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