kahn System. In this system, which is somewhat 
 similar to the Hennebique, the distinguishing feature is the care 
 taken to provide against shear, or against that combination of 
 tension and shear which tends to cause failure in a beam by 
 cracks that extend diagonally upward toward the center of 
 span from near the points of support. The steel plates forming 
 the tension members are sheared longitudinally at intervals, 
 and short ends are bent up at an angle of forty-five degrees 
 with the horizontal. These ends, which may be compared to 
 the tension diagonals of a truss, are thus a part of the main 
 steel member, and the stress is transferred directly to the 
 latter without dependence on the concrete. 
 
 The advantages are the great resistance offered by the bar 
 to being pulled out of the concrete and the thorough manner 
 in which all tension stresses may be provided against. The 
 main disadvantages would seem to be the necessity of detailed 
 shop work for each size of girder, the inconvenience of shipping 
 the steel in its complete form and the difficulty of thoroughly 
 tamping the concrete around the diagonals. 
 
 574. The Ransome System. One of the earliest patents to 
 be issued in this country for a method of using concrete and 
 iron in combination was that issued to Mr. E. L. Ransome in 
 1884. The valuable and distinctive feature of this system is 
 the use of a square bar that has been twisted cold. This twist- 
 ing not only insures a good bond between the concrete and iron, 
 but actually somewhat increases the strength of the bar. 
 
 In building beams with twisted bars as tension members, the 
 latter are given a slight inclination from the center upward 
 toward the ends. For use in buildings, as in floors and columns, 
 and for covers to areaways, and similar uses, this system is 
 largely employed. 
 
 575. Roebling System. As its name implies, wire forms 
 the main feature of this system, and in a general way it resem- 
 bles the Monier. Its application thus far is found principally 
 in floor construction, two distinct methods being used. In the 
 arched floor a wire netting, stiffened by round steel rods woven 
 through it is sprung between the lower flanges of the main 
 
STRENGTH 387 
 
 I-beams of the floor. This netting, further stiffened and held in 
 place by iron rods running parallel to the axis of the arch, forms 
 a permanent center for the placing of the concrete, which fills 
 all of the space to the level of the top of the I-beams. A level 
 veiling below is obtained by a similar netting laid flat against 
 the under side of the I-beam and fastened thereto. This acts 
 as a wire lath to receive a coat of plaster. If the level ceiling 
 is not necessary, the plaster may be applied to the under side 
 of the arch netting, in ..which case the lower flange of the I- 
 beam should be encased in concrete to protect it from corrosion 
 and fire. 
 
 576. For lighter loads, flat bars are placed at suitable in- 
 tervals above and below the I-beams and clamped to the flanges. 
 To these bars the wire netting is attached, a thin layer of con- 
 crete laid on the upper wire incasing the bars, and plaster ap- 
 plied to the lower netting forming the ceiling. Cinder concrete 
 is usually employed with this system. 
 
 577. Expanded Metal. The use of whac is commonly known 
 as expanded metal lath has been extended to concrete-steel 
 construction. As in the Monier and Roebling systems, the 
 strength and stiffness of the structure are increased by the use 
 of steel rods in connection with the expanded metal, the chief 
 duty of the latter, where great strength is required, being that 
 of a distributing member. Expanded metal is made from 
 sheet steel by shearing short slits parallel to the grain, and 
 extending the sheet at right angles to the slits, resulting in a 
 network o diamond shaped openings. The metal used is of all 
 weights up to one-quarter inch thick with meshes six inches long. 
 
 578. The steel bars used in connection with expanded metal 
 by the St. Louis Expanded Metal Fireproofing Co. are square, 
 with frequent corrugations surrounding the bar. These corru- 
 gations serve only to prevent the slipping of the bars in the 
 concrete without adding to the strength. 
 
 The applications of this system include conduits, sewers, 
 and walls of buildings, as well as floors and roofs. 
 
 ART. 68. THE STRENGTH OF COMBINATIONS OF CONCRETE AND 
 
 STEEL 
 
 579. While we have in this country been somewhat slow in 
 acknowledging the worth of concrete-steel construction, there 
 
388 CEMENT AND CONCRETE 
 
 is now a strong interest displayed in the subject; many experi- 
 ments are being made in our educational and commercial labo- 
 ratories and the theory of the action of concrete and steel in 
 combination is being rapidly developed. It is natural that in 
 the investigation of a form of construction permitting so many 
 variations in methods of preparation, that the opinions now 
 advanced, based on insufficient data, should be more or less 
 conflicting. 
 
 580. Experiments. The experiments of M. A. Considere, 
 made in France between 1898 and 1901, which have been made 
 more available to us through the translation and collection oi 
 his articles on the subject by Mr. Moisseiff, 1 are exceedingly 
 valuable. The effect of the quality of the steel and the con- 
 crete, of repeated loads, of changes in volume in hardening, 
 and many other points are carefully analyzed by experiment 
 and theory. 
 
 One of the most important deductions drawn by M. Con- 
 sidere is that fibers of concrete within what may be called the 
 sphere of influence of a reinforcing rod of iron or steel, is capable 
 of enduring very much greater elongations without visible frac- 
 ture than similar concrete without reinforcement. The expla- 
 nation advanced for this is that the steel so distributes the 
 stress throughout the length of the concrete in tension that 
 the development of insipient fractures or excessive elongations 
 at the weaker sections of the concrete is prevented until each 
 section has taken its maximum load. The conclusion to which 
 this theory leads is that the resistance of the concrete through- 
 out the area of influence of the steel reinforcement, is main- 
 tained far beyond that degree of deformation which, in concrete 
 not reinforced, would cause its rupture. 
 
 581. Neglect of Tensile Strength. Notwithstanding these 
 conclusions, it is believed that it is sufficient in most cases of 
 design to neglect the tensile strength of the concrete in concrete- 
 steel combinations. This course may be defended by the fol- 
 lowing considerations. The tensile strength of concrete is, at 
 best, not usually above two hundred to four hundred pounds 
 per square inch. If the stress on the extreme fibers of a beam 
 
 1 " Reinforced Concrete," by Armand Considere, McGraw Publishing Co., 
 New York. 
 
STRENGTH 389 
 
 is three hundred pounds, and we consider that this stress de- 
 creases uniformly toward the neutral axis, the mean stress is 
 but one hundred fifty pounds per square inch. Again, if we 
 disregard M. Considered conclusions, we find that since the 
 modulus of elasticity of steel is, say, fifteen times that of con- 
 crete, the former is only stressed to forty-five hundred pounds 
 per square inch when the imbedding concrete has reached its 
 ultimate strength. 
 
 582. The resistance of concrete to tension may easily be 
 destroyed or impaired by accident, especially when fresh. The 
 properties of concrete vary so much with the materials, the 
 proportions, and the manipulation, and the investigation of 
 the behavior of concrete and steel under stress is as yet so in- 
 complete, as to make refinements in theoretical treatment not 
 only unwarranted but really undesirable for practical purposes, 
 since they lead to the appearance of greater accuracy than is 
 in reality attainable. 
 
 It is true that by the judicious selection of values for the 
 constant appearing in formulas for the strength of concrete- 
 steel beams, the results of such formulas sometimes show a re- 
 markable agreement with the results of that series of actual 
 tests for which the constants have been selected; but one has 
 only to recall his experience in other lines, hydraulics for in- 
 stance, to realize the importance of the almighty constant. 
 The opinion sometimes advanced, that the strength of a given 
 concrete-steel beam may be as accurately derived by formula 
 as can the strength of a steel beam, the writer does not believe 
 to be tenable, at least in the present state of our knowledge 
 concerning the behavior of concrete. 
 
 583. To neglect the tensile strength of the concrete will 
 result in a slight increase in the required area of steel reinforce- 
 ment, and, in so far as the tensile strength of the concrete 
 may be developed, will tend to make the compression side of 
 the beam weaker than the tension side. The only objection 
 to this is that the failure of the beam, though at a higher 
 load, may be more sudden. This possibility, however, seems 
 less serious than the error of depending on the tensile strength 
 of the concrete only to find it lacking at the critical mo- 
 ment. 
 
 Since the aim here is to develop a formula that may be used 
 
390 CEMENT AND CONCRETE 
 
 with safety in the design of structures, and since to neglect 
 the tensile strength of the concrete is to add an unknown, 
 though probably small, factor of safety, the tensile strength 
 will not be considered in the following analysis. 
 
 ART. 69. CONCRETE-STEEL BEAMS WITH SINGLE 
 REINFORCEMENT 
 
 584. Definitions. In this discussion the word strain has 
 its technical meaning, the relative change in length of a piece 
 under stress. It is usually expressed as the ratio of the elonga- 
 tion (or shortening if in compression) to the original length of 
 the piece. But for our purpose it is the ratio of the increment 
 of change in length, occasioned by a given increment of stress, 
 to the length of the piece before the increment of stress was 
 applied. These two expressions for strain are usually consid- 
 ered equivalent, since, according to Hooke's law, the ratio be- 
 tween stresses and corresponding strains, for a given material, 
 is constant within the elastic limit. But in dealing with con- 
 crete it is found that, even before the stresses become excessive, 
 Hooke's law does not hold true. Bearing in mind, then, the 
 meaning of the word strain, we represent as usual the ratio of 
 stress to strain by E, the modulus of elasticity, or 
 
 ^ _ stress 
 strain 
 
 Let E a = modulus of elasticity of steel. 
 
 E c = modulus of elasticity of concrete in compression. 
 / = tension in steel, fbs. per sq. in. 
 / = compression in concrete, Ibs. per sq. in. 
 a = thickness of steel considered as a flat plate, or the area of imbedded 
 
 steel bars per inch of width of beam z. 
 yi = distance the extreme fiber of concrete in compression is from the 
 
 neutral axis. 
 2/2 = distance the center of the steel reinforcement in the tension side 
 
 of the beam is from the neutral axis. 
 i = depth of concrete below reinforcement. 
 d = y\ + 7/2 and h = d + i. 
 
 Xi = unit compression of extreme fibers of concrete in compression, 
 X 2 = unit elongation of steel in tension. 
 
 Tjl 
 
 Represent by R, and by r. 
 
 &c fc 
 
SINGLE REINFORCEMENT 
 
 391 
 
 585. Formulas for Constant Modulus of Elasticity. The 
 cross-section of the beam, the graphical representations of the 
 strains and of the stresses are shown in the following diagrams: 
 
 FIG. 10. 
 CROSS-SECTION. 
 
 STRAINS 
 
 FIG. 13. 
 STRESSES. 
 
 Figure 12 shows the conditions when the stresses are so small 
 that the modulus of elasticity of the concrete may be considered 
 constant, and this case will be first considered. 
 
 In the strain diagram, A t represents the deformation of the 
 extreme fiber of concrete in the compression side of the beam, 
 and Ag the deformation of the steel. Since a section plane be- 
 fore bending is considered to be plane after bending, the steel 
 is considered not to slip in the concrete, and NN is the neutral 
 axis, 
 
 r^ = ; but E t = ~, and 
 
 ^2 2/2 A. 2 
 
 r - f- 
 
 LJC x 
 
 or 
 
 and 
 
 A.J = ^ and A! = 
 
 1 _ /l ^ /C *-''* 
 
 fc &s ft 
 
 (Eq. 1.) 
 
 In the stress diagram the triangle NAB represents the 
 total compressive stress on the concrete for unit width of beam, 
 
 and is equivalent to a single force i^H. applied at the center of 
 
 gravity of the triangle. 
 
 The total compressive stress for section of width z is 
 
 The total tension in the steel is T = zaf s . 
 
392 CEMENT AND CONCRETE 
 
 As we disregard the tensile strength of the concrete, and as the 
 total normal compression and total normal tension on a section 
 must be equal, as they are the two forces of a couple, we have 
 
 P = T, or z^f c =zaf,, 
 whence a = f = f^- (Eq. 2.) 
 
 Js 4 ^f 
 
 2 
 
 The point of application of the force P is - y l above the neu- 
 
 o 
 
 tral axis, while the point of application of T is y 2 below the 
 
 (9 \ 
 
 ^ ?/l + 2/ 2 ), 
 
 and the moment of resistance is equal to either force into this 
 arm, 
 
 H.r*. + 
 
 substituting the value of y 2 given in (1) and reducing, 
 
 (Eq - 3 -> 
 
 586. Formulas for Varying Modulus of Elasticity. The fore- 
 going formulas are based on the supposition that the compres- 
 sive stress in the extreme fiber of the concrete has not passed 
 the point beyond which equal increments of stress no longer 
 produce equal increments of strain or deformation. They are 
 based, in other words, on the common theory of flexure, except 
 so far as we have departed from the application of this theory 
 in neglecting the tensile strength of the concrete. It is well 
 known that even for steel and wooden beams this common 
 theory does not, and is not meant to, apply outside the elastic 
 limit. In the case of concrete, however, it has been found that, 
 even for quite moderate stresses, the modulus of elasticity is 
 not constant (Art. 56), but that after a certain stress is reached 
 the modulus decreases with increasing stress. The effect of 
 this upon the internal forces may be illustrated by the curve 
 N B in Fig. 13. The extreme fiber is supposed to be subjected 
 to the stress f c ; the fibers nearer the neutral axis have a smaller 
 stress per square inch, and the modulus of elasticity for this 
 smaller stress is greater; but in order that a section 1 that is 
 
SINGLE REINFORCEMENT 393 
 
 plane before flexure shall be plane after flexure, the strain must 
 be proportional to the distance from the neutral axis. It fol- 
 lows, then, that the stresses in the inner fibers do not decrease 
 accord' ng to the ordinates of the triangle, but are greater than 
 indicated by such ordinates. The exact form cf the curve 
 B N is not known, but the examination of a number of de- 
 formation curves has indicated that it is parabolic, and for the 
 purpose of this discussion it may be considered a parabola 
 with axis A B without serious error, although it is known the 
 axis does not coincide with A B for stresses below the elastic 
 limit of the concrete. 
 
 587. While the formulas derived in 585 may representation, 
 the conditions existing in a beam subjected to very moderate 
 stresses, it appears that beyond the limit of stress at which 
 the modulus of elasticity of concrete becomes variable, they 
 should be so modified as to take into account this variable 
 modulus. 
 
 Then if A B in Fig. 13 now represents f c and MS = /, we have 
 as before, 
 
 The total stress on the concrete above the neutral axis is 
 
 2 
 now represented by the area within the parabola, or f c j/ lt and 
 
 the total compression on section of width z is 
 
 and the total tension 
 
 T r = zal. 
 As these are the two forces of a couple 
 
 3 22/i/c = z/ 8 ; 
 whence a = | = | &.. (Eq. 5.) 
 
 The point of application of P' is on a line through the center 
 
 5 
 of gravity of the parabola, or - y^ from the neutral axis, while 
 
 o 
 
 the point of application of T' is at distance ?/ 2 below the neutral 
 
394 CEMENT AND CONCRETE 
 
 axis; the arm of the couple is, therefore, ^ y l + y z , and the 
 
 o 
 
 moment of resistance 
 
 2 
 
 = ^ '+?/ 
 
 Substitute value of y z given in (4) 
 
 59 f 77 
 
 . .f . <>. i ^ .. .. / **e 
 
 In applying these formulas, it must be remembered that 
 (1), (2), and (3) are applicable where the stresses are below 
 the point at which the modulus of elasticity of the concrete 
 begins to diminish, while (4), (5), and (6) illustrate the con- 
 ditions for stresses above that limit. 
 
 588. Example. Design a beam of 10 foot span to carry a 
 load due to 20 feet head of water. 
 
 Load per square foot = 20 X 62.5# = 1250#. 
 Total load per foot width of beam = 125,000# = W^ 
 First, using Eqs. 1,2, and 3. 
 
 M = ~ = 187,500 inch-lbs. on beam 1 ft. wide, (z = 12). 
 
 o 
 
 Assume 
 
 /, = 12,500, f e = 500, r = f s = 25; 
 
 /c 
 ' ft 1 ' 
 
 E a = 28,000,000, E c = 2,000,000, R = ~ = 14. 
 From (3) 
 
 M = 187,500 = 12 X 500 + 
 
 y* = 25.5, 2/ t = 5.05 inches. 
 From (2) 
 
 az = .101 X 12 = 1.21 sq. in. of steel for beam 12 in. wide. 
 
SINGLE REINFORCEMENT 395 
 
 From (1) 
 
 r 25 
 
 7/2 = -= v/j = - 7 X 5.05 = 9.02 inches, 
 it 
 
 If i = thickness of concrete below center of steel bars = 2 inches, 
 k = total depth beam = 5.05 + 9.02 -f- 2.00 = 16.07 inches. 
 
 Second, using Eqs. 4, 5, and 6. 
 Assume 
 
 /. = 50,000, f e = 2,000, r = '- = 25; 
 
 Jc 
 
 E s = 28,000,000, E c = 1,400,000, R = (' = 20. 
 
 1C 
 
 As the stresses per square inch given above are approxi- 
 mately the breaking strengths of the materials, we must supply 
 a factor of safety, say 4; i.e., design the beam to withstand four 
 times the required bending moment before the stresses assumed 
 above are attained. 1 
 From (Eq. 6) 
 
 M = 4 M = 4 X 187,500 = 1 2 X 2000 (^ + | X ?j 
 
 whenc 2 _ 187,500 X 4 x 12 _ 
 
 ' y * ' 12 X 2000 xl5~ 
 
 or, y v =5. 
 
 From (Eq. 5) 
 
 2 5 
 a = 3 25 = ' 133 inch) 
 
 and az = 1.6 square inches of steel for 12-inch width of beam. 
 
 1 The method of using the breaking strengths of the materials, and com- 
 puting the ultimate resistance equal to a certain number of times the desired 
 strength, is considered inferior to that of assuming safe working stresses and 
 computing directly the safe load. These safe working stresses should be 
 fixed with reference to the elastic limit of the materials, rather than with 
 reference to ultimate strength. The use here of the term factor of safety is 
 for the momentary purpose of emphasizing the fact that the conditions 
 assumed in deriving equation (6) are such as are supposed to exist under 
 comparatively high stresses; but the formulas may evidently be applied to 
 the safe working stresses the same as equations (1), (2) and (3), and in the 
 present example the same size beam will result by eliminating "factor of 
 safety" and using working stresses equal to one-fourth the values of the 
 stresses assumed. 
 
396 CEMENT AND CONCRETE 
 
 From (Eq. 4) 
 
 7* 25 
 
 2/2 = ft-Ui = go t/i = L25 x 5 " = 6 ' 25 inches - 
 
 If i = 2 inches as before, 
 
 h = total depth beam = 5.00 + 6.25 + 2.00 = 13.25 inches. 
 
 It is seen that equations 4, 5 and 6 give, for the assumption 
 made, a lesser depth of beam with more reinforcement than 
 is given by equations 1, 2 and 3 with the corresponding as- 
 sumptions as to stresses and moduli. 
 
 589. An inspection of the equations shows that to increase 
 the amount of steel reinforcement in the tension side of the 
 beam tends to move the neutral axis nearer to the tension 
 side, and bring a greater area of cross-section of concrete into 
 compression. If we arbitrarily decrease the depth of the beam 
 which must withstand the same bending moment, it will in- 
 crease the required area of reinforcement, and if carried too 
 far will eventually raise f c beyond a safe value. On the other 
 hand, if we take the beam as designed in accordance with equa- 
 tions 1, 2 and 3 and subject it to a greater bending moment 
 than that for which it is designed, then so long as R remains 
 constant, r also remains constant, that is, the steel and con- 
 crete are equally overstressed ; but since R increases with the 
 load, r will also increase, that is, the increment of stress in 
 steel will be relatively greater than that in concrete. 
 
 590. Excessive Reinforcement. In the solution of the above 
 example if we introduce the requirement that the total depth 
 of the beam shall be but 12 inches, while the quality of the con- 
 crete is not improved, we may assume, as before, E s 
 28,000,000 and E c = 1,400,000. Let us introduce the same 
 factor of safety, 4, by using f e = - -- = 500 pounds instead 
 of designing the beam for four times the required bending 
 moment; as we have seen, this does not affect the result. 
 Since the depth of the beam is fixed, /, and r cannot be as- 
 sumed, but must be found, together with a. 
 
 We have 
 
 d 2/i + 2/2 = 12 2 = 10 inches, and y 2 = 10 y r 
 From (6 a) 
 MO = T % x 12 X 500^' + f X 12 X 500^(10 - y t ) = 187,500, 
 
EXCESSIVE REINFORCEMENT 397 
 
 Solving, we have y t = 6 inches nearly, 
 and ?/ 2 = 10 6 = 4 inches. 
 
 From (4) **- = ^ ' 
 
 ?/i fcE. 
 
 Substituting values of y^y^ f e , E c &nd E t , we have 
 /, = 6,667 Ibs. per sq. in. 
 
 F rom( 5) .-!*- fxjjjjgx 6 -JOin. 
 
 and az 3.6 sq. in. of metal to each foot width of beam. This 
 is more than double the amount of reinforcement required for 
 a 13.25 inch beam, while the steel is stressed only 6,667 Ibs. 
 per square inch. 
 
 It may be asked why not use a smaller area of metal, say 
 2 sq. in., stressed to 12,000 Ibs. per square inch, giving the same 
 total tension; but a moment's consideration shows that in order 
 that the metal should assume this higher stress, its elongation 
 must increase proportionally, involving a corresponding in- 
 crease of strain in the concrete in compression with an accom- 
 panying increase in stress beyond the assumed safe limit of 
 500 Ibs. per sq. in. 
 
 591. To pursue this subject of excessive reinforcement a 
 little further, let us examine some tests of concrete-steel beams 
 made by Prof. Gaetano Lanza and reported in Trans. Am. 
 Soc. C. E. for June, 1903. 
 
 In these beams the width z = 8 inches, h = 12 and d = 10 
 inches nearly. The span was 11 feet. Proportions in concrete 
 by volume 1 part Portland cement, 3 parts sand, 4 parts broken 
 trap that would pass 1 inch ring, and 2 parts of the same rock 
 that would pass % inch ring. Both plain and twisted square 
 steel bars were used as reinforcement, the plain bars having a 
 tensile strength of about sixty thousand pounds per square 
 inch and the twisted steel about eighty thousand pounds per 
 square inch. 
 
 If we assume the ultimate strength of the concrete to be 
 2,000 pounds per square inch, the modulus of elasticity at this 
 high stress to be 1,400,000 and the modulus of the steel to be 
 28,000,000, we have, 
 
 28,000,000 
 
 R 
 
 1,400,000 
 
398 CEMENT AND CONCRETE 
 
 and for twisted bars, 
 
 80,000 
 
 r = ^ooo- = 40 ' 
 
 ' From Eq. (4) ?/ 2 = ^y, = 2 y 1} 
 
 ' 37/j = 10 inches, y^ inches. 
 
 o 
 
 From E(i. (5)a= -^ =^X ^-X -^ =^ = .055, and az =.444. 
 or o o 4U lo 
 
 That is, .444 sq. in. of twisted steel reinforcement is required 
 in the beam 8 inches wide in order that the stresses in concrete 
 and steel shall simultaneously reach the values of 2,000 and 
 80,000 Ibs. per square inch, respectively. 
 
 From (6) M = 8 X 2000 X g + f X ^ 
 
 = 311,100 inch-pounds. 
 
 One beam having .56 sq. in. reinforcement, or an area very 
 close to the theoretical amount called for above, broke under a 
 bending moment of 470,000 inch-lbs. Eight other beams hav- 
 ing a greater area of reinforcement gave moments of 355,000 
 to 443,000 inch-lbs., and the average of the nine bars was 403,000, 
 or 30 per cent, greater than the value derived by formula. 
 
 592. Included in the series of .tests were three beams, in 
 each of which were placed two 1^ inch twisted rods. As we 
 have seen, the correct amount of steel to develop the full strength 
 of both steel and concrete is about .444 sq. in.; the three bars 
 mentioned had 3.12 sq. inches of steel, or a large excess of 
 reinforcement. To determine the theoretical moment of re- 
 sistance of these beams, assume as before: 
 
 E s = 28,000,000, 
 E c = 1,400,000, 
 f e = 2,000. 
 
 From( 4) *=f; t'^i 
 
 1.25 s X 2 
 
 a = -^=.39. 
 
STRENGTH OF BEAMS 399 
 
 2 2 000 
 
 From (5) a = .39 = - -y lt (6) 
 
 d / 
 
 2/2=10 2/1. (c) 
 
 Solving (a), (6) and (c), we obtain 
 /., = 22,000, ?/! = 6.45 inches, and // 2 = 3.55 inches; 
 
 whence r =-' = 11, 
 
 7 C 
 and from ((>), 
 
 \l t> = 8 X 2000 f - -f X J(6.45) a = 522,000 inch-pounds. 
 
 These three beams developed the following moments of re- 
 sistance: 553,550, 663,700 and 783,500, mean 667,000 inch-lbs., 
 or 28 per cent, greater than that derived by formula. None of 
 them failed, however, by crushing of the concrete at the top 
 of the beam, but by longitudinal shearing "at or a little above" 
 the reinforcing rods. 
 
 593. It appears, then, that by increasing the area of steel 
 reinforcement over 600 per cent., or from .44 sq. in. to 3.12 
 sq. MIS., the strength of the beams was increased about 68 per 
 cent, by theory, or 66 per cent, according to the few tests cited. 
 The cost of the beam, however, was increased about one hun- 
 dred per cent. 
 
 This method of increasing the moment of resistance of a 
 beam is not economical; it is better to improve the quality of 
 the concrete. It may, however, be necessary at times to use 
 excessive reinforcement on account of restrictions on the size 
 of beam, but one may easily carry this so far that he passes 
 outside the true theory of concrete-steel construction, and it 
 becomes a question of the steel being sufficient to carry the 
 entire load. In such cases double reinforcement may be adopted. 
 
 594. Tables of Strength. In Table 160, equations (5) 
 and (6) have been reduced to simpler forms by the introduc- 
 tion of values of E s and /.,. Selecting in the table the division 
 corresponding to the modulus of elasticity of the concrete 
 which is to be used, and the line opposite the assumed stress in 
 the concrete, M = quantity in column a times the square of the 
 depth of beam, d; and the area of steel in a beam of 12 inch 
 width, i.e. 12 a, equals quantity in column b times the depth 
 
400 
 
 CEMENT AND CONCRETE 
 
 1 
 
 CO 
 
 
 . 
 
 ^S* 3 
 
 N 2 
 
 
 
 O CO 1 'M t-H Tti 
 
 O i i CO O OS <M 
 
 Tl 
 
 
 o 
 
 a 
 
 OS 
 
 f 
 
 s 
 
 02- 
 
 Jill 
 
 .0 
 
 IllSli 
 
 
 o 
 N 
 
 5 
 
 
 02 f* 3 ^ 
 
 
 
 
 S I S. a 
 
 lull 
 
 
 < 
 
 I 
 
 > o3 
 
 : * 
 J| a 
 
 s 
 
 523222 
 
 0) 
 
 *S .S 8 
 
 
 
 PI 
 
 
 5 CO t- I-H 
 O (N ^ t- i i >O 
 
 c 
 
 i 34 s 
 
 
 1 
 
 
 
 T ( T 1 
 
 >j 05 
 
 * 1 J3 ? 
 
 fl ~CC QJ "* 03 
 
 "3 *8 I ~ J ^ 
 
 1 
 
 
 02 
 
 ^ 
 
 - 
 
 r^ i^ o 71 co O 
 
 8<N iC OS CO GO 
 _ 0_ O r-H ^-t 
 
 S o 
 
 oS ^ 
 
 ; lili 1 
 
 J ii '$ c ^ 
 
 SO -^ % -g JJ OH 
 
 
 H> 
 | 
 ( 
 | 
 
 ill 
 
 e 
 
 OS OS <M 1-1 i-i O 
 O (N GO l-' OS 
 <N -^ O:> <M 
 
 iS 
 
 ^ *^ 
 
 So" 
 
 CB O 
 
 H ,5 S >,.S ^ 
 
 STS J " S - ^ 
 a f "S TJ 8 
 n rt "o ^ J 
 
 
 
 iig 
 ii 
 
 PH Oc 
 
 
 
 rH GO O CO t- 00 
 
 i-H CO t~ i-H O rH 
 
 ' I-H i-i sq 
 
 gnCO 
 
 II 
 
 g . 
 
 ^ Iff 11 | 
 
 I s 1 ^1 - 
 
 i 
 
 02 
 
 2 ce~ o>' 
 
 -0 
 
 CO O O (M O <N 
 i i ^ OS -<t O O 
 O O O i i (M CM 
 
 o &3 
 
 
 IS 
 
 ai ^ x S ^3 
 'S & ~ - ^ 
 
 i 
 
 
 w; e 
 
 
 
 r 2 
 
 S e .S i, 
 Jel^ | | 
 
 a 'S "o S ^ "o P 
 
 
 : 
 i 
 j 
 
 3^g g 
 ^ ""CS 
 
 e 
 
 <N <M t- CM O O 
 
 00 GO Tti ?D O ~~O 
 <N O GO CM O 
 
 II 
 
 0) OH 
 
 3 S 
 
 J 3 
 
 .S 5 _n 8 
 Sffi *B 
 r-" rj t_ _. -t- 3 OU 
 
 o 73 5 .9 "S 
 
 ^ P* -ia - 'S \> J 
 
 s in I c i 
 
 
 
 Jll 
 
 
 
 un o >o 10 O' cc 
 
 T-H tO OS ^ O ifi> 
 ' rH <M' (N 
 
 K* 
 
 O bfi 
 
 - s 
 
 1 1 
 
 1 ' 3 T| 1 1 i 
 
 I "O $ OJ 
 
 3 s3 6 3 
 
 j 
 
 02 
 
 W! 
 
 iO 
 
 CO O rt< rh O OS 
 
 t 1 ?O r- 1 t-- Tf( O 
 
 z 
 
 op 
 
 1 
 
 
 H) 
 
 ! 
 i 
 i 
 
 : * s 
 
 |5 
 
 e 
 
 TH T-l II O O O 
 
 TJH CO 00 iO U5 O 
 
 TH ^ t^ i-H O OS 
 
 
 S 
 
 
 [MOD.ELAS. 
 CONCRETE 
 
 1 Working 
 
 Stress 
 
 s-S | . 
 
 _" d) p t fl 
 
 
 SHIII 
 
 
STRENGTH OF BEAMS 401 
 
 of beam, d. Column c gives the area of cross-section of steel 
 expressed as the per cent, of the area of section above the center 
 of steel reinforcement. 
 
 595. For example, suppose we wish to know the strength of 
 a beam ten inches deep (d = h i = 10 in.) and the amount 
 of steel required to develop a stress in the concrete of 400 Ibs. per 
 square inch when the stress in steel is 10,000 Ibs. per sq. in., 
 and the modulus of elasticity of the concrete is assumed at 
 3,000,000. In column a under 3,000,000 modulus, and opposite 
 400 Ibs. stress, we find 68.1 ; then the moment of resistance of a 
 beam one inch wide is 68.1 inch-lbs. X 10 X 10 = 6,810 inch- 
 Ibs., and the resistance of a beam 12 inches wide is 6,810 foot- 
 Ibs. The area of steel required in 12 inches width of beam is 
 .092 d or 0.92 sq. in. This beam is reinforced with .77 of one 
 per cent, steel. Similar tables may be prepared for other values 
 of .7, and /, if desired. 
 
 596. In Table 161 the equations have been completely 
 solved for certain typical values of E c and / assuming the 
 values for E s and /, of thirty million and ten thousand respec- 
 tively, as in Table 160. Having computed the bending mo- 
 ment, and fixed upon the probable safe working stress and 
 modulus of elasticity of the concrete which it is proposed to 
 use, it is only necessary to take from the table the required 
 depth of beam and the amount of steel reinforcement required. 
 
 For example, a girder 10 feet long supported at the ends 
 carries two loads of 5,000 pounds, each load being 2.5 feet from 
 a support. 
 
 If the width of girder is 15 inches, working stress of concrete 
 300 Ibs. per sq. in. and modulus of elasticity of concrete 1,500,000, 
 what is the required depth of girder and area of steel in tension 
 side? 
 
 The maximum bending moment (neglecting weight of beam) 
 is 12,500 ft.-lbs. throughout the central five feet. The required 
 
 12 
 
 moment of resistance for twelve inches in width is of 12,500 
 
 = 10,000 ft.-lbs. Looking in the table for this bending moment 
 under 300 Ibs. stress and 1,500,000 modulus, we find it is be- 
 tween d = 12 and d = 14, or at about d = 13 inches. If we 
 allow 2 inches below center of steel reinforcement, we have 
 total depth of beam, h = 13 + 2 = 15 inches. In the same 
 
402 
 
 CEMENT AND CONCRETE 
 
 rcemen 
 width) . 
 
 .3 
 
 02 
 
 g i 
 S B 
 
 o 
 
 * -8 *:? 
 
 <* 
 
 ^ O 02 
 
 2 S.S 
 
 85 
 
 8 
 
 O H 
 
 WORK 
 ONCR 
 
 ^Hi i^'N'NCOCOCO 
 
 O 10 to O >-0 O C 
 i-tW'*?Ot-ONCO 
 
 o o o o o p o o o o 
 
 r-l 73| CM' (N CO CO rji -*' t*' O 
 
 hjj r-i 
 
 4^ 
 
 
 ' < 
 COOO-fCN 
 
 OOOO 
 rp C^ O O 
 
 
 O' i^'OTjHi ICOOOOOOOOOOOOO 
 SOiOpl>'^OOCO / NOCCO^OOO!|.t"-OOO 
 
 OGO'O-^iOGOCOOCi'M^tiCOOOOi i i iGCCO 
 
 r-H (M CO Tt^ 5O GO OS "* O: O (7<l C7 t"^ CN O5 1^ GO 
 
 i lrH<MCO-^iOl-^OOO(N 
 
 SGO 
 i-( 
 
 
 i irJHOOiOTfCO 
 
 OOOOOOO 
 *^COO5T iCi^U7>i 
 
 
 - n 
 
 x-^ 
 
 1 g 
 
 03 -f-J 
 
 |^o 
 
 II I 
 l 
 
 Pco <u 
 
 Ci C<l CO Oi T-t CO >-O 
 
 t- O I-H O T* CO l^ CO ^ 7^1 
 
 GOCOC<l"*OTtit^OiOiGO 
 
 oc6e^oiHt^<NtHt-i 
 
 T-H i l(NCNCOTtiOO 
 
 5 
 
 0) ^3 
 
 IH 09 eo ic co 
 
 O^ 
 
 ^T- 
 rH i 1 T 1 !?^ CO CO Tt* 
 
STRENGTH OF BEAMS 403 
 
 lines we find area of steel for 12 inch width between 1.08 and 
 1.26, or, say, 1.17; then for 15 in. width the required area is 
 
 X 1.17 = 1.46 sq. in. The bars should not be more than 3 to 
 
 9 
 6 inches apart. We may use, then, 5 bars inch square or 
 
 -inch diameter, spaced three inches apart. In large beams it is 
 
 o 
 
 necessary to consider the bending moment occasioned by the 
 weight of the beam after making a first approximation to the 
 size required. 
 
 597. The above tables are prepared on the assumption that 
 the stress in concrete shall be equal to the value selected when 
 the stress in the steel reinforcement reaches 10,000 Ibs. per sq. 
 in. From the equations, other tables may be prepared if 
 desired, in which the working stress in steel shall be 12,500, 
 16,000 or any other assumed value. The tables are not suited 
 to the computation of beams in which excessive reinforcement 
 is used. 
 
 As to actual tests of the performance of concrete and steel in 
 combination, the possible variations in material are so diverse 
 and the cost of experiments so great that the results thus far 
 obtained appear somewhat fragmentary, but each investigator 
 has selected a small branch of the subject for experiment. 
 Among the more valuable tests in this line may be mentioned 
 the following : 
 
 Tests at Massachusetts Institute Technology, Prof. Gaetano 
 
 Lanza, Trans. Amer. Soc. C. E., vol. 50, p. 486. 
 Tests at Purdue University, Prof. W. K. Hatt, Jour. Western 
 
 Soc. Engrs., June, 1904. 
 Tests at Rose Polytechnic Institute, Prof. Malvard A. Howo, 
 
 Jour. Western Soc. Engrs., June, 1904. 
 Tests at University of Illinois, Prof. A. N. Talbot, Proc. Amer. 
 
 Soc. for Testing Materials, 1904. 
 
 Tests at University of Wisconsin, Prof. F. E. Turneaure, Proc. 
 Amer. Soc. for Testing Materials, 1904. 
 
 ART. 70. CONCRETE-STEEL BEAMS WITH DOUBLE REINFORCE- 
 MENT 
 
 598. We have seen that when the depth of. a beam is limited 
 by structural considerations we may increase the normal load 
 
404 
 
 CEMENT AND CONCRETE 
 
 by excessive reinforcement, but that this method results in low 
 stresses in the steel and is not usually economical. We may 
 now consider the effect of placing reinforcing rods in the com- 
 pression side of the beam as well as in the tension side. 
 
 FIG. 14. FIG. 15. FIG. 1C. 
 
 CROSS-SECTION CROSS-SECTION STRAIN DIAGRAM. 
 
 (Single Reinforcement.) (Double Reinforcement.) 
 
 Let Fig. 14 represent the cross-section of a beam reinforced 
 on the tension side with sufficient steel, area a, to develop the 
 proper working stresses in the materials, and let the position 
 of the neutral axis be N N. If at distance x from the neutral 
 axis we add an area of steel A' in the compression side, the 
 position of the neutral axis would be changed for similar load- 
 ing; but if at the same time we place in the tension side an ad- 
 
 A x 
 ditional area of steel A such that -p = > the position of the 
 
 neutral axis will be unchanged. Let // = stress in steel in 
 compression; then since the steel must suffer the same deforma- 
 tion as the surrounding concrete p = . Multiplying the last 
 
 / S " 
 
 two equations, we have, f 8 A = /', A', that is, we have added 
 equal forces to the two sides of the beam, and have increased 
 the moment of resistance by f, A (x + ?/ 2 ) inch-pounds. 
 
 599. To illustrate the application of this principle we may 
 take the beam considered in 591, in which z = 8, R = 20; 
 
 444 20 
 
 r = 40, az = .444, a = ^- = .055, /. = 80,000, y t = y inches, 
 
 y l= -- in., and M = 311,100 inch-pounds. 
 
 o 
 
 When the area of reinforcement in the tension side of this 
 beam was increased to az = 3.12 sq. in. or a = .39, the theo- 
 retical bending moment was increased to 522,000 inch-pounds 
 ( 592). What will be the result of a similar increase in steel 
 distributed between the two sides of the beam? 
 
DOUBLE REINFORCEMENT 405 
 
 Let k distance from top of beam to center of reinforce- 
 ment on compression side = 2 inches, 
 
 10 4 
 
 then x y^ 2" = ^ 2 = - inches, 
 
 o o 
 
 \. JL> ^i J. vJ A At 
 
 f = -=--*- = 0.4 or A = 0.4 A'. 
 A 2/2 & * 
 
 A + A' = .39 - .055 = .335 
 
 1.4 A' = .335 A' = .24 ,4 '2 = 1.92 
 
 A = .095 4s = .76 
 
 9 
 
 whence a = .055 az = .44 
 
 A + a = .150 Total steel, 3.12 sq. inches. 
 
 // = ^ = .4 /. = 32,000. 
 
 2/2 
 
 Added moment of resistance equals 
 
 Az (x + y 2 ) /. = .095 x 8 X ^x 80,000 = 486,400 in.-lbs. 
 
 o 
 
 And total moment of resistance eq uals 
 
 311,100 + 486,400 = 797,500 inch-pounds. 
 
 None of the bars in the series mentioned in 591 had as large 
 an area of reinforcement as 1.92 sq. in. on the compression side. 
 
 It is noticed, first, that the double reinforcement gives bet- 
 ter results than such excessive reinforcement on the tension 
 side; second, that the stress in steel on the compression side is 
 less per square inch than that in tension; and third, that in 
 case a large addition of steel is made, this results in a greater 
 area of steel in compression than the total area of steel in ten- 
 sion. In practice the area of steel in compression is usually 
 made equal to, or less than, the area in tension, but beams with 
 double reinforcement are seldom accurately designed. 
 
 ART. 71. SHEAR IN CONCRETE-STEEL BEAMS 
 
 600. There are several methods of failure of concrete-steel 
 beams other than those considered above, direct tension in the 
 steel or direct compression in the concrete due to the bending mo- 
 ment. These other methods of failure are popularly called failures 
 in shear, although some of them cannot properly be so classed. 
 
 601. We have seen that the shearing stress of concrete is 
 usually considered to be somewhat in excess of the tensile 
 
406 CEMENT AND CONCRETE 
 
 strength (458) and that the latter is one-fifth to one-tenth the 
 compressive strength. With a beam having only a normal 
 amount of reinforcement, then, there is little danger to be 
 feared from simple vertical shear, and as a matter of fact, tests 
 have not developed instances of such weakness. In compara- 
 tively short spans, however, failures have occurred near the 
 quarter points, in cracks starting at the under side of the beam 
 and extending upward in a direction inclined toward the center. 
 This method of failure has the appearance of being due to a 
 combination of shear and tension in the lower section of the 
 beam, since the cracks are approximately at right angles to the 
 theoretical " lines of direct tension." Such failures, however, 
 are almost always accompanied by a slipping of the steel bar in 
 the concrete, and may frequently be prevented by taking 
 proper precautions against such slipping. 
 
 602. A more frequent cause of failure is a longitudinal shear 
 in the plane near the steel reinforcement and on that side of it 
 lying nearer the concave side of the beam. 
 
 It is evident that a failure caused by slipping of the bar in 
 the beam, although caused primarily by shearing forces, is 
 really a failure in adhesion, yet the two forms of weakness are 
 so closely connected that it is simpler to consider them together. 
 
 603. Comparison with Plate Girder. In a steel plate girder 
 the lower flange is considered to carry the tension, the upper 
 flange the compression; the web connects the two flanges, caus- 
 ing them to act together as one beam, and we may think of 
 the web as preventing the ends of the compression flange sliding 
 beyond the ends of the tension flange. When the web is not 
 able to accomplish this without buckling, it is stiffened by 
 vertical angles. 
 
 In a concrete steel beam we have considered the entire 
 tension to be carried by the steel reinforcement, and the entire 
 compression to be carried by the concrete on the other side of 
 the neutral axis. The connecting web is also concrete. This 
 web is thick and not liable to buckle, but it may shear in a lon- 
 gitudinal plane as a wooden beam may do when short and deep. 
 All of the tension in the steel reinforcement must be trans- 
 mitted through the surrounding concrete. If there are no pro- 
 jections on the steel bar, the adhesion of the concrete to it 
 may, under certain circumstances, be not strong enough to 
 
SHEAR IN BEAMS 407 
 
 safely carry this stress; and if the adhesion is sufficient, then 
 the shearing strength of the concrete may be too low to transmit 
 the stress to contiguous fibers or layers. 
 
 604. Illustration. Let us consider a concrete-steel beam 
 twelve inches wide, twelve inches deep and of ten foot span, 
 supported at the ends; reinforcement, one square inch of metal 
 properly distributed in a plane two inches above the bottom 
 of the beam. Let us suppose this beam carries a uniform load 
 of 600 pounds per foot, giving a maximum bending moment of 
 90,000 inch-lbs., and a stress in steel of 10,000 pounds at the 
 center. The ends of the steel bars are of course without stress. 
 Since the bending moment at any section of such a beam is 
 proportional to the product of the segments into which the 
 section divides the span, the bending moment one foot from 
 the ends will be 
 
 ^ X 90,000 = 32,400 inch-pounds, 
 o X o 
 
 Let us consider the neutral axis in the same position at the 
 end of the beams as near the center. (This is not strictly true, 
 because of the lighter stress near the ends of the beam, but 
 the error made by such an assumption will be unimportant for 
 our present purpose.) Then the tension in the steel will have 
 the same proportion, or, tension in steel one foot from the end 
 = 2\ X 10,000 = 3,600 pounds. 
 
 The stress in steel, then, which is zero at the end, has in- 
 creased to 3,600 Ibs. in one foot of length. To provide against 
 poor contact near the end, consider two-thirds of this length, 
 or eight inches, to be operative. If the reinforcement consists 
 of four one-half-inch square bars, the necessary adhesion per 
 
 S600 
 
 square inch is - = 57 Ibs. per sq. in. ; but if only one bar 
 o X o 
 
 is used one inch square, the required adhesion is 114 Ibs. per 
 sq. in. The latter would not be good practice, not only be- 
 cause of high adhesion required, but because the steel is not 
 properly distributed. 
 
 Where the stress in adhesion is greater than can be safely 
 relied upon for plain rods, it is necessary to use some kind of 
 deformed bar, or to anchor the bar securely at the end. This 
 may be done by passing the end of the tension bar around a 
 
408 CEMENT AND CONCRETE 
 
 rod transverse to the beam near the end. Care should be taken 
 that the safe value of adhesion is not assumed too high. 
 
 605. Value of Shear. The same total stress of 3,600 Ibs. 
 must be transferred through the concrete immediately above 
 the bar. If the reinforcement is so distributed that the entire 
 width of the beam has practically the same stress, and we 
 consider, as before, that two-thirds of the length of the end 
 
 3 600 
 
 foot is operative, we have mean shear = ~ = 37.5 Ibs. 
 
 LZ X o 
 
 per sq. in. The value of stress in shear should not exceed one- 
 tenth the safe value in compression, and there is a general ten- 
 dency to use not more than one-twentieth. 
 
 If the same form of beam had a span of but five feet with 
 same bending moment, the value of the shearing stress by this 
 method becomes 75 Ibs. per sq. in., and it will be necessary to 
 provide against this stress coming upon the concrete. 
 
 Another approximate method is the ordinary one for rect- 
 angular beams, viz. to consider the shear in horizontal plane 
 just above the steel reinforcement to be f times the total shear 
 at any section, divided by the area of vertical section of the 
 beam. 
 
 606. Provision is sometimes made for relieving the concrete 
 of all shearing stresses. In this case the beam is divided into 
 imaginary panels of length equal, say, to the depth of the 
 beam, and the diagram of maximum shear is drawn. The 
 shear in each imaginary panel is then provided for by a vertical 
 or inclined bar of the proper dimensions. Or, what is usually 
 better, the shear bars are all of one size and the proper num- 
 ber of them are distributed throughout each panel length; the 
 spacing of the shear bars thus becomes wider near the center 
 of the beam. 
 
 607. Resistance to Shear. When provision against shear 
 is made by using small steel rods placed either vertical or in- 
 clined downward toward the center of the beam, as mentioned 
 above, these rods may well be made in the form of inverted 
 U-shaped stirrups, with their ends securely fastened to the 
 reinforcing metal in tension. 
 
 In many cases all the provision necessary is given by the use 
 of two longitudinal bars, parallel and close together near the cen- 
 ter of the span, but one of them leading to a plane near the top of 
 
SHEAR IN BEAMS 409 
 
 the beam at the supports. This system is very conveniently 
 applied in concrete slabs supported by I-beams, one bar of the 
 pair being hooked over the upper flanges of the I-beam and 
 sagging toward the center. The Hennebique system (571) is 
 a combination of the inclined bar and U-shaped stirrups. 
 
 608. A modification of the single inclined bar is the Cum- 
 mings system, wherein there are several pairs of bars of vary- 
 ing lengths; these are all horizontal and near the bottom along 
 the center of the beam; a short distance from the center the 
 shortest pair turns up at an angle of about forty-five degrees; a 
 littVe farther toward the end a second pair of bars is turned up, 
 and so on, leaving a single pair to go through straight to the 
 support. 
 
 Another, arid more radical modification, is the Kahn system 
 (573), in which the bar is square with wings of metal on oppo- 
 site corners which are sheared and bent up at angles of forty- 
 five degrees, so that the outline of the steel work in a beam 
 resembles the tension members of a Pratt truss. 
 
CHAPTER XX 
 
 SPECIAL USES OF CONCRETE: BUILDINGS, WALKS, 
 FLOORS AND PAVEMENTS 
 
 ART. 72. BUILDINGS 
 
 609. While the use of concrete and steel for the walls and 
 floors of buildings is about fifty years old, yet it is only in com- 
 paratively recent years that its value has become generally 
 known. It is now applied to all classes of structures, ware- 
 houses, factories, residences, station and office buildings, and 
 it is anticipated that in the next twenty years concrete-steel 
 will be as familiar in architecture as steel skeleton, stone, and 
 brick are now. 
 
 610. It happens that at present the concrete-steel building 
 industry is largely in the hands of companies who are exploiting 
 some particular form of steel rods or bars applied according to 
 some one of the many " systems" of reinforcement. This con- 
 dition has both good and bad features. A reputable concern of 
 this kind will have in their employ engineers who should sat- 
 isfy themselves that each design is a safe one, for the failure 
 of a building will cast disrepute on their particular system. It 
 is this fact that leads the companies to keep the construction 
 entirely, and the design largely, in their own hands. Another 
 advantage is that these concerns are able to perfect methods of 
 construction by experience, and to lessen the expense of one 
 structure by making use of the concrete plant and the molds 
 that have been used on another. 
 
 611. In making plans for a building, the owner is usually 
 represented in the first instance by an architect whose business 
 it is to dictate the design. If concrete-steel is considered, the 
 architect may call an engineer in consultation and they may 
 together harmonize the features of utility and appearance with 
 economy and strength, but in letting the contract it is found 
 that the competition is limited to one or two companies using 
 the particular system which the engineer considers the best 
 adapted to the particular conditions in question. 
 
 410 
 
BUILDINGS 411 
 
 On the other hand, the architect will hesitate to go to the con- 
 struction company for assistance, since he must first select the 
 system he shall use, a question upon which his ideas may be neither 
 clear nor well grounded, and he is then having the prospective 
 contractor assist in the design. Under these circumstances the 
 architect will usually consider concrete-steel construction as 
 something he wishes to avoid if possible. But this condition 
 will correct itself in time, for owners will demand a considera- 
 tion of this form of construction, engineers will become fa- 
 miliar with its use and will be employed to design the engineer- 
 ing features, while reliable contractors in every city will obtain 
 permission to build in accordance with any " system" under 
 the supervision of a competent engineer. 
 
 612. Roof. While a pitch roof is sometimes built of con- 
 crete-steel, this form of construction is particularly adapted to 
 so called flat roofs. The roof is constructed much the same as 
 a floor slab (Art. 65-67), except that expansion joints are some- 
 times provided, and the roof is covered with tar and gravel, 
 or some of the patent roofings ordinarily used. While the 
 roof loads are usually light, permitting a greater span of slab 
 between beams than for floor construction, it will seldom be 
 economical to introduce these longer spans because of the 
 changes necessary in the molds. In most buildings it is neces- 
 sary to provide against condensation, and for this purpose a 
 flat ceiling may be suspended at the level of the under side of 
 the beams giving an air space. 
 
 613. Floor System. The floors may be constructed in con- 
 formity with the principles stated in Chapter XIX. The 
 strength of short span arches, such as are used for floors, 
 where the haunches are built up level with the top of crown 
 of arch, is a matter of experiment and cannot be accurately 
 determined theoretically. Empirical formulas may be derived 
 for a certain system based on a sufficient number of tests. 
 The principles underlying the strength of slabs may be con- 
 sidered the same as those applying to beams (Art. 69), although 
 if the length of slabs is not much greater than the span, they 
 are not strictly applicable, but will err on the safe side. 
 
 614. A decision must first be made as to the size of bays into 
 which the floor space is to be divided. This will of course de- 
 pend on the use of the building, the engineering features con- 
 
412 CEMENT AND CONCRETE 
 
 forming to requirements of utility. If the bays are not square, 
 the girders should usually take the shorter span between columns. 
 This length is then divided into the number of slab spans that 
 will give maximum economy. The shorter these spans the less 
 the amount of material required in slabs and the greater the 
 number and cost of floor beams. Computations should . be 
 made, therefore, for two or three arrangements to determine 
 this point. As this distribution for maximum economy will 
 vary with the loads to be provided for, it is well, if the floors 
 are not all to carry the same load, to take for this computation a 
 load intermediate between the heaviest and lightest, and use if 
 possible the same arrangement of spans throughout the building. 
 The strength of slabs for given bending moments may be 
 taken directly from Table 161, after deciding upon the working 
 stress to be allowed in the concrete and the probable modulus 
 of elasticity. The beams and girders, if single reinforcement 
 is used, are taken from the same table or computed by the 
 methods of Art. 69. 
 
 615. In some instances it may be found economical to use 
 concrete-steel slabs for floors supported by concrete protected 
 steel beams and girders. One advantage of this system is that 
 the forms for building the protecting concrete and for the 
 floor slabs may be hung from the steel girders and beams. For 
 this method of construction the enveloping concrete should 
 not be less than one and one-half inches thick over the edges of 
 flanges, and wire fabric or metal lath wrapped about lower 
 flanges of beams will insure the concrete remaining in place. 
 This is not properly concrete-steel construction, but simply 
 concrete protected steel, and except in case the concrete ex- 
 tends well above the steel, forming an independent compres- 
 sion flange, no added strength should be computed for the con- 
 crete covering. 
 
 616. Columns. In the foundations of buildings of moder- 
 ate height the supporting columns may be built entirely of 
 concrete. Since, however, the pressure on the concrete, even 
 when it is constructed with the greatest care, should not ex- 
 ceed two hundred to three hundred pounds per square inch, 
 the required area of cross-section in the lower stories is usually 
 so great as to preclude the use of columns built entirely of 
 concrete. 
 
BUILDINGS 413 
 
 617. Concrete Filling and Covering. A steel column of 
 any of the ordinary styles, built up of steel shapes may be 
 used, and protected from corrosion and fire by filling and cover- 
 ing with concrete. This not only serves. as a protection against 
 rust, but materially increases the stiffness and permits the use 
 of a somewhat higher working stress in the steel. The concrete 
 filling should be mixed quite wet in order that it shall work 
 into all angles. The edges of the metal should not approach 
 nearer than one and one-half inches to the exterior of the con- 
 crete, and flat surfaces of metal should have a covering of at 
 least two and one-half inches. Where it is necessary to cover 
 large, flat surfaces, they should be first covered with expanded 
 metal or wire fabric, locked on by twisting around the edges 
 of the plate or channel. 
 
 618. COLUMNS OF CONCRETE STEEL. Concrete-steel col - 
 umns differ from the above in that the main dependence is 
 placed on the concrete rather than on the steel. For such 
 columns longitudinal reinforcement has generally been employed. 
 Steel bars extending from end to end of the column are dis- 
 tributed throughout the cross-section, and are tied together at 
 intervals of four to twelve inches by smaller bars forming loops 
 to hold them in place. The splicing of the bars is effected by 
 placing a small tube over the upper end of the lower bar and 
 projecting above it, and then setting the lower end of the upper 
 bar within the tube resting on the lower bar. Where this is 
 done it is essential that the two ends be planed perfectly square, 
 and it is much better to avoid splices in a column between 
 lateral supports. In a building the reinforcing rods project up 
 through the floor above and are spliced into the bars of the 
 columns in the next story. 
 
 619. Strength of Columns. When a column reinforced with 
 longitudinal bars is subjected to pressure, the concrete and 
 steel must shorten together. The relative stresses in the two 
 materials will then be proportional to their moduli of elasticity. 
 From this follows the formula, 
 
 P = / c (C 
 
 where P = total pressure on column, 
 
 f c = stress in concrete, 
 C and S = areas of concrete and steel respectively, 
 
414 CEMENT AND CONCRETE 
 
 W 
 
 and R = ^, or ratio of the modulus of elasticity 
 
 &C 
 
 of the steel to that of the concrete. 
 
 In a series of tests of twenty-one columns made by Prof. 
 Gaetano Lanza, 1 but three failed under a lower stress than that 
 computed by the above formula. The columns were eight to 
 ten inches square, six to seventeen feet long and reinforced 
 with either one or four bars, the latter being from f inch to 1J 
 inches square. 
 
 The lowest breaking load was fifty tons on an 8 by 8 inch 
 column with one bar one inch square, and the strongest column, 
 10 by 10 inches,- with four J inch longitudinal bars, was not 
 crushed with a load of one hundred fifty tons, the limit of the 
 testing machine. The lowest result was twenty per cent, less 
 than that given by the formula, and the greatest excess strength 
 over the theoretical was fifty per cent. 
 
 620. While longitudinal reinforcement undoubtedly strength- 
 ens a long column against flexure, as well as adds to the resist- 
 ance to crushing, yet the added strength is gained at the ex- 
 pense of considerable additional cost. Suppose we have a ten 
 inch square column, twelve feet long, made of concrete with a 
 breaking load of 1,800 Ibs. per square inch, or 180,000 Ibs. 
 total breaking load. Suppose eight } inch square bars to be 
 built into this column as longitudinal reinforcement, and that 
 the modulus of elasticity of the steel is ten times that of the 
 concrete. Then the strength of the reinforced column would 
 be, by the formula above, 
 
 P = 1,800 (95.5 + (10 X 4.5)) = 252,900. 
 
 The longitudinal reinforcement has thus resulted in an increase 
 of strength of 40 per cent., while by the addition of 180 
 pounds of metal, the cost of the column has risen from about 
 $3.00 to say $8.50, an increase of about 180 per cent, without 
 counting the cost of lateral ties, and the additional trouble in 
 building a reinforced column. 
 
 621. Hooped Concrete. In extended experiments on what 
 he has called " hooped concrete," M. Considere 2 has shown that 
 
 1 Trans. A. S. C. E., Vol. 1, p. 487. 
 
 2 Comptes Rendus de I' Academic des Sciences, 1898-1902. Translation, 
 "Reinforced Concrete," by Armand Considere, translated by Leon S. Mois- 
 seiff, McGraw Publishing Co., New York, 
 
BUILDINGS 415 
 
 reinforcement is much more important and beneficial in a 
 transverse or circumferential direction than if longitudinal. 
 This may be accounted for by the fact that the natural method 
 of failure of concrete prisms, is by splitting along planes parallel 
 to the direction of pressure, and the ordinary method of failure 
 by shear along inclined surfaces is induced by the friction of 
 the plates transmitting the pressure to the prism. It was also 
 shown that while concrete reinforced by longitudinal bars with 
 the ordinary amount of lateral ties breaks suddenly, hooped 
 concrete fails gradually under a much heavier load. 
 
 622. M. Considere concluded from his experiments that the 
 circumferential ties should not be farther apart than one- 
 seventh to one-tenth the diameter of the column, even when 
 longitudinals were used to assist in completing the network, 
 and that the results were more successful the nearer together 
 the hoops or ties were placed. He found that spirals were 
 better than individual single ties and that longitudinals were of 
 value chiefly in assisting to confine the concrete, transmitting 
 the bursting pressure at a given plane to the contiguous spirals 
 above and below. 
 
 623. M. Considere says 1 that the " Compressive resistance of 
 a hooped member exceeds the sum of the following three ele- 
 ments : 
 
 "1. Compressive resistance of the concrete without rein- 
 forcing. 
 
 "2. Compressive resistance of the longitudinal rods stressed 
 to their elastic limit. 
 
 "3. Compressive resistance which could have been produced 
 by imaginary longitudinals at the elastic limit of the hooping 
 metal, the volume of the imaginary longitudinals being taken 
 as 2.4 times that of the hooping." 
 
 To subject hooped concrete to a practical test, M. Considere 
 constructed, in 1903, a truss bridge of sixty-five foot span with 
 parabolic top chord of seven and one-half feet rise, 2 the com- 
 pression members being of hooped concrete, and the tension 
 members of concrete-steel with longitudinal reinforcement, or 
 concrete protected steel. A central panel of the truss was con- 
 
 " Reinforced Concrete," p. 159. 
 2 Engineering News, May 5, 1904. 
 
416 CEMENT AND CONCRETE 
 
 structed with a reduced section of top chord about eight inches 
 diameter reinforced by eight longitudinal bars .43 inch in 
 diameter and a helix 6J inches in diameter of .43 inch metal 
 coiled to a pitch of about one inch. This reduced top chord 
 section showed signs of failure when the computed stress reached 
 about 5,000 pounds per square inch. 
 
 624. FORMS FOR BUILDINGS. One of the most serious prob- 
 lems in the construction of concrete-steel buildings is the de- 
 signing of the forms. They must be as light as is consistent 
 with strength to facilitate handling. They should be of simple 
 construction so that they may be set up and removed without 
 too much supervision, and they should be so assembled with 
 bolts and screws that they may be used repeatedly. In erect- 
 ing a large building sufficient forms are usually provided to set 
 up one floor complete, including columns, beams, girders and 
 floor slabs. After placing the reinforcement, the concrete is 
 filled in as rapidly as possible, making the slabs, girders and 
 columns practically monolithic. 
 
 The forms for the girders usually rest upon the column 
 molds and are supported at intermediate points by posts rest- 
 ing on the completed floor below. While column molds are 
 sometimes filled from the top, better work is assured by having 
 one side of the mold built up as the concrete is filled in from 
 the side. 
 
 The mold to receive the concrete forming the floor slab is 
 either a part of, or is supported by, the pieces forming the 
 sides of the girders and beams. Provision is sometimes made 
 for leaving supports at intervals under the completed beams 
 and girders after removing the forms from the sides of the beams 
 and the bottom of floor slabs. This is done by making the 
 bottom piece of the girder mold separate, and attaching the 
 side pieces to it by screws which may be removed without dis- 
 turbing the bottom. The caps of the supporting posts are then 
 made long enough to permit the lower edges of the side pieces 
 to rest directly on them. This method was adopted in build- 
 ing the Central Felt and Paper Company's factory at Long 
 Island City. 1 
 
 1 Wight-Easton-Townsend Company, Contractors, Engineering Record, 
 Jan. 16, 1904. 
 
BUILDINGS 
 
 625. In the same building the walls were built with molds 
 three feet high and sixteen feet long, placed in pairs on oppo- 
 site sides of the wall. When one section was completed, the 
 molds were "lifted until the lower edges were two inches below 
 the top of the concrete. In the new position they were sup- 
 ported by horizontal bolts through their lower edges, across the 
 top of the concrete; the upper edges were tied together by 
 transverse wooden strips nailed to them about three feet apart, 
 and they were braced to the false work supporting the roof 
 and column molds." "The bolts passed through sleeves which 
 were left permanently embedded in the walls. At first, iron 
 pipes were used for this purpose, but afterwards it was dis- 
 covered that pasteboard tubes were equally efficient and much 
 easier to trim and point after the molds were removed." 
 
 626. An excellent system of molds was used in the con" 
 struction of the Kelley and Jones Company's factory at Greens- 
 burg, Pa. 1 The floor molds were especially convenient, being 
 made collapsible by a hinge joint at the top along the longi- 
 tudinal center line. These floor molds were in reality cores 
 between adjacent floor beams; when in place the top surface 
 was horizontal, to form the under side of the floor slab, and 
 the vertical side pieces formed the sides of the floor beams. 
 When the concrete had set sufficiently, the lower edges of the 
 form were made to approach each other, thus coming away 
 from the concrete gradually. A special light wooden framework 
 or tower, with a working platform six feet below the floor, and 
 a rope sling to receive and lower the floor mold, permitted of 
 removing the molds rapidly and without injury. A special truck 
 was also used for moving the floor molds about the building. 
 
 627. A convenient adjunct for the construction of concrete 
 wall forms consists of a short section of I-beam having a width 
 between flanges equal to the thickness of the plank to be used. 
 These plank holders are laid in pairs, with web horizontal, one 
 on either side of the wall, and connected by a bolt passing 
 through them and through the wall. 2 Two rows of planks on 
 edge are first placed around the building so as to inclose the pro- 
 
 1 Mr. E. L. Ransome, Architect and Engineer, Engineering Record, Feb. 
 6 and 13, 1904. 
 
 1 Patented by Thomas G. Farrell, Washington, N. J. 
 
418 CEMENT AND CONCRETE 
 
 posed wall. At the upper side of each junction between two 
 planks in the same horizontal row is placed one of these plank 
 holders. Another horizontal row of planks may now be placed, 
 with the iron plank holders at the joints as before. As the 
 wall is built up, the lower planks and holders may be removed 
 and placed on top, and thus few forms are required. Tees and 
 L-forms are provided for partition walls and corners. 
 
 When an air space is desired in a wall a special terra cotta 
 tile or building block may be built into the wall, but this is 
 quite expensive, and an interior collapsible form may be made 
 of timber by the use of two planks held apart by a wooden 
 brace which may be knocked out. Special means of handling 
 the interior plank should be provided, and the building of a 
 high wall cannot be continuous with this method. 
 
 628. New York Building Regulations. While city building 
 regulations are not always criteria of good practice, yet the 
 Regulations of the Bureau of Buildings of the Borough of 
 Manhattan concerning the use of concrete-steel construction are 
 exceptional. Emanating from a bureau that has been dis- 
 tinctly hostile to concrete-steel, they are naturally conservative, 
 but are, on the whole, excellent, and work conscientiously done 
 in accordance with them will not bring discredit on concrete 
 construction. 
 
 It is specified that the cement shall be only high grade 
 Portland standing certain tests, that the sand shall be clean 
 and sharp, aggregate, broken trap, or gravel of a size that will 
 pass a three-quarter inch ring, and that the proportions used 
 shall be one cement, two sand and four of stone or gravel, or 
 that the concrete shall have a crushing strength of two thou- 
 sand pounds per square inch in twenty-eight days. Only the 
 best quality of concrete is thus permitted. 
 
 629. The Regulations concerning the design are then stated 
 as follows : 
 
 " Concrete-steel shall be so designed that the stresses in the 
 concrete and the steel shall not exceed the following limits: 
 
 Extreme fiber stress on concrete in compression, 500 Ibs. per sq. in. 
 
 Shearing stress in concrete 50 " " 
 
 Concrete in direct compression 350 " " 
 
 Tensile stress in steel 16,000 " " 
 
 Shearing stress in steel 10,000 " " 
 
BUILDINGS 419 
 
 "The adhesion of concrete to steel shall be assumed to be 
 not greater than the shearing strength of the concrete. 
 
 "The ratio of the moduli of elasticity of concrete and steel 
 shall be taken as one to twelve. 
 
 "The following assumption shall guide in the determination 
 of the bending-moments due to the external forces: Beams and 
 girders shall be considered as simply supported at the ends, no 
 allowance being made for the continuous construction over 
 supports. Floor plates when constructed continuous and when 
 provided with reinforcement at top of plate over the supports, 
 may be treated as continuous beams, the bending-moment for 
 
 W L 
 
 uniformly distributed loads being taken at not less than - ; 
 
 W L 
 
 the bending-moment may be taken as - - in the case of square 
 
 floor plates which are reinforced in both directions and sup- 
 ported on all sides. The floor plate to the extent of not more 
 than ten times the width of any beam or girder may be taken 
 as part of that beam or girder in computing its moment of 
 resistance. 
 
 "The moment of resistance of any concrete-steel construc- 
 tion under transverse loads shall be determined by formulas 
 based on the following assumptions: - 
 
 "(a) The bond between the concrete and steel is sufficient 
 to make the two materials act together as a homogeneous solid. 
 
 "(6) The strain in any fiber is directly proportionate to the 
 distance of that fiber from the neutral axis. 
 
 "(c) The modulus of elasticity of the concrete remains con- 
 stant within the limits of the working stresses fixed in these 
 Regulations. 
 
 "From these assumptions it follows that the stress in any 
 fiber is directly proportionate to the distance of that fiber from 
 the neutral axis. 
 
 "The tensile strength of the concrete shall not be considered. 
 
 "When the shearing stresses developed in any part of a 
 construction exceed the safe working strength of concrete, as 
 fixed in these Regulations, a sufficient amount of steel shall be 
 introduced in such a position that the deficiency in the resist- 
 ance to shear is overcome. 
 
 "When the safe limit of adhesion between the concrete and 
 
420 CEMENT AND CONCRETE 
 
 steel is exceeded, some provision must be made for transmitting 
 the strength of the steel to the concrete. 
 
 " Concrete-steel may be used for columns in which the ratio 
 of length to least side or diameter does not exceed twelve. 
 The reinforcing rods must be tied together at intervals of not 
 more than the least side or diameter of the column. 
 
 "The contractor must be prepared to make load tests on 
 any portion of a concrete-steel construction, within a reasonable 
 time after erection, as often as may be required by the Super- 
 intendent of Buildings. The tests must show that the con- 
 struction will sustain a load of three times that for which it is 
 designed, without any sign of failure." 
 
 ART. 73. CONCRETE WALKS 
 
 630. One of the most important uses of concrete is in the 
 construction of street and park walks. It has not only driven 
 stone nagging almost out of use, but it is being employed to a 
 large extent in towns and villages where board walks have 
 formerly been used almost exclusively. 
 
 A concrete walk is made up of a sub-base or foundation, a 
 base, and a wearing surface. 
 
 631. Foundation. As in other structures, one of the most 
 important essentials for success lies in the preparation of the 
 foundation, and the care that must be bestowed on it will 
 depend upon the character of the soil and the climate. In the 
 higher latitudes of the United States, frost may soon destroy a 
 walk the foundation of which is not well drained. 
 
 The excavation should be made to the sub-grade previously 
 determined upon, any objectionable material such as loam or 
 organic matter being removed, and the bottom of the excava- 
 tion smoothed and well rammed. Upon this is laid the sub- 
 base, its thickness varying from nothing to twelve inches. In 
 a sandy soil with good natural drainage and little danger from 
 frost, and where light traffic is expected, it may be unnecessary 
 to provide any special sub-base, since the soil itself furnishes a 
 good foundation for the concrete, but in clay soil in northern 
 climates, twelve inches of sub-base may be required. The best 
 material for this sub-base is broken stone varying in size from 
 one-half inch to two and one-half inches. Usually broken 
 stone is considered too expensive, and gravel, coarse sand, 
 
CONCRETE WALKS 421 
 
 cinders, or broken brick is employed. A layer four inches thick is 
 usually sufficient for good materials, but six to twelve inches of 
 cinders are sometimes required. It should be well rammed to 
 a level surface, and when completed should be firm but porous. 
 The most important point is that this course shall have 
 good drainage, otherwise it may be a menace to the walk. If 
 it is more porous than the retaining soil, it will naturally drain 
 this soil, and if the water is not able to escape into the sewer 
 or elsewhere, it may be frozen and heave the walk. An ex- 
 cellent plan sometimes adopted is to lay at intervals of twenty 
 to twenty-five feet, a blind stone drain from the walk founda- 
 tion to the foundation of the curb. In exceptional cases it may 
 be necessary to lay a tile drain in the sub-base to lead the water 
 away from the walk. 
 
 632. Base. The base is the body of the walk giving stiff- 
 ness to the structure. Its functions are to furnish a solid 
 foundation for the wearing surface and to give transverse 
 strength to the walk, transmitting the pressure uniformly to 
 the sub-base. The base is of concrete, which need not be very 
 rich for ordinary traffic. A proportion of one part packed 
 Portland cement to two and one-half volumes of dry sand 
 and six volumes broken stone is excellent, and proportions of 
 one, three and seven parts cement, sand and stone, respectively, 
 will usually be found sufficient, though the richer the concrete in 
 the base the better will the top dressing adhere to it. 
 
 The broken stone for this concrete should be of a size not 
 exceeding one and one-half inches in any dimension, some cities 
 requiring three-quarters inch or less. Crushed granite and trap 
 are excellent, though limestone or any other moderately hard 
 rock may be used that is suited to making concrete for ordinary 
 purposes. If of a hard rock, the screenings may well be left 
 in the broken stone, and when this is done, the dose of sand 
 should be diminished. (See Art. 37.) 
 
 The thickness of the layer of concrete should not be less 
 than three inches. Four inches is much better and is recom- 
 mended for general use in sidewalks, while in exceptional cases 
 six inches is required. The top of the concrete base should 
 be finished to a plane parallel to the proposed surface of the 
 walk and at a distance below it equal to the proposed thickness 
 of the top dressing. 
 
422 CEMENT AND CONCRETE 
 
 633. Wearing Surface. The preparation and application 
 of the wearing surface require much care if satisfactory results 
 are to be obtained. The most evident service of this layer is 
 to withstand wear, and it should therefore be made of rich 
 Portland cement mortar. With a sand consisting principally 
 of quartz particles, it is found that a mortar composed of equal 
 parts cement and sand gives about the best results in tests of 
 abrasion. If the mortar is used richer than this, it is likely to 
 check or crackle in setting, marring the appearance of the walk. 
 Mortar containing two parts cement to three parts sand gives 
 nearly as good results, and two parts sand or fine crushed 
 granite to one of Portland cement is usually satisfactory. The 
 sand for the mortar should be quartz if possible, or crushed 
 granite or trap. It should be screened through a quarter 
 inch mesh, and there should not be a large proportion of 
 very fine particles. 
 
 The thickness of the layer of top dressing is usually about 
 one inch, and this is probably the maximum thickness ever 
 required. One-half inch of top dressing is believed to be suf- 
 ficient when the wear is not excessive, provided the base has 
 been carefully leveled. 
 
 634. The Construction of the Walk. If the walk has not a 
 considerable longitudinal slope, it should be given a transverse 
 slope of about a quarter inch to the foot to provide for draining 
 the surface. 
 
 Stakes for grade and line having been given, a maitre cord 
 is stretched along the line stakes to mark the sides of the exca- 
 vation. After the material has been excavated to the proper 
 sub-grade and all soft material in the bottom removed, the 
 bottom of the trench is well rammed. If tile drain is necessary, 
 it is laid with open joints on this foundation. The material to 
 form the sub-base is now wheeled in and rammed to the proper 
 thickness, water being used freely if it facilitates the packing. 
 The top of the sub-base is brought to a level plane at the proper 
 distance below the grade stakes. 
 
 The molds for the walk are now to be laid. These are made 
 of two by four or two by six inch scantling, sized and dressed 
 on at least one side and one edge. Stakes are first securely 
 driven, about five or six feet apart, with their faces two inches 
 back from the side lines of the proposed walk, and their tops 
 
f ' 
 
 CONCRETE WALKS 423 
 
 at grade. Against these stakes the scantlings are placed on 
 edge with dressed side toward the walk, and smooth edge level 
 with the grade stakes. These molds are held in place by nail- 
 ing through the supporting stakes into the scantling, and if 
 these nails are not driven "home," they* may easily be pulled 
 to release the mold when the work is completed. On the upper 
 edges of the mold are then marked off the sizes of blocks de- 
 sired, being careful that the marks defining a joint are exactly 
 opposite each other on the two scantlings. 
 
 635. The concrete materials having been previously deliv- 
 ered near the work, the concrete is mixed, either by hand or 
 machine, according to the methods already given, and rammed 
 in place after the sub-base has been well wet down to receive 
 the concrete. The concrete should be just short of quaking, 
 and in ramming care must be taken not to disturb the molds. 
 For tamping next the molds, the makers of cement working 
 tools offer a light rammer with square face at one end and 
 blunt, chisel shaped tamper at the other. The surface of the 
 base is brought to a plane parallel to the proposed finished sur- 
 face of the walk, and at a distance below it equal to the thick- 
 ness of the top dressing. A straight edge, long enough to span 
 the walk and notched out at the ends so that when placed on 
 the molds the straight edge will define the correct grade of the 
 base, is a convenience here. 
 
 636. The concrete is now cut into blocks exactly corre- 
 sponding to the proposed blocks in the top dressing. For this 
 purpose a straight edge is laid across the walk in line with 
 marks previously made on the molds to define the joints, and 
 with a spade or special tool the concrete base is cut entirely 
 through to the sub-base. This division is necessary to allow for 
 expansion and contraction, and prevent cracks in the top dressing 
 elsewhere than at the joints. This joint in the base should 
 then be filled with clean sand. If preferred, these joints in the 
 base may be made by placing thin steel strips across the molds 
 to be removed after the concrete for the next block is in place. 
 
 The end block made from a given batch of concrete should 
 be limited by a cross mold set exactly on line of a proposed 
 joint. When the base is continued, this cross mold is removed. 
 A part of a block should never be molded and then built on 
 after having stood long enough to begin to set. Any concrete 
 
424 CEMENT AND CONCRETE 
 
 left over from finishing a block should either be mixed in with 
 the next batch, if this is to follow in a very short time, or it 
 should be wasted. A disregard of this rule will probably result 
 in a crack in the top dressing above the line of division between 
 adjacent batches. 
 
 637. When a block of base is finished, the top dressing or 
 wearing surface should be applied immediately. The lack of 
 adhesion between the base and wearing surface is one of the 
 most frequent causes of failure in cement walks. The mortar 
 should not merely be laid on in a thick layer and then struck 
 off to grade, but it should be worked and beaten into close con- 
 tact with the concrete at every point. The mortar should be 
 tamped with a light rammer and beaten with a wooden batten, 
 and to accomplish this properly the mortar must not be very 
 wet. The surface is then to be struck off with a straight edge 
 bearing on the top of the mold planks. Some hollows or rough 
 places will remain, and the straight edge should be run over a 
 second or perhaps a third time, a small amount of rather moist 
 mortar, made from thoroughly screened sand, having been first 
 applied to such places. 
 
 When the surface film of water is being absorbed, the surface 
 is worked with a wooden float. The exact time when the work 
 should be floated will soon be known by experience. After the 
 floating is completed, the trowel may be used to give a smoother 
 surface, but this makes the walk so slippery that it is not usually 
 desirable. 
 
 638. If the top dressing is worked too long, the cement is 
 brought to the surface, robbing the next lower layer of its ce- 
 ment and resulting in scaling. The top dressing is now cut 
 entirely through on exact line above the joints in the base. 
 This may be done by a trowel working against a straight edge, 
 but special tools are made for cutting through the mortar and 
 rounding the edges of the joint at one operation. A quarter- 
 round tool is also run along the edges of the mold to give a neat 
 finish. When desired, an imprint roller run over the walk 
 gives it the appearance of having been bushhammered. 
 
 It is important that the top dressing be applied before the 
 concrete has begun to set, and it must not be applied to a por- 
 tion of a block and then some time allowed to elapse before 
 applying the remainder. The edge of the top dressing must 
 
CONCRETE WALKS 425 
 
 be cut off squarely at the end of the block. If desired, the 
 wearing surface may be colored by the use of lamp black in the 
 mortar, giving a uniform gray color to the walk. ( 535.) 
 
 639. When the walk is completed, it should be fenced off 
 so that animals may not walk over it while still fresh, and it 
 should be protected from a hot sun. The surface should be 
 kept moist, and this may be done after the first twenty-four 
 hours by spreading a layer of damp sand over the walk and 
 wetting the sand with a rose nozzle as often as may be needed. 
 The walk may be opened to light travel after about four days, but 
 it is better to remain covered with the damp sand for a week. 
 
 640. Cost of Concrete Walk. The cost of concrete walks 
 varies from ten cents to twenty-five cents per square foot. A 
 fair price for a walk of average quality where there are no 
 special difficulties is twelve to eighteen cents per square foot. 
 
 As an instance of a walk built with special care, the one 
 constructed about the top of the bank of the Forbes Hill Reser- 
 voir may be mentioned. 1 The sub-base of this walk was of 
 stone and twelve inches thick, the layer of concrete was five 
 inches thick at the center of the walk and four inches at the 
 sides. The top was of granolithic finish one inch in thickness. 
 The walk was laid in separate blocks about six feet square. 
 The average gang employed on the concrete consisted of six 
 men and one team, while the finishing was done by two masons 
 and one tender. The amount laid per day was about forty 
 square yards. The cost per square yard was as follows: - 
 
 $ cu. yd. stone in foundation or sub-base, at $.40 per cu. yd. . $0.133 
 Labor, placing stone at $1.50 per day ......... . .502 
 
 Total cost stone foundation per sq. yd. of walk . . . $0.035 
 
 .158 bbl. cement, at $1.53 per bbl. ... ........ $0.242 
 
 .065 cu. yd. sand, at $1.02 per cu. yd ............ 060 
 
 .109 cu. yd. stone, at $1.57 per cu. yd ............ 170 
 
 Labor, mixing and placing concrete ............ 450 
 
 Total cost concrete base per sq. yd ......... " $0.028 
 
 .11 bbl. cement, at $1.53 per bbl ............. $0.168 
 
 .022 cu. yd. sand, at $1.02 per cu. yd ............ 022 
 
 Lamp black ...................... 008 
 
 Labor, preparing and finishing surface ........... 140 
 
 Total cost top dressing or wearing surface ..... " $0.347 
 
 Total -t walk per sq. yd. | J^ ;?$ ; ; ; ; 
 
 C. M. Saville, M. Am. Soc. C. E., Engineering News, March 13, 1902. 
 
426 CEMENT AND CONCRETE 
 
 641. The following is given as an estimate of cost of items 
 in a walk built with six inch cinder sub-base, four inch concrete 
 base and one inch top dressing . 
 
 COST PER SQ. YD. OP WALK 
 
 MATERIALS LABOR 
 
 Preparation of foundation, excavation and ramming . . . $0.20 
 
 Sub-base, 6 in. cinders cu. yd., at $0.40 cu. yd $0.07 
 
 Placing and ramming cinders 0.04 
 
 ^ cu. yd. concrete, at $3.00 per cu. yd. for materials alone . 0.33 
 
 | cu. yd. concrete, placing, at $1.80 per cu. yd 0.20 
 
 Top dressing V cu. yd. mortar, at $9.00 per cu. yd. . . . 0.25 
 
 Placing top dressing and finishing walk 0.25 
 
 Superintendence and molds 0.10 
 
 Totals $0.65 $0.79 
 
 Total cost per sq. yd., $1.44, or 16 cents per sq. ft. 
 
 642. As an example of a low priced walk, the concrete walks 
 in San Francisco 1 are but three inches thick, two and one-half 
 inches of concrete composed of one part Portland cement, two 
 parts beach gravel, and six parts of crushed rock of size not 
 exceeding one inch; the top dressing being one-half inch thick 
 of equal parts Portland cement and beach gravel. With ce- 
 ment $2.50 per bbl., crushed rock and gravel from $1.40 to 
 $1.75 per cu. yd., and wages twenty cents an hour for laborers 
 and forty cents for finishers, this walk is constructed at from 
 nine to ten cents per square foot. It is stated that a gang of 
 three or four men will lay 150 to 175 square feet per day. 
 
 ART. 74. FLOORS OF BASEMENTS, STABLES AND FACTORIES 
 
 643. The principles governing the laying of walks apply also 
 in a general way to the construction of floors that rest directly 
 on the ground. 
 
 For residences, basement floors may be laid with three inch 
 base of concrete and one-half inch wearing surface. The thick- 
 ness of sub-base will depend upon the character of the soil. 
 Where natural conditions do not assure good drainage of the 
 foundation, this should always be provided for by either a blind 
 stone or tile drain laid around the outer edge of the building 
 and leading to the sewer or other outlet. The finished surface 
 of the floor should always have a slight slope toward the center 
 
 Engineering News, March 4, 1897. 
 
FLOORS 427 
 
 or one corner of the basement, and a trapped sewer connection 
 set at this lowest point in such a way that it is accessible for 
 repairs and cleaning. 
 
 644. Wet Basements. Where much ground water is en- 
 countered, and especially where a basement is subjected to a 
 head of water from without, special precautions must be taken 
 in building the floor. The concrete must be made thick enough 
 so that its weight and the arch action set up, shall be able to 
 withstand the upward pressure of the water. In building such 
 a floor it is necessary to keep a sump hole, preferably in the 
 center, towards which the construction proceeds from the sides. 
 A pipe placed in the sump hole permits pumping until the con- 
 crete is laid about the pipe, when the latter may be filled with 
 rich cement mortar. In such cases the side walls of the base- 
 ment should be plastered with Portland cement mortar on the 
 outside and special care taken in joining the floor to the wall. 
 
 645. Size of Blocks. As the changes in temperature in a 
 building are usually much less than in open air, the blocks of 
 concrete may be of much larger size, say ten feet square, and 
 many basement floors are laid without any joints, though 
 sooner or later they will probably crack if so laid. In factories 
 for certain purposes, however, the floors may be subjected to 
 greater changes in temperature than walks laid in the open 
 air. In such cases the blocks should not be more than three 
 or four feet on a side, and the joints may well be filled with 
 asphalt, especially if water-tightness is desired. 
 
 646. Stable floors may be made of six inch cobble or broken 
 stone sub-base, six inches of concrete made with mortar con- 
 taining three parts sand to one cement, and one inch of top 
 dressing containing three parts sand (mixed sizes) or crushed 
 granite to two parts cement. 
 
 Factories having heavy machinery with much vibration re- 
 quire strong floors. Such a floor may be made of six inches 
 of cobble stone sub-base covered by six inches of a lean concrete 
 made with one-to-four mortar, and above this, three to five 
 inches of rich concrete made with mortar containing two and 
 one-half parts sand to one cement, and one inch of top dressing, 
 equal parts cement and sand or cement and crushed granite. 
 
 647. Example and Cost. In the construction of the new 
 printing building for the Government Printing Office at Wash- 
 
428 CEMENT AND CONCRETE 
 
 ington, the basement floor is nine inches thick, made as fol- 
 lows: * 
 
 1. Concrete sub-base, six inches thick of one part natural 
 cement, two parts sand and four and one-half parts broken 
 brick. 
 
 2. Concrete base, two and one-half inches thick of Portland 
 cement one part, sand two parts and fine broken gneiss four 
 parts. 
 
 3. Top dressing, one-half inch in thickness, of two parts 
 sand to one part Portland cement. 
 
 The cost of this floor was about $1.50 per square yard, or 
 about seventeen cents per square foot. 
 
 ART. 75. CONCRETE IN PAVEMENTS AND DRIVEWAYS 
 
 648. PAVEMENT FOUNDATIONS. The principal use of con- 
 crete in connection with city pavements has been as a founda- 
 tion, the wearing surface being of some other material, as brick, 
 asphalt, cedar blocks, etc. 
 
 Concrete for pavement foundations should not be less than 
 six inches in thickness, and a greater thickness will be required 
 where the ground is insecure. The excavation having been 
 made to the required sub-grade, and all loose soil removed and 
 the places refilled with broken stone, the earth is thoroughly 
 rolled to a smooth surface parallel to the surface of the proposed 
 pavement. Drainage for the foundation should be provided 
 where necessary by broken stone or tile drains beneath the 
 curb. Before beginning the placing of concrete, stakes may be 
 driven in the foundation, with their tops at grade, at intervals 
 of five to ten feet over the entire pavement, to assist in securing 
 the proper grade of concrete surface. 
 
 649. The stone for the concrete should be broken so that no 
 piece is larger than two and one-half inches in its greatest di- 
 mension. If the stone is of good quality, it need not be screened 
 except to remove the finest dust, if this is present in consider- 
 able quantities. Sufficient mortar should be used to fill the 
 voids in the stone, this mortar being composed of about two 
 parts sand to one of natural cement, or better, two and one- 
 half or three parts sand to one of Portland cement. This con- 
 
 Report of Capt. John S. Sewall, Report Chief of Engineers, 1896. 
 
PAVEMENTS 429 
 
 crete is thoroughly rammed in place, care being taken that 
 adjacent batches as laid in the street mingle with each other 
 so as to show no line of demarcation. In stopping work for 
 the night, the concrete should cut off sharply on a straight 
 line parallel to the direction of the proposed joints in the wear- 
 ing surface. Joints extending across the street should be left 
 at intervals of thirty to forty feet to allow for expansion and 
 contraction. 
 
 650. The concrete is finished to a surface parallel with the 
 proposed street surface, a templet being employed to secure 
 this. The concrete should be kept damp for a few days, and 
 no traffic allowed upon it until the wearing surface is laid. If 
 the wearing surface is of brick or wooden blocks, a layer of 
 sand about one inch thick is first spread over the concrete. 
 
 The advantages of a concrete foundation for street pave- 
 ments are its strength and durability and water-tightness. 
 
 651. CONCRETE PAVEMENT. Concrete has not been a popu- 
 lar material for a street surface except for short driveways arid 
 in courts where both vehicles and pedestrians must be accom- 
 modated. One reason for this is that concrete is slippery, and 
 another, that owing probably to carelessness or ignorance, the 
 wearing qualities have not been good. The first objection may 
 be largely removed by cutting the surface into blocks, four by 
 eight inches, by deep grooves, or by the use of a deep imprint 
 roller on the wearing surface. As to wearing qualities, there 
 seems to be no good reason why a concrete cannot be made 
 tough enough to withstand heavy traffic. It will of course be 
 necessary to divide the work into blocks of twenty to twenty- 
 five square feet, with expansion joints of sand, asphalt, or 
 tarred paper between. A third objection is the glare of the 
 surface in summer. A partial remedy for this may be had by 
 placing some coloring matter, such as lamp black, in the top 
 dressing. 
 
 652. The sub-base may consist of a six inch layer of broken 
 stone, or twelve inches of cinders, well drained and thoroughly 
 compacted by rolling. For exceptionally heavy wear it may be 
 advisable to use a five inch layer of lean concrete for the sub- 
 base, after rolling the bottom of the excavation and providing 
 drainage. 
 
 Upon the sub-base should be laid a base, composed of four 
 
430 CEMENT AND CONCRETE 
 
 inches of concrete made with first class stone, such as granite, 
 trap or hard limestone crushed to pass a ring one and one-half 
 inches in diameter, and containing enough mortar, one part 
 Portland cement to two or three parts sand, to fill the voids in 
 the stone. The top dressing, a layer of granolithic one and one- 
 half or two inches thick, should then be immediately applied. 
 This mortar should be made with one or two parts granite, trap, 
 or other hard rock crushed to pass a five-eighths inch screen, to 
 one part Portland cement. 
 
 These two layers are placed in much the same manner as 
 that described for laying concrete sidewalks, but the joints in 
 base and top dressing should run at angles of forty-five degrees 
 with the curb to prevent ruts following the lines of the joints. 
 A roller making deep imprints is then run over the finished 
 surface to furnish a foothold for horses, or, for this purpose a 
 special roller may be used to mark the top dressing into blocks 
 approximately four by eight inches, with deep (one-half inch) 
 grooves. 
 
 When completed, the pavement should be kept moist, pref- 
 erably by a layer of damp sand, and no traffic should be al- 
 lowed upon it for at least a week or ten days. 
 
 653. Concrete pavement laid in Bellefontaine, Ohio, was 
 found to be in good condition after ten years' service; 1 the 
 only serious defect apparent being that, since the blocks were 
 marked off parallel to the curb, ruts have sometimes formed 
 along these joints. This pavement was made with four inches 
 base concrete, laid directly on sub-grade where foundation 
 is gravel, sand or porous soil; or if soil is impervious, the 
 base was laid on four inches of broken stone or cinders. The 
 top layer was two inches thick, equal parts cement and sand or 
 pea granite. Sub-drains of three inch tile were laid inside each 
 curb line, and the curb is formed as part of the outer blocks. 
 Both the base and top dressing were cut through in squares, 
 five feet on a side. The cost of the pavement is said to have 
 been $2.15 per square yard, and very few repairs have been 
 found necessary. 
 
 In Germany a cement macadam, made with six inch sub- 
 
 1 Municipal Engineering, December, 1900, and Engineering News, Jan. 7, 
 1904. 
 
CURBS AND GUTTERS 431 
 
 base of broken stone or gravel, with a wearing surface of hard 
 macadam stone mixed with cement, has been successfully used. 
 
 ART. 76. CURBS AND GUTTERS 
 
 654. The use of concrete for curbs and gutters is rapidly 
 increasing. Curbing is sometimes molded and afterward put in 
 place like stone curbing, but the greatest advantages in the use 
 of concrete for this purpose are only attained by molding in 
 place the curb and gutter as one structure. 
 
 The Parkhurst combined curb and gutter is a patented form 
 that has proved very satisfactory. This form has a projection 
 of about one inch at the back and another along the bottom 
 just below the curb, this feature being patented. 
 
 A combined curb and gutter may consist of a curb four to 
 six inches wide at the top, and five to seven inches at the bot- 
 tom, and have a face of six to seven inches above the gutter. 
 The upper face corner of the curb and the angle between curb 
 and gutter should be rounded with a radius of one and one- 
 half to two inches. The gutter is sixteen to twenty inches 
 wide, and from six to nine inches thick, with top surface con- 
 forming to the grade of the street. 
 
 655. The sub-base should consist of a layer of broken stone 
 six inches thick, or six to twelve inches of cinders thoroughly 
 rammed. The preparation of the foundation should be similar 
 to that required for a pavement, care being taken that the 
 sub-base be thoroughly drained, tile being used if necessary. 
 Forms to receive the concrete are held in place by stakes, the 
 molds being carefully set to grade. The sub-base may now be 
 covered by a layer of four to six inches of Portland concrete of 
 only moderate richness, as one to three to six, and the concrete 
 to form the curb and gutter placed upon it before it has set, 
 or a six inch layer to form the gutter may be placed directly 
 on the sub-base. 
 
 656. Concrete to form the curb and gutter should be of good 
 quality, not more than two and one-half parts sand to one part 
 Portland cement being used for the mortar, and sufficient mor- 
 tar used to entirely fill the voids in the stone. The broken 
 stone for this concrete should be rather fine, with few, if any, 
 pieces larger than one inch in greatest dimension. The ex- 
 posed faces receive a top-dressing, or wearing surface, of one- 
 
432 CEMENT AND CONCRETE 
 
 half inch to one inch of granolithic containing not more than 
 one and one-half parts of trap or granite, pea size, to one part 
 Portland cement. This coating is applied as soon as possible 
 after the concrete is placed, as in sidewalk work. The surface 
 is troweled or floated, but a smooth, glossy finish is avoided. 
 
 The curb and gutter may well be laid in alternate blocks 
 about six feet long, but a somewhat neater appearance is se- 
 cured by making the work continuous, and cutting it entirely 
 through at intervals of six feet to provide for slight movement. 
 As the molds may be used repeatedly, they should be sub- 
 stantially made. Special forms are of course required at corners, 
 catch basins, etc. As in other concrete construction, the work 
 should be protected from injury and kept moist for at least a 
 week. 
 
 657. On business streets it is desirable to build the sidewalk 
 close to the curb, with only a joint between, the grade of the 
 walk conforming to the curb and sloping up toward the build- 
 ing line one-quarter inch to the foot. On residence streets the 
 walk should be separated from the curb by a park strip, the 
 walk being high enough to give drainage toward the curb. 
 
 Steel facing is sometimes used for curbs subjected to excep- 
 tional wear, as in front of shipping warehouses and freight 
 sheds. Where these are applied, they should cover the top 
 and the upper part of the face of the curb and must be well 
 anchored, by bolts or special webs, to a substantial mass of 
 concrete, otherwise they will work loose and defeat the object 
 for which they are used. 
 
 658. Cost of Concrete Curb and Gutter. At Champaign, 
 111., 1 a curb was built seven inches high and five inches thick, 
 the gutter, six inches thick, extending nineteen inches into the 
 roadway from the face of the curb. The foundation consisted 
 of six inches of gravel or cinders well rammed. The concrete 
 was composed of one part Portland cement to five parts fine 
 gravel, and the finishing coat, one inch thick, was of one part 
 Portland cement to one part clean, sharp, coarse sand. The 
 cost per foot was thirty-nine cents, including all excavation. 
 
 A similar curb at Urbana, 111., was 4 inches thick at the 
 top, 5 inches at the base and 7 inches high; the gutter being 
 
 1 W. H. Tarrant, Engineer, Proc. 111. Soc. Engr. and Surveyors, 1899. 
 
STREET RAILWAY FOUNDATIONS 433 
 
 5 inches thick and extending 18 inches into the roadway. The 
 foundation was composed of eight inches of cinders or gravel. 
 The concrete was of one part Portland cement to five parts 
 clean gravel, and the finishing coat was one inch thick, com- 
 posed of one part Portland cement to two parts sharp sand. 
 The price per linear foot, including the excavation, removal of 
 old curbing, and refilling, was forty-six cents. 
 
 At South Bend, cement curb alone, 6 inches wide at top, 
 7 inches at bottom and 16 inches depth, with the upper half 
 composed entirely of one to two Portland cement mortar, has 
 been constructed for eighteen cents per linear foot. 
 
 ART. 77. STREET RAILWAY FOUNDATIONS 
 
 659. The heavy motor cars used on city and urban electric 
 railways subject the track to very severe service. As the head 
 of the rail must be practically flush with the pavement on city 
 streets, cross- ties, when used, are so far beneath the surface 
 that they decay rapidly and their renewal entails the tearing 
 up of the pavement. As there is not the same necessity for a 
 cross-tie on street tracks as on railroads, since the rails are held 
 to gage by the pavement, these objections to the cross-tie have 
 led to the adoption of a concrete girder under each rail. The 
 rails and ties (if ties are used) should not only rest upon the 
 concrete, but should be imbedded in it. Track in which the 
 rails rested upon concrete, but were not imbedded in it, has 
 been found to yield laterally and get out of alinement, while 
 on the other hand, if the ties rest upon earth or gravel and are 
 filled between with concrete, the track is likely to settle, break- 
 ing the bond of the concrete. 
 
 660. The method of placing concrete beams for street rail- 
 way tracks in Minneapolis was as follows : l The rails were first 
 spiked to cross-ties at intervals of six to eight feet, and the rail 
 joints cast-welded. In laying the street pavement foundation 
 of natural cement concrete, a rough groove, fifteen inches wide 
 at the bottom and eighteen to twenty inches at the top, was 
 left under each rail. This groove was immediately filled be- 
 tween ties with concrete made of one part Portland cement, 
 
 1 F. W. Cappelen, M. Am. Soc. C. E., Engineering News, Oct. 14, 1897; 
 Municipal Engineering, November, 1896. 
 
434 CEMENT AND CONCRETE 
 
 two and one-half parts sand, and four and one-half parts broken 
 stone. 
 
 The rails were tied together every ten feet with wrought 
 iron tie bars, three-eighths inch by two inches, set on edge. These 
 tie bars were rounded at the ends, threaded and attached to 
 the web of the rail by two nuts, one on either side of the web. 
 The rails were then spiked to the concrete beam, the temporary 
 wooden ties removed, and the spaces left by them filled with 
 concrete, completing the beam. As the concrete beam was 
 eight inches thick and the rail five inches, the sub-grade was 
 thirteen inches below the top of the rail. 
 
 On the gage side of the rail were placed toothing blocks of 
 granite, 3^ by 9 inches by 4 inches deep, held away from the 
 rail 1J inches by temporary wooden strips. After removing 
 these strips, cement grout was poured into the groove to fill 
 2J inches over the base of the rail, the remaining 2^ inches to 
 the top of the rail being filled by asphaltic cement which re- 
 mained soft enough to permit a flange groove to be made by the 
 first car over the track. The asphalt wearing surface was laid 
 against the rail on the outer side. Mr. Cappelen, in describing 
 this construction, says that a rail six inches high with six-inch 
 base should be used, with granite toothing blocks, six by nine 
 inches by five and one-half inches deep. 
 
 The cost per foot of rail for the concrete beam construction 
 only, was twenty-six to twenty-seven cents, and for the filler, 
 five cents per foot. The cost per mile of double track, exclusive 
 of rails and pavement, was about $8,670.00. 
 
 Somewhat similar methods have been employed in Toronto 
 and Montreal. Canada, Indianapolis, Ind., and Scranton, Pa., 
 Denver, Detroit and Cincinnati. 
 
 661. At Scranton, Pa., 1 the rails were laid directly on the 
 six-inch concrete base of the pavement. This thickness was 
 increased to twelve inches under the joints (which were rein- 
 forced by an inverted rail four feet long) and under steel cross- 
 ties spaced ten feet centers and formed of old girder rails in- 
 verted and riveted through the flanges at the intersection. 
 Flat steel tie bars, threaded at the ends, spaced ten feet centers, 
 were also used here as at Minneapolis. 
 
 1 Description of the systems employed in several cities are given in En- 
 gineering News, Dec. 26, 1901. 
 
STREET RAILWAY FOUNDATIONS 435 
 
 The concrete mixing plant was mounted on a car running 
 on the track; the materials were delivered to the machine by 
 hand measuring boxes, and the Drake mixer deposited the con- 
 crete directly into the trench. The total cost per foot of track 
 is given as $2.65, $1.17 of which was for grading, rolling, con- 
 creting and brick paving at $1.97 per square yard, and for extra 
 concrete at joints and ties at $0.72 per square yard. 
 
 662. At Toronto, Canada, the six-inch concrete base of the 
 pavement is increased to eight inches in thickness for twenty 
 inches width under each rail, and the base of the latter is im- 
 bedded one inch in the concrete. A 6J-inch grooved girder 
 rail is used, with mortar rammed between the web and the 
 adjacent paving blocks. 
 
 663. At Cincinnati the bottom of the concrete stringer is 
 nine inches below the base of the nine-inch grooved girder rail, 
 and the concrete is built up from three to six inches on the web, 
 according to the thickness of the wearing surface of the pave- 
 ment. The space between the upper part of the web and the 
 adjacent paving is then filled with cement mortar, thus sup- 
 porting the head of the rail as well as protecting the web from 
 corrosion. 
 
CHAPTER XXI 
 
 SPECIAL USES OF CONCRETE (CONTINUED). SEWERS, SUB- 
 WAYS, AND RESERVOIRS. 
 
 ART. 78. SEWERS 
 
 664. There seems to be no very good reason why concrete is 
 not more generally employed in the construction of all large 
 sewers. With sizes less than two or two and one-half feet in 
 diameter the difficulty of removing the centers prohibits the use 
 of concrete in the ordinary way, and although some appliances 
 have been devised for building these small sewers as monoliths 
 by a mold that advances as fast as the concrete is tamped in 
 place, they have not proved popular. The difficulty of obtain- 
 ing a perfect grade, and the undesirable feature of leaving the 
 green concrete unsupported, are probably reasons sufficient for 
 this lack of popularity. 
 
 For the larger size sewers concrete has several advantages 
 over brick. First may be mentioned the very smooth finish 
 that may be obtained on the invert, appreciably increasing the 
 velocity of flow over that usually obtained with brick inverts. 
 Cheaper labor may be employed in concrete work with less 
 danger of annoyances from strikes. The cost is from one- 
 third to one-half less than for brick. 
 
 665. METHODS OF CONSTRUCTION. The City of Washing- 
 ton was one of the - first to use concrete extensively in sewer 
 construction 1 . For sizes up to twenty-four inches internal diam- 
 eter the concrete is used only as a foundation and bedding 
 for the ordinary sewer pipe. For a twenty-four inch sewer 
 the pipe rests in a bed of concrete twenty-seven inches wide 
 at the bottom, flaring to forty inches wide at the level of the 
 center of the pipe, and then carried up with plumb sides for 
 six inches, and finally finished by planes tangent to the upper 
 
 curve of the pipe. At the joints there are bands of concrete 
 
 1 Described by Capt. Lansing H. Beach, Corps of Engrs., U. S. A. Report 
 Operations District of Columbia, 1895. 
 
 436 
 
SEWERS 437 
 
 extending over the top, so that at these places the pipe is en- 
 tirely inclosed. Similar forms are used for the smaller sizes 
 with corresponding decreased dimensions. For all sewers be- 
 tween ten inches and twenty-four inches the sub-grade is six 
 inches below the exterior of the pipe, and in all cases the band 
 about the joint is four inches thick at the top. 
 
 666. The method of laying these sewers is as follows: The 
 trenches are 2^ to 3 feet in width, with " headers" about 2 feet 
 wide, left at intervals of 10 to 16 feet, which are tunneled 
 through. The grade and line pegs are placed in the headers 
 at the ground siarface, and a cord is stretched on the sewer line 
 over at least four stakes, at a convenient height above the 
 grade, and thus parallel to the bottom of the sewer. 
 
 When the trench is to the required grade, a six inch layer of 
 concrete, made with one barrel natural cement, two barrels 
 sand and four barrels gravel, is placed. This concrete is rammed 
 with iron rammers weighing sixteen pounds, and having eighteen 
 square inches ramming surface. The pipe is then laid upon 
 this bed and each section is tested for line and grade. For the 
 former, a plumb bob is used with its cord held against the 
 grade cord already mentioned, and for testing the grade a grad- 
 uated pole is used, with a projection at the bottom which sets 
 on the interior of the pipe, just within the open end. 
 
 Concrete is then lowered in buckets, deposited on top of 
 the pipe and allowed to fall down on the sides so as not to 
 disturb the alinement. When enough concrete to secure the 
 pipe has been thus placed, it is rammed and the concreting 
 continued until the required form is obtained, as already de- 
 scribed. The concrete in the bands carried over the joints is not 
 rammed but is beaten with wooden paddles and heavy trowels to 
 compact it and bring it to the desired form, four inches thick 
 anc^ four inches wide at the top, and flaring to twelve inches 
 wide (in the direction of the sewer) at the top of the pipe. 
 
 667. Cost. The quantities of concrete materials required 
 to lay one hundred linear feet of pipe sewers as described 
 above are given as follows : 
 
 Size of sewer .8 inch 12 inch 18 inch 24 inch 
 
 Cement, bbls 6.76 10.58 14.77 19.14 
 
 Sand, cu. yd 2.07 3.23 4.52 5.85 
 
 Gravel, cu. yd 4.16 6.47 9.04 11.70 
 
438 CEMENT AND CONCRETE 
 
 With natural cement costing $0.79 per barrel in sacks, sand 
 $0.47 per cu. yd., gravel $0.75 per cu. yd., and laborers $1.50 
 to $1.75 per day, foremen, masons and inspectors $4.00 per 
 day, the average cost of laying pipe sewers in this manner was 
 approximately as follows, exclusive of the cost of the pipe: 
 8-inch, $1.11; 12-inch, $1.14; 15-inch, $1.46; 18-inch, $1.60; 
 21-inch, $1,67; 24-inch, $2.32 per foot. 
 
 668. Sewers at Chicago. In the construction of some 
 17,000 feet of sewers for the Chicago Transfer and Clearing 
 Yards/ concrete was used for all sewers of thirty-six inches 
 diameter and over. The excavation was mostly in blue clay 
 and done by steam shovel to a depth of twenty feet, the re- 
 mainder being removed by hand shovels and swing derrick. 
 The material was such that in general the bottom of the trench 
 could be trimmed to the form of the exterior of the sewer. 
 The thickness of the ring of concrete was 8 inches for 36 and 
 42-inch sewers, 10 inches for 48-inch, and 12 inches for 84 and 
 90-inch sewers. 
 
 The concrete was composed of one part " Steel Pozzolana" 
 (slag) cement, three parts safld and five parts broken stone. 
 The cement was of course very finely ground and showed high 
 seven-day tests. The cost was $1.30 per barrel delivered. 
 The sand was the Chicago " torpedo" sand, coarse and of good 
 quality, and cost about ninety cents per cubic yard delivered. 
 The stone was a limestone from Summit, 111., crushed in two 
 sizes, namely, 1 to 2J inches and to 1J inches. These two 
 sizes of stone were mixed in proportions one part of the coarser 
 to two of the finer. The cost of stone was about $0.80 per 
 cubic yard delivered. 
 
 The concrete was mixed by a rotary mixer of the continuous 
 type provided with radial blades. The mixer was mounted on 
 a flat car, with engine and upright boiler. Three cars of stone, 
 the mixer car, two cars of sand and one of cement made up the 
 concrete train, which ran on a track laid close to the trench 
 and was kept near the work by a small locomotive. The mixer 
 was supplied by wheelbarrows running from the material cars 
 on plank runways attached to the cars. The concrete was also 
 transported in wheelbarrows from the mixer to the trench. 
 
 E. J. McCaustland, Trans. Assoc. C. E., Cornell University, 1902. 
 
SEWERS 439 
 
 The bottom of the trench being cut to form, the con- 
 crete for the invert was laid directly on the sub-grade, tamped in 
 layers carried up until the invert occupied about one hundred 
 forty degrees of arc. The form of the inner face of the invert 
 was maintained by template, grade stakes being set 12 \ feet 
 apart along the trench. The remainder of the sewer was laid 
 on centers resting on the invert. The ribs for this centering 
 were made in a complete circle, of three thicknesses of one by 
 twelve inch boards nailed together and cut to a true circle. 
 Ribs were placed four feet center to center, and covered with 
 lagging two inches thick and three inches wide, planed to radial 
 joints. The strips of lagging were held in place at each end of 
 a section by a j\ by 2 inch iron band passing over all of the 
 strips, and turned in at the ends, forming a hook in which rested 
 the lower lagging strip, the other strips being supported by this 
 one. The lower part of each rib rested on the invert, the upper 
 portion being cut to a diameter four inches less (that is, smaller 
 by twice the thickness of the lagging). While the trench was 
 near enough to the outside of the sewer ring not to measurably 
 increase the amount of concrete over and above the desired 
 thickness, the trench served as the outside form. Above this 
 point, planks were inserted and braced to the sides of the trench. 
 From the haunches to the crown the exterior was finished with 
 a template. 
 
 When completed, the exterior form planks were removed, 
 and a light covering of earth placed on the surface to protect 
 it from drying too rapidly. This was especially necessary in 
 this case on account of the kind of cement used. The centers 
 were removed usually after forty-eight hours, by swinging the 
 ribs about the vertical diameter and removing the lagging. As 
 soon as the centering was removed, the inner surface was plas- 
 ered with a mortar composed of three parts lake sand to one 
 part cement. 
 
 670. Cost. The company furnished the materials used in 
 the sewer ring and manholes, and delivered it on the work, 
 while the contractor furnished all tools and labor to dig the 
 trenches, complete sewer and manholes, and do the back filling. 
 The contract prices per foot are given by Mr. McCaustland, the 
 resident engineer, as follows : 
 
440 
 
 CEMENT AND CONCRETE 
 
 36-inch sewer in trench averaging 11 feet deep, 3,340 feet, at $2.30. 
 42 " " " 14 " 2,660 " 3.00. 
 
 48 " " " 17 " 4,540 " 3.57. 
 
 84 " " " 22 " 1,000 " 5.91. 
 
 90 " " " 24 " 5,400 " 6.68. 
 
 From the data given we have computed the approximate 
 quantities of concrete per foot of sewer, and assuming the cost 
 of the materials for a cubic yard at $3.00, we obtain the follow- 
 ing approximate costs: 
 
 
 
 MATERIALS. 
 
 
 
 SIZE 
 SKWER. 
 
 DEPTH 
 TRENCH. 
 
 
 CONSTRUCTION 
 CONTRACT 
 PRICE PER 
 FOOT. 
 
 ESTIMATED 
 TOTAL COST 
 PER FOOT. 
 
 Approximate 
 Cubic Yards 
 
 Approximate 
 Cost 
 
 
 
 Concrete. 
 
 Concrete. 
 
 
 
 36 in. 
 
 11 feet. 
 
 .285 
 
 $ .85 
 
 $2.30 
 
 $3.15 
 
 4'2 " 
 
 14 " 
 
 .325 
 
 .97 
 
 3.00 
 
 3.97 
 
 48 " 
 
 17 " 
 
 .47 
 
 1.41 
 
 3.57 
 
 4.1)8 
 
 84" 
 
 22 " 
 
 .93 
 
 2.79 
 
 5.91 
 
 8.70 
 
 90" 
 
 24 " 
 
 .99 
 
 2.97 
 
 6.68 
 
 9.65 
 
 671. Special Molds for Small Sewers. In the construction 
 of a thirty inch sewer at Medford, Mass., 1 Mr. William Gavin 
 Taylor invade use of a very convenient form. The lower 240 
 degrees of the sewer was of concrete, the upper 120 degrees 
 being of brick. To construct the concrete portion as a mono- 
 lith, the forms were constructed in lengths of ten feet, separat- 
 ing on a vertical line into two halves. The two halves were 
 connected by clamps, and held at the proper distance apart by 
 dog irons in the end ribs of each form. After smearing the 
 forms as usual, the concrete was deposited and rammed. When 
 it had partially set, the dog irons were removed and turn-buckles 
 used to slowly pull the two halves together. This method pre- 
 vented the green concrete being broken, although the concrete 
 extended up on the sides thirty degrees above the horizontal 
 diameter. 
 
 672. The centers used for the brick arch were also ingen- 
 iously arranged, and since they might have been used for a con- 
 crete arch they may be described here. These centers were 
 also in ten foot lengths. The ribs, of two inch plank, were 
 
 1 Abstract from Annual Report of City Engineer, Engineering Record, 
 Nov. 7, 1903. 
 
SEWERS 441 
 
 spaced two feet centers, with lagging J inch thick by 1 inch 
 wide, with one bevel edge to make a tight upper surface. The 
 rear end of each center was supported by wedges securely 
 fastened to the outer end of the preceding section, the forward 
 end being supported by a screw jack. 
 
 After turning the arch, these centers were removed by the 
 aid of a special truck the axles of which were bent at such an 
 angle as to make the cast iron wheels fit the concrete invert. 
 The axle of a roller was first fastened to the outer rib of the 
 center to be removed; the truck was then run back a foot or so 
 under the center and the screw jack supporting the forward 
 end of the center released. This allowed the forward end to 
 drop a short distance, the roller resting on the running board 
 of the truck. The latter was then pulled into the sewer far 
 enough to let the roller run off the end of the truck and lock 
 itself. The truck being then pulled out of the sewer toward 
 the finished end, drew the center away from the wedges sup- 
 porting the rear end, allowing the form to drop on the truck 
 and be wheeled out of the sewer. By this method the centers 
 were successfully removed without injuring the concrete. 
 
 673. Cost. From data given, the cost of this sewer 
 about sixteen hundred feet in length is approximately as 
 follows, labor costing twenty-five cents an hour: - 
 
 1.25 cu. yds. excavation and back fill, at $.59 $0.74 
 
 .15 cu. yd. concrete, at $6.70 1.00 
 
 .037 cu. yd. brick masonry, at $12.05 44 
 
 Cost of linear foot, exclusive of manholes, estimated at . . $2.18 
 The total cost per linear foot is given as $2.39 
 
 674. New York Sewers. In connection with the construc- 
 tion of the New York Rapid Transit Railway, some of the 
 sewers were built of concrete. This work was done with ex- 
 ceptional care, and on a large scale, and it was found that the 
 concrete sewers cost one-third less than similar sewers of brick. 
 
 The method of construction of one section may be described 
 as follows : l The forms for the invert of the straight lengths 
 of sewer were twelve feet in length, consisting of a strong frame- 
 work covered with closely matched lagging, planed smooth and 
 
 Engineering News, March 6, 1902. 
 
442 CEMENT AND CONCRETE 
 
 greased with machine oil. After the trench was prepared, con- 
 crete was placed and rammed until the top of the concrete was 
 within about one-half inch of the flow line of the invert. To 
 accomplish this, a straight edge was used, bearing on the fin- 
 ished invert in the rear and a template secured to the trench 
 timbering just ahead of the section under construction. 
 
 The invert centers were then placed, resting on the finished 
 invert at the rear and on a solid foundation accurately set to 
 grade at the forward end. Mortar composed of equal parts 
 Portland cement and sand was then tamped between the invert 
 form and the bottom concrete already laid. When the flow 
 line had been thus accurately formed, the center was braced 
 and vertical planking set to form the outside of the walls. The 
 concrete was then rammed in place. 
 
 Joists of two inch by four inch scantling laid along the 
 center of the top of each side wall of the invert section, formed, 
 when removed, a mortise into which the fresh concrete of the 
 arch section was rammed to form a bond. Similar mortises 
 were also made in the forward end of each section as built. 
 After twenty-four hours or more the forms were removed, and 
 a thin cement wash was applied to the interior, sufficient only 
 to fill any slight imperfections in the surface. 
 
 The arch centers, similar in construction to the forms for 
 the invert, were put in place and plastered with one inch of 
 rich Portland mortar. Concrete was then placed sufficient to 
 make the arch eight inches thick, the outside of the walls being 
 formed by inclined boards braced to the trench, and the top of 
 the extrados was formed by hand. 
 
 675. Steel Forms. Two novel types of centering have been 
 devised, in which the surface next the concrete is of steel. In 
 one of these * the forms are in sections about three feet long. 
 Two of the pieces of steel are of a width suitable to reach from 
 the bottom of the sewer to just above the spring line of the 
 arch, while a third piece forms the arch center. The strips are 
 bent at an acute angle at the sides, thus projecting into the 
 sewer along an element of the surface where the plates join ; the 
 two sides of adjacent plates, which flare away from each other, 
 are then connected by a continuous U-shaped clip of steel slipped 
 
 Engineering Record, Jan. 9, 1904. 
 
SUBWAYS AND TUNNELS 443 
 
 on from the end of a three foot section, and the intervening 
 space in the clip filled with clay or melted paraffin. The form is 
 assembled outside the trench, and after the paraffin is in place, 
 the center may be handled. When the sewer is completed, 
 the paraffin is melted by a suitable heater, or the clay is washed 
 out, and the form may be collapsed and removed. 
 
 676. In the other form l the steel plates are in continuous 
 strips about six inches wide and are applied by setting up the 
 wooden form on an improvised axis, revolving the form and 
 wrapping the steel sheet about it as it is revolved. The wooden 
 form is in two parts, upper and lower, firmly connected while 
 in use, but the two parts may be made to approach each other 
 by driving out the wedges between them. After the winding, 
 the center, with its sheet steel jacket, is lowered into the trench. 
 When the concrete is completed, the form is collapsed and 
 removed, leaving the spiral of steel in place to support the con- 
 crete until the latter is well set. The steel is then removed by 
 simply pulling on one end. As it comes away from the concrete 
 it is wound into a coil, and is then ready to be rewound on the 
 wooden form. Both of the above styles have been patented. 
 
 ART. 79. CONCRETE SUBWAYS AND TUNNEL LINING 
 
 677. The advantages of concrete in subway construction 
 and in tunnel lining are now well established. In subways 
 built in open cut, the side walls and invert are of concrete built 
 in place, while the roof is frequently made with I-beams with 
 concrete arches turned between them. The I-beams are sup- 
 ported directly on the side walls, which are usually made mono- 
 lithic with the invert. 
 
 678. Special precautions have to be taken to exclude water 
 from a subway, and for this purpose tarred felt and Portland 
 cement plaster are employed. 
 
 The specifications for the New York Rapid Transit Subway 2 
 were carefully framed to secure a waterproof construction. On 
 the sub-grade was placed a layer of concrete, smooth and level 
 on top. This was covered by alternate layers of hot asphalt 
 and felt, from two to six layers of each being used as deemed 
 
 1 Engineering News, Feb. 18, 1904. 
 
 2 Abstracted in Engineering News, Feb. 13, 1903. 
 
444 CEMENT AND CONCRETE 
 
 necessary for the conditions encountered. The remainder of 
 the concrete forming the floor was then laid upon the top layer 
 of asphalt. In dry, open soil the felt was not required, and in 
 dry rock excavations above water level both the asphalt and 
 felt were omitted. Similar provisions were made for water- 
 proofing the side walls and roof, resulting in a complete layer 
 of asphalt and felt imbedded in concrete about the entire tun- 
 nel, the waterproofing being protected both inside and out by 
 concrete. 
 
 679. In the construction of the Boston Subway l the por- 
 tion built in open cut was made as follows: The work was di- 
 vided into sections of convenient length, about twelve feet, 
 so that work on a section could be carried on continuously until 
 completed. Upon the prepared grade were laid three thick- 
 nesses of tarred felt with six-inch lap joints, well pitched be- 
 tween the layers, and the top of the upper layer thoroughly 
 covered with the pitch. When the latter had hardened, the 
 invert was laid over the entire width of the section. 
 
 At each side a back wall six inches thick was built up to a 
 convenient height and braced. The forms were then removed 
 and the face of this back wall was plastered with rich Portland 
 cement mortar. The main side walls were then built up be- 
 tween this layer of plaster and the forms defining the interior 
 face of the wall. This portion of the subway had an arch 
 roof, two feet thick at the crown, which was laid on wooden 
 centers. The exterior of the roof was plastered like the side 
 walls, and then covered with four inches of concrete to protect 
 the plaster from injury. The centers were removed after from 
 ten to thirty days; the span of the arch was about twenty- 
 three feet. 
 
 680. Tunnel Lining in Firm Earth. In building tunnels in 
 earth that is sufficiently firm not to require extensive timber- 
 ing, concrete is well adapted for lining. An instance of this is 
 furnished by the extensive system of tunnels constructed for 
 telephone and telegraph service under the streets of Chicago. 2 
 The trunk conduits for this system are about thirteen by four- 
 
 1 Annual Report Boston Transit Commission, 1900; also described in 
 Engineering News, April 4, 1901. 
 
 2 Mr. George W. Jackson, Engineer, Proc. W. Soc. Engrs., 1902; also in 
 Engineering News, Feb, 19, 1903. 
 
SUBWAYS AND TUNNELS 445 
 
 teen feet inside, and the laterals about six by seven feet, all of 
 the five center horseshoe form. 
 
 The excavation was in hard clay which stood up well. Shafts 
 were located in basements of buildings rented for the purpose, 
 and in these basements were placed the compressed air plants, 
 material bins, concrete mixers, etc. The large air locks, some 
 of which would hold ten small construction cars, were placed 
 at the bottoms of the shafts. Work was done in three shifts, 
 working eight hours each. The two night shifts could excavate 
 about twenty-one feet of lateral tunnel in the sixteen hours, 
 and the day shift placed the lining. 
 
 681. The concrete was in general composed of five parts of 
 broken stone and screenings, or of mixed gravel and sand, to 
 one part Portland cement. For intersections but four parts 
 aggregate were used. This should make a very strong concrete. 
 The centers for the smaller conduits were made of three-inch 
 channels, each rib being in five parts bent to the proper form 
 and connected by flange plates bolted to the inside of the chan- 
 nels at the ends. These ribs were placed three feet apart, and 
 two-inch plank used for lagging. 
 
 The ribs for the trunk sewers were of similar construction, 
 but with heavier channels braced with angles. Steel lagging 
 was used, made of plates about twelve by thirty-six inches, 
 stiffened by 1^ inch angles on four edges. There were also 
 provided bulkheads or steel end plates of voussoir shape, twelve 
 inches along the intrados and twenty inches high, for the pur- 
 pose of retaining the end of each section of lining and permit 
 thorough tamping. These bulkhead sheets, or "end flights," 
 were also stiffened along three edges, and could be attached to 
 the webs of the channel ribs by short bolts. 
 
 The concrete was mixed at the shaft head and conveyed to 
 the work in cars twenty inches wide and four feet long, running 
 on a fourteen-inch gage track. The floor of the tunnel was 
 first laid in the excavation, the steel ribs then put in place on 
 the floor, and the lagging placed at the bottom and built up 
 the sides just ahead of the concrete. When near the crown, 
 short pieces of lagging three feet in length covering but two 
 ribs were used, and the concrete rammed in from the end of 
 these short sections until they were complete, and then another 
 row of short pieces placed and the operations repeated. 
 
446 CEMENT AND CONCRETE 
 
 The concrete floor of laterals was designed to be thirteen 
 inches, and the sides and arch ten inches thick, but in all cases 
 the entire space between the lagging and the sides of the ex- 
 cavation was filled with concrete. 
 
 682. In such work as this only the best materials should .be 
 used, and, as early strength is desired, the use of Portland 
 cement is general in order that the centers may be removed 
 within a reasonable period. The ends of the sections into 
 which the work is divided should, if possible, be brought up 
 square, the bulkhead sheets described above being an ingenious 
 and effective method of providing for this. Where it is not 
 practicable to finish with a square end over the entire area of 
 section, then the work on the sides should be stepped back 
 from the bottom toward the crown, each step being bounded 
 by planes corresponding to coursing and heading joints in a 
 masonry arch. 
 
 683. Tunnel Lining in Soft Ground. For tunnels in soft 
 ground requiring the use of a shield, some difficulties in using a 
 concrete lining are apparent. The principal one of these lies 
 in the fact that the fresh concrete is not capable of taking the 
 thrust of the jacks used in forcing the shield ahead. Attempts 
 have been made to overcome this difficulty by so constructing 
 the centers that the jacks may bear against them instead of 
 on the fresh concrete. 
 
 Another difficulty is that in materials requiring almost con- 
 tinuous support, the temporary timbering is in the way of the 
 centering for the concrete construction; and still another is the 
 difficulty of properly tamping the arch at the crown where the 
 tail of the shield confines the working space. Concrete blocks 
 were tried in the construction of sewers in Melbourne, but 
 without entire success. Such blocks were successfully em- 
 ployed in the underground road system of Paris, though at- 
 tempts to use fresh concrete in shield tunneling for this work 
 proved a failure. 
 
 684. East Boston Tunnel. In the construction of the East 
 Boston Tunnel Extension of the Boston Subway, however, a 
 monolithic concrete lining has been successfully built, the 
 tunnel being excavated by shield. 
 
 This tunnel is about twenty by twenty-four feet for double 
 track electric line, The arch ring and the walls are thirty- 
 
SUBWAYS AND TUNNELS 447 
 
 three inches in thickness, while the invert is twenty-four inches. 
 Two side drifts, eight feet square, were first driven a certain dis- 
 tance and timbered. The bottoms of these drifts were then 
 excavated, and the side foundations of concrete were placed in 
 lengths of sixteen to twenty feet. When the foundations had 
 set, the interior forms for the side walls were placed upon them, 
 supporting the caps, the exterior plumb posts removed, and 
 the concrete side walls, three feet thick, built up to within 
 sixteen inches of the springing line of the arch. This work was 
 kept about one hundred feet in advance of the shield. 
 
 The shield, provided with live rollers, rested upon these side 
 walls, the rollers running in a flanged plate placed on top of the 
 walls. The shield was forced ahead thirty inches at a time, 
 and sections of the arch thirty inches in length were turned 
 directly behind the shield. 
 
 685. The centers of the arch were of curved, ten inch steel 
 channels spaced thirty inches apart, and the lagging, four 
 inches thick, was placed from the bottom toward the key as 
 the concrete was built up. Each section of arch is keyed with 
 concrete pressed through two holes in the rear girder at the 
 top of the shield, special rammers being used to tamp the con- 
 crete into the space at the crown of the arch, the concrete being 
 directed into place by curved sheet-iron troughs. 
 
 In each section of arch sixteen cast iron bars, three and one- 
 quarter inches in diameter and thirty inches long, are built 
 into the concrete in position to receive the thrust of the shield 
 jacks. Wooden bulkheads on the jack plungers serve to con- 
 fine the fresh concrete, but the reaction is taken on the cast 
 iron bars which, being butted end to end in successive sections 
 of the arch, carry the stress back to concrete that is able to 
 sustain it. As the shield advanced, the space left over the 
 completed arch by the tailpiece of the shield was filled with 
 grout under pressure. The centers remained in place thirty 
 days. The invert was excavated and laid in ten-foot sections 
 about twenty-five feet in the rear of the shield. The concrete 
 was mixed at the bottom of the shaft and passed through the 
 air lock on cars. The concrete cars ran on a higher level than 
 the muck cars, in order not to interfere with the excavation. 
 
 686. Lining Tunnels in Rock. If the rock through which 
 a tunnel is driven is seamy and insecure, concrete is in most 
 
448 CEMENT AND CONCRETE 
 
 cases the cheapest and best lining. The cost of the lining is, 
 of course, less if it can be built in connection with the excava- 
 tion, but it is frequently difficult to foresee how a given rock 
 will stand exposure to the air and water, and it becomes an 
 exceedingly nice question to determine at the time of building 
 a tunnel whether lining is required. In many cases this ques- 
 tion is settled in the affirmative by other considerations than 
 the character of the rock, as the resistance to flow, in water- 
 works and sewers, or the ease of ventilation and the necessity 
 of a good appearance, as in street or steam railway tunnels. 
 
 687. New York Subway. In the construction of portions 
 of the rapid transit subways of New York, a traveling center 
 which served also to support a working platform was carried 
 on six wheels running on a track laid on the footing courses of 
 the side walls. This center carried at the side, sections of lag- 
 ging curved to the required form of the side walls. This lagging 
 was adjusted in place, and braced from the platform or center 
 by means of wedges. Directly behind this traveling center was 
 a similar platform carrying a derrick; and behind this, the 
 traveling center carrying the lagging for the roof. This third 
 platform was jacked up to place the roof lagging at the correct 
 elevation, and firmly supported by wedges. 
 
 The concrete was brought in skips on cars that ran on the 
 floor level and stopped beneath the derrick platform. The 
 derrick hoisted the skips through a hole in the platform and 
 placed them on cars on either the side wall or the roof platform, 
 so that the concrete was' delivered either to the side wall forms 
 in advance, or the roof forms in the rear as required. The 
 concrete was rammed in a direction transverse to the tunnel 
 axis until the roof was completed, except for a space about 
 five feet wide at the crown. The arch was then keyed by 
 tamping the concrete in from the end of the form. The two 
 platforms carrying the forms were each forty feet long, and 
 the derrick platform was eighteen feet. 
 
 688. The excavated rock was crushed for the concrete on a 
 working platform erected over and around the shaft head. 
 Cars delivered the excavated material at the shaft in steel skips, 
 which were hoisted to the working platform, set on push cars 
 and dumped into bins, from which stone was delivered to the 
 crusher; these cars then passed under the crushed stone bins, 
 
SUBWAYS AND TUNNELS 449 
 
 were loaded with broken stone, run back to the shaft head, 
 and the broken stone dumped into bins mounted over the 
 mixer. The skips were then lowered into the shafts by the 
 derricks, to be run to the headings and reloaded. The stone 
 and sand were fed to a measuring box by means of a hopper, 
 the measuring box discharging directly into a cubical mixer, 
 which was high enough above the tunnel floor to dump directly 
 into skips on the cars. 
 
 689. Cascade Tunnel. In the construction of the Cascade 
 Tunnel of the Great Northern Railway a somewhat different 
 arrangement was used. 1 The working platform in the tunnel 
 was erected five hundred feet in length, and cars hauled by 
 cable up an incline to the platform. The side walls were built 
 in alternating sections, eight to ten feet in length, the support 
 of the arch timbering being thus gradually transferred from 
 the plumb posts to the concrete of the side walls. Arch sections 
 were built in twelve foot lengths, the centers being made of 
 four by sixteen inch plank without radials, so as to leave a 
 clear way for concrete cars on the working platform. The 
 latter were high enough to allow the material cars to run be- 
 neath them. 
 
 690. Concrete vs. Brick. There are frequent instances in 
 engineering construction where brick masonry might well have 
 been replaced by concrete, and the use of brick for tunnel lining 
 is still adhered to in many cases. This is partly because some- 
 what less elaborate centers can be used for brick arches, and 
 the centers may be struck somewhat earlier, and partly be- 
 cause of extreme conservatism on the part of the designer, 
 although without doubt there are cases where the use of brick 
 is entirely warranted. 
 
 An interesting instance of the greater adaptability of con- 
 crete under unforeseen conditions, however, is presented by the 
 Third Street concrete and brick lined tunnel at Los Angeles, 
 Cal. 2 This tunnel was excavated mostly through an argilla- 
 ceous sandstone. The side walls were of concrete up to the 
 haunches, the upper part of the arch being of six courses of 
 brick. A streak of yellow clay was encountered, and it "was 
 
 1 Mr. John F. Stevens, M. Am. Soc. C. E., Engineering News, Jan. 10, 
 1901. 
 
 2 J. H. Quinton, M. Am. Soc. C. E., Engineering News, July 18, 1901. 
 
450 CEMENT AND CONCRETE 
 
 soon demonstrated that the six ring brick arch, which occupies 
 the central portion of the roof, was not strong enough to hold 
 up the immense weight above it, and the temporary timbering 
 was crushed and broken in a most alarming way." The strength 
 of the arch was increased by using nine rows of brick instead 
 of six until the clay seam was passed. In such portions of the 
 six ring arch as had cracked, it was found that the inner ring 
 of brickwork had separated from the second ring, and in places 
 the second .ring had separated from the third. The concrete 
 walls had shown no evidence of weakness. 
 
 To repair the brickwork, steel concrete beams or arches were 
 inserted in the brickwork at intervals of four feet, and extend- 
 ing from one concrete wall to the other. These beams were 
 tw r elve inches wide and eight to twelve inches deep, made of 
 rich concrete, and had imbedded in each beam two pieces of 
 three inch by three-quarter inch steel. The steel ribs were set 
 in recesses cut out of the brickwork, and rested at the ends upon 
 the concrete of the side walls. Substantial centers were used 
 for building the concrete beams, and when the latter had set, 
 the defective brickwork between adjacent beams was cut out 
 and replaced by rich concrete. 
 
 691. Aspen Tunnel. Another illustration of the adaptabil- 
 ity of concrete when unexpected difficulties arise, is furnished 
 by the construction of the Aspen Tunnel on the Union Pacific 
 Railroad. l The original design provided for sets of timbers to 
 support the excavation, spaced about three feet, center to center, 
 but for nine hundred feet of the tunnel such pressures were 
 encountered that in places a solid wall of twelve by twelve inch 
 timber was forced in. For a portion of this section the lining was 
 built of a combination of concrete with steel ribs. The latter 
 were 12-inch, 55-pound I-beams spaced from twelve to twenty- 
 four inches, center to center, curved to conform to the interior 
 of the tunnel. The concrete was built up around and between 
 the beams, the inner flange being covered by from four to seven 
 inches, and the total thickness of the walls two to three feet. 
 
 692. The Perkasie Tunnel of the Philadelphia and Reading 
 Railroad was constructed through a firm rock, which, however, 
 was intersected by several strata of seamy rock. As trouble 
 
 1 W. P. Hardesty, Engineering News, March 6, 1902. 
 
SUBWAYS AND TUNNELS 451 
 
 was experienced from rock falling from these strata, it was 
 decided to line the tunnel at such places. This lining had a 
 minimum thickness of eighteen inches at the crown and twelve 
 inches at the sides. Traffic through the tunnel was not ob- 
 structed during the work of placing the lining. In laying 
 about five hundred cubic yards of concrete, the cost was about 
 ten dollars and eighty cents per cubic yard, exclusive of cost 
 of centering and dry filling. 1 
 
 693. Water Works Tunnel. The lining of portions of the 
 Beacon Street Tunnel of the Sudbury River Aqueduct was 
 undertaken some fourteen years after its excavation, and at a 
 time when it was necessary to use the tunnel intermittently to 
 supply water to the city of Boston. The methods employed 
 are described by Mr. Desmond FitzGerald in Transactions 
 American Soc. C. E. for March, 1894. 
 
 A substantial track of 2 feet 1J inch gage was laid from a 
 manhole furnishing access to the sewer to the portion of the 
 tunnel to be lined. The rails', weighing thirty-six pounds to 
 the yard, were supported on small but substantial trestles, 
 built of three by four inch spruce joists, and placed eight feet 
 between centers. Every third trestle was braced from the 
 sides and roof of the tunnel to prevent the track being floated 
 when the tunnel was in use. The trestles also carried five rows 
 of planks for the workmen to walk on in pushing the cars. 
 The track was elevated by these trestles, so the work was not 
 seriously interfered with by a small amount of water in the 
 tunnel. The track cost about eighty-seven cents a foot. 
 
 Cars to run on these tracks to deliver materials and concrete 
 had frames five feet by one foot nine inches, with twenty inch 
 wheels, and cost about fifty-six dollars each. 
 
 694. Centers. The centers were in three parts, two for 
 side walls and one for roof. The ribs were of three thicknesses 
 of two by ten spruce plank, without interior bracing for the 
 roof section. The side sections had each an inclined brace. 
 Wedges were inserted between the tops of the side sections 
 and the bottoms of the roof ribs to hold the latter in place. 
 The lagging was two by four inch spruce, in eight foot lengths, 
 with beveled edges and planed both sides. The centers .were 
 
 P. D. Ford, M. Am. Soc. C. E., Trans. A. S. C. E., March, 1894. 
 
452 CEMENT AND CONCRETE 
 
 spaced four feet apart, and seventy-five full centers were built; 
 these, with the lagging, contained 14,000 feet B. M. of lumber, 
 and cost $1,460.55, or $104.30 per thousand feet B. M. 
 
 695. Methods of Work. Broken stone, sand and cement 
 were stored in shanties over and around the manhole leading 
 to the tunnel, and arrangements made by which the materials 
 could be delivered through chutes down the manhole to the 
 cars As it was found more convenient to work in winter, 
 special provision was made for storing large quantities of ma- 
 terial in the shanties. The sand was piled around an iron lined, 
 wooden bulkhead, in the center of which was a large stove. 
 
 The concrete was mixed within the tunnel as close to the 
 work as possible, and in places where the cross-section had 
 been sufficiently enlarged by falls of rock to permit easy work- 
 ing. The materials, delivered to the material cars down the 
 chutes already mentioned, were pushed to the mixing platforms 
 and combined in the proportions of 18.56 cubic feet of crushed 
 stone and 7.35 cubic feet of sand to one barrel of Portland 
 cement, being approximately 1 to 2 to 5J. The above quanti- 
 ties of materials made 20 to 21 cubic feet of concrete. When 
 mixed, the concrete was shoveled into cars, conveyed to the 
 work and then shoveled into place. 
 
 The tamping was done principally with oak rammer five 
 inches square, twelve inches long, with a short wooden handle 
 in one end. In tamping the key of the arch, long-handled iron 
 rammers were used. Much care was requisite here to prevent 
 the aggregate separating from the mortar and lodging next 
 the lagging, as it always has a tendency to do, thus resulting 
 in voids in the face of the work when the lagging is removed. 
 The concrete was built up on the sides in horizontal layers and 
 stepped back by inserting bulkheads, so that the adjacent 
 sections bonded together. 
 
 696. Cost. The cost of this concrete lining, which was 
 built under great disadvantages, amounted to $16.15 per cubic 
 yard. This cost must be considered reasonable in view of the 
 fact that the materials had to be transported an average dis- 
 tance of more than one-half mile on small push cars, and the 
 work in the tunnel was suspended for three days of each week 
 to allow the tunnel to be used to maintain the water supply of 
 the city. 
 
RESERVOIRS 453 
 
 ART. 80. RESERVOIRS: LININGS AND ROOFS 
 
 697. Although the choice of the material with which to 
 construct a reservoir may in some cases be varied by local 
 conditions, it is found that under ordinary circumstances con- 
 crete offers the greatest advantages for a minimum cost. For 
 the side walls of small reservoirs, concrete furnishes the requi- 
 site strength and water-tightness with a moderate thickness; 
 earthen embankments and floors may be made practically im- 
 pervious with concrete and mortar, combined with asphalt 
 when considered necessary; while for the roofs, groined arches 
 or beams and slab construction, with supporting piers, all of 
 this material, make a neat, permanent, and altogether satis- 
 factory covering, at a smaller expense than would be required 
 for brick or stone masonry. 
 
 698. Details of Construction. In the walls and floors, 
 water-tightness is a prime consideration, and this is best at- 
 tained by a layer of mortar on the inner surfaces or between 
 two layers of concrete. 
 
 As in floors, walks, etc., the necessity of providing for ex- 
 pansion and contraction will depend upon the extremes of 
 temperature to which the surface is to be subjected. In covered 
 reservoirs which are to be almost constantly filled with water, 
 or in very equable climates, the blocks may be large, say twenty 
 feet square, while under more severe conditions the blocks 
 may not contain more than twenty square feet. The joints 
 between the blocks may well be wide enough to be filled with 
 asphalt. This furnishes an elastic joint which is compressed 
 as the blocks expand, and swells when the blocks again con- 
 tract. 
 
 699. Reservoir Floors. One of the principal difficulties ex- 
 perienced in the construction of floors is from settlement of the 
 foundation. The floor should, therefore, have strength enough 
 to bridge any small irregularities in the foundation that may 
 result from inequalities in settlement. For a similar reason, it 
 is not well to make the blocks too large, as smaller blocks with 
 compressible joints will more readily conform to an uneven 
 surface without permanent injury. In order that the reser- 
 voir shall not leak even if the foundation settles, the concrete 
 and mortar may be covered with one or more layers of asphalt. 
 
454 CEMENT AND CONCRETE 
 
 In building the floor lining, alternate blocks are sometimes 
 placed first in molds and the intermediate blocks built in later. 
 In other cases the blocks are laid consecutively. The advan- 
 tage of the former method seems to lie principally in the ease 
 of construction, as access may be had to all sides of the 
 block. 
 
 700. In hard clay soil not liable to settlement, four inches 
 is sufficient thickness for the floor, the concrete to be covered 
 before it has set with a half-inch layer of rich Portland mortar, 
 troweled to a smooth surface. If the reservoir when empty 
 will be subjected to hydrostatic pressure from without, the 
 floor must be designed to resist this pressure. In this case, if 
 seepage from without into the reservoir is objectionable, a layer 
 of mortar may be placed over the first layer of concrete and 
 protected by the concrete laid upon it. This outside pressure 
 may be provided for in a covered reservoir by making the floor 
 of inverted arches between piers, the weight of the floor, piers, 
 roof, and earth filling over the roof, being made sufficient to 
 balance the upward pressure on the floor. If there is no ob- 
 jection to the water from without being led into the reservoir, 
 a porous layer of broken stone or gravel beneath the floor may 
 be connected with the interior of the reservoir through pipes 
 provided with check valves, and the outside pressure be thus 
 removed. Where it can be accomplished, it will usually be 
 better to lead this ground water through a pipe to a sewer or 
 a lower level rather than into the reservoir. 
 
 701. Walls. The thickness of the wall is determined by 
 methods similar to those used in designing a retaining wall or 
 a dam according as the pressures are greater from the embank- 
 ment without or the water pressure within. In the case of a 
 covered reservoir, the thrust of the roof arches may convert 
 any vertical section of the wall into a beam, the earth pressure 
 from without being supported by the floor at the bottom and 
 the roof at the top. Or in case there is no back pressure from 
 earth filling, the thrust of the roof may be added to the inner 
 water pressure. In circular covered reservoirs the arch thrust 
 is usually taken by steel bands laid in the concrete and en- 
 circling the reservoir near the top of the wall. In narrow 
 reservoirs rectangular in plan, tie rods may be used, or the 
 wall may be buttressed to take the roof thrust. Concrete side 
 
RESERVOIRS 455 
 
 walls are usually built vertical, or nearly so, on the inside, and 
 with a batter on the outside. 
 
 702. Linings. Linings of sloping earthen embankments are 
 laid the same as the floors, and similar precautions are required. 
 There is greater danger of settlement of embankments than of 
 the floor foundation, and the blocks, therefore, may well be made 
 smaller. Some difficulty may be experienced with laying hori- 
 zontal asphalt joints on a sloping face, and some sliding of the 
 lining may be expected under ordinary conditions, the asphalt 
 joints being compressed. For this reason it would seem to be 
 better to use asphalt in the inclined joints only, and a mastic 
 in the horizontal joints. Another method which would probably 
 prove satisfactory is to lay first a tier of blocks next the floor, 
 and when these have set, apply a very thin coat of asphalt to 
 the upper edges of these blocks, following with another tier, 
 and so on. 
 
 703. ROOFS. Where it is necessary to cover a reservoir, 
 either to prevent the formation of ice, or the growth of algse, 
 or for other reasons, the groined arch is an excellent design for 
 the roof on account of the small amount of concrete required, 
 the clear head room given, and the ease of ventilation. The 
 extending use of reinforced concrete will also probably enter 
 this field to a greater extent in the future than it has here- 
 tofore. 
 
 The determination of the stresses in a groined arch roof is 
 complicated not only by the peculiar form of the arch itself, 
 but by the fact that the spandrels of the arches are filled with 
 concrete over the piers to the level of the extrados at the crown. 
 This evidently results in making of any given unit of the roof, 
 having a pier as its center, a cantilever, and the arch action is 
 interfered with. Unless, however, tension members of steel are 
 laid in the concrete near its upper surface, it is not wise to count 
 on the strength of the cantilever except to consider it a factor 
 of ignorance on the safe side. If one wishes to depart from 
 the ordinary and tried dimensions for groined arches in concrete, 
 such departure had better be based on some special experi- 
 ments and tests on full sized sections. Some of the dimensions 
 that have been used are given in the examples cited below. 
 
 704. Forms. The preparation of forms or centers for 
 groined arches is one of the most difficult and expensive details 
 
456 CEMENT AND CONCRETE 
 
 of the construction of such a roof. It will probably be best to 
 have each section of the form cover the space, square in plan, 
 between four piers. The ribs of the centering may well be 
 built up of planks, nailed together and sawed to proper form. 
 The lagging should be planed to size, and have radial joints to 
 make a smooth and even top surface. Care is necessary to make 
 a neat fit along the valley extending diagonally between piers, 
 and a small fillet may well be fitted into this valley to avoid 
 a sharp corner on the finished concrete, as well as to cover up 
 possible imperfections in the joints. The forms should, of 
 course, be designed to take the thrust of the adjacent com- 
 pleted arches, and if sufficient forms are not built to cover the 
 entire reservoir, and thus transmit the thrust to the walls, the 
 piers at the border of the forms must be thoroughly braced to 
 the opposite side walls or the piers will be toppled over and 
 the roof wrecked. This accident occurred to one reservoir 
 roof during construction, the pier braces having been removed 
 without the knowledge of the engineer. 
 
 705. In laying the concrete, joints between the work done 
 on consecutive days should cut the arches at right angles to 
 their axes, and bulkheads should be used to make such a joint 
 a vertical plane. The covering of each unit between four piers 
 is made monolithic, and care is necessary to prevent the stones 
 working to the bottom of the mass and thus becoming exposed 
 when the forms are removed. This may be prevented by plas- 
 tering the forms with mortar and placing the concrete upon it 
 before the mortar has begun to set. 
 
 706. A roof consisting of a network of concrete-steel beams 
 intersecting at right angles, supported by piers and covered by 
 concrete-steel slabs, makes a very simple design. The forms 
 are much easier to construct, and forms for only a limited area 
 need be erected at one time. An excellent article on " Covered 
 Reservoirs and Their Design," by Mr. Freeman C. Coffin, M. 
 Am. Soc. C. E., is contained in the July, 1899, number of the 
 Jour, of the Assn. of Engr. Soc. An article on the " Groined 
 Arch," by Mr. Leonard Metcalf, Assoc. M. Am. Soc. C. E., ap- 
 pears in Trans. A. S. C. E. for June, 1900; and Mr. Frank 
 L. Fuller presents an article on " Covered Reservoirs," in Jour. 
 Assn. Engr. Soc. for Sept., 1899. 
 
 707. Examples of Concrete Reservoirs. Wellesley. The 
 
RESERVOIRS 457 
 
 reservoir at Wellesley, Mass., 1 a part of the water supply sys- 
 tem, was designed by Mr. Freeman C. Coffin. It is eighty-two 
 feet in diameter, walls fifteen feet high, four feet thick at bottom 
 and two feet at top. The walls are of concrete and rubble 
 masonry. In the construction of the walls, concrete was used 
 containing three parts sand and five parts of stone to one 
 of cement, one cubic yard of concrete containing about 1.2 
 barrels of cement. The bottom of the walls, which were de- 
 signed to be built of concrete three feet four inches thick, were 
 actually built of rubble four feet thick, as a large quantity of 
 bowlders was at hand. The excavation was in hard clay con- 
 taining but little water, and the floor was made only four inches 
 thick, of concrete of the same quality as that used in the 
 walls. 
 
 The floor and side walls were plastered with two coats, the 
 first, one-half inch thick, of mortar containing two parts sand 
 to one of Portland cement, and a coat about one-eighth inch 
 thick, of neat Portland carefully rubbed and smoothed with 
 trowels. Such a plaster coat should be applied before the con- 
 crete has set. The two plaster coats cost twenty cents per 
 square yard. 
 
 708. The piers to support the groined arch roof were two 
 feet square, and built of brick. The span of the arches was 
 12 feet, rise 2.5 feet, and the concrete 0.5 foot thick at the 
 crown. A channel iron ring or band was set in the concrete 
 walls at the springing of the roof arches to take the thrust of 
 the latter. The centers were placed over one-fourth of the 
 area at a time, the piers being braced to take the thrust of the 
 arches until the roof was completed. The concrete in the roof 
 was composed of two and one-half parts sand and four and one- 
 half parts broken stone to one part Portland cement. The 
 centering cost twenty-two and one-half cents per square foot of 
 area covered. The spandrels were filled in level with top of 
 concrete at crown. On top of the concrete roof was placed six 
 inches of clean gravel for drainage and to prevent the earth 
 freezing to the concrete. This gravel was drained by four 
 inch vitrified pipe discharging at the toe of the slope wall. 
 
 1 Engineering News, Sept. 30, 1897; Jour. Assn. Engr. Societies, July 
 1899; Trans. A. S. C. E., June, 1900. 
 
458 CEMENT AND CONCRETE 
 
 One foot of earth filling and one foot of loam were placed upon 
 the gravel. 
 
 709. Astoria. The reservoir for the Astoria City Water 
 Works * was designed and built by Mr. Arthur L. Adams, M. 
 Am. Soc. C. E. The reservoir has a capacity of six and one- 
 fourth million gallons, walls twenty feet high. The excavation 
 was in hard clay and sand mixed with clay, which in some places 
 resembled a soft sandstone. The embankment was in general 
 about five feet, the remainder of the depth being in excavation. 
 
 The floor consisted of six inches of concrete, f inch cement 
 mortar, one coat liquid asphalt and one coat harder asphalt. 
 The slope lining was of six inches concrete, one coat asphalt, 
 one layer of brick dipped in asphalt and laid flat, and a final 
 finishing coat of asphalt. The concrete was composed of one 
 barrel Portland cement, one-tenth cubic yard sand, five-tenths 
 cubic yard gravel and nine-tenths cubic yard of crushed stone, 
 these quantities of the ingredients making one cubic yard of 
 concrete. Here we have an instance of the use of a mixture 
 of broken stone and gravel, a practice which has already been 
 commended as resulting in a small amount of voids. 
 
 The concrete of the floor was laid in blocks twenty feet on a 
 side, molds of two by six inch plank forming the outside edges 
 of a block, and serving as a guide to the straight edge used in 
 finishing, as in concrete walk construction. The finishing coat 
 was of two parts fine sand to one of Portland cement and was 
 applied, before the concrete base had begun to set, by two 
 finishers with smoothing trowels. When the next block was to 
 be laid, the plank were replaced by one-half inch weather 
 boarding. When the concrete had thoroughly set, these boards 
 were removed and the joints so formed were run full of asphalt, 
 when the first layer of this material was spread. 
 
 The concrete on the sides was also six inches thick and laid 
 in sheets ten feet wide, extending up and down the slopes, 
 expansion joints being provided on the inclined joints only. 
 The finishing coat of mortar was not used here, but all inequali- 
 ties in surface were smoothed by using a little mortar from the 
 next batch of concrete. 
 
 710. Each concrete gang was composed of twenty men and 
 
 1 Trans. A. S. C. E., December, 1896. 
 
RESERVOIRS 459 
 
 one water boy. All concrete was mixed by hand on movable 
 platform in half-yard batches. On the entire work 1.84 cubic 
 yards of concrete per day were mixed and placed per man 
 employed, and on the floor alone this quantity was increased 
 to 2.oo cubic yards, an excellent showing for this class of work. 
 The cost of concrete per cubic yard, without profit, was as 
 follows: 
 
 On Slopes: Cement, at $2.45 per bbl $2.82 
 
 Other materials 1.94 
 
 Labor 1.07 
 
 Total per cubic yard for 600 yards $5.83 
 
 On Floor: Cement, at $2.45 $2.64 
 
 Other materials 1.92 
 
 Labor .68 
 
 Total cost per cubic yard for 680 yards $5.24 
 
 The costs of the slope lining and floor complete, per square foot, 
 are given as follows: - 
 
 Slope: 6 inches concrete $0.1187 
 
 .649 inch asphalt 0100 
 
 Brick in asphalt 0889 
 
 .851 inch asphalt 0131 
 
 Chinking crevices 0030 
 
 Ironing 0036 
 
 Total cost per square foot of slope $0.2373 
 
 Bottom : 6 inches concrete $0.1031 
 
 Cement mortar finish 0113 
 
 .537 inch coat asphalt 0077 
 
 .573 inch coat asphalt 0082 
 
 Total cost of bottom per square foot $0.1303 
 
 711. Forbes Hill. The Forbes Hill reservoir * forms a part 
 of the distribution system of the Metropolitan Water Works of 
 Boston and was built under the direction of Mr. Dexter Brack- 
 ett, M. Am. Soc. C. E. The reservoir is two hundred eighty by 
 one hundred feet, partly in embankment. The soil under the 
 embankment was first stripped to a depth of two and one-half 
 
 1 Described by Mr. C. M. Saville, M. Am. Soc. C. E., Division Engineer, 
 before the N. E. Water Works Assn. Abstracted in Engineering News, March 
 13, 1902. 
 
460 CEMENT AND CONCRETE 
 
 feet at the toe, increasing to five feet stripping at the inner edge 
 of the slope. The material was hard pan, and the embank- 
 ments were built in four inch layers, rolled with four thousand 
 pound rollers, so made as to leave a slightly corrugated surface. 
 The bank was extended one foot inside of the finished line to 
 assure a compact face, and afterward trimmed to grade. 
 
 712. The bottom of the reservoir was covered first with a 
 layer of concrete about four and one-half inches thick, com- 
 posed of one part natural cement, two parts sand, and five 
 parts stone. The sand was of good quality; the stone came from 
 the excavation and was washed before crushing. This layer of 
 natural cement concrete was covered by a layer of Portland 
 cement mortar one-half inch thick, made of two parts sand to 
 one cement, and finished with a richer mortar, one part sand to 
 four of cement. 
 
 This half-inch layer was laid in strips four feet wide and 
 finished like a cement sidewalk. Although this mortar coat 
 was kept well moistened, some cracks developed which were 
 filled with grout before applying the second layer of concrete. 
 If no joints were used in the lower layer or base concrete, and 
 joints in the coat of mortar were provided in one direction only, 
 as appears to have been the case, the cracking should have been 
 anticipated. At any rate, the value of the mortar coat be- 
 tween the two concrete layers was greatly impaired by this 
 cracking, and the experience points to the advisability of plac- 
 ing the upper layer of concrete on the mortar before the latter 
 has set, thus avoiding the expense of finishing the mortar layer. 
 
 The upper layer of concrete was composed of one part Port- 
 land cement, two and one-half parts sand and four parts broken 
 stone, and was laid in blocks ten feet square. These blocks 
 were laid alternately each way. 
 
 The slope was first lined with Portland concrete of 1 to 2J 
 to 6^, then one-half inch layer of mortar as for the bottom. 
 The upper layer of concrete on slope was same as the upper 
 layer of the bottom lining, but the blocks were eight by ten 
 feet and finished with one inch of granolithic, in which stone 
 dust and particles smaller than three-eighths inch were sub- 
 stituted for the one and one-half inch stone of the concrete. 
 
 713. Cost. The cost of lower layer of concrete on bottom, 
 natural 1 to 2 to 5, was as follows : 
 
RESERVOIRS 461 
 
 1.25 bbl. natural cement, at $1.08 $1.350 
 
 .34 cu. yd. sand, at $1.02 347 
 
 .86 cu. yd. stone, at $1.57 1.350 
 
 Materials in concrete $3.047 
 
 Forms, lumber, at $20.00 per M ... $0.090 
 Forms, labor 0.100 
 
 Total forms .190 
 
 General expenses $0.08 
 
 Mixing and placing 1.17 
 
 1.250 
 
 Total cost per cu. yd $4.487 
 
 Cost of lower layer on slopes, Portland 1 to 2J to 6J, was as 
 follows: 
 
 1 . 08 bbls. Portland cement, at $1.53 $1.652 
 
 .37 cu. yd. sand, at $1.02 377 
 
 .96 cu. yd. stone, at $1.57 1.507 
 
 Materials in concrete .......... $3.536 
 
 Forms, lumber, at $20.00 per M $0.016 
 
 Forms, labor 0.121 
 
 Total forms .137 
 
 General expenses $0.177 
 
 Mixing and placing 1.213 
 
 1.390 
 
 Total cost per cubic yard $5.063 
 
 The cost of the upper layer on bottom and slopes, including 
 the finish on slopes, Portland 1 to 2 to 4, was as follows: - 
 
 1.37 bbls. Portland cement, at $1.53 $2.09 
 
 .47 cu. yd. sand, at $1.02 48 
 
 .745 cu. yd. stone, at $1.57 1.17 
 
 Materials in concrete $3.74 
 
 Forms, lumber, at $20.00 per M $0.25 
 
 Forms, labor 0.26 
 
 Total forms .51 
 
 General expenses $0.15 
 
 Mixing and placing 1.53 
 
 1.68 
 
 Total cost per cu. yd $5.93 
 
462 CEMENT AND CONCRETE 
 
 The cost of the half-inch plaster coat between the layers of 
 concrete was twenty cents per square yard. 
 
 714. Rockford. A reservoir for the city of Rockford, 111., 1 
 was built almost entirely of concrete after plans prepared by 
 the City Engineer, Mr. Chas. C. Stowell. The soil was a loose 
 gravel, and after excavation was completed, parallel lines of 
 drain tile were laid in trenches nine to ten feet centers and 
 leading to a fifteen inch vertical sewer pipe carried to the sur- 
 face of the street and capped. This sewer pipe served as a 
 sump for a pump should it be found necessary at any time to 
 repair the bottom. These trenches were filled with broken stone 
 and the whole area of the foundation covered with three inches 
 of clay. The concrete bottom was in two layers, eight inches 
 and seven inches thick, respectively, and composed of two 
 parts sand and five parts stone to one of Portland cement. 
 
 The walls were of similar concrete for the bottom twelve feet, 
 natural cement being used in the upper eight feet of the walls. 
 The thickness at the bottom was 4^ feet, walls being straight on 
 outside with one to ten batter on inside. The entire inner 
 surface of floor and walls was plastered with one-half inch of 
 Portland mortar, one to one. The cost of concrete in the work 
 averaged $6.50 for Portland and $4.00 for natural, and the 
 finishing coat cost seventy-five cents a square yard. 
 
 715. The roof was of concrete, expanded metal lath, and 
 steel rods, the finished thickness being but two inches. This 
 was supported by ribs of concrete, each twelve inches thick at 
 the crown and having a seven-inch channel on the under side. 
 The springing line of the ribs was eight feet below the top of 
 the walls, giving a good depth at the skew back. Ribs were 
 spaced about seven feet centers. The span of the roof was 
 about fifty-five feet, and rise about eleven feet. The cost of 
 roof was less than twenty-five cents a square foot. 
 
 716. Concord. The groined arch roof of the Concord, 
 Mass., 1 sewage storage well, designed by Mr. Leonard Metcalf, 
 Assoc. M. Am. Soc. C. E., was fifty-seven feet diameter and 
 contained about one hundred cubic yards of masonry. The 
 cost of the roof per square foot of surface was as follows: 
 
 Described in Engineering News, Feb. 22, 1894. 
 
RESERVOIRS 463 
 
 Centering $0.18 
 
 Concrete materials .15 
 
 Labor and supervision .05 
 
 Total $0.38 
 
 717. Albany. The Albany filter plant roof, 1 designed by 
 Mr. Allen Hazen, Assoc. M. Am. Soc. C. E., was also of the 
 groined arch type, the arches having the same span and rise as 
 the Wellesley reservoir. The cost of the roof per square foot 
 of area was as follows: - 
 
 .029 cu. yd. concrete, at $6.30 $0.182 
 
 Piers 054 
 
 Earth filling and seeding 014 
 
 Manholes, entrances, etc .027 
 
 Total cost per sq. ft $0.277 
 
 For a list of reservoirs and filter beds, in the roofs of which 
 groined arches have been used, giving in tabular form the 
 general dimensions, the proportions used in the concrete, and 
 in several cases the cost of the roof per square foot of reservoir, 
 the reader is referred to Engineering News of December 24, 
 1903. 
 
 Trans. A. S. C. E., June, 1900. 
 
CHAPTER XXII 
 
 SPECIAL USES OF CONCRETE (CONTINUED) 
 BRIDGES, DAMS, LOCKS, AND BREAKWATERS 
 
 ART. 81. BRIDGE PIERS AND ABUTMENTS AND RETAINING 
 
 WALLS 
 
 718. The use of concrete in large bridge piers was at first 
 confined to the hearting or backing of stone masonry shells. 
 It was soon found, however, that in many cases the concrete 
 was able to withstand the effects of frost and ice better than 
 was the variety of stone available for building the masonry 
 shell, and many important bridges are now supported by piers 
 built entirely of concrete. 
 
 As an example of this use may be mentioned the bridge 
 across the Red River * in Louisiana, which has six concrete 
 piers of heights from forty-four to fifty-three feet. The pivot 
 pier is twenty-seven feet in diameter, with vertical sides. The 
 draw rest piers are seven feet wide under the coping, nineteen 
 feet between shoulders and twenty-six feet long over all. The 
 sides have a batter of one-half inch to the foot. The coping is 
 of limestone. 
 
 719. In the construction of the Arkansas River Bridge 2 of 
 the K. C. P. & G. R. R., ten piers and two abutments were 
 built of concrete. The piers varied in height from twenty to 
 sixty-five feet, some of them containing over six hundred cubic 
 yards of concrete. The entire work was completed in eleven 
 months, although many difficulties were met. The concrete 
 was composed of one part Portland cement, two and one-half 
 parts coarse, sharp sand, and five parts of clean, broken stone. 
 
 The lagging for the forms was of two-inch yellow pine, sur- 
 faced one side and sized to one and three-quarters inches. On 
 
 1 George H. Pegram, Consulting Engineer. Walter H. Gahagan, Engi- 
 neer for Contractors. 
 
 2 Engineering News, Aug. 25, 1898. 
 
 464 
 
BRIDGE PIERS 465 
 
 the straight part of the pier this lagging was laid horizontal and 
 supported by four by six vertical posts set four-feet centers, 
 posts on opposite sides of the pier being tied together by three- 
 quarter inch bolts passing through one and one-half inch gas 
 pipes spaced five feet vertically. The gas pipe was allowed to 
 remain in the finished pier, the bolts being withdrawn. 
 
 The lagging for the semicircular ends was of two by six with 
 bevel joints, placed vertical, and supported at five-foot inter- 
 vals by segmental ribs of double two by twelve planks. At 
 the ends of the ribs were bolted short pieces of eight by eight 
 inch angle irons with edge horizontal. These angle irons were, 
 in turn, bolted to four by six verticals at the corners or shoul- 
 ders of the pier. 
 
 720. The foundation piers of the Lonesome Valley Viaduct, 1 
 thirty-six piers and two abutments, are entirely of concrete. 
 The piers are four feet square on top with batter of one inch 
 to the foot, and are from five to sixteen feet in height. The 
 total concrete laid was 926 cubic yards at a contract price of 
 about $7.00 per cubic yard. The piers were finished on top with 
 a steel plate, four feet square and one-half inch thick, taking 
 the place of coping stones. Where rock foundations were not 
 found, the lower portion of the pier was given an increased 
 batter to secure such a cross-sectional area at the bottom that 
 the unit pressure on the earth did not exceed one ton per square 
 foot. The cost of the concrete for this work has already been 
 given ( 322). 
 
 721. Steel Cylinders. Steel shells filled with concrete have 
 been used to good advantage, especially for bridge approaches. 
 Such shells are usually in pairs placed abreast, one under each 
 truss of the bridge or viaduct. The two cylinders of a pair 
 are usually connected by lateral bracing, and if desired in heavy 
 work, this bracing may be inclosed in a concrete wall and thus 
 protected from injury by running ice, etc. The thickness of 
 metal in the shells need not be great, three-eighths of an inch 
 usually being sufficient, though this depends on the height, the 
 stresses, and the liability to injury. In soft ground requiring 
 piles, most of the piles are sawn off below the limit of scour, or 
 below the water line for land piers, but one or more may be 
 
 Gustave R. Tuska, Trans. A. S. C. E., September, 1895. 
 
466 CEMENT AND CONCRETE 
 
 allowed to project up into the cylinders and the concrete filled 
 in around the heads, thus anchoring the pier. In foundations 
 on rock if the cylinders require an anchorage, this may be pro- 
 vided with bolts fox-wedged or cemented in the rock and pro- 
 jecting into the cylinder. (For details of methods adopted in 
 this class of construction, see " Bridge Substructure and Foun- 
 dations in Nova Scotia/' by Martin Murphy, Trans. A. S. C. E., 
 September, 1893.) 
 
 722. Repair of Stone Piers. Where masonry piers are being 
 destroyed by the abrading or expansive action of ice, or by 
 other causes, concrete is successfully used to arrest such action, 
 the entire pier being incased in a layer, one to three feet 
 thick, of Portland cement concrete of good quality. 
 
 The piers of the Avon River bridge, 1 originally built of ashlar 
 masonry, failed entirely to withstand the deteriorating in- 
 fluences of an extreme range in tide coupled with the severe 
 temperature of a Nova Scotia winter. Five of them were sub- 
 sequently incased in concrete, as follows: A form was made of 
 ten by ten inch spruce timber surrounding the ashlar masonry 
 of the piers and forming a mold to receive the concrete and 
 retain it in place until set. The thickness of the concrete casing 
 was two and one-half feet at the bottom and one and one- 
 third feet at the top, which was three feet above high water. 
 The concrete was composed of one barrel Portland cement, 
 one and one-half barrels clean sand, one barrel of clean gravel, 
 and in it was placed by hand four parts of common field stone 
 weighing from eight to twenty pounds each. This treatment 
 was entirely successful in preventing further disintegration. 
 
 723. An efficient cutwater for bridge piers is made by placing 
 old rails vertically on the upstream nose of the pier, anchoring 
 them to the masonry and filling between with concrete, leaving 
 only the wearing surface of the*rail head exposed. 
 
 724. Pneumatic caissons are usually filled with concrete, the 
 filling over the working chamber being carried up fast enough 
 to keep the work above water as the caisson is sunk. The 
 filling of the working chamber calls for special care in tamping 
 under and around the longitudinal and cross-timber braces. A 
 space of about three inches next the roof of the chamber is 
 
 1 Trans. A. S. C. E., December, 1893. 
 
RETAINING WALLS AND ABUTMENTS 467 
 
 filled with a rich concrete, containing no stone larger than one 
 inch, mixed quite dry and solidly tamped with special edge 
 
 rammers. 
 
 725. RETAINING WALLS AND ABUTMENTS. Concrete is used 
 
 very largely for constructing retaining walls and bridge abut- 
 ments. The foundations of a retaining wall should be of ample 
 width, and if the wall is not founded on rock, some settlement 
 and outward movement may be expected if the common for- 
 mulas are used in computing the dimensions. 
 
 If this movement is not equal throughout the wall, cracking 
 is likely to take place, and to confine these cracks to prede- 
 termined vertical planes, it is well to construct the wall, if a 
 long one, with vertical joints at intervals of fifteen to thirty 
 feet. Such a joint is made by building one section and fol- 
 lowing with another, without special precautions to make a bond 
 between. 
 
 If there is an opportunity for water to accumulate, care 
 should be taken to drain the earth back of the wall, either by 
 drains leading around the ends, or by pipes passing through 
 the wall. The latter may result in discoloration of the face. 
 
 726. Coping. The face of a retaining wall or abutment is 
 preferably given a batter, and a coping is provided to improve 
 the appearance. The coping should have a slight inclination 
 toward the back to prevent the discoloration of the face by 
 dripping. It should be divided by vertical joints into blocks, 
 not more than six to eight feet in length. The projection of 
 the coping will depend upon the dimensions of the wall. Wing 
 walls are preferably built with a sloping top or coping, but this 
 should be made monolithic with the wall by special molds 
 (729). 
 
 727. Rules for Use of Concrete in Abutments. In the use 
 of concrete for abutments and piers, the practice of the Illinois 
 Central Railroad, as set forth in their specifications, can hardly 
 be improved upon. The engineer of bridges and buildings on 
 this road, Mr. H. W. Parkhurst, M. Am. Soc. C. E., is one of 
 those engineers who early recognized the value of concrete in 
 bridge work, and as the result of his extensive experience along 
 this line, he is widely known as an able and conservative au- 
 thority. 
 
 These specifications are printed nearly in full in Engineering 
 
4G8 CEMENT AND CONCRETE 
 
 News of July 18, 1901, from which the following extracts are 
 made: 
 
 728, Natural and Portland Cement : Where used : - 
 
 Natural cement concrete "may be used where foundations 
 are entirely submerged below low-water mark, or where there is 
 no risk of the same being exposed to the action of the weather 
 by cutting away the surrounding earth. It, however, shall be 
 used only where a firm and uniform foundation is found to 
 exist* after excavations are completed. In all cases where 
 foundations are liable to be exposed to the action of the water, 
 or where the material in the bottom of excavations is soft or 
 of unequal firmness, Portland cement concrete must be em- 
 ployed for foundation work. 
 
 "The natural cement concrete shall usually be made in the 
 proportions (by measure) of one part of approved cement to 
 two parts of sand and five parts of crushed stone, all of char- 
 acter as above specified. For Portland cement concrete foun- 
 dations, one part of approved cement, three parts of sand and 
 six parts of crushed stone may be used. Wherever in the 
 judgment of the engineer or inspector in charge of the work, a 
 stronger concrete is required than is above specified, the pro- 
 portions of sand and crushed stone employed may be reduced, 
 a natural cement concrete of 1, 2 and 4, and a Portland cement 
 concrete of 1, 2 and 5 being substituted for those above speci- 
 fied. 
 
 "Portland Cement Concrete: Concrete for the bodies of 
 piers and abutments, for all wing- walls for same, and for the 
 bench walls of arch culverts, shall generally be made in the pro- 
 portions (by measure) of one part of cement, two and one-half 
 parts of sand and six parts of crushed stone. Where special 
 strength may be required for any of this work, concrete in the 
 proportions of 1, 2 and 5 may be used; but all such cases shall 
 be submitted to the judgment of the engineer of bridges, before 
 any change from the usual specification is to be allowed. 
 
 "For arch rings of arch culverts and for parapet head walls 
 and copings to same, Portland cement concrete, in proportions 
 of 1, 2 and 5, shall generally be used. Concrete of these pro- 
 portions shall also generally be used for parapet walls behind 
 bridge seats of piers or abutments, and for the finished copings 
 (if used) on wing-walls of concrete abutments, also for arch 
 
USE OF CONCRETE AND ABVTMENTS 469 
 
 work in combination with I-beams or in combination with iron- 
 work for transverse loading. 
 
 " Bridge seats of piers and abutments and copings of con- 
 crete masonry which are to carry pedestals for girders or longer 
 spans of ironwork, shall generally be made of crushed granite 
 and Portland cement, in the proportion (by measure) of one 
 part of approved cement, two parts of fine granite screenings, 
 and three parts of coarser granite screenings, the larger of which 
 shall not exceed three-quarters inch in greatest dimension." 
 
 729. After specifying the method of building molds, which 
 is treated elsewhere (Art. 62), the specifications proceed: - 
 
 "The planking forming the lining of the molds shall in- 
 variably be fastened to the studding in perfectly horizontal 
 lines ; the ends of these planks shall be neatly butted against 
 each other, and the inner surface of the mold shall be as nearly 
 as possible perfectly smooth, without crevices or offsets be- 
 tween the sides or ends of adjacent planks. Where planks are 
 used a second time, they shall be thoroughly cleaned, and, if 
 necessary, the sides and ends shall be freshly jointed so as to 
 make a perfectly smooth finish to the concrete. 
 
 "The molds for projecting copings, bridge seats, parapet 
 walls, and all finished work shall be constructed in a first-class 
 workmanlike manner, and shall be thoroughly braced and tied 
 together, dressed surfaces only being exposed to the contact of 
 concrete, and these surfaces shall be soaped or oiled if necessary, 
 so as to make a smoothly finished piece of work. The top 
 surfaces of all bridge seats, parapets, etc., shall be made per- 
 fectly level, unless otherwise provided in the plans, and shall 
 be finished with long, straight edges, and all beveled surfaces 
 or washes shall be constructed in a true and uniform manner. 
 Special care shall be taken in the construction of the vertical 
 angles of the masonry, and where I-beams or other ironwork 
 are not used in the same, small wooden strips shall be set in the 
 corners of the mold, so as to cut off the corners at an angle of 
 45, leaving a beveled face about one and one-half to two inches 
 wide, instead of a right-angled corner. 
 
 "Where wing- walls are called for, which have slopes corre- 
 sponding to the angle of repose of earth embankments, these 
 slopes shall be finished in straight lines and surfaces, the mold 
 for such wing-walls and slopes being constructed with its top 
 
470 CEMENT AND CONCRETE 
 
 at the proper slope, so that the concrete work on the slope may 
 be finished in short sections, say from three to four feet in 
 length, and bonded into the concrete of the horizontal sections 
 before the same shall be set, each short section of sloped sur- 
 face being grooved with a cross-line separating it from adjacent 
 sections. It will not be permitted to finish the top surface of 
 such sloped wing-walls by plastering fresh concrete upon the 
 top of concrete which has already set, but the finished work 
 must be made each day as the horizontal layers are carried up, 
 to accomplish which the mold must be constructed complete at 
 the outset; or, if the wing- wall is very high, short sections of 
 the mold, including the form for the slopes, must be completed 
 as the horizontal planking is put in place." 
 
 730. This is followed by directions concerning foundation 
 work; the following is given relative to building steel into the 
 masonry : - 
 
 "Iron rails to be furnished by the railroad company shall be 
 laid and imbedded in such manner as may be specified in such 
 foundation concrete as in the opinion of the engineer of bridges 
 needs such strengthening, and no extra charge, except the 
 actual cost of handling the same, shall be made by the contrac- 
 tor for such work, but the volume of such iron shall be esti- 
 mated as concrete. 
 
 "Where I-beams are to be placed in the angles of concrete 
 piers as a protection against ice, drift, etc., these shall be set 
 up and securely held in position so that they will extend one 
 foot or more into the foundation concrete. The planking of 
 molds shall be fitted carefully to the projecting angles of these 
 I-beams, and small fillets of wood shall be fitted in between 
 the inner faces of the mold and the rounded edges of the I-beam 
 flanges, so that no sharp projecting angle of concrete will be 
 formed as the work is constructed. 
 
 "These fillets may be made in short pieces and fastened 
 neatly into the mold as the layers of concrete are carried up. 
 Such I-beams will generally be furnished of sufficient length to 
 extend at least six inches above the top of the battered masonry 
 into the concrete coping, and special pains shall be taken to 
 tamp the concrete thoroughly around the I-beams, and to 
 finish the coping above and around the ends of the same, so as 
 to make a compact and solid bearing against the ironwork. 
 
CONCRETE PILES 471 
 
 " Where anchor bolts for bridge-seat castings are required, 
 they shall be set in place and held firmly as to position and 
 elevation, by templets, securely fastened to the mold and 
 framing. Such I-beams and anchor bolts shall be imbedded 
 in the* concrete work without additional expense beyond the 
 price to be paid per yard for the several classes of concrete in 
 which such iron is placed, the volume of iron being estimated 
 as concrete. 
 
 "After the work is finished and thoroughly set, all molds 
 shall be removed by the contractor. They shall generally be 
 allowed to stand not less than forty-eight hours after the last 
 concrete work shall have been done. In cold weather, molds 
 shall be allowed to stand a longer period before being removed, 
 depending upon the degree of cold. No molds shall be re- 
 moved in freezing weather, nor until after the concrete shall 
 have had at least forty-eight hours, with the thermometer at 
 or above 40 Fahr.. in which to set.'' 
 
 731. After giving in detail the methods to be followed in 
 placing and ramming concrete and the use of facing mortar, 
 the following paragraph is especially applicable to the subject 
 in hand: 
 
 " Layers of concrete shall be kept truly horizontal, and if, 
 for any reason, it is necessary to stop work for an indefinite 
 period, it shall be the duty of the inspector and of the contractor 
 to see that the top surface of the concrete is properly finished, 
 so that nothing but a horizontal line shall show on the face of 
 the concrete, as the joint between portions of the work con- 
 structed before and after such period of delay. If for any reason 
 it is impossible to complete an entire layer, the end of the layer 
 shall be made square and true by the use of a temporary plank 
 partition. No irregular, wavy or sloping lines shall be per- 
 mitted to show on the face of the concrete work as the result 
 of constructing different portions of the work at different 
 periods, and none but horizontal or vertical lines shall be per- 
 mitted in such cases." 
 
 ART. 82. CONCRETE PILES 
 
 732. Piles may be made of concrete either with or without 
 steel reinforcement. In the former case they are built in place, 
 but where steel is used the piles are usually driven after they 
 
472 CEMENT AND CONCRETE 
 
 have been prepared in suitable molds. Concrete is also em- 
 ployed to protect from decay, or from the ravages of the teredo, 
 wooden piles already in service. 
 
 Concrete piles are much more durable than wooden piles, 
 and may be used without reference to the water line. A sav- 
 ing may thus be made under certain conditions, as the use of 
 concrete piles may obviate the necessity of excavating to the 
 water line and building up with masonry resting on a wooden 
 pile foundation. As the diameter of the pile is not limited, 
 a much greater load per pile may be provided for. There are 
 of course many places where piles of concrete are not as suit- 
 able as wooden piles; they are not as well adapted to with- 
 stand certain kinds of hard usage, such as violent shocks, and 
 they are much less flexible. 
 
 733. Building in Place. In certain kinds of soil, such as 
 stiff clay, a wooden pile, or dummy, of the proper length may 
 be driven and withdrawn, the hole left being at once filled with 
 concrete. The application of this crude method is very lim- 
 ited, as it is seldom that the soil will stand until the hole is 
 filled with concrete. 
 
 For the building of piles without reinforcement, Mr. A. A. 
 Raymond l has patented a system by which a thin steel shell 
 or casing is driven to the desired depth and then filled with 
 concrete in place. A shell is first slipped over a steel pile core 
 made to fit it, and the shell and core are driven by a pile driver 
 in the ordinary manner. The core is then slightly shrunken 
 in diameter, by a simple device, and withdrawn, leaving the 
 shell in the ground. The core is hoisted in the pile driver 
 leaders, another shell is lowered into the one just driven and 
 then slipped up on the core, after which the driver is shifted 
 to the next location, and this shell is driven in the same manner 
 as the first. The filling of the shells with concrete is clone as 
 soon as convenient. While the shape of the shells may be 
 varied to suit conditions, the ordinary size is about twenty 
 inches diameter at the top and six inches at the bottom, and 
 such a shell twenty feet long weighs about seventy pounds. 
 
 734. The same company has a system of sinking shells 
 in sand by the water jet. For this purpose the shells are in 
 
 Raymond Concrete Pile Co., Chicago, 111. 
 
CONCRETE PILES 473 
 
 conical telescopic sections about eight feet in length. A two 
 and one-half inch pipe with three-quarter inch nozzle is attached 
 to the center of a cast iron point fixed to the inner section. 
 Water forced through the pipe causes the shell to settle, and 
 as the inner shell descends, its upper end engages with the lower 
 end of the second section, so that when fully lowered the sec- 
 tions form a continuous cone. The concrete is filled in simul- 
 taneously with the sinking, imbedding the two and one-half 
 inch pipe which remains permanently in the center of the con- 
 crete pile. 
 
 735. Concrete-Steel Piles : Molding. Piles of concrete-steel 
 usually have three or more steel rods of about one square inch 
 cross-section imbedded longitudinally in the pile, and connected 
 by smaller rods or wires at intervals of six to ten inches. 
 Molds are so made that they may readily be detached and used 
 again. At least one side of the mold should also be in short 
 sections that may be put in place as needed, in order to facil- 
 itate placing the concrete. The molds should be set up verti- 
 cally with the longitudinal steel rods in position. Enough con- 
 crete is put in the molds to fill six to ten inches in length, when 
 a set of transverse tie rods or wires is placed, then another 
 layer of concrete, etc. The concrete, which is of Portland 
 cement, should be mixed rather wet, as thorough tamping is 
 difficult in the confined space. The piles should be provided 
 with a cast iron shoe at the bottom, or a steel plate covering 
 to protect the point. At the top, one of the main rods is bent 
 over to form a ring to facilitate handling the piles. 
 
 736. Driving. When the concrete has hardened suffi- 
 ciently, say- at the end of four to eight days, the mold should 
 be removed, and the pile allowed to remain in its original posi- 
 tion twenty to twenty-five days longer, sprinkling it occasionally. 
 When thoroughly set, they may be driven with an ordinary 
 pile driver, using a heavy hammer and short drop. A steam 
 hammer is preferred, however, and a special cap must be used 
 to prevent injury to the pile head. Such a special cap may 
 well be made of cast steel, fitting over the head of the pile 
 like a helmet. The space between the lower end of the cap 
 and the side of the pile is calked with clay and rope yarn or 
 other, suitable material. Through a hole provided in the top of 
 the helmet, the space between the pile and cap is then com- 
 
474 CEMENT AND CONCRETE 
 
 pletely filled with dry sand. Such a cushion cap effectually 
 protects the pile head, distributing the pressure to the entire 
 head. Caps in the form of a steel ring filled with sawdust sur- 
 mounted by a wooden block, and also caps made of alternate 
 layers of lead, wood and iron plates have been successfully 
 used. 
 
 ART. 83. ARCHES 
 
 737. The use of concrete in the construction of arch bridges 
 is becoming so extended and diversified that it would require 
 a volume by itself to adequately cover the subject, and such 
 a treatment of it is well merited. All that can be attempted 
 here is to describe briefly one or two examples of well propor- 
 tioned arches, and to give a few hints on methods of design and 
 construction. 
 
 738. DESIGN. Concrete arches may be built with or with- 
 out steel reinforcement, but for long spans concrete-steel is 
 usually employed. The design of a concrete arch without steel 
 is entirely similar to that of a stone masonry arch, except that 
 planes of weakness corresponding to joints between voussoirs 
 in a masonry arch, may be somewhat more arbitrarily arranged 
 in the former. 
 
 In fixed concrete-steel arches, the arch ring is continuous, 
 and is capable of resisting a bending moment. The compu- 
 tations are therefore somewhat more complicated, and until the 
 action of concrete and steel in combination has been more 
 carefully determined, it may be said, in the words of a promi- 
 nent engineer, that "the development of the system of arches 
 of concrete must necessarily be largely based upon empirical 
 information coupled with sound judgment and wo'rk executed 
 with great care." ' Fortunately, the saving effected by this con- 
 struction over a masonry arch is usually so great that it is 
 possible to use a large factor of ignorance, and it is to be hoped 
 that the use of concrete-steel for arches of long span will not be 
 given a serious check by the failure, perhaps under unforeseen 
 conditions, of some of the web-like structures that have been 
 built of it. 
 
 739. Where the span and rise of the arch are not fixed by 
 the local conditions, the comparative economy of different 
 
 1 L. L. Buck, Trans. A. S. C. E. ; April, 1894. 
 
ARCHES 475 
 
 arrangements and the appearance of the completed structure 
 must govern. Shortening the spans decreases the amount of 
 concrete required in the arches, but increases the pier work, 
 which is usually the most expensive part of the structure. 
 
 These points having been decided, the form to be given the 
 arch ring is next to be considered. While it is desirable that 
 the neutral axis of the arch ring should nearly correspond with 
 the line of pressures for a full load, there is still considerable 
 choice allowed the designer as to the actual form to be given 
 the intrados without serious changes in the amount of material 
 required. As the semicircular arch can usually be adopted 
 for very short spans only, the choice must lie between the seg- 
 mental, the elliptical, and the polycentered arch approaching 
 more or less closely the ellipse, the parabola, or the transformed 
 catenary. 
 
 The segmental arch, the parabola and the catenary do not 
 give a pleasing effect at the junction of the arch ring and the 
 abutment, and the curve is sometimes departed from near the 
 springing to make the intrados tangent to the face of the abut- 
 ment. The final choice will thus usually lie between a true 
 ellipse and the basket-handled arch. 
 
 Mr. Edwin A. Thacher, M. Am. Soc. C. E., designer of the 
 Topeka bridge, considers that "arches with solid spandrel fill- 
 ing should be flat at the center and sharper at the ends, ap- 
 proaching an ellipse; while arches with open spandrel spaces 
 should be sharp at the center and flatter at the ends approach- 
 ing a parabola, or, which is better, sharp at the ends and center 
 and flat at the haunches." 1 
 
 The form of the intrados having been fixed, the depth of key- 
 stone for an arch without reinforcement is derived, tentatively, 
 from the rules of either Rankin or Trautwine, to be corrected 
 later if necessary. The form of the extrados is then so chosen 
 as to give the required depth of arch ring to confine the line of 
 pressure within the middle third. 
 
 740. Concrete-steel Arch. The computation of a concrete- 
 steel arch is, as already stated, more involved. The graphical 
 analysis is much the simplest method of deriving the bending 
 moment, direct thrust and shear. The experience of Mr. 
 
 Engineering News, Sept. 21, 1899. 
 
476 CEMENT AND CONCRETE 
 
 Thacher has led him to endeavor to have the line of pressure lie 
 within the middle third of the arch ring, although this is not 
 absolutely necessary in reinforced concrete. The same author- 
 ity considers it good practice to design the steel work to be 
 capable of taking the entire bending moment with a unit stress 
 of about one-half the elastic limit of the steel. 
 
 The thrusts, bending moments and shears at successive sec- 
 tions of the arch ring having been determined, both for full 
 and half span loads, by the graphical methods explained in 
 Greene's " Arches" or Cain's " Elastic Arches," or by the analy- 
 sis given in Howe's " Treatise on Arches," the dimensions of the 
 arch ring and the steel reinforcement are to be derived by 
 the aid of such formulas as are given by Mr. Thacher 1 involving 
 the allowable unit stresses in steel and concrete, and their 
 respective moduli, of elasticity. 
 
 741. General Considerations. In short spans, parallel 
 spandrel walls with earth filling between, may be used, but 
 for long spans the spandrels are usually open, that is, built 
 of vertical piers or walls running parallel to the axis of the 
 soffit, and arched over at the top to support the pavement or 
 ballast. This treatment has the following advantages: only 
 vertical forces are transmitted to the arch ring; decreased 
 loads on arch and abutments; increased waterway in case of 
 unusual floods; and better architectural effects. 
 
 The beauty of the structure is an important consideration, 
 inasmuch as the decision in favor of a concrete arch as against 
 a steel bridge is usually affected quite as much by considera- 
 tions of aesthetic effect as of cheapness and durability. In 
 this connection it may be said that in concrete-steel construc- 
 tion there may be little difference in the thickness of arch ring 
 required at the crown and near the springing, but the appear- 
 ance of the structure will usually be improved by accentuating 
 a little, if necessary, the increased thickness at the springing, 
 except in the case of the semicircular arch in which the eye is 
 accustomed to a nearly uniform thickness of the voussoirs. 
 The appearance is also frequently improved by molding the 
 concrete at the crown to represent a keystone, projecting a 
 little beyond the face of the rest of the arch ring. 
 
 1 Engineering News, Sept. 21, 1899. 
 
ARCHES 477 
 
 The beauty of a concrete arch may easily be marred by 
 faulty design, and some very ugly, as well as some very beau- 
 tiful, arches have been erected. 
 
 742. Stone Facing. The practice which has been followed 
 to some extent of facing the spandrel walls with cut stone 
 masonry, is considered questionable. The cost of ashlar facing 
 is likely to be so great as to discourage the use of headers of 
 sufficient length to give a good bond with the concrete, and it 
 is next to impossible to make this equal to monolithic concrete 
 construction. Again, since concrete is frequently used to pro- 
 tect ashlar masonry that has started to disintegrate, it is rather 
 a reversal of what has been found good practice to face con- 
 crete with a thin layer of cut stone. No criticism is intended 
 of the method of building a pier of large dimension stone with 
 concrete hearting, as this is a different matter. But a thin 
 parapet or spandrel wall faced with a mere shell of cut stone, 
 however beautiful it may be when built, is likely to take on a 
 somewhat dilapidated appearance after ten years' service, espe- 
 cially if it is called upon to pass through one or two floods of 
 unusual violence. 
 
 743. Quality of Concrete. As already intimated, the cost 
 of a concrete arch, especially where reinforcement is used, is, 
 under ordinary circumstances, considerably less than a masonry 
 arch of equal appearance and strength. The only exceptions 
 to this rule are where the facilities for obtaining stone suitable 
 for masonry are exceptional, and where the work is far re- 
 moved from cement-producing regions and from the coast. 
 The ability to employ common labor for much of the construc- 
 tion work in a concrete arch is an advantage only partially 
 offset by the necessity of having somewhat more careful work 
 done upon the arch centers and more careful supervision of 
 construction. 
 
 The concrete of the arch ring should be of the best quality, 
 especially if steel reinforcement is not used. For this purpose, 
 the stone, broken to a size not exceeding two inches in any 
 dimension, should be mixed with a quantity of mortar a little 
 more than sufficient to fill the voids, and composed of one part 
 Portland cement to two parts sand. Interiors of piers and 
 abutments may be made of a poorer mixture, such as one 
 Portland cement to three of sand and six of broken stone, or 
 
478 CEMENT AND CONCRETE 
 
 even in some cases where abutments are massive, one to four 
 to eight concrete may be employed. 
 
 744. Centers. Substantial centers must be provided for 
 concrete arches, and the lagging should be sized, dressed on the 
 upper side, and laid with radial joints parallel to the arch axis. 
 Two inch plank sized to one and three-quarters inches is usually 
 employed for lagging, and the supporting ribs should be from 
 three to four feet centers. For spans up to forty feet a braced 
 wooden rib with one center support and two end supports is 
 used, but for longer spans a trussed center with supports ten to 
 eighteen feet apart is employed. The centers should be made 
 rigid and the camber need be very slight, say from T^TT to 
 ^^ of the radius at the crown. Not less than twenty-eight 
 days should be allowed to elapse after building the arch before 
 striking the centers. 
 
 745. Construction. A method that has been largely em- 
 ployed in building the arch ring is to divide the arch into lon- 
 gitudinal rings by planes at right angles to the arch axis. It 
 is believed to be better practice, however, to build the arch as 
 a series of voussoir courses beginning with the spring course, 
 but not necessarily proceeding in order from the springing to 
 the crown. The advantages of this method of building the 
 arch, in transverse courses parallel to the axis of the intrados, 
 are that the planes of weakness may be made at right angles 
 to the line of pressure; the unequal loading, and consequent 
 settlement of the centers, has less tendency to crack the sec- 
 tions or to separate one section from another. In cases of 
 failure of concrete arches under excessive floods, the tendency 
 of the arch to separate along a longitudinal joint forming a 
 plane of weakness has been clearly shown. 
 
 746. The tendency of the center to rise at the crown as the 
 arch ring is built up on the haunches is sometimes overcome by 
 temporarily loading the crown. In constructing the ring in 
 voussoir courses, the order of the work may be so arranged as 
 to distribute the loading on the centers in any manner desired. 
 Such an expedient was adopted in the construction of the 
 Illinois Central R. R. arch across Big Muddy River, where the 
 arch ring was divided into nineteen voussoirs. The two spring- 
 ers were built first, then the fifth row of voussoirs towards the 
 crown on each side, followed by the ninth row, the third and 
 
ARCHES 479 
 
 seventh. The intermediate blocks were then built in order 
 toward the crown, the second, fourth, sixth and eighth, and 
 finally the keystone. In this way the weight was well dis- 
 tributed on the centers, and the load on the two sides of the 
 crown was kept symmetrical. The monolithic blocks forming 
 the voussoirs that were built in molds had recesses on either 
 side, which were made by securing planks to the interior of 
 the mold. When the intermediate blocks were built, the con- 
 crete thus keyed into the blocks first made. 
 
 The division of the work into voussoir courses will usually 
 admit of such size molds or blocks that two, one on either side 
 of the center, may be completed in a day. If it becomes ne- 
 cessary to interrupt the laying of a block, however, a vertical 
 bulkhead should be constructed in the mold, with key or dowel 
 pins if desired, to assist in making a bond when the block is 
 completed. 
 
 747. Finish and Drainage. To provide a smooth face, a 
 thin facing mortar of one part Portland to two parts sand is 
 desirable, laid at the time of building the concrete in accord- 
 ance with methods already described. A thicker layer of 
 granolithic may be used on the soffit and will somewhat more 
 effectually prevent the broken stone of the concrete settling on 
 the lagging, which is always likely to occur to the detriment 
 of the appearance of the finished work. 
 
 The division between adjacent voussoirs should be clearly 
 marked on the face, and additional joints may be indicated if 
 desired, by lines in a plane approximately perpendicular to the 
 line of pressure. Such lines are obtained by securing triangular 
 strips on the inner face of the molds. When spandrel walls are 
 used, these may be similarly marked on the face by horizontal 
 and vertical joints. On long spans the spandrels should have 
 expansion joints, and the coping and parapet, when of concrete, 
 should also have vertical joints to provide for changes in length 
 due to loading or thermal variations. 
 
 The arches over the spandrels should be provided with a 
 waterproof covering, either of Portland cement grout or an 
 asphalt mixture to prevent the percolation of water to the arch 
 ring. Pipe drains should be provided to carry the water to a 
 point over the piers where it may be discharged. Care should 
 be taken that such pipes have their outlets so located that the 
 
480 CEMENT AND CONCRETE 
 
 drip shall not disfigure the wall. Open spandrels may be drained 
 by pipes built into the arch ring at suitable places. 
 
 748. Highway Arch without Reinforcement. A good ex- 
 ample of a highway bridge built of concrete without reinforce- 
 ment is the monolithic arch spanning San Lea'idro creek, be- 
 tween Oakland and San Leandro, Cal. 1 This arch has a five 
 centered, elliptical intrados, with span of 81| feet, rise of 26 
 feet and width of about 60 feet. At the crown the thickness 
 of the arch ring is 3 feet, the radius of the intrados 61 i feet and 
 of the extrados 88 feet. 
 
 As the arch rests directly on a bed of clay containing some 
 gravel, the footings are made 30 feet wide, and they extend 
 5 feet below the creek bed. The lagging for the forms was 
 of 2 by 6 inch scantling laid transverse to the axis of the 
 structure or parallel to the axis of the intrados. The ribs of 
 the centering were built of two 1 by 12 inch boards and the 
 braces of 4 by 6 inch timbers, converged to three short 12 by 
 12 inch timbers supported by wedges bearing on 12 by 12 
 inch longitudinals. 
 
 749. The concrete was composed of one barrel Portland 
 cement, two barrels sand to seven barrels of broken stone of 
 varying sizes. 
 
 When the haunches had been built up about one-third the 
 way, as flooding of the work was anticipated, an arch ring one 
 foot thick was first completed, the remainder being placed as 
 a second layer. There is a parapet wall three feet six inches 
 high on either side of the bridge. The spandrel walls show a 
 solid face, and are paneled to bring out the outlines of the 
 extrados and parapet. The centers were struck ten days after 
 the completion of the second arch ring, and the settlement at 
 the crown was about one and one-half inches. The forms con- 
 tained 90,000 feet, B. M., of lumber, and 3,384 cubic yards of 
 concrete were used. The cost of the bridge was $25,840.00, 
 or less than $8.00 per cubic yard of concrete. The contractors 
 were the E. B. and A. L. Stone Co. of Oakland, Cal., and the 
 plans were prepared by the County Surveyor's office of Alameda 
 County, Cal. 
 
 750. A Three Span Arch. The three span arch spanning a 
 
 1 Described by Mr. William B. Barber, Engineering News, Aug. 27, 1903. 
 
ARCHES 481 
 
 mill pond on Anthony Kill, near Mechanicsville,^ N. Y., 1 is 
 worthy of notice on account of some peculiarities in the center- 
 ing and because of the location of the plant on a side hill, so 
 that the concrete was delivered on the work with very little 
 labor. Two of the arches were of 100 ft. span, with rise of 
 20 feet, and the remaining arch was of 50 ft. span. The width 
 is but 17 feet, and the piers are founded on rock at a depth not 
 exceeding 12 feet. 
 
 For the centering, piles were first driven, six feet centers, 
 in bents ten feet apart, and the bents capped with ten by twelve 
 inch timbers. Stringers of the same dimensions were then laid 
 longitudinally, and eight by ten inch posts were erected on the 
 longitudinals and spaced three feet center*. These posts, 
 which were cut to proper length, so that their tops conformed to 
 the curve of the intrados, were then capped with eight by ten 
 inch timbers parallel to the axis of the intrados, and the lagging 
 laid upon them transverse to this axis or parallel to the center 
 line of the bridge. This lagging was of two thicknesses of one 
 inch boards sprung into place and nailed, the upper layer being 
 of dressed lumber to give a smooth surface to receive the con- 
 crete. 
 
 The concrete was of one part Portland cement, three parts 
 sand, three parts gravel and three parts broken stone, except 
 for the arch ring, in which but two and one-half parts each of 
 gravel and stone were used. The concrete plant was so ar- 
 ranged that the stone could be passed from the crusher to the 
 mixer by gravity. The concrete was delivered on the arch in 
 cars of three feet gage drawn by cable. From fifty to sixty 
 cubic yards of concrete were placed in ten hours with but nine 
 laborers. 140,000 feet B. M. lumber was used in centers. 
 The entire work consumed about 2,500 cubic yards of concrete. 
 
 751. Railroad Arch without Reinforcement. The concrete 
 bridge carrying the Illinois Central R. R. over the Big Muddy 
 River furnishes an excellent example of a long span arch, built 
 without reinforcement so far as the arch ring is concerned. 
 The bridge is very fully described by Mr. H. W. Parkhurst, 
 Engineer of Bridges and Buildings I. C. R. R., in Engineering 
 News of Nov. 12, 1903. There are three spans, each 140 feet 
 
 1 Described in Engineering News, Nov. 5, 1903. 
 
482 CEMENT AND CONCRETE 
 
 in the clear, with 30 feet rise above springing lines. The arch 
 ring proper is five feet thick at the crown, but as the spandrels, 
 which are built open over the haunches, have near the crown 
 only a false opening on the face, the actual thickness of concrete 
 at the crown is seven feet. 
 
 The piers and abutments already in place for the three 
 Pratt trusses formerly in use, were surrounded with new con- 
 crete masonry, making the piers 21 ft. 6 in. wide at the top. 
 As rock was found only at considerable depth, the piers rested 
 on piles. To relieve the load on foundations as much as possible, 
 as well as to avoid cracking, which would be likely to occur in 
 heavy longitudinal spandrel walls from temperature strains, 
 transverse spandrel arches were adopted. Since in case of de- 
 railment of trains these spandrel arches would be subjected to 
 shock, the concrete in this portion of the structure was rein- 
 forced by a self-supporting skeleton structure built of steel 
 rails. Longitudinal rails were laid horizontally, three feet 
 center to center, connected at frequent intervals by one inch 
 rods and held in place by vertical posts, which in turn rested 
 upon transverse horizontal rails laid in recesses left in the arch 
 rib. 
 
 752. Expansion joints were provided in the spandrel arches 
 at the ends, two at each pier and one at each abutment, to al- 
 low some movement due to changes in temperature. The ex- 
 pansion joints were made by placing in the joint several thick- 
 nesses of corrugated asbestos board protected by a J-inch lead 
 plate folded into the joint, forming a trough at the top. The 
 lead plate lies flat on top of the concrete for five inches from 
 the joint, and about two inches at each end of the plate is bent 
 down at right angles and set into the concrete. An asphaltic 
 composition is then laid over the lead plate, entirely covering 
 it and filling the trough. 
 
 The centering was erected on pile bents spaced about 14 
 feet centers, the calculated pressure on each pile being about 
 eighteen to twenty tons. For the center span, five 60 foot 
 deck plate girders resting on pile piers were used over the deep- 
 est portion of the channel to provide for possible floods bring- 
 ing large amounts of drift. 
 
 753. The arch ring was laid in voussoir courses as described 
 in 746. Face joints were made by securing triangular shaped 
 
ARCHES 483 
 
 pieces to inner face of the molds in lines approximately at 
 right angles to the line of pressure. All exposed work was 
 faced with a layer of about 1^ inches of Portland cement mortar 
 placed and rammed with the concrete. The surfaces were not 
 given, in general, any further finish, no attempt being made to 
 remove or conceal the usual marks left by the mold boards. 
 
 Portland cement was used throughout, the quality of the 
 concrete being varied by the amount of cement used to given 
 quantities of the aggregates. In the centers of large masses 
 the poorer mixtures were employed, while the richer concretes 
 were used in those places subjected to the most trying conditions. 
 
 In making the concrete the principle followed seems to have 
 been to keep the mixer as near the work as practicable, moving 
 the mixer and carrying materials to it, rather than to transport 
 the mixed concrete from a certain fixed location of the mixing 
 plant. Much of the concrete was handled in barrows, but 
 derricks were also used in portions of the work. As traffic on 
 the old bridge had to be maintained during the erection of the 
 new structure, considerable extra handling of concrete was 
 necessary and additional work was involved in ramming the con- 
 crete in places difficult of access. The concrete was mixed rather 
 wet, so that but little tamping was required to make it quake. 
 
 754. Cost. The total amount of concrete was over 12,000 
 cubic yards, which was placed at an average cost of $5.43 per 
 cubic yard. In cofferdams and centers 400,000 feet B. M. of 
 timber was used, and about 300,000 pounds of steel was em- 
 ployed in the skeleton structure of the spandrels. This steel 
 cost 1.2 cents per pound, the punching, fitting and erecting 
 costing but about 0.61 cent per pound. The total cost of the 
 bridge is estimated to have been $125,000.00, or about the 
 same as the estimated cost of a steel structure designed for 
 the same duty. 
 
 755. The Melan Arch Bridge at Topeka, Kan., is one of the 
 most important concrete-steel structures yet erected in the 
 United States. It consists of one span of 125 feet, two of 
 110 feet each, and two of 97.5 feet each. The foundations for 
 piers and abutments are piles in soft sand. The steel rein- 
 forcement is in the form of a latticed member. The bridge is 
 fully described in Engineering News of April 2, 1896, and En- 
 gineering Record, April 16, 1898, 
 
484 CEMENT AND CONCRETE 
 
 756. Concrete-Steel Viaduct. A viaduct of ten concrete- 
 steel arches, of about 65 foot span, carries a double track elec- 
 tric line across West Canada Creek near Herkimer, N. Y. 1 The 
 piers rest on piles driven into hard blue clay, the surface of 
 which is 6 to 12 feet below the creek bed. The segmental 
 arches have a rise of 12 to 14 feet, with thickness of 21 inches 
 at the crown and 4^ feet at the springing; the radius of intrados 
 is about 46 feet, and of extrados about 57 feet. The stresses 
 were computed for full load and for live load on half span, 
 Prof. Cain's graphical method being employed. The maximum 
 stresses allowed were six hundred pounds per square inch com- 
 pression in concrete and ten thousand pounds per square inch 
 tension in steel. The stresses caused by a variation of fifty 
 degrees in temperature were allowed for. The tensile strength 
 of the concrete was disregarded. Thacher bars, 1J inches dia- 
 meter, were used for the reinforcement, being placed eleven 
 inch centers near both intrados and extrados. 
 
 757. Expansion joints were provided in spandrel walls by 
 nailing to the sides of the forms for arch pilasters a narrow 
 strip of timber, thus forming a groove into which the spandrel 
 wall is tongued. These joints show some motion and allow 
 some water to leak through. 
 
 The concrete was mixed three parts sand and seven parts 
 gravel to one volume packed cement for foundations and piers, 
 and two and one-half parts sand and five of gravel to one ce- 
 ment for the arch rings and spandrel walls. All concrete was 
 mixed wet and by hand. The work was faced with mortar 
 composed of one part cement to two and one-half parts sand, 
 and after the removal of forms the face was brushed with thin 
 mortar wash and rubbed with sandstone blocks, giving a uni- 
 form color to the surface. 
 
 ART. 84. DAMS 
 
 758. Concrete vs. Rubble. Concrete has been employed to 
 some extent in most of the important masonry dams of recent 
 construction, and has formed the main portion of some of the 
 largest dams yet built. 
 
 The relative value of concrete and uncoursed rubble masonry 
 
 Engineering News, Feb. 27, 1904. 
 
DAMS 485 
 
 laid in Portland cement mortar is perhaps still an open ques- 
 tion, though it is believed that the former will eventually be 
 preferred by engineers who are familiar with both. Concrete 
 will require in general a larger proportion of cement than does 
 the masonry, so that in localities difficult of access, the ma- 
 sonry may for this reason be cheaper. Usually, however, con- 
 crete will be the cheaper, and less skilled labor will be required 
 in the building. With the same amount of inspection, concrete 
 of good materials properly proportioned will form at least as 
 impervious a wall as will rubble. 
 
 759. Quality of Concrete. The up-stream face of the dam 
 should be made as nearly water-tight as possible, and therefore 
 a rich concrete employed in which the mortar is in excess of 
 the voids in the stone, and the mortar itself contains about 
 two parts sand to one cement. The body of the wall, however, 
 may be made of a poorer mixture, one to three to six usually 
 being sufficient. Bowlders may also be imbedded in the mass 
 to cheapen the concrete without any serious detriment. Such 
 bowlders should, of course, be sound and clean, and well wet 
 before being placed. They should be kept well back from the 
 face of the wall and should be separated one from another by 
 at least six inches, to allow of thoroughly tamping the concrete 
 between them. 
 
 760. Building in Sections. In a wall of rubble the con- 
 traction and expansion are taken care of by minute cracks 
 between the stone and mortar which frequently are not notice- 
 able. In a concrete wall, unless provision is made for this, 
 these signs of movement may be concentrated in cracks at in- 
 tervals of thirty to sixty feet; these are always unsightly, and 
 may in exceptional cases be a serious defect. The remedy 
 evidently lies in so building the dam that if these cracks appear, 
 they shall be confined to predetermined planes where they will 
 not do any serious harm. Such contraction cracks will be 
 very much less likely to occur in a dam arched in plan than 
 in a straight dam, since in the former a slight movement of the 
 masonry up or down stream changes the length of the wall 
 and relieves the tension strains. 
 
 761. Joints. The joints in a concrete dam should not be 
 unbroken planes for any great distance. That is, the concrete 
 should be so placed that the joints between work of different 
 
48G CEMENT AND CONCRETE 
 
 days are not planes extending through the wall. The wall 
 may well be kept higher on the down-stream side and step down 
 toward the up-stream side. The vertical joints should also be 
 broken by right-angled off-sets, but the wisdom of using a dove- 
 tail joint in such work is very questionable. The joining of 
 one day's work to another necessarily forms a plane of weak- 
 ness, and therefore the work should be carefully planned to 
 the end that these planes shall be, in direction and location, 
 where they will not unnecessarily weaken the structure or render 
 it pervious to water. 
 
 762. Examples: St. Croix Dam. A dam at St. Croix, Wis., 1 
 was built of sandstone masonry of uncoursed rubble in one-to- 
 three mortar, and faced with concrete of one Portland cement 
 to three parts sand to four parts broken stone of 1 to 3^ inch 
 size. The concrete was rammed in place between the stone- 
 work and the concrete forms. The selection of the uncoursed 
 rubble was probably made on account of the site being five 
 miles from the railway and the consequent difficulty of getting 
 cement. The dam was arched in plan, and in preparing the 
 foundation, several grooves or trenches were cut in the rock in 
 a longitudinal direction, to avoid, as usual, a through course at 
 the bottom, and these trenches were also filled with concrete. 
 Had the concrete for the facing- contained five parts of broken 
 stone having maximum size of 2 or 2J inches, it would have 
 been more nearly in conformity with the best practice. 
 
 763. Massena Dam. In the construction of the dam at the 
 forebay of the Massena Water Power Company, Massena, N.Y., 2 
 it was sought to take up the tension stresses due to contrac- 
 tion by imbedding in a longitudinal direction in the concrete, 
 T-rails two feet apart horizontally and four feet apart ver- 
 tically. 
 
 764. Butte Dam. The dam built in connection with the 
 Butte, Montana, water system is 120 feet high, 350 feet long, 
 10 feet wide at the top and 83 feet wide at the 120 foot point. 
 The bed rock was granite, which was first covered with four 
 inches of concrete made with small sized stone. In the body 
 of the dam, granite bowlders were thickly imbedded in the 
 
 1 Engineering News, June 13, 1901. 
 
 2 Engineering News, Feb. 21, 19.01. 
 
DAMS 487 
 
 concrete, care being taken that each bowlder was entirely en- 
 veloped in concrete and that there were no horizontal or nearly 
 horizontal courses either of concrete or bowlders. 
 
 765. San Mateo Dam. The San Mateo Dam of California, 
 one of the highest dams in existence, is built entirely of concrete, 
 170 feet high. It is 126 feet thick at the base and is arched up- 
 stream with a radius of 637 feet. The dam was constructed in 
 blocks of 200 to 300 cubic yards each, of irregular heights, so as 
 to bond the courses together and have no through joints. Con- 
 crete, one, two to six, was delivered in small push cars on a 
 high trestle over the dam, and was dropped through iron pipes 
 16 inches in diameter to the place of deposition. In some 
 cases this drop was 120 feet, and it is stated that the concrete 
 appeared not to be injured by this method of handling. 
 
 766. Barossa Dam. The Barossa Dam in South Aus- 
 tralia 1 is of a bold arch design. The arch has a radius of 200 
 feet, and the chord is 370 feet subtending an angle of 135 degrees, 
 20 minutes, and the length of the arc 472 feet. The height of 
 the dam is 94 feet above the ground line, yet the greatest thick- 
 ness above the foundation is only 34 feet, with a top width of 
 only 4 feet. 
 
 Special care was taken in selecting the materials and fixing 
 the proportions. The cement was aerated fourteen days before 
 use. Test cubes of concrete two feet on a side were prepared 
 with different proportions of materials and subjected to a 
 hydrostatic pressure of two hundred feet before deciding upon 
 the proportions to use in the concrete. As a result of these 
 tests, the aggregate was made up of one part screenings J to 
 J inch, two parts "nuts" J to 1J inch, and four and one-half 
 parts "metal" 1J to 2 inch. This mixture contained about 35 
 per cent, voids. The mortar was made of one part Portland 
 cement to one and one-half parts sand, and was from seven and 
 one-half to fifteen per cent, in excess of voids in aggregate. 
 Plumbs were used in the clam to within fifteen feet of the top, 
 and above this level iron tram rails were placed in string courses. 
 The success accompanying the use of concrete in structures of 
 this magnitude is sufficient evidence of its value and adapta- 
 bility. 
 
 1 Mr. A. B. Moncrieff, Engineer in Chief, Engineering News, April 7, 1904. 
 
488 CEMENT AND CONCRETE 
 
 ART. 85. LOCKS 
 
 767. The use of concrete in the construction of canal locks 
 is comparatively recent, but it has met with much favor, and 
 its use is extending. The requirements for a lock wall are that 
 it shall be reasonably water-tight, that its strength shall be 
 sufficient to withstand the thrust of the gates and support the 
 earth filling behind it (or in a river wall, the difference in water 
 pressure on the two sides), and that it shall withstand the 
 impact and abrading action of boats using the canal. In all 
 of these respects concrete is believed to be the equal of a good 
 class of stone masonry. At St. Marys Falls Canal, portions 
 of the lock walls which have been injured by boats and re- 
 paired with concrete have given entire satisfaction, although 
 in such cases the concrete had to be patched on, and some- 
 times in places difficult of access for work of this character. 
 
 768. Methods of Building. The present accepted method 
 of concrete lock construction is to build the walls in alternate 
 sections, filling in the intermediate sections after the others 
 have set. It is sometimes thought necessary to make the 
 work on a section continuous from time of starting the con- 
 creting to its completion. That the exterior appearance of the 
 work may be somewhat better if such a course is followed, is 
 true, but it is very questionable whether the attainment of this 
 desirable result is worth the additional expense and the addi- 
 tional liability of having poor work done under the cover of 
 darkness when work at night is necessitated by such a rule. 
 With proper precautions, such as making steps in the top 
 surface of work left for the night, as already detailed elsewhere, 
 and being careful that the limit of work on exposed faces is 
 bounded by true horizontal and vertical lines, the plane of 
 weakness occasioned by a horizontal joint extending only par- 
 tially through the work cannot be a serious defect in a con- 
 crete wall. 
 
 769. The molds, so far as the walls alone are concerned, are 
 comparatively simple and have already been described under 
 the head of forms (Art. 62). Cable passages, gate recesses, 
 hollow quoins, culverts, etc., call for special carpentry work, 
 sometimes of quite intricate character. While the efficiency of 
 the machinery and the lock as a whole should not be sacrificed 
 
LOCKS 489 
 
 to obtain easy construction, yet sharp corners should always 
 be avoided, and simplicity of outline should be the constant 
 aim. Linings of hollow quoins (when steel quoins are consid- 
 ered necessary), gate anchorages, cable sheaves and other parts 
 built into the masonry, are in general placed with greater 
 difficulty in concrete forms than in stone masonry. Aside from 
 such special constructions, the walls may be built up much 
 more rapidly of concrete than of stonework. 
 
 As to the proportions to be used in concrete for locks there 
 is no rule of thumb. As a guide, the stresses in each part of 
 the structure should be determined as well as the knowledge of 
 the forces will permit, but the proportions will depend on the 
 question of water-tightness and freedom from deterioration quite 
 as much as upon required strength. It may be said, however, 
 that in a considerable portion of the cross-section of the walls, 
 weight is the main consideration and the concrete need not be 
 very rich. The concrete surrounding the culvert, however, 
 should be of good quality, as the stresses which may be devel- 
 oped here do not admit of close analysis. 
 
 770. The walls should be faced with mortar made of one 
 and one-half or two parts sand, or, better, two parts of granite 
 screenings one-half inch and smaller, to one part of the same 
 kind of cement used in the body of the concrete. This facing 
 need not be more than three inches thick, and if made of sand 
 and cement, it will probably be better if not more than one inch 
 thick, though this may depend on the materials and local con- 
 ditions. In any case this facing should be laid with the con- 
 crete by means of a removable steel plate similar to that de- 
 scribed in 528. The top of the wall should be finished with 
 mortar or granolithic similar to a concrete walk or driveway. 
 While the walls should in general have a vertical face, a slight 
 batter is allowable at the top, starting at about upper pool level, 
 to protect the concrete from being chipped by the impact of 
 boats, and for a similar purpose the outer corner of the wall 
 should be rounded with six to twelve inch radius. 
 
 Special care must be taken in lining the culverts, particularly 
 in silt-bearing streams, and in such places as a change is made 
 in the direction of the flowing water. For high heads it may 
 be necessary to line the culverts with cast iron for a portion of 
 their length. Granite and hard burned bricks have also been 
 
490 CEMENT AND CONCRETE 
 
 used for this purpose, but in locks of moderate lift, granolithic 
 lining will usually be found sufficiently resistant. 
 
 All necessary irons and bolts should be built into the masonry 
 as the work progresses, as they will be much more secure than 
 if set later in recesses left for them. 
 
 771. Cascades Lock. The large lock in the canal at the 
 Cascades of the Columbia was one of the first in the United 
 States to be designed of concrete in this country. In this lock 
 the walls, wells, copings and portions of culverts were faced 
 with stone. The foundation rock was covered with eight inches 
 of rich concrete, one part Portland cement, two parts sand to 
 four parts gravel. Fourteen feet of the chamber walls and 
 ten feet of gate abutments or wide walls were of concrete, one 
 to three to six, while balance of masonry was of one to four to 
 eight concrete. 
 
 The molds were of four by six posts four feet apart, and 
 lagging of two-inch lumber, dressed to size for exposed faces. 
 The work was carried up in horizontal layers, not more than 
 two feet being placed in one day. The set concrete was picked 
 and washed when fresh concrete was to be laid upon it so as to 
 get as good a bond as possible. The inlet pipes to the turbines 
 to operate the machinery were built in the lock walls, and as it 
 was not desirable to place an iron pipe in this location, the pipe 
 was molded of concrete and afterwards laid in the wall. The 
 pipe was thirty-nine inches diameter, walls six inches thick and 
 contained about 0.22 cubic yard of concrete per foot. It was 
 made in three foot lengths in vertical molds, and the cost of 
 about six hundred feet of it was at the rate of $3.56 per foot, 
 or $16.19 per cubic yard. 
 
 772. Hennepin Canal. In the locks for the Illinois and 
 Mississippi Canal the walls are entirely of concrete, and were 
 built in alternate sections about thirty feet long. Work on a 
 given section once commenced was continued to completion 
 without intermission. The top was finished without any plas- 
 ter or wet coat, the excess concrete being simply cut off with 
 a straight edge and rubbed smooth and hard with a float. 
 Vertical wells one foot square were left in the walls at intervals, 
 and these were kept filled with water for about three weeks 
 after the completion of the section, and then filled with concrete. 
 To avoid weak places due to single batches made from cement 
 
LOCKS 491 
 
 of poor quality which might have passed inspection, the ce- 
 ment was mixed in lots of five to ten barrels before being used 
 in the concrete. 
 
 The quoins of these locks were of cast iron. The founda- 
 tions and the spaces in rear of lock walls are cut off from upper 
 pool by cross-walls, and are underdrained to the lower pool to 
 prevent the action of water pressure due to the upper pool 
 level tending to force up the foundation. Ten inch and twelve 
 inch tile drains were used for this purpose. 
 
 The proportions used in general were one part Portland 
 coment, three to three and one-third parts gravel, and four 
 parts broken stone, the concrete containing about one and four- 
 tenths barrels of cement per yard. The average cost of con- 
 crete in quantities of two thousand to four thousand yards was 
 from $8.50 to $9.15 per cubic yard, distributed approximately 
 as follows: 
 
 Materials $5.00 to $6.00 
 
 Molds 82 to 1.42 
 
 Mixing and placing 1 .64 to 1 .82 
 
 Miscellaneous .12to .47 
 
 773. Herr Island. In the Herr Island Locks, Alleghany 
 River, the failure of the cofferdam to exclude water from the 
 lock pit on account of porosity of the river bed, led to the adop- 
 tion of a concrete foundation, laid "under water, of sufficient 
 weight to balance the hydrostatic pressure. After this founda- 
 tion was in place, the cofferdam was pumped out and the con- 
 crete side walls built in the dry. 
 
 The concrete was placed in one foot courses covering the 
 entire area of the wall, the forms being made of one course of 
 two by twelve inch plank set on edge and halved at the ends 
 to form two inch lap splices. Iron rods one-quarter inch diam- 
 eter were placed six feet eight inches apart to tie face and 
 back plank together. A two by twelve inch cross-plank was 
 placed on edge beside each tie rod, dividing the work into short 
 sections. After completing the concreting to the top of the 
 forms throughout, the cross-planks were removed and the space 
 filled with concrete, thus making a vertical joint. The forms 
 for the next course were then put in place in a similar manner. 
 The size of stone used as aggregate was first two inches in one 
 dimension, but this size was afterward reduced to one and 
 
492 CEMENT AND CONCRETE 
 
 one-half inches, and finally to one inch, the smaller size stone 
 being preferred. 
 
 774. Mississippi River. The lock in the Mississippi River 
 between Minneapolis and St. Paul was founded on a soft sand- 
 stone rock having many water-bearing seams. The lock was 
 surrounded on three sides by a cut-off wall. A trench two 
 inches wide and ten feet deep was cut in the soft rock by jet- 
 ting a series of holes in close juxtaposition and then breaking 
 out the intervening wall with a drill and saw of special con- 
 struction. In this trench was first laid a double thickness of 
 three-quarter inch boards and the remaining space was grouted 
 full. Sections of this wall afterward uncovered, showed the 
 method to have been very effective. Similar methods of seal- 
 ing open seams in rock by the use of grout under pressures 
 have been used elsewhere. 
 
 The forms for the construction of this lock were of excellent 
 design 1 and have been described under the head of "forms" 
 ( 514). The walls were built in alternate blocks, twelve feet 
 long. At the ends of the blocks are left vertical spaces five by 
 seven inches, to be filled with mortar and other water-tight 
 composition. The forms are lined with sheet iron, and to 
 obtain a smooth face the concrete is thrown against the lining, 
 the stones rebound, leaving only mortar on the face. The 
 face is rammed with tampers of special form, wedge shaped, 
 and measuring } inch by 5 inches on the lower edge. This is 
 followed by a flat rammer. The finish is said to be excellent. 
 
 775. Sand-cement was used quite largely in the lock con- 
 struction. It was prepared at the site of the work, of equal 
 parts Portland cement and siliceous sand ground together in a 
 tube mill. 
 
 Proportions in the concrete were varied somewhat from time 
 to time, though in general it was mixed one part silica cement, 
 two and one-third parts sand and six and two-thirds parts of 
 crushed stone without screening. Tests showed that about ten 
 per cent, of this crusher product was fine enough to be consid- 
 ered sand, and account of this fact was taken in fixing the pro- 
 portions as above. The cost of the concrete, over 11,000 yards, 
 was as follows : 
 
 1 Mr. A. O. Powell, Asst. Engr., Report Chief of Engrs., 1900, p. 2778. 
 
BREAKWATERS 493 
 
 Cement $2.76 
 
 Stone $1.29 
 
 Breaking stone for crusher .38 
 
 Crushing stone .82 
 
 Total stone $2.49 
 
 Sand , .52 
 
 Total materials . . . 
 
 Forms 
 
 Mixing and placing concrete 
 
 Total cost per cubic yard concrete . $8.42 
 
 ART. 86. BREAKWATERS 
 
 776. The use of concrete in the construction of breakwaters 
 in the United States was suggested as early as 1845. In recent 
 years it has been employed quite extensively, especially for 
 harbor improvements on the Great Lakes, where it has with- 
 stood the rigorous winters, the severe storms, the attrition of 
 ice, and the impact of boats, in a highly satisfactory manner. 
 Its use has been confined largely to the construction of a super- 
 structure on timber cribs, the concrete work being in the form 
 of blocks set with derricks, or of monolithic blocks molded in 
 place, or more frequently composed of a combination of these 
 two forms. 
 
 Since in breakwater construction weight is of prime impor- 
 tance, it is not necessary, in general, to use an exceptionally 
 strong concrete, as the increased expense had better be in- 
 curred in increasing the cross-section. 
 
 777. Buffalo Breakwater. In the construction of the ex- 
 tensive breakwaters at Buffalo, 1 concrete has been used in large 
 quantities and according to various plans. In 1887 the super- 
 structure of some 750 feet of timber-crib breakwater was re- 
 newed, mainly with natural cement concrete. 250 feet of this 
 superstructure was built with a facing of Portland cement 
 concrete, while 500 feet of it was faced with stone masonry. 
 The concrete started two feet below mean lake level. The 
 cross-section of the superstructure was about 350 square feet, 
 and the cost of concrete, exclusive of materials, was about $2.36 
 per cubic yard. 
 
 1 Described by Mr. Emile Low, U. S. Asst. Engr. Trans. Am. Soc. C. E., 
 December, 1903. 
 
494 CEMENT AND CONCRETE 
 
 During the following year concrete footing blocks were used 
 on both the lake and harbor faces, since it was found that the 
 cement was washed out of the concrete laid in place below 
 water. The blocks contained about 3J cubic yards and cost on 
 the average a little more than $30.00 each, or $37.35 each in- 
 cluding the setting, or at the rate of $11.29 per cubic yard. 
 The molds or forms, which were used repeatedly, cost about 
 $40.00 each. 
 
 778. Another style of concrete superstructure developed at 
 Buffalo is that recommended by Major F. W. Symons. It 
 consists of three longitudinal walls, connected at intervals by 
 cross-walls, filled between with rubble stone and provided with 
 heavy parapet and banquette decks. The longitudinal wall on 
 the lake side is founded on heavy concrete blocks 5 feet high, 
 8 feet thick at the base and 7.2 feet long ; the two minor walls 
 are formed by smaller blocks, 4 feet by 4.5 feet by 12 feet. 
 The total width at base is 36 feet. The space between lake face 
 blocks and center row is 14 feet, and between center row and 
 harbor face blocks is about 5 feet. The cross-wall blocks are 
 7 by 6 by 4 feet under the parapet, and 4 by 3 by 4 feet under 
 the banquette, all spaced 36 feet centers. All concrete blocks 
 have their bases set two feet below mean lake level and have 
 panels in their upper surfaces to provide a bond with the 
 concrete laid in place. 
 
 The lake wall above the concrete block is 8 to 4 feet thick, 
 with batter on face, and the decks are 3 to 4 feet thick, built of 
 concrete in place. The forms for the harbor face wall and cross- 
 walls were of J inch matched pine, with vertical posts two to 
 three feet centers tied through the wall with one-half inch tie 
 rods. 
 
 The concrete was composed of the following volumes: one 
 part Portland cement, one part screened gravel (about f inch), 
 two parts sand grit (nearly half of which was J inch to \ inch 
 gravel), and four parts unscreened broken limestone (about 11 
 per cent. dust). The cost of the concrete in blocks was $10.00 
 per cubic yard, and that in place cost $9.40 per cubic yard. 
 
 779. Cleveland Breakwater. Several forms of concrete su- 
 perstructure have been employed in the work at the Cleve- 
 land breakwater. One section on a thirty-two foot crib has 
 three rows of concrete blocks, one each on lake and harbor 
 
BREAKWATERS , 495 
 
 sides and one in center of the crib, extending three feet below 
 mean lake level. The concrete in place is started at mean lake 
 level and is composed of a base five feet thick, with vertical 
 faces over the entire crib, and surmounted on the lake side by 
 a parapet five feet high and about twelve feet wide. The stone 
 filling of the cribs was covered with a cheap decking of wood 
 before laying the concrete in place. 
 
 780. Marquette Breakwater. In the construction of the 
 superstructure of the breakwater at Marquette, Mich., the con- 
 ditions were peculiar in that it was desirable to provide a pas- 
 sageway within the superstructure through which the lighthouse 
 on the outer end might be reached in stormy weather. This 
 was accomplished by leaving near the harbor face a conduit, 
 6 feet 3 inches high and 2 feet 10 inches wide, the entire length 
 of the structure. 
 
 The old timber structure having been removed to about one 
 foot below mean lake level, a foundation course two feet thick 
 of Portland cement concrete was laid on a burlap carpet placed 
 over the stone filling of the crib. Upon this the monolithic 
 blocks were built in place, substantial molds being set up for 
 alternate blocks ten feet apart. After these had set, the molds 
 were removed and other molds set up to form the two faces of 
 the intervening blocks, the ends of the blocks already com- 
 pleted taking the place of end molds. The monolithic blocks 
 were of natural cement concrete in proportions of 489 pounds 
 of cement to one-half cubic yard of sand and one cubic yard of 
 broken stone. About twenty per cent, of these monoliths was 
 composed of rubble stone ranging in size from one-half to three 
 cubic feet, care being taken that no rubble should be placed 
 nearer than one foot to any outside surface. The standard 
 block was twenty-three feet wide on the base, which was one 
 foot above mean lake level. The lower five feet of the face had 
 a 45 slope. There was then a nearly level berm, 7.5 feet wide, 
 forming the banquette deck; from the back of this deck the 
 face sloped at an angle of 45 to the parapet deck, which was 
 6 ft. 4 inches wide. The harbor side of the block was vertical, 
 9.4 feet high. Since the structure proved very stable and free 
 from vibrations in heavy seas, the horizontal dimensions of the 
 block were reduced as the shore was approached. 
 
 781. The method of placing the Portland cement concrete 
 
496 CEMENT AND CONCRETE 
 
 foundation was modified as described under the head of the 
 block and bag systems of concrete constructions (Art. 64). 
 
 The cost of the monolithic blocks of natural cement concrete 
 was as follows : 
 
 490 Ibs. cement, $1.04 per bbl $1.815 
 
 .5 cu. yd. sand, $0.50 per cu. yd .25 
 
 1.0 cu. yd. stone, $1.58 " " " 1.58 
 
 Materials in one cubic yard concrete $3.645 
 
 80 per cent, concrete in the finished block, .80 of 
 
 $3.645 $2.91 
 
 Loading materials .33 
 
 Mixing concrete .52 
 
 Depositing concrete .41 
 
 Handling rubble .09 
 
 Finishing blocks .09 
 
 Moving and setting forms .25 
 
 Timber waling, anchor boJts, etc .13 
 
 Total cost in place per cu. yd $4.73 
 
 Very interesting and detailed accounts of the construction 
 of this breakwater, which was carried out with special care as 
 to all details, were made by Mr. Clarence Colenlan, Asst. Engr., 
 and may be found in the reports of Major Clinton B. Sears, 
 Reports Chief of Engineers, U. S. A., 1896 and 1897. 
 
INDEX 
 
 Abrasion 
 
 Resistance to, 329. 
 Tests of, 94. 
 Abutments, 467. 
 
 Accelerated Tests (see Soundness), 77. 
 Acceptance of Cement, 153. 
 Accuracy Obtainable in Tests, 137. 
 Acid- 
 
 Sulphuric, in Cement, 34. 
 Use on Concrete Surface, 368. 
 Adhesion 
 
 Cement to Brick, 273, 278. 
 Glass, 273. 
 Iron, 273. 
 Steel Rods, 284. 
 Stone, 272. 
 Terra Cotta, 273. 
 Effect of Character Surface, 276. 
 Plaster Paris, 277. 
 Regaging, 276. 
 Richness Mortar, 274. 
 Neat and Sand Mortars, 279. 
 Portland to Natural, 270. 
 Results of Tests,. 270. 
 Tests of Cement, 92. 
 Adulteration, 4, 43. 
 Age and Aeration of Cement 
 Effect on Time Setting, 68. 
 
 Specific Gravity, 42. 
 Strength, 235. 
 Aggregate 
 
 Bowlder Stone, 322. 
 Brick, 186, 324, 335. 
 Cinders, 302, 309, 338. 
 Clean, 188. 
 Cost, 195. 
 Crushing, 194. 
 Fireproof Concrete, 335. 
 
 Aggregate 
 
 Granite, 322. 
 
 Gravel as, 192, 298, 303, 309. 
 
 Material for, 186. 
 
 Sand in, 202. 
 
 Sandstone, 294, 322. 
 
 Sea Water, 350. 
 
 Size and Shape of Fragments, 188. 
 
 Tests of, 298, 322. 
 
 Trap, 298, 309. 
 
 Voids in, 190. 
 
 Weight of, 189. 
 Air Hardened Mortars, 122, 232. 
 
 260. 
 
 Alum and Soap Washes, 344. 
 Alumina in Cement, 33. 
 Aluminous Natural Cement, 25. 
 Amount of Mortar in Concrete, 200. 
 
 Effect on Compressive Strength, 
 293. 
 
 Effect on Transverse Strength, 
 
 318. 
 Analysis 
 
 Methods, 35. 
 
 Materials, 11. 
 
 Natural Cement, 8. 
 
 Portland Cement, 6. 
 Anchor Bolts, 284, 471. 
 Arch- 
 
 Big Muddy River, 481. 
 
 Highway, 480. 
 
 Mechanicsville, 480. 
 
 Melan, 384, 483. 
 
 Monier, 382. 
 
 Plain Concrete, 480, 481. 
 
 San Leandro, 480. 
 
 Thacher, 385. 
 
 Three Span, 480. 
 
 497 
 
498 
 
 INDEX 
 
 Arch 
 
 Topeka, Kan., 483. 
 
 Wiinsch, 383. 
 Arches 
 
 Centers, 478, 482. 
 
 Construction, 478. 
 
 Cost, 483. 
 
 Design, 474. 
 
 Drainage, 479. 
 
 Finish, 479. 
 
 Viaduct, 484. 
 
 Bag for Depositing Concrete, 369, 
 
 373, 377. 
 
 Bag Method, 374. 
 Bags of Concrete to Form Face, 377. 
 
 - to Prevent Scour, 377. 
 Baker, Classification of 
 
 Hydraulic Products, 3. 
 Ball Mills for Grinding, 20. 
 Barrels, Cement 
 
 Capacity, 172. 
 
 Records, 146. 
 Basement Floors, 426. 
 Base of Concrete Walk, 421. 
 Beams 
 
 Concrete-Steel, 390. 
 
 for Street Railway Tracks, 433. 
 
 Steel, Protected, 412. 
 
 Strength, Experiments, 313, 403. 
 Formulas for, 391, 393. 
 Tables of, 400, 402. 
 Belt Conveyor for Concrete, 358. 
 Blast Furnace Slag 
 
 Cement, 22, 23. 
 
 Sand, 159. 
 
 Block System, 351, 378. 
 Blocks, Concrete, in Breakwaters, 
 
 379, 493. 
 
 Blowing of Cement (see Soundness). 
 Board, Mixing, for Concrete, 204. 
 Bohme Hammer Apparatus, 114. 
 Boiling Test, 77. 
 
 Bolts, Adhesion of Mortar to, 284. 
 Boston Elev. R.R. Tests Concrete, 
 
 292, 308. 
 Boston Subway, 444, 446. 
 
 Bowlder Stone as Aggregate, 322. 
 Box Mixer (see Cubical). 
 Braces for Forms, 354. 
 Breaking Briquets, 123. 
 Breaking Stone by Hand, 194. 
 Breakwater, 493. 
 
 Buffalo, 214, 493. 
 Cleveland, 494. 
 Concrete in, 379, 493. 
 Marquette, 375, 379, 495. 
 Brick - 
 
 Adhesion of Cement to, 272. 
 as Concrete Aggregate, 186, 324, 
 
 335. 
 
 Dust with Cement, 258. 
 Bridge 
 
 Abutments, 467. 
 Piers, 464. 
 
 Forms for, 464. 
 Bridges (see Arches). 
 Briquets 
 
 Area Breaking Section, 109. 
 Breaking, 123. 
 Form of, 108. 
 Machine for Making, 1 14. 
 Methods of Making, 113. 
 Records, 147. 
 Storing, 117. 
 Broken Stone (see Aggregate). 
 
 vs. Gravel, 192. 
 
 Brushing Concrete Surface, 361, 366. 
 Buffalo Breakwater, 214, 493. 
 Buffalo, Concrete Mixing at, 214. 
 Buhr Millstones, 20. 
 Building Regulations, New York, 418. 
 Buildings of Concrete, 410. 
 Burlap Bags for Placing Concrete, 
 
 369, 377. 
 Burning 
 
 Natural Cement, 26. 
 Portland Cement, 16. 
 Bushhammering Concrete, 367. 
 
 Caisson Filling, 466. 
 Calcium Chloride 
 
 Effect on Setting, 70. 
 
 Test for Soundness, 81. 
 
INDEX 
 
 499 
 
 Calcium Sulphate 
 
 Effect on Strength, 249. 
 
 Time Setting, 69. 
 Canal Locks 
 
 Concrete for, 224, 357, 488. 
 
 Forms for, 357. 
 
 Capacity Cement Barrels, 172. 
 Carbonic Acid, 34. 
 Cars 
 
 Concrete Plant on, 216. 
 
 for Transporting Concrete, 213. 
 Cascades Canal, Concrete for, 224. 
 Centers (see also Forms) 
 
 for Arches, 478, 480. 
 
 for Tunnel Lining, 451. 
 Chamber Kilns, 17. 
 Chemical Tests, 31. 
 Chicago Drainage Canal, Concrete on, 
 
 224. 
 Cinder Concrete 
 
 Strength, 302. 
 
 Modulus Elasticity, 309. 
 Cinders, Sulphur in, 338. 
 Classification Hydraulic Products, 1. 
 Clay- 
 
 for Cement Manufacture, 10. 
 
 in Concrete, 305. 
 
 in Mortar, 253. 
 Clip for Breaking Briquets, 124. 
 
 Cock, 128. 
 
 Form Suggested, 133. 
 
 Gimbal, 130. 
 
 Requirements for Perfect, 132. 
 
 Russell, 129. 
 
 Tests of, 131. 
 Clip Breaks, 126. 
 
 Cause, 126. 
 
 Prevention, 127. 
 
 Strength, 127. 
 
 Coarse Cement and Fine Sand Com- 
 pared, 57, 62. 
 Coarse Particles (see Fineness) 
 
 Effect of, 52. 
 
 on Time Setting, 69. 
 Cock Clip, 128. 
 
 Cockburn Concrete Mixer, 211. 
 Coefficient Expansion, 332. 
 
 Cohesion and Adhesion Compared, 
 
 275, 279. 
 
 Cold, Effect on Cement, 260. 
 Color for Concrete Finish, 367. 
 of Cement, 36. 
 
 of Concrete Surface, 365, 367. 
 Columns, 412, 415. 
 
 Concrete-Steel, 413. 
 
 Steel, Filled and Covered, 413. 
 
 Strength of, 413. 
 Comparative Tests 
 
 Natural Cements, 138. 
 
 Portland Cements, 138. 
 Compression Tests, 89. 
 Compressive Strength 
 
 Concrete, 291. 
 
 Mortar, 288. 
 Compressive and Tensile Strength 
 
 Compared, 288, 313. 
 Compressive and Transverse Strength 
 
 Compared, 313. 
 Composition, Chemical, 6, 8. 
 
 Effect on Specific Gravity, 42. 
 Concrete 
 
 Amount of Mortar in, 200. 
 
 Compressive Strength of, 291. 
 
 Construction, Rules for, 467. 
 
 Cost, 218. 
 
 Definition, 186, 200. 
 
 Deposition in Water, 326, 369. 
 
 Making, 200. 
 
 Mixers, 207, 212. 
 
 Mixing, Cost, 212. 
 
 Mixing by Hand, 203. 
 
 Mixing Plants, 212. 
 
 Proportions in, 200. 
 
 Thorough Mixing, 203. 
 Concrete-Steel, 381. 
 Conductivity of Concrete, 333. 
 Considered Experiments, 388. 
 Consistency Concrete, Effect on 
 
 Strength, 293, 296, 319. 
 Consistency Mortar,, 176. 
 
 Determination, 97. 
 
 Effect on Adhesion, 274. 
 
 Tensile Strength, 99, 
 232, 314. 
 
500 
 
 INDEX 
 
 Consistency Mortar 
 
 Effect on Time Setting, 71. 
 
 Transverse and Compressive 
 
 Strength, 314. 
 
 Effect in Low Temperatures, 268. 
 Constancy of Volume (see Soundness) . 
 Contraction Concrete in Setting, 331. 
 Coosa River Concrete Plant, 212. 
 Coping for Retaining Wall, 467. 
 Corners of Concrete Forms, 354. 
 Corrosion, Action of, 336. 
 Cost 
 
 Aggregate, 195. 
 Concrete, 218. 
 Arch, 483. 
 
 Curb and Gutter, 432. 
 Floor, 427. 
 Mixing, 212. 
 Tunnel Lining, 452. 
 Walk, 425, 426. 
 Mortar, 182. 
 Sand, 171. 
 Sand Washing, 170. 
 Cracks in Concrete, 361. 
 Crushing Strength (see Compression). 
 Cubes, Concrete, Tests of, 292. 
 Cubical Concrete Mixer, 208, 212. 
 Curb and Gutter, 431. 
 Cut Stone Facing, 477. 
 Finish, 366. 
 Cylinder, Steel, Bridge Pier, 465. 
 
 Dams, 484. 
 
 Barossa, 487. 
 
 Butte, 486. 
 
 Concrete vs. Rubble, 484. 
 
 Massena, 486. 
 
 San Mateo, 487. 
 
 St. Croix, 486. 
 Definitions, 1. 
 Delivery of Cement, 144. 
 Density, Apparent, 37. 
 Deposition Concrete in Running 
 
 Water, 326, 369. 
 Deterioration of Cement, 235. 
 Deval, Test for Soundness, 78, 81 . 
 
 Diary, Use of, 153. 
 
 Dietsch Kiln, ] 7. 
 
 Drake Concrete Mixer, 211, 216. 
 
 Dromedary Concrete Mixer, 209. 
 
 Efflorescence, 346. 
 Estimates, Cost Concrete, 218. 
 
 Mortar, 182. 
 
 Excessive Reinforcement, 396. 
 Expanded Metal, 387. 
 Expansion 
 
 Coefficient of, 332. 
 
 Concrete in Water, 331. 
 
 Joints, 482, 484. 
 Experiments 
 
 Columns, 413. 
 
 Concrete-Steel, 388, 397, 403. 
 
 Considered, 388. 
 
 Hooped Concrete, 414. 
 
 Face of Concrete (see also Finish) 
 
 Bushhammer, 367. 
 
 Colors for, 365. 
 
 Cut Stone, 366, 477. 
 
 Efflorescence, 346. 
 
 Lock Walls, 489. 
 
 Mortar, 363. 
 
 Pointed or Tooled, 367. 
 Face Pressed in Compressive Tests, 
 
 292. 
 Faija, Mortar Mixer, 107. 
 
 Tests for Soundness, 78. 
 Failure of Concrete in Sea Water, 348. 
 Farrel's Wall Molds, 417. 
 Filtration through Concrete, 340, 342. 
 Fineness Cement 
 
 Effect on Specific Gravity, 52, 59. 
 Strength, 54, 60. 
 Time Setting, 52, 60,69. 
 Weight, 59. 
 
 Importance, 45. 
 
 Specifications, 51. 
 
 Tests, 45. 
 Fineness of Sand, 97. 
 
 in Freezing Weather, 268. 
 Finish of Concrete Surface, 363. 
 
 Colors, 367. 
 
INDEX 
 
 501 
 
 Finish of Concrete Surface 
 Mortar, 363. 
 Pebble-dash, 366. 
 Plaster Paris, 365. 
 Rubbed, 365. 
 Shovel, 363. 
 Tooled or Pointed, 367. 
 Fire, Resistance Concrete to, 332. 
 Fireproof Buildings, 332. 
 Fireproof Concrete, Aggregate for, 335. 
 Flexure, Concrete-Steel Beams, 390. 
 Tests Concrete, 314. 
 
 Mortar, 90, 313. 
 Floor, Systems of Concrete-Steel, 381 , 
 
 411. 
 Floors 
 
 Basement, 426. 
 Buildings, 411. 
 Reservoirs, 453. 
 Fonns, Concrete, 351. 
 
 for Buildings, 416, 417. 
 Bridge Piers, 464. 
 Columns, 416. 
 Lock Walls, 488, 492. 
 Piles, 473. 
 
 Reservoir Roofs, 455. 
 Subways, 445, 448. 
 Tunnel Lining, 447, 449, 451. 
 Oiling, 354. 
 Time Left in Place, 352, 439, 471, 
 
 478. 
 Formulas for Concrete-Steel Beams, 
 
 391, 393. 
 Foundation 
 
 Concrete Walks, 420. 
 Pavements, 428. 
 Piles, 471. 
 
 Free Lime in Cement, 31, 76, 83. 
 Freezing Weather 
 
 Use of Cement Mortar in, 260. 
 Use of Concrete in, 326. 
 
 Gage of Wire for Sieves, 46, 47. 
 Gaging Mortar 
 
 by Hand, 105. 
 
 Effect of Thorough, 236. 
 
 with Hoe and Box, 106. 
 
 Gaging Concrete (see Mixing). 
 
 German Normal Sand, 96. 
 
 Gilmore Wires for Time Setting, 66. 
 
 Gimbal Clip, 130. 
 
 Glass, Adhesion of Cement to, 274. 
 
 Granite as Aggregate, 322. 
 
 Granolithic, Facing, 365. 
 
 Top Dressing, 422. 
 Granulometric Composition 
 
 Aggregate, 189. 
 
 Sand, 163. 
 Gravel as Aggregate, 186, 192, 298, 
 
 303, 309. 
 
 vs. Broken Stone, 192. 
 Gravity Concrete Mixer, 212. 
 Griffin Mill, 21. 
 
 Grinding Cement (see Fineness), 20. 
 Grout, to Seal Cracks, 492. 
 
 on Surface Concrete, 363, 365. 
 Gutters and Curbs, 431. 
 Gypsum (see Plaster Paris). 
 
 Hammer, Bohme, 114. 
 Heat, Effect on Concrete, 332. 
 Heating Materials in Cold Weather, 
 
 267, 452. 
 
 Hennebique System, 385, 409. 
 History, Hydraulic Products, 1. 
 Hoe and Box for Mortar Mixing, 106. 
 Hoffman Kiln, 17. 
 Hooped Concrete, 414. 
 Hot Materials in Cold Weather, 267. 
 Hot Tests (see Soundness). 
 House Walls, 417. 
 Hydraulic Limes, 2. 
 
 Immersion of Briquets, 119. 
 
 Impervious Concrete, 340, 343. 
 
 " Improved " Cement, Strength of, 
 
 244. 
 
 Impurities in Sand, 168. 
 Ingredients 
 
 in Cubic Yard Concrete, 218. 
 
 Mortar, 179. 
 Portland Cement, 5. 
 Interpretation Tensile Tests, 137. 
 
502 
 
 INDEX 
 
 Iron 
 
 Adhesion Cement to, 274, 284. 
 
 Corrosion in Concrete, 336. 
 Iron Oxide, 33. 
 
 Jig for Mortar Mixing, 107. 
 Johnson Bar, 387. 
 Joints 
 
 Expansion, 482, 484. 
 in Concrete, 361. 
 Blocks, 378. 
 Dam, 485. 
 Molds, 354. 
 Walks, 423, 424. 
 
 Kahn System, 386, 409. 
 Kilns, Cement, 16. 
 Output, 19. 
 
 Lagging for Forms, 352. 
 
 Tongue and Groove, 352. 
 Laitance, 370. 
 Lamp Black, in Concrete, 365, 367. 
 
 Surface Finish, 368. 
 Laying Fresh Concrete on Set Con- 
 crete, 361. 
 Le Chatelier, Apparatus for Specific 
 
 Gravity Test, 40. 
 Test for Soundness, 81. 
 
 Time Setting, 66. 
 Lime, Classification, 3. 
 
 Hydraulic, 3. 
 Lime in Cement, 31, 245. 
 Lime Paste, Effect on Adhesion, 280. 
 Lime, Slaked, with Cement, 245, 345. 
 Limestone, Adhesion Cement to, 274, 
 
 277. 
 Limestone, Crushed as Aggregate, 
 
 297, 322, 335. 
 Limestone Dust with Cement, 160, 
 
 187, 258, 325. 
 Lining of Forms, 353. 
 
 Reservoirs, 455. 
 Loam in Sand, 168. 
 Lock- 
 
 Cascades, 490. 
 Hennepin Canal, 490. 
 
 Lock 
 
 Heir Island, 491. 
 
 Mississippi River, 492. 
 Locks, 488. 
 
 Culvert Lining, 489. 
 
 Facing, 489. 
 
 Methods Building, 488. 
 
 Molds, 488, 490. 
 
 Louisville and Portland Canal, Con- 
 crete on, 223. 
 
 Machine for Breaking Briquets, 
 123. 
 
 Concrete Mixing, 207. 
 Mortar Mixing, 107, 178. 
 Maclay, Test for Soundness, 78. 
 Magnesia in Cement, 32. 
 Magnesian Natural Cements, 24. 
 Manufacture Natural Cement, 24. 
 Portland Cement, 10. 
 Marking Briquets, 117. 
 Materials 
 
 for Cubic Yard Concrete, 218. 
 
 Mortar, 179. 
 
 Natural Cement Manufac- 
 ture, 24. 
 
 Portland Cement Manufac- 
 ture, 10. 
 Melan System, 384. 
 
 Arch, Topeka, 483. 
 Microscopical Tests, 36. 
 Mills - 
 
 Ball, 20. 
 Griffin, 21. 
 Tube, 20. 
 
 Mixing Concrete 
 by Hand, 204. 
 
 Cost, 206. 
 by Machine, 207. 
 
 Cost, 212. 
 
 Necessity of Thorough, 303, 319. 
 Mixing Mortar 
 for Tests, 105. 
 
 Use, 177. 
 
 Necessity of Thorough, 236. 
 Mixing Natural and Portland Cement, 
 243. 
 
L\'J)KX 
 
 Modulus of Elasticity 
 
 Concrete, 308. 
 
 Mortar, 306. 
 Modulus of Rupture in Flexure 
 
 Concrete Prisms, 314. 
 
 Mortar Prisms, 313. 
 Moist Closet for Briquets, 119. 
 Moistening Concrete, 362. 
 Moisture, Effect on Volume Sand, 
 
 166. 
 
 Mulder's iN-rord, 1 17. 
 Molding 
 
 Bolnne, Hammer, 114. 
 
 Hand, 115. 
 
 Jamieson Machine, 114. 
 
 Machine, 114. 
 
 Methods, 113. 
 Molds - 
 
 Briquet, Cleaning, 113. 
 Forms of. 108. 
 Kinds of, 112. 
 
 Concrete (see Forms). 
 Blocks, 378. 
 Sewers, 439, 442. 
 Walks, 422. 
 Walls, 417. 
 
 Monier Arch, Test, 382. 
 Monier System, 381. 
 Mortar 
 
 Amount in Concrete, 200. 
 
 Cost, 182. 
 
 Definition of, 155. 
 
 Facing, 363. 
 
 for Plastering Concrete, 363. 
 
 Ingredients for Cubic Yard, 179. 
 
 Mixing, 105, 177. 
 
 Varying Richness, 227. 
 
 Natural Cement 
 
 Analysis, 8. 
 
 Definitions, 8. 
 
 Manufacture, 24. 
 Natural Cement Concrete, Strength 
 
 of, 300. 
 
 Neat vs. Sand Tests, 95. 
 Needle Test for Time Setting, 66. 
 Numbering Briquets, 117. 
 
 Oiling Forms or Molds, 334. 
 
 Painting Concrete, 368. 
 Pan Mixer 
 
 for Cement, 14. 
 
 Concrete, 210. 
 
 Paper Sacks for Concrete, 377. 
 Pat Test (see Soundness). 
 Pavement, Concrete, 429. 
 Pavement Foundation, 128. 
 IVbblc-Dash Finish, 366. 
 Permeability of Mortars, .",10, 343. 
 Piers, Bridge, -Nil. 
 
 Forms for, 461. 
 Piles, Concrete, 471. 
 
 Protection by Concrete, 383. 
 Pipe, Sewer, in Concrete, 436. 
 Placing Concrete under Water, 326, 
 
 369. 
 Placing Consecutive Layers Concrete, 
 
 361 . 
 
 Plant, Portland Cement, 14. 
 Plants, Concrete, 212. 
 Plaster Paris 
 
 Effect on Adhesion, 277. 
 Strength, 249. 
 Soundness, 250, 251. 
 Time Setting, 69. 
 Plastering Concrete Surface, 363. 
 Platform, Mixing, 204. 
 Plums in Concrete, 361, 485, 495. 
 Point, Dressing Surface Concrete, 367. 
 Pointing Mortar, 347. 
 Porosity of Mortars, 340. 
 Portland and Natural Compared, 279, 
 
 282. 
 Portland Cement 
 
 Composition, 5. 
 
 Definition, 4. 
 
 Manufacture, 10. 
 Posts for Forms, 354, 356. 
 Pot Cracker for Grinding, 26. 
 Pozzolana Cement -(see Slag Cement), 
 
 7. 
 
 Pozzolana with Cement, 365. 
 Preservation of Iron and Steel, 336. 
 Proportions in Concrete 
 
 Theory of, 200. 
 
504 
 
 INDEX 
 
 Proportions in Concrete 
 
 Effect on Strength, 295, 301, 317. 
 Modulus of Elasticity, 
 
 309. 
 Proportions in Mortar, 173. 
 
 Effect on Strength, 227. 
 Puzzolana (see Pozzolana). 
 
 Qualities, Desirable, in Cement, 28. 
 
 Rails Imbedded in Concrete, 470. 
 Hammers for Concrete, 360, 492. 
 Hamming Concrete, 359. 
 
 Effect on Strength, 297. 
 Ransome Bars, 284, 386. 
 
 Concrete Mixer, 209. 
 System, 386. 
 Rate of Applying Tensile Stress, 
 
 133. 
 Ratio Compressive to Tensile 
 
 Strength, 289. 
 Records of Tests, 140. 
 Regaging Mortar, 237. 
 
 Effect on Adhesion, 276. 
 Hegrinding Cement (see Fineness). 
 Reinforced Concrete (see Concrete- 
 Steel). 
 
 Reinforcement, Double, 403. 
 Excessive, 396. 
 Longitudinal, 413. 
 Single, 390. 
 
 Repair of Stone Piers, 466. 
 Reservoirs, 453. 
 
 Examples, 456. 
 Floor, 453. 
 Lining, 455. 
 Roof, 455. 
 Walls, 454. 
 
 Results of Tests, Treatment of, 135. 
 Retaining Walls, 467. 
 Retardation of Setting of Cement, 69. 
 Richness of Concrete, Effect on 
 
 Strength, 296, 317. 
 Rods, Adhesion of Mortar to, 284. 
 
 Tie, for Forms, 356. 
 Roebling System, 386. 
 Roman Cement, Definition, 2. 
 
 Roof, Concrete, for Building, 411. 
 
 for Reservoir, 455. 
 Rosendale Cement (see Natural). 
 Rubbed Finish for Concrete, 365. 
 Rubble Concrete, 360. 
 Rubble vs. Concrete, 484. 
 Rules for Concrete Construction, 467. 
 Russell Clip, 129. 
 Rust, Prevention of, 336. 
 
 Sacks of Concrete, 374, 377. 
 Salt, Effect on Mortars, 263. 
 
 Time Setting, 70. 
 Use in Freezing Weather, 260, 
 
 326. 
 Sampling, Method, 145. 
 
 Per cent, of barrels, 144. 
 Sand 
 
 Character, 154, 157. 
 Cost, 171. 
 Damp 
 ' Mortars Hardened in, 278. 
 
 Volume of, KM). 
 Detecting Impurities in, 168. 
 Fineness, 97, 159. 
 for Tests - 
 
 Comparison of, 96. 
 Fineness, 97. 
 German Normal, 96. 
 Natural, 96. 
 
 for Use in Sea Water, 159. 
 Heating in Winter, 452. 
 Impurities in, 168. 
 in Aggregate, 202. 
 Quality, 170. 
 Shape and Hardness Grains, 155, 
 
 159, 162. 
 Slag, 159. 
 
 Varying Amounts of, 227. 
 Voids in, 162. 
 
 Measuring, 164. 
 vs. Neat Tests, 95. 
 Washing, 169. 
 Weight, 170. 
 Sand-Cement 
 
 Manufacture, 21. 
 Use in Locks, 492, 
 
INDEX 
 
 505 
 
 Sandstone 
 
 Adhesion of Cement to, 274. 
 as Aggregate, 294, 322. 
 Sawdust in Mortar, 359. 
 Screenings hi Broken Stone, 187, 
 
 325. 
 
 Screw Concrete Mixer, 211. 
 Sea Wall, Concrete in, 216. 
 Sea Water 
 
 Cements in, 348. 
 Concrete in, 318. 
 Storing Briquets in, 121. 
 Section, Breaking, of Briquets 109. 
 Setting, Process of, 65. 
 Setting, Rate or Time of, 66. 
 
 Approximate Method Determ n- 
 
 ing, 67. 
 
 Effect of Aeration, 68. 
 Age, 68. 
 
 Composition, 67. 
 Consistency, 71. 
 Fineness, 69. 
 Gaging, 73. 
 Gypsum, 69. 
 Medium, 74. 
 Plaster Paris, 69. 
 Salt and Sugar, 70, 71. 
 Temperature, 72, 73. 
 Gilmore Wires, 66. 
 in Air and Water, 74. 
 Mortar and Neat Cement, 72. 
 Requirements as to, 74. 
 Variations in, 67. 
 Vicat Needle, 66. 
 Sewers 
 
 Cost, 437, 439. 
 Forms, 439, 441. 
 
 Steel, 442. 
 
 Methods Construction, 436, 443. 
 Pipe, in Concrete, 436. 
 Shear 
 
 in Concrete-Steel Beams, 405. 
 Strength in, 328. 
 Tests of, 90. 
 Sheathing for Forms, 352. 
 
 Tongue and Groove, 352. 
 Shoefer Kiln, 17. 
 
 Short Time Tests, Interpretation, 
 
 137. 
 
 Shrinkage in Setting, 331. 
 Sidewalk, Concrete, 420. 
 Base, 421. 
 Construction, 422. 
 Cost, 425. 
 Drainage, 420, 422. 
 Foundation, 421. 
 Wearing Surface, 422. 
 Sieves for Cement, 46, 51. 
 
 Value of Coarse, 63. 
 Sifting (see also Fineness). 
 
 Mechanical and Hand, 49. 
 Time of, 50. 
 Silica, 10. 
 Silica Cement 
 
 Manufacture, 21. 
 Use in Locks, 492. 
 Skip for Placing Concrete, 372. 
 Slaked Lime with Cement, 245, 280. 
 Slag Cement 
 Definition, 7. 
 Manufacture, 23. 
 Slag Sand, 159. 
 
 Smith Concrete Mixer, 210. 216. 
 Soap and Alum Solutions, 344. 
 Soundness, 76. 
 Tests for 
 
 A. S. C. K., 76. 
 Boiling, 77. 
 Chloride Calcium, 81. 
 Deval, 79. 
 Discussion, 82. 
 Faija, 78. 
 
 German Normal, 77. 
 Hot, for Natural, 87. 
 Hot Water, 78. 
 Kiln, 77. 
 Le Chatelier, 81. 
 Records of, 151. 
 Warm Water, 78. 
 Spandrels, Arch, 476. 
 Special Test Records, 153. 
 Specific Gravity Cement, 39. 
 Effect Aeration, 42. 
 
 Coarse Particles, 52. 
 
506 
 
 INDEX 
 
 Specifications for Concrete Work, 
 
 467. 
 
 Specimens, Marking, 146. 
 Steel Beams, Concrete Covered, 412. 
 Steel Facing for Curbs, 432. 
 Forms for Sewers, 442. 
 Lining for Forms, 353. 
 Shell for Bridge Piers, 465. 
 Steel with Concrete, 387. 
 Steinbriich Mortar Mixer, 107. 
 Steps in Concrete Construction, 362. 
 Stone, Broken (see Aggregate) 
 
 vs. Gravel, 192. 
 
 Character Surface of, 276. 
 
 Crushers, 194. 
 
 Crushing, 195. 
 
 Facing for Concrete, 477. 
 
 Finish for Concrete, 367. 
 Stop Planks, 362. 
 Storage for Cement, 144. 
 Storing Briquets, 117. 
 
 before Immersion, 117. 
 
 in Air, 122, 232, 246, 260. 
 
 in Sand, 123, 278. 
 
 in Water, 119. 
 Storing Concrete Cubes, Effect of 
 
 Medium, 293. 
 
 Street Railway Foundations, 433. 
 Strength (see Tensile, Transverse, 
 etc.). 
 
 Compressive, of Concrete, 291. 
 Mortar, 288. 
 
 of Concrete-Steel, 390, 403. - 
 
 Tensile, of Mortar, 227. 
 
 Transverse, of Concrete, 313. 
 Stringers for Street Rails, 433. 
 Subways, Concrete, 443. 
 
 Boston, 444, 446. 
 
 Chicago Telephone, 444. 
 
 New York, 443, 448. 
 Sugar, Effect on Time Setting, 71. 
 Sulphuric Acid, 34, 368. 
 Summary of Tests, Record, 147. 
 Surface Concrete (see Finish). 
 Surface Stone, Effect on Adhesion, 
 
 276. 
 Sylvester's Process, 344. 
 
 Tamping Concrete, 359. 
 Temperature Cement and Water 
 
 Effect on Tensile Strength, 103. 
 
 Time Setting, 72. 
 Temperature, Low 
 
 Use of Concrete in, 326. 
 
 Mortar in, 260. 
 Tensile and Compressive Strength 
 
 Compared, 288, 313. 
 Tensile Strength 
 
 Effect Sand, 227. 
 
 Neglect of, in Concrete-Steel, 388. 
 Tensile Tests Cohesion, 95. 
 Terra Cotta, Adhesion of Cement to, 
 
 274. 
 
 Dust with Cement, 260. 
 Test Monier Arch, 382. 
 Testing Machine, Tensile, 123. 
 Testing, Uniform Methods, 30. 
 Tests (see also Tensile, Transverse, 
 etc.) - 
 
 Abrasion, 94, 329. 
 
 Adhesion, 92, 270. 
 
 Chemical, 31. 
 
 Cohesion, 95. 
 
 Compression, 89, 288. 
 
 Concrete, 291, 314. 
 
 Fineness, 45. 
 
 Sand, 96, 155. 
 
 Shear, 90. 
 
 Soundness, 76. 
 
 Specific Gravity, 39. 
 
 Tensile, 95. 
 
 Time Setting, 65. 
 
 Transverse, 90. 
 
 Weight per Cubic Foot, 37. 
 Tetmajer, Boiling Test, 77. 
 
 Kiln Test, 77. 
 Thacher System, 385. 
 Theory of Concrete-Steel Beams, 387, 
 390, 403. 
 
 of Proportions in Concrete, 200. 
 Thermal Expansion Cement, 332. 
 Tile, Pulverized, Use of, 260. 
 Time Required to Sift, 49. 
 Time Setting (see Setting, Rate of). 
 Tooling Concrete Surface, 367. 
 
INDEX 
 
 507 
 
 Top Dressing, Concrete Walks, 422, 
 
 424. 
 
 Topeka Bridge, 483. 
 Transporting Concrete, 358. 
 Transverse Strength 
 
 Comparison with Tensile, 313. 
 
 Concrete, 314. 
 
 Mortar, 313. 
 
 Tests of Cement, 90. 
 Tremie for Placing Concrete, 371. 
 Trussed Posts, 356. 
 Wales, 357. 
 Tube Mill, 20. 
 Tunnel Lining - 
 
 Brick vs. Concrete, 449. 
 
 Cost, 452. 
 
 Forms for, 447. 
 
 in Firm Earth, 444. 
 
 in Rock, 447. 
 
 in Soft Ground, 446. 
 Tunnels 
 
 Aspen, 450. 
 
 Cascades, 449. 
 
 East Boston, 446. 
 
 Perkasie, 450. 
 
 Sudbury River Aqueduct, 451. 
 Twisted Rods 
 
 Adhesion to, 284. 
 
 Ransome, 386. 
 
 Uniformity in Methods Testing, 30. 
 
 Viaduct, Concrete-Steel, 484. 
 Vicat Needle for Time Setting, 66. 
 Voids in Aggregate, 190, 201. 
 Voids in Sand, 162. 
 
 Effect Moisture, 166. 
 
 Voids in Sand 
 
 Effect Shape Grains, 162. 
 
 Size Grains, 163. 
 Volume, Proportions by, 173, 200. 
 
 Changes in, During Setting, 
 331. 
 
 Wales, Trussed, 357. 
 
 Walks of Concrete^ 420. 
 
 Wall Molds, Buildings, 417. 
 Parrel's, 417. 
 
 Warehouse for Cement, 144. 
 
 Washes for Concrete Walls, 344. 
 
 Washing Sand, 169. 
 
 Water in Mortar and Concrete (see 
 Consistency). 
 
 Water, Deposition Concrete in, 326, 
 369. 
 
 Water of Immersion for Briquets, 
 119. 
 
 Water, Stale, for Immersing, 121. 
 
 Waterproof Construction in Sub- 
 ways, 443. 
 
 Waterproof Mortar and Concrete, 
 340, 343. 
 
 Waterproof Work in Reservoirs, 453. 
 
 Wearing Surface of Walks, 422. 
 
 Wedge Rammers for Concrete, 492. 
 
 Weight of Concrete, 299, 305. 
 
 Weight per Cubic Foot Cement, 37. 
 
 Wells in Concrete, 362, 490. 
 
 Wheelbarrows for Conveying Con- 
 crete, 359. 
 
 White Finish for Concrete, 365. 
 
 Wire in Sieves, 47. 
 
 Wires for Testing Time of Setting, 66. 
 
 Wunsch System, 383. 
 
UNIVERSITY OF CALIFORNIA LIBRARY 
 BERKELEY 
 
 Return to desk from which borrowed. 
 This book is DUE on the last date stamped below. 
 
 ISFeb'SOCS 
 
 LD 21-100m-ll,'49(B7146sl6)476 
 
YC 13714