REESE LIBRARY OF THK UNIVERSITY OF CALIFORNIA. Deceived ,190 Accession No. 93477 ... Class No. m m CAST IRON: A RECORD OF ORIGINAL RESEARCH. BY WILLIAM J. KEEP, Member American Society of Mechanical Engineers, and of its Committee on Standard 1 ests and Methods of Testing Materials ; Member American Institute of Mining Engineers ; Member Iron and Steel Insti- tute ; Member international Association for Testing Materials ; Fellow American Association for the Advancement of Science; t Honorary Member Rensselaer Society of Engineers^ etc.. etc. FIRST EDITION. FIRST THOUSAND. NEW YORK; JOHN WILEY & SONS. LONDON : CHAPMAN & HALL, LIMITED. 1902. Copyright, 1902, BY WILLIAM J. KEEP. ROBERT DRUMMOND, PRINTER, NEW YORK. PREFACE. THIS volume has been prepared in response to many requests that the author's researches might be presented in a convenient form . In May, 1885, the author discovered that there was a relation between shrinkage and the composition of a foundry mixture, but not until the publication of Professor Turner's discovery that the condition of carbon depended upon the proportion of silicon was it discovered that shrinkage varied inversely as silicon. Since that date the author, by his method of tests, has endeavored to discover the influence of the chemical elements in cast iron, and the results were recorded in the Transactions of the American Institute of Mining Engineers prior to 1894. In 1894 it became evident that the physical qualities of cast iron were not understood. Professor Turner on page 232 of his " Metallurgy of Iron and Steel " (1895) says regarding shrinkage: < The subject has since been carefully investigated by W. J. Keep of Detroit, whose -experiments embody the whole of the trust- worthy data available. ' ' The author, as member of the Testing Committee of the American Society of Mechanical Engineers, made extensive experiments to determine the physical properties of cast iron, the results of which are recorded in the Transactions of that society. These were such a surprise that the committee was requested to prove the author's conclusions by tensile tests. Fortunately five series of Dr. Richard Moldenke's extensive tensile and compres- 93477 iv PREFACE. sive tests were completed before this volume was prepared, and have been used to meet this request. This volume contains the results of this whole line of research. Decided opinions are advanced regarding the best methods for a founder to obtain the best results ; and that he may be able to use the shrinkage test at once the whole subject is summarized on page v. WILLIAM J. KEEP. DETROIT, MICH. MECHANICAL ANALYSIS TO REGULATE A FOUNDRY MIXTURE. FOR shop routine read pages 182 and 183. Measure the shrinkage of a -in. D test-bar from your iron mixture, when you consider it satisfactory, and use it for your standard, For stove-plate and small castings it will be .115 to .140; for ordinary machinery castings .150 to .160. If the shrinkage is greater than your standard, use more soft iron (increase silicon). If it is less, use more scrap or cheap iron. The strength of a -in. Q test-bar should be above 400 Ibs. With high shrinkage and high strength of a \ -in. Q test-bar, heavy castings will be strong, but castings in. thick may be brittle. With low shrinkage and high strength of a J-in. n test-bar, large castings will be weak and thin castings will be strong. With uniform shrinkage, an increase in the strength of a J-in. n test-bar will increase the strength of all castings proportion- ately. For ordinary foundry work, and for all irons that will run gray in a J-in. test-bar, that size gives better results than any other. ACKNOWLEDGMENTS. PROFESSOR THOMAS TURNER, formerly Lecturer on Metal- lurgy at Mason College, Birmingham, and now Director of Tech- nical Instruction for Staffordshire County, England, described the action of silicon in cast iron in 1885.* The author adopted the method of Professor Turner in his work (Method 4), p. 14. In 1894, as a member of the Committee on Testing, of the American Society of Mechanical Engineers, the author made nine- teen series of tests, using three tons of ' ' Iroquois ' ' and three tons of " Hinkle " pig iron. The first twelve series were molded and cast by the Detroit Stove Works, and series 1 3 to 1 5 by the Michigan Stove Company. The other four were made by C. G. Bretting & Co., Ashland, Wis., light machinery iron ; Bement, Miles & Co., Philadelphia, heavy machinery; A. Whitney & Sons, Philadelphia, carwheel iron ; and Michigan Malleable Iron Co., white iron (air-furnace). The strength tests of these nineteen separate series were made by Professor R. C. Carpenter, Sibley College, Cornell University, and by Professor C. H. Benjamin of Case School of Applied Sciences, Cleveland, Ohio. These records are given in Chapter XIII. The shrinkages are given in Chapter VII. The complete analysis of each size of these test-bars was made by Messrs. Dickman and Mackenzie, "The Rookery," Chicago. The carbons are given in Chapter XIII, and the other records are given in the chapters on Silicon, Phosphorus, Sulphur, and Manganese. *Journ. Chem. Soc., 1885, pp. 577, 902. vii viii ACKNOWLEDGMENTS. Mr. H. S. Fleming and Mr. Edward Orton, Jr., did nearly all of the analytical work prior to 1894. Dr. C. F. Mabery, Professor of Chemistry at Case School, made the determinations of Chapter XXIII. Messrs. Cary and Moore and Mr. George H. Ellis of Chicago and several others made many determinations. Dr. R. Moldenke, as chairman of the Committee on Testing of the American Foundrymen's Association, procured thirteen series of test-bars from as many founders, and made nearly all tests himself. A portion are given of the results of the first five series in Chapter XIII. Mr. Gus. C. Henning, as secretary of the Testing Committee of A. S M. E., aided largely in pre- paring results of the work done. The dates at which the author published the results of his investigations are as follows : Description of Apparatus and Methods; Journal U. S. Asso- ciation of Charcoal Iron Workers, 1887 (the editor suggested the name "Keep's Test"); also in the Journal of South Stafford- shire (Eng.) Institute of Iron and Steel Managers, 1888. Ferro- silicon and the Economy of Its Use ; Silicon in Cast Iron ; Trans. American Institute of Mining Engineers, vol. XVII., 1888; Phos- phorus ; Aluminum in Cast Iron ; Aluminum in Wrought Iron ; Aluminum in Steel; Aluminum and Other Metals Compared; ibid., vol. XVIII., 1889. Aluminized Iron and Steel; Journal of the Iron and Steel Institute (London), vol. I., 1890. Manganese; Trans. A. I. M. E., vol. XX., 1891. Sulphur; ibid., vol. XXIII., 1893. Shrinkage (" Relative Tests, etc."); Report of Committee on Tests; Transverse Strength; Keep's Cooling Curves; Trans. American Society of Mechanical Engineers, vol. XVI., 1895. Strength; ibid., vol. xvil., 1896. Impact; ibid., vol. XIX., 1898, and vol. xxi., 1900. Plardness; ibid., vol. xxn., 1901. The author proposed the name " Mechanical Analysis " in a discussion before the Foundrymen's Association, Philadelphia, in 1894. CONTENTS. CHAPTER I. DEFINITIONS. Chemical Test, Physical Test, Direct Physical Test, Relative Tests, i. Transverse Test, Tensile Test, Compression or Crushing Test, Iron, 2. Wrought Iron, Ingot Iron, Steel, Cast Iron, 3. Lime, Fuel, 4. Mechanics, Stress, Strain, Fracture, Deflection, 5. The Spring-line, Set, Elasticity, The Measure of Elasticity or Elastic Deflection, Rigidity or Stiffness, Perfect Rigidity, No Rigidity, The Diagram of Rigidity, The Measure of Rigidity, 6. Strength, Ultimate or Maximum Breaking Strength, Dead Load, Measure of Useful- ness, Chill, Grain, 7. CHAPTER II. GRAPHIC RECORDS. Average Prices of No. i Foundry Pig Iron at Philadelphia for Forty Years, 8. Specific Gravities, 10. CHAPTER III. METHODS OF INVESTIGATION. Description of Materials used, n. Test-bars, Preparing for a Test, 12. Cupola Iron (Methods i to 3), 13. Crucible Tests (Methods 4 to 14), 14. CHAPTER IV. CRYSTALLIZATION OF CAST IRON. Form of Crystals, 19. Can Cast Iron Expand at the Instant cf Solidifica- tion? Does each Crystal Expand as it Forms? 20. Shrink-holes or CONTENTS. Spongy Spots, 21. General Arrangement of Crystals and Lines of Weakness, 22. Cooling Strains, 24. CHAPTER V. CARBON IN CAST IRON. Description of Carbon, 25. Carbon in Cast Iron, 25. Origin of Carbon in Cast Iron, 26. Quantity of Carbon in Cast Iron, 27. Saturation of Carbon, 27. Condition of Carbon in Cast Iron, 27. Diffusion of Carbon in Cast Iron, 28. Bulk of Carbon in Cast Iron, 28. Car- bon and Chill, 29. Carbon and Hardness, 29. Carbon and Fusibil- ity, 29. Fluidity, 29. Influence of Different Percentages of Carbon, 30. Arbitrary Values of Inch-pounds for Impact Tests, 32. An- nealing Ordinary Gray Iron Castings, 33. Influence of Remelting on Carbon, 33. CHAPTER VI. SILICON IN CAST IRON. Silicon, 34. Silicon in Cast Iron, 34. Silicon added to White Iron changes it to Gray Iron, 36. Silicon added to Gray Iron Low in Silicon will make it more Gray, 36. It is the Influence of Silicon, not the Percentage, that Produces the Results, 37. The Influence of Silicon is Indirect, acting through Carbon, 38. By changing Silicon in an Iron Mixture we can Control the State of Carbon and the Chill, 41. Diffusion of Silicon in Cast Iron, 41. Analysis of Silicons of A. S. M. E. Tests, 42. Loss of Silicon in Remelting, 42. Silicon and Hardness, 44. Silicon and Chill, 44. Silicon and Fluidity, 44. Silicon and the Surface of a Casting, 45. The Relation of Carbon to Silicon in Pig Iron, 45. CHAPTER VII. SHRINKAGE OF CAST IRON. One Size of Test-bar must be Used when the Influence of Variations in Chemical Composition is being Observed, 46. In the Examination of the Variation in Shrinkage caused by a Variation in Size of. Test- bar, Iron of Unvarying Composition should be put into all the Sizes of Test-bars, 46. Carbon and Shrinkage, 46. Silicon Reduces Shrinkage by Changing Combined Carbon into Graphite, 47. Shrink- ages of Test-bars of A. S. M. E. and of A. F. A. Tests, 48. Shrink- age Decreases as Silicon Increases, 48. Variation in Shrinkage indicates the Variation in Silicon, 49. Shrinkage Decreases as the Size of a Casting Increases, 49. CONTENTS. CHAPTER VIII. KEEP'S COOLING CURVES. A STUDY OF MOLECULAR CHANGES IN METALS DUE TO VARYING TEMPERATURES. Autographic Record of Shrinkage, 50. Shrinkage Curves, 51. Yokes for Chilling and Fixing the Length of Test-bars, 51. Curves from Cast Iron, 53. First, Second, and Third Expansion, 54. Solidifying of Cast Iron, 54. Hard and Soft Cast Iron, 55. Silicon is a Softener and Lessener of Shrinkage, 56. Phosphorus, Sulphur, and Man- ganese in Cast Iron, 58. Size of Casting and Expansion, 58. Effect of Hot or Dull Iron on Shrinkage, 59. Temperatures at which the Three Expansions Take Place, 61. When does Carbon Combine when Heated towards Fusion? 62. Floating of Solid Iron on the Surface of Fluid Iron while it is Melting, 63. Curves from Heated Rolled Steel, 63. Relation of these Expansions to the Critical Points of Iron and Steel, 65. CHAPTER IX. PHOSPHORUS IN CAST IRON. Phosphorus, Red and White, Sp. Gr., Phosphide of Iron, 66. Fusibility of Phosphorus Iron, Cold-shortness, 66. Tests to find Influence of Phosphorus, 67. Phosphorus and Blow-holes, 72. Additional Tests, 72. Analysis of Phosphorus, A. S. M. E. Tests, 76. Influence of Remelting on Phosphorus, 77. Influence of Phosphorus upon the Grain of Cast Iron, 77. Phosphorus and Carbon, 78. Phosphorus and Shrinkage, 79. Influence of Phosphorus on the Strength of Cast Iron, 80. Influence of Phosphorus upon the Chill of Cast Iron, 80. Phosphorus and Hardness, 80. Phosphorus and Fluidity, 81. General Remarks, 81. CHAPTER X. SULPHUR IN CAST IRON. Sulphur, 82. Sulphur in Cast Iron is considered Injurious, 82. Vari- ous Experiments, 83. Brimstone added to Gray Cast Iron, 86. In- fluence of Iron Cooling in Ladle, 88. Iron Sulphide added to Gray Iron, oo. Iron Sulphide added to White Iron, 91. Smith and Weston's Experiments, 93. Influence of Remelting on Sulphur, 96. Analysis of Sulphur, A. S. M. E. Tests, 97. Iron Pyrites in Lime- stone-flux makes Iron Hard, 98. xii CONTENTS. CHAPTER XI. MANGANESE IN CAST IRON. Ferromanganese, 99. Influence of Silicon when added to Spiegeleisen, 100. Tests with Manganese, 101. Manganese and Carbon, 102. Swedish Manganese Irons, 103. Manganese and Shrinkage, 104. Manganese and Strength, 104'. Manganese and Chill, 104. Analy- sis of Manganese, in A. S. M. E. Tests, 105. Manganese and Hard- ness, 105. Influence of Remelting on Manganese, 106. CHAPTER XII. SEGREGATION. Segregation of Swedish Pig Iron, 108. Remarkable Examples, 109. Segregations in Test-bars, 112. Ordinary Commercial Impurities a Benefit, 113. Chill as Explained by Segregation, 114. CHAPTER XIII. STRENGTH OF CAST IRON. Records of Tests for Committee, A. S. M. E., 115. Records of Tests of Committee, A. F. A., 115. Average Maximum Dead Load (Transverse) for Nominal Size of Test-bars, A. S. M. E. Tests, 116. Transverse Strength of Test-bars, A. F. A., 117. Analysis of Test- bars, A. F. A. Tests, 119. Strength generally Increases with each Increase of Silicon (up to 3%), 119. The Lower the Silicon the Weaker the Small Castings and the Stronger the Large Castings, and the Higher the Silicon (up to 3%) the Stronger the Small Cast- ings and the Weaker the Large Castings, 120. The Strength of a l /2-in. n Section of each Test-bar Decreases as the Size of the Cast- ing Increases, 120. The Strength of a ^2-in. rj Section of each Test-bar Decreases more rapidly with each Increase in Size in Small Sizes of Test-bars than in Large Sizes, 121. Decrease in Strength due to each Increase in Size of a Casting is greater and more rapid with each Increase of Silicon, 121. The Centre of a Casting has a Coarser Grain and is Weaker than the Surface of a Casting, 122. Removing the Surface of a Test-bar or Casting Diminishes its Strength per Unit of Cross-section, 122. Casting Test-bars in Horizontal Molds gives Stronger Castings and more Uniform in Size than in Vertical Molds, 123. The Variation to be provided for in Ordinary Castings is at least 50%, 124. The Variation in Strength is due to the Naturally Uneven Structure of Cast Iron, and not in any great degree to Varying Temperature of Iron entering a Mold, CONTENTS. xiii or to Varying Chemical Constitution, or to the Character of the Mold, 125. The Strength of any Size of Casting cannot be Calculated, by any Mathematical Formula, from Data obtained from Testing a Test-bar of another Size, 125. Tensile Strength of A. F. A. Test- bars, 126. Compression Tests, 128. Stronger Castings are made in Green-sand than in Dry-sand molds, 128. Square Test-bars are Stronger than Round Test-bars with Equal Areas of Cross-section, 128. Test-bars 2" X i" X 36" Cast Flat and Tested Flat, and also on Edge, 130. Methods for Producing the Strongest Castings, 131. Total Carbon, 132. Combined Carbon, 133. Combined Carbon Weakens Cast Iron, 134. Combined Carbon may Decrease as C-S.- ings are Larger, but the Strength always Decreases. This Decrease of Combined Carbon and of Strength is Caused by the Slow Cooling, and the Decrease of Combined Carbon has Nothing to do with the Decrease in Strength, 134. Mr. A. L. Colby's Tests of Machine and Sand Cast Pig, 136. Graphitic Carbon, 138. Annealed Castings, 139. CHAPTER XIV. IMPACT. The Test-bars, 140. The Recording Apparatus, 140. Dead-load Records, 140. Dead-load Diagrams, 141. Impact Testing Machine with a Swinging Hammer, 142. Diagrams from Test-bars which are not Perfectly Elastic, 145. Impact Diagrams made with a Hammer having a Direct Drop, 149. Impact Tests, 152. Influence of Sh^ck on Cast Iron, 152. Influence of Shock on Strength, 153. Cause of Increased Strength of Tumbled Test-bars, 154. CHAPTER XV. GRAPHIC METHOD FOR CLOSELY APPROXIMATING THE PERCENTAGE OF SILICON, THE SHRINKAGE, AND STRENGTH OF ANY OTHER SIZE O7 CASTING THAN THE ONE TESTED. Shrinkage, 155. Keep's Shrinkage Chart, 156. Classes of Castings, 159. Keep's Strength Chart, 160. To Find the Strength of a Section l / 2 " square by 12 ins. long of a Large Test-bar from Fig. 82, 162. To Find the Actual Strength, of the Test-bar of the Required Cross- section, 162. CHAPTER XVI. HARDNESS OR WORKABILITY OF METALS. Professor Turner's Definition of Hardness, 164. Diagrams of Hardness, 165. Drillings for Chemical Analysis, 165. Hardness and Chemical xiv CONTENTS. Composition, 166. Hardness Tests, A. S. M. E. and A. F. A. Tests, 166. Hardness, Strength, and Chemical Analysis of Various Irons, 167. CHAPTER XVII. MECHANICAL ANALYSIS OR CHEMICAL ANALYSIS FOR REGULATING FOUNDRY IRON. Mechanical Analysis, 168. Chemical and Mechanical Analysis, 168. Ad- vantages of Mechanical Analysis, 169. Chemical Analysis, 170. Drawbacks to Chemical Analysis, 171. CHAPTER XVIII. CHEMICAL ANALYSIS WILL NOT ACCOUNT FOR ALL PHYSICAL PROPERTIES OF CAST IRON. Influence of Iron standing in a Ladle, 173. Influence of Temperature of Cupola, 175. Influence of Dry and Green Ladles, 175. The In- fluence of Wet and Dry Molds, 175. Silicon is more Effectual as found in some Irons than in others, 176. Tests and Analyses of Various Irons, 177. Cupola Irons Chemically Alike, Physical Un- like, 178. CHAPTER XIX. TEST-BARS. Test-bars y 2 " and i" n, 2" X i", 2", 3", and 4" n, 179- Advantages and Disadvantages of i" G Bars as a Test of Strength, 180. Variation in Size of Test-bars due to Poor Molding, 180. CHAPTER XX. KEEP'S TEST APPARATUS. For Mechanical Analysis. 182. Test-bar Patterns y 2 " D X 12", 182. Flask for Molding Test-bars, 182. Measuring Shrinkage and Chill, 183. Taper Scale, 183. Keep's Dead-load Testing Machine No. 10, 183. Test-bar Patterns i" n X 24". 184. Keep's Dead-load Testing Machine No. 40, 184. Advantages, 184. Keep's Hardness Testing Machine, 186. Keep's Wet Grinder for Y%" Drills, 187. Keep's Impact Testing Machine No. 2, 189. Recorder, 189. Professor Thomas Turner's Hardness-test Machine, 190. Keep's Cooling- curve Machine, 190. Test-bar Patterns with Yokes, 190. Record Paper, 191. CONTENTS. xv CHAPTER XXI. PIG IRONS AND SILICON IRONS. Ferrosilicon, Flaky Silvery, Nos. i, 2, and 3 Silvery, 192. Tennessee Coal, Iron & R. R. Co., Various Irons, Tests, and Analyses, 195. Southern Silvery, Nos. i and 2 Soft, 196. Nos. i, 2, 3, and 4 Foundry, 198. Gray Forge, Mottled and White, 201. Weakness of Silicon Irons, 201. History of the Use of Silicon Iron, 201. CHAPTER XXII. TESTING SMALL SAMPLES OF PIG IRON. Crucible Furnace, 203. A Small Cupo 1 a, 203. CHAPTER XXIII. ALUMINUM IN CAST IRON. Analysis and Tests of Aluminum Mixtures, F L M and White Test-bars. 205. Influence of Aluminum on the Grain of Cast Iron, 207. Aluminum in Steel Castings, 210. Aluminum in Wrought-iron Castings, 210. CHAPTER XXIV. INFLUENCE OF VARIOUS METALS IN CAST IRON. Nickel, Copper, Zinc, Tin, Lead, 212. Chromium in Cast Iron, 213. CAST IRON. CHAPTER I. DEFINITIONS. A Test consists in subjecting a material to conditions that disclose its true character. As applied to cast iron the chemical composition and the physical quantities are to be disclosed. Chemical Test. This determines the percentage of pure iron and of all other elements which are present. Physical Test. The subjection of cast iron to conditions which shall disclose its physical characteristics, which are; appearance of grain, shrinkage, depth of chill, hardness, strength, change of shape while under stress, set. Changes in chemical composition and peculiarities of treat- ment influence the physical character of cast iron, and the only way to determine this influence is by physical tests. Direct Physical Test. This is breaking a casting which is an exact duplicate of the one whose strength is desired. The only time when a direct test would be preferable would be when all castings made in any given foundry were exactly alike and did not vary in section in any of their parts. Relative Tests are such as are applicable to every case. For such a test any size of test-piece might be selected; but the same must be used afterwards; and having made one test-record, every other record by the same method is so much greater or less than 2 CAST IRON. the original, which is regarded as standard. There is a direct relation between such test-results and the composition of the iron, the size of casting, and the shape. It would be well to fix upon a given size of test-piece, which should be used by all, and a definite routine in producing it should be prescribed so as to prevent variations in conditions as much as possible. The only variable would then be composition. Such test-results would in regular foundry practice indicate changes due to variations in composition. The records from bars of the same dimensions and the same composition are the only ones that can be averaged or compared directly with each other. Test-records of bars 2'' X i", whether tested flat or on edge, can be compared with each other, but cannot be compared with bars 2 ins. square, even if made from the same iron, because the latter cooled more slowly and have a looser grain, and are therefore proportionately weaker. Transverse Test. As cast iron is commonly used to resist cross-breaking and because of the ease with which it can be made, this test is the most frequently used. Tensile Test. We do not wish to use cast iron for a tensile member in any case, unless in a steam-engine cylinder, therefore the tensile strength of cast iron is not of very much importance. If it were important it would be a difficult thing to determine it accurately, because it is difficult to hold the specimens in a tensile machine so that they can be broken fairly and truly. Impact Test. This strikes a blow and records the behavior of the test-piece. The stress exerted by impact cannot be expressed in pounds avoirdupois, but in inch-pounds, which is the weight of the hammer in pounds avoirdupois multiplied by the height of fall in inches. Compression or Crushing Test. When applied to cubes or short cylinders the stress required to crush cast iron is so great that this test is not often used. Iron. Pure iron is never found in nature. The purest iron that has been made had a specific gravity above 7.84, some say DEFINITIONS. 3 8.OO. It is of silvery lustre, is more tough but softer than ordinary commercial wrought iron. The purest iron of commerce is often nearly as malleable as copper, has considerable hardness and lustre, and its fracture is of a bluish-white or bluish-gray color. Its specific gravity is about 7.75. The temperature required for fusion is greater in proportion as the iron is pure. The most valuable property of commercially pure iron is its power of becoming soft and pasty before fusion, which allows of its being welded or squeezed into various shapes. Wrought Iron is the name given to iron which is manu- factured without complete fusion. It is made up of fibres inter- spersed with more or less slag, which has been partially squeezed, forged, or rolled out. It is therefore not only chemically impure, but it has impurities mechanically intermixed. Ingot Iron, being manufactured by fusion, is practically free from slag. Ingot iron is also called mild or machinery steel, but cannot take temper. After wrought iron has been melted, the slag should have separated from the metal and the product should not materially differ from ingot iron, if the slag was the only difference to begin with. Steel. Excepting the name machinery or low-carbon steel sometimes given to ingot iron, iron with a small percentage of carbon, in such a state as to take a temper, is called tool-steel. The carbon is imparted to it in various ways and the special method of manufacture gives to the steel a special name, as cement, blister, crucible, cast steel, etc. Cast Iron, as ordinarily understood, is iron which contains all the carbon that it could absorb during its reduction in the blast- furnace, and it cannot be welded or forged, nor can it take temper. It also contains other impurities which were originally in the ores or taken from the fuel used in its production. Cast iron fuses at about 2075 F. for white and 2230 F. for gray. Cast iron is not a simple metal, nor is it an alloy, but it is an aggregation of compounds combined chemically and mechanically. 4 CAST IRON. Any change in the proportion of the compounds and of the elements of which it is composed, in the conditions attending its production in the blast-furnace, in remelting, or in its solidifica- tion, changes its character so much that it becomes a material of different qualities. Cast iron is a comprehensive term covering any iron with carbon too high to be classed with steel, and in the different fur- nace-yards it is separated into more than twenty different grades, on account of the differences in the appearance of its exterior surface or of its fracture, and when sold by its chemical com- position each run of iron may be different from any other on account of unequal diffusion of the elements. Lime. Carbonate of lime is obtained by burning limestone. By using limestone in the cupola the iron is freed from slag, and the cupola is cleared from cinder when large quantities of iron are melted at one time. The surface of American pig iron is covered with sand, and the gates and old castings that are returned to the cupola often contain much sand. Lime and silica (sand) are both infusible, but together, with a little alumina, form a fusible fluid slag. About one part of caustic lime to two of silica, and about one tenth as much alumina as lime is the best proportion ; or from 20 to 40 Ibs. of limestone to each ton of iron charged in the cupola. Lime added to the cupola seems to improve the quality of the iron, perhaps by taking care of a part of the sulphur in the fuel, and it unites with the ash of the fuel and carries it out of the cupola. Fuel. Iron is melted in the cupola with anthracite coal or with coke. It is considered that i tons of coke equal I J tons of coal, but it is only the large pieces of coal that reach the melting-point that add to the heat. Small fragments are an injury. Coal breaks up by heat and the small particles are lost. Considerable coke can be recovered from the dump, while no coal is saved. It is not, therefore, the heat-units that decide the econ- omv but it is largely dependent on these physical characteristics. DEFINITIONS. 5 The value of a fuel from a chemical standpoint depends upon the percentage of carbon. The more carbon and the less ash the more heat, but about 10$ of ash is necessary to give the coke strength. The very best anthracite may have 89$ carbon and 6% ash, but the best in actual use will be nearer 88$ carbon and io# ash and the average not over 83$ carbon. Coke will rarely give better averages than 84$ to 87$ carbon and \\% to 13$ ash. The sulphur in coke is nearly all in combination with iron that was originally in the coal. If any sulphur is in a form that can be volatilized it will be likely to escape before it reaches the point where the iron melts. The most of the remainder would be likely to join the cinder. It would be difficult for any large quantity to be taken up by gray cast iron, especially after the cupola has been thoroughly heated. Although anthracite coal does not contain as much sulphur as coke, yet it is in a form (pyrites) to be more readily taken up by iron. Mechanics is that science which treats of the action of forces on bodies. Stress is the name applied to a force which acts on a body. Its intensity is usually expressed in pounds avoirdupois. Strain is the effect of stress OP a body, that is, its alteration of volume and figure. Fracture occurs when the strain is so great as to separate a solid body into parts. Fig. I is an autographic diagram from a test -bar to which the stress is applied transversely at _a ' h the centre of the test-bar supported at the ends. A pencil attached to the centre of the test-bar bears against a paper which is moved to FIG. i. the right in proportion as stress is applied. Deflection of a test-bar broken transversely is the distance in inches that the centre of the test-bar moves when the stress is CAST IRON. applied, and represents the strain measured on ac, due to the stress as measured on ab. The Spring-line is the record of stresses and strains of a per- fectly elastic test-bar. For materials like cast iron, which are not perfectly elastic in their original condition, the spring-line is slightly curved. A spring-line can be formed by removing the load and then gradually applying it a second time as hg. By drawing a line from a, parallel to hg, a spring-line of is formed. Set. When a stress is removed the centre of the test-bar will to some extent recover its original position. The amount ah that it comes short of complete recovery is the set. The distance between the lines a/and ag shows set. Elasticity is the property by virtue of which a body tends to regain its original volume or shape after it has been distorted. Elasticity is perfect or imperfect according as volume or shape is wholly or partially regained. The Measure of Elasticity or Elastic Deflection is the dis- tance in one hundredths of an inch that the centre of a test-bar moves towards regaining its original position on the base-line when the stress is removed. The distance between ab and af, Fig. I , is the measure of elasticity. Rigidity or Stiffness is the ability of a material to withstand stress and to retain its original form. Perfect Rigidity would require no change of form with any stress that could be applied. No Rigidity would cause the form of a material to change without limit on the application of stress. The Diagram of Rigidity will be a line starting from the zero-point, and will lie between the lines ab and ac, and is the same as the spring-line af. The Measure of Rigidity which I have proposed is the angle that the line or diagram of rigidity makes with ac. To apply this measure all dimensions of the diagram must be the same as those used here, i.e. 100 Ibs. on ab, and -f^ of an inch on ac\ both measure I inch. DEFINITIONS. 7 Strength is the ability of a material to resist rupture. Ultimate or Maximum Breaking Strength is the greatest stress which a piece of material will resist. It is the number of pounds avoirdupois of stress that a material resists when rupture takes place. Dead Load is stress applied so gradually as not to produce shock, and is measured in pounds avoirdupois. Each increment is added to that already applied so slowly that each molecule of the material tested shall have time to adjust itself to the stress. When the stress is so great that some of the individual molecules are separated from others, the test-bar would ultimately break if time were given even if no greater stress were applied. After tJie question of ultimate strength is decided, ability to withstand stress without taking set is tJie measure of usefulness for materials as ordinarily used. To be fit for most structural purposes, a material should be able to bear, without change of form, the stress to which it is subjected. To do this, it must be either perfectly rigid or perfectly elastic. In any construction, if any part remains out of shape after stress is removed, such dis- tortion will either throw too much stress on some other part and thus imperil the structure, or (in such a case as that of an instru- mental application, for example, of an engineer's transit) will throw all parts out of adjustment. In making a compression test the stress should be observed the instant that the material begins to take set, and in a tensile test at the instant that the material begins to elongate. Chill is the depth, in inches, of the peculiar white grain, caused by the iron running against an iron chilling surface. Grain is the granular appearance of a fracture. CHAPTER II. GRAPHIC RECORDS. THE graphic method conveys to the eye at a glance by means of a pictorial representation the general and local relationship of the different parts of a record with very little mental effort, and leaves the mind free to make comparisons or draw conclusions which it would be very difficult to do while endeavoring to remember the relative values of exact figures. As an illustration : The average prices of No. I foundry pig iron for forty years is given by both methods in Table I and Fig. 2. TABLE I. AVERAGE PRICES NO. I FOUNDRY PIG IRON AT PHILADELPHIA FOR FORTY YEARS. The average price from 1860 to 1899, was $25.95 per ton. Highest price touched Aug., 1864, $73.62 per ton. Lowest " " July, 1898, 11.25 Average price for ten years, 1860 to 1869, was $37.84 per ton. Highest, Aug., 1864, $73.62. Average 1864, $59.25 per ton. Lowest, Oct., 1861, 18.62. Average 1861, 20.25 " Average price for ten years, 1870 to 1879, was $29.60 per ton. Highest, Sept., 1872, $53.87. Average 1872, $48.88 per ton. Lowest, Nov., 1878, 16.50. Average 1878, 17.63. " Average price for ten years, 1880 to 1889, was $21.58 per ton. Highest, Feb., 1880, $41.00. Average 1880, $28.50 per ton. Lowest, May, 1889, 17.00. Average 1889, 17.75 " Average price for nine years, 1890 to 1898, was $14.30 per ton. Highest, Jan., 1890, $19.90. Average 1890, $18.40 per ton. Lowest, July, 1898, 11.25. Average 1898, n.66 " Average price for the year 1899 was $19.36 per ton. Highest, Dec., 1899, $25.00 per ton. Lowest, Jan., 1899, 12.12 " Oct., 1900, price $16.00 per ton. 8 GRAPHIC RECORDS. 9 In Fig. 2 the general tendency of price is down. The /erage decrease each decade is uniform. 1 sro.oo 60.00 50.00 40.00 30.00 20.00 10.00 I 1 1 1 1 1 1 1 1 ^ ! \% ! _S i \ -458SS*. AC Ay ft r * / \ J AVE RAGE FOR 40 YEARS *v~.^^_ ^.___.. >_^_ - 5 "~ 5 - i _________ . i \ 7 V _ ^^^^ ^^ FIG. 2. Any number of records can be placed on this same chart, as the cost of fuel and the cost of labor, a different kind of line being used for each so as to more easily distinguish each record. Relative values are often represented graphically by the length of lines, or by differences in areas; for example: the volumes of the substances of Table II are inversely as their specific gravity. TABLE II. Sp. gr. Gold. Lead. Silver. Copper. Brass. Wr'ght Iron. Tin. Cast Iron. Zinc. Alumi- num. 19 26 11.45 10.50 8.87 8.33 7.84 7.29 7.18 7-00 2.72 Fig. 3 shows the relative volume of a given weight of each metal by the area of a surface, each heavier metal is laid upon the lighter ones. Fig. 4 shows the same thing by the length of a line. In this case each volume is supposed to extend from the base-line upward. Table III gives the analysis of a No. 3 foundry pig iron, also the specific gravity and the bulk of each element. In Fig. 5 is shown graphically, by the parallelogram on the left, 100$ of pure IO CAST IRON. iron, on the right, 93.68$ of pure iron, and above this is added the bulk of each element that the pig iron contains. Though the 100.00$ IRON FIG. 3. FIG. 4. FIG. 5. right and left parallelogram each contains the same weight, yet on account of the bulk of the elements, that on the right repre- senting cast iron is much larger. TABLE III. Gr.C. Cd.C. Si. P. S. Mn. Iron. 2 2$ i OO 2.40 2 14 2.O5 8.00 7.84 2. 1^ .61 2.61 . 7O .06 .21 93.68 Bulk 7.4.C I.4Q 8.12 2. "iQ .24 .21 93.68 CHAPTER III. METHODS OF INVESTIGATION. THIS chapter will describe details of experiment, to avoid repetition of the description each time that a test is mentioned. The value of an experiment depends largely upon a detailed description of each step. Description of Materials Used. The composition of each \vas determined by analysis. F L M. Gray Pig Iron. I procured in 1885 one ton of this iron, made with charcoal at Laxa Iron Works (Lt.), Carlsdal, near Kortfors, Sweden; chemical composition: TC. 3.55, GC. 3.22, CC. 0.33, Si. 1.249, P- 0.084, S. 0.040, Mn. 0.187. The iron was in small pieces of about 4 Ibs. (see Fig. 49, Chap. XII), and was a perfectly even gray and remarkably uniform. It never runs white in a -inch square bar, and is exceedingly sensitive under any treatment. " Gay lord" White Pig Iron, I procured half a ton of white charcoal pig as near like F L M. as a white iron was likely to be. It contained TC. 2.53, GC. 0.49, CC. 2.12, Si. o. 18, P. 0.26, S. 0.03, Mn. 0.09. Iroqnois Furnace Company of Chicago sent me three tons of No. 3 " Iroquois " Mall. Bessemer coke pig iron, of clear uniform gray fracture, very strong and tough in the pig. The furnace analysis was: TC. 4.07, GC. 3.15, CC. 0.92, P. 0.23, Si. 0.88, S. 0.035, Mn. 0.50. Dickman & Mackenzie's analysis: TC. n 12 CAST IRON. 4.05, GC. 3.20, CdC. 0.87, Si. 0.98, P. 0.225, S. 0.035, Mn. 0.49. The Ashland Iron and Steel Company of Ashland, Wis., also sent me three tons of " Hinkle " charcoal pig iron Furnace analysis: TC. 3.507, GC. 2.69, CC. 0.817, P. 0.13, Si. 1.09, S. 0.015, Mn. 0.72. D. & M. analysis: TC. 3.50, GC. 2.73,* CC. 0.87, Si. 1.03, P. 0.129, S. 0.012, Mn. 0.70. As I took drillings from twenty-five pigs of both " Iroquois " and ''Hinkle" after receiving fchem, and did not let Messrs. Dickman & Mackenzie know the furnace analysis, the close- ness of results is remarkable. To find loss in remelting con- sult analysis of test-bars, Series I and 7, Tables XII, XIII, XXVI, XXXIX, XLVI, LXI, LXII, and LXIV; also see LXIIL " Pcncost " High-silicon Iron. I procured one ton with TC. 2.79, GC. 2.04, CdC. 0.75, Si. 11.00, P. 0.487, S. 0.015, Mn. 0.670; also 200 Ibs. each with Si. 4.37, 6.54, 8.08, 9.42, 10.34,* 11.34, and 12.08$.* I had also 100 Ibs. each of irons with Si. 1.97, 3-57, 3-64, 4.05, 4.66, 5.15, 5.66, 6.82,6.86, and 6.99. Also 100 Ibs. each of imported ferro-silicons : Si. 9.87, 10.45, II -99> an d 16.27. For phosphorus, sulphur, and manganese irons used in experi- ments see chapter on these materials. Test-bars. In all early experiments I cast together one bar " n X 12." long and one bar T y X i" X 12". In some cases I made six pairs and in others four to get an average. Preparing for a Test. I calculated the exact composition of each casting and the amount of each item which was to enter into it. The items for one cast were weighed, and tied in sep- arate bundles, and all these put in a single bundle, to which was tied a tag, containing the list of items, which tag was afterwards to be tied to the test-bars. * For analysis see Table XIV. METHODS OF INVESTIGATION. 13 CUPOLA IRON. (Method No. I.) Separate heats in a small cupola are very satisfactory. When this cannot be done the following are quite satisfactory. (Method No. 2.) As many ladles as there are to be mix- tures, of a size to hold about 15 Ibs. when two thirds full, are lined as usual. These ladles are marked on the outside with chalk, I, 2, 3, etc., the same as the bundles to be put into them. One flask containing a square and a flat bar, and two or three flasks containing only square bars, are numbered No. I, No. 1-2, No. 1-3, etc., all to be filled from ladle No. I. Another set of flasks for the iron from the second ladle is numbered No. 2, No. 2-2, No. 2-3, etc., and so on for as many separate compositions as are to be made. These several groups of flasks are located near the cupola. The ladles are dried and heated by having melted iron poured in them and then poured out, and immediately afterwards the contents of each bundle is placed in the ladle which has the corresponding number marked on it. This addition should be heated if possible before it is put in the ladle. Enough iron has just been caught by other men in a large ladle from the cupola to fill the small ladles with as near 15 Ibs. as can be estimated. Each small ladle is held by a molder who knows the set of flasks he is to fill, and each one pours his set of flasks, beginning with flask No. I. One ladle and set of flasks has no addition and is for comparison to show the influence of the additions placed in the other ladles. As much scrap iron is added to the first and other ladles as is necessary to make the total amount to be melted by the fluid iron in each ladle the same. This is to make the cooling influence the same in each, and to give the same temperature to the metal in each mold. The additions are sometimes cemented to the bottom of the ladle, and sometimes are powdered and made a part of the lining. In such case the ladle cannot be dried with hot iron as U CAS7 IRON. advised above, and the one used for comparison must have an ordinary lining treated the same as the others. It is necessary to weigh the whole bars, gates, etc., of each cast to calculate the percentage of each element in the casting, because the iron caught in the ladles will not weigh exactly 1 5 Ibs. as figured in making up the mixture. (Method No. 3.) When additions of pure metal of a fusible nature are to be made, make four sets of three flasks each, and pl^ce them conveniently for pouring. Twenty-eight pounds of iron, as near as possible, is caught from the cupola and the set of test-bars marked I and having no addition is poured. Then with a pair of tongs stir in the first addition and pour the second set of flasks. Stir in the next addition and pour the third set. Stir in the last addition and pour the fourth set. Th? remaining iron can be weighed to see what correction of percentages is needed on account of the iron in the ladle not weighing exactly 28 Ibs., or each cast not weighing 7 Ibs. CRUCIBLE TESTS. (Method No. 4.) After preparing the bundles of addition, which in this case contain the pig iron which is to form the bulk of the mixture, we place the first bundle in a crucible and when melted pour the set of flasks. A separate heat is required for each bundle. In this case the product is exactly as calculated unless some element volatilized or escaped in other ways. If there is danger of this, the pig iron may be melted first and the addition fed to the pot while in the fire, or the pot can ,be withdrawn and the addition made and then returned until all is fluid. A separate cast of the pig iron alone must be made for com- parison. (Method No. 5.) It is not possible to prevent some variation of carbon and silicon. A series for comparison may be made like METHODS OF INVESTIGATION, o Method No. 4, without any element addition, and add enough fine iron wire to make the carbon in each cast the same as in each cast of the original series. The influence of the carbon dilution is thus determined to compare with the former series. Subtract- ing the test-records of the latter from the former gives the influ- ence due to the variation of the element under consideration. ^Method No. 6.) To show the exact influence of the dilution of carbon alone in each heat proceed as in Method No. 5 and along with the wire add enough ferro-silicon to keep the silicon uniform. (Method No. 7.) When the additions are very volatile or very readily oxidized the following method may be pursued: Melt the required amount of iron in a crucible. Cut from an inch-and-a-half pine plank a round cover for the crucible; in the centre of this bore a two-inch hole. Take a stick two inches in diameter, fit, and split one end and wedge it in the hole of the cover. In the lower end of this stick bore an inch-and-a-quarter hole to hold the addition, and fit to it a square plug three quarters of an inch long. When the cover is on the crucible the lower end of the stick should reach to within one half inch of the bottom. Drive a heavy nail in the top of the cover near one edge. When the pig iron is melted place in the hole of the wooden stick the material to be added and drive in the square plug. Seize the nail in the cover with a long-handled tongs and while the crucible containing the fluid iron is still in the furnace place the cover on the crucible, which forces the end of the stick containing the addition to the bottom of the melted iron. The iron can enter the hole and the addition can melt and run out at the four sides of the square plug and boil up through the melted iron. A piece of fire-brick as a weight should be put on the cover, which in charring will form a seal to the top of the crucible. The objection to this is the presence of oxygen in the hole of the stick and in the wood. To find what influence this exerts, make a comparison heat with the hole empty. (Method No. 8.) We may entirely fill the crucible with the T < CAST IRON. metal to which the additions are to be made, melt it, and pour one set of test-bars for comparison. Break off and clean the gates and return them to the crucible, then return this to the furnace. Before beginning, however, calculate the weight of each set of test-bars, and how many sets can be made from the metal and the additions; calculate how much of the addition will be neces- sary to give the correct percentage to the second cast. Then when the gates are put back, calculate the addition for the next cast, and so on, until only 5 Ibs. will be left. Tie each of these additions in a bundle with a tag containing a record of percentages in the casting so that all there is to do after beginning the work is to clean gates and return them to the crucible along with the required bundle. In this way casts can be made as fast as moulds can be prepared. The percentages of the castings are sure to be very nearly as estimated. By weighing what is left, if it is more or less than calculated, a correction may be made for the per- centage in each cast. (Method No. 9.) When we have not enough material to trust the last method, especially when we have a pig iron contain- ing the highest percentage of the element under consideration, the following is satisfactory: Find a pig iron having substantially the same composition except the smallest possible percentage of the element under examination. Knowing the percentage which we wish each intermediate member of the series to contain, and knowing how much we have of the material containing the highest percentage of the element, weigh out in bundles the amount of the pig iron necessary to dilute the element to the desired percentage for each heat. Having done this, melt all the material which contains the high percentage of the element, and cast a set of test-bars. Return the gates and add the portion of the pig iron by which the percentages of the element is to be diluted for the next heat, and make all fluid and pour another set of bars. Proceed in this way until we have the series complete. A last cast will be made from the iron having the low percentage. METHODS OF INVESTIGATION. 17 (Method No. 10.) If we have a very small amount of ma- terial containing the highest percentage of the element, melt this and cast a set of bars. Test this first set of bars and return all of the bars and gates to the crucible except one half of a single bar. Dilute by additions as in Method No. 9, and make another cast, test it, and so on. (Method No. 11.) Begin with the crucible full of pig iron and cast a set of test-bars for comparison. Arrange to return the gates each time with the addition which contains the element under consideration, and a new portion of the original pig iron. This addition of pig iron each time allows each heat to be as large as is desirable, and the series of heats can be continued during the day. They can be resumed another day, but more variation will probably be found between the two heats that unite the day's work than between any of the others for the reason that we end with a very poor fire but a very hot furnace, and begin next day with opposite conditions. The weight of what is left will tell how near the weight of each heat has come to that calculated, and correction is made to get the correct percentages. (Method No. 12.) To find what the variation of carbon and silicon will be in Method No. 1 1 , run through a complete series,, returning gates and adding pig iron each time a set of test-bars is cast, but with no addition of any element. The influence due to the variation for each number of a series can be found by this last method. I call this last remelt series, as it is almost identical with ordinary foundry cupola practice. (Method No. 13.) It often occurs that the element is com- bined with pure iron only, as for example, sulphide or phosphide of iron. In such case the addition of pure iron along with the element dilutes the carbon and silicon, in addition to that due to the continued heat. To find the influence of this dilution, run through a series as in Method No. 1 2 and to each cast add of fine iron wire the same weight as the pure iron added with the element. A comparison 18 CAST IRON. of this series with the remell series shows the changes due to the dilution of carbon and silicon and the additions of pure iron. By subtracting the results of the tests of this series from those of the series when the varying element was added will give the influ- ence of the element. To prevent oxidation of the wire when put in the crucible the bundle of wire should be dipped into a strong solution of silicate of soda and dried. The wire, or any element, had better be added when the contents of the crucible is fluid, and while in the furnace. (Method No. 14.) This is exactly like Method No. 3, only 28 Ibs. of F L M. is melted in a crucible. CHAPTER IV. CRYSTALLIZATION OF CAST IRON. IRON which has been poured into a mold, in solidifying becomes a mass of crystals more or less irregular, both in shape and size, but the form towards which they tend is that of a reg- ular octahedron, an eight-sided figure, each side of which is an equilateral triangle. It is formed by two pyramids with their square bases together. FIG. 6. Aggregations of Octahedral Crystals of Cast Iron. The crystals are too small to examine closely, but we may get an idea of the form and arrangement of single crystals by an examination of aggregations of crystals, as in Fig. 6, which is actual size. Crystals begin to form against a cooling surface and interlock with each other, holding between themselves minute flakes of graphite. That the free carbon is in irregular flakes is proved from the nature of graphite, and because flakes of graphite may be picked out of a casting. 19 20 CAST IRON. In a perfect crystal of iron all the axes, that is, the lines join- ing the opposite angles, are of equal length and at right angles to each other, but very few crystals are perfect, for they are more or less tangled with, or pressed out of shape by, those next to them. Can cast iron expand at the instant of solidification ? There is no such instant. Each crystal becomes solid while other parts are fluid, and it is not until such crystals are numerous enough to form a rigid shell that the casting can shrink or expand. Does each crystal expand as it forms ? When cast iron enters a mold a thin skin of solid iron is instantly formed by the cooling action of the sides of the mold. This is proven by breaking a casting which is still fluid; the central portion will run out. (Fig. 6 was procured in this way.) A shell having once formed, the heat of the metal can never melt it again, though, at first, it has no rigidity. New crystals form on the interior of this skin very rapidly, and, as the mold would prevent any expansion out- ward, if there was any expansion of individual crystals, it would be inward, which would lessen the holding capacity of the interior. The fluid interior is contained in a rigid shell of the same metal at very nearly the same temperature as the melted portion. If a hole is broken through the upper surface of a partially solid casting, the currents of the molten metal can be seen, but no metal ever exudes. On the contrary, if the casting is of any considerable size, the fluid will sink. This proves that the fluid metal does not expand as it loses heat, and it also proves that each crystal does not expand, at least not so fast as to overcome the general shrinkage from loss of heat. One thing which led to the opinion that cast iron expanded as it solidified was that a piece of dry, cold cast iron thrown into a ladle of molten iron would float, and the inference .was that the specific gravity was greater for molten than solid cast iron. Mr. Robert Mallet found, however, by two independent methods, CRYSTALLIZATION OF CAST IRON. 21 that the specific gravity of melted iron was 6.650, while that of the same cold iron was 7.170. He thought that the probable reason for the cold iron floating was the fact that its surface could not be easily wetted by the liquid iron, also that a thin layer of steam and expanded air may have been held between the surfaces of the solid and fluid metals. On page 63 of Chapter VIII the true explanation of the phenomenon is given for the first time. Another phenomenon which led to the mistaken idea that liquid iron was more dense than crystallized iron, was that when molten iron is caught in a foundry ladle a circulatory movement immediately commences, the iron against the sides of the ladle rising to the top, flowing to the center, and then dropping to the bottom, flowing across the bottom to the sides, and so on. This led to the belief that it was the cooling of the iron at the sides which caused it to rise. The true reason for the motion is that the hot iron against the side lining drives out the moisture that remains, and also causes gas to be evolved from the material of the lining. It is this steam and the streams of gas which give the upward motion to the iron. That this is the true cause is proved by the fact that when the ladle is refilled, the metal will not move at all, or if so, it will be in the opposite direction. Fig. 10 is a casting of a mill-roll cast on end, the fluid metal being fed from the bottom. The sides cool first and the upper portion of the casting solidifies next. As the upper central portion becomes solid the crystals pull away from each other towards the sides, forming along the cen- tral axis a loose spongy casting. As these cavities form, the metal will flow in from the top as long as it is fluid, causing the upper part to sink ; but as soon as all above is solid, cavities may form, although too small to be visible. The iron having entered from the bottom, the last part to become solid is the lower part of the roll, and at this point there will not be enough metal to fill the space when cold, and the crystals pull away toward the surfaces until a large shrink-hole a is the result, 22 CAST IRON. This partly explains why a large casting does not shrink as much as a small one. In both large and small castings a rigid shell forms at once. In the small casting there are no cavities and all crystals are small and lie close together. The larger the casting the more cavities, and the larger the crystals and the more "loosely they join each other. This also explains why a large casting is not proportionally as strong as a small one. It is customary to carry the upper part or head of the casting high enough to cause a considerable ferrostatic pressure in the mold, and then with a wrought-iron rod churn the center of the casting so as to prevent the upper part from becoming solid, and to allow the fluid iron to enter all cavities throughout the interior of the casting. If a channel can be kept open to the lower part of the casting, and fresh molten metal is fed to the top, the casting will be practically solid. The extra head is afterwards removed. Robert Mallet, in his work on Ordnance, gives an admirable description of the method of crystallization of cast iron. Some thirty years ago the author added to Mr. Mallet's description his own experience, substantially as follows: The crystals assemble or group themselves in lines parallel to the direction in which the heat leaves the metal. This direc- tion is always perpendicular to the cooling surface. As heat leaves all surfaces of a casting, each surface will have its lines of assemblage of crystals perpendicular to it. Fig. 7 has four sides and four systems of crystallization at right angles to each other, which meet in lines connecting the corners of the casting. In meeting, each line of crystals tends to pull away from the other. Some run past the line into the opposite system and do not form a homogeneous casting. These lines are lines of weakness, and the casting will bear less stress without fracture at these points with more metal than at any others with less. Fig. 8 shows a flat casting in which is another line of weakness connecting the diagonals. Fig. 9 shows another arrangement of perpendicular systems 01 crystallization. The angle at a being a re-entrant angle, the CRYSTALLIZATION OF CAST IRON. 2 3 perpendicular lines continue past the angle and form a line of weakness from the .angle to the center- of the casting where it meets the lines of weakness from the opposite corner and from the parallel sides. If stress is exerted in the direction of the arrows, the casting FIG. 13. FIG. 14. FIG. 15. FIG. 16. FIG. 17. will crack from the inner corner through the outer corner, follow- ing the weak line, although there is more metal in this part of the casting. Mr. Mallet refers to a case where the lower end of the cylin- der of a hydraulic press, Fig. II, was driven out by the pressure 2 4 CAST IRON. of water, and that instead of breaking when there was least metal, it gave way nearly on the lines of weakness. He also gives a drawing, Fig. 12, of the breech of a cast-iron cannon, and another of a section of a trunnion showing weak lines. The gun will burst at one of these weak lines. Let us now look for a means of remedy for Figs. 7 and 8. Fig. 13 is a section of a round casting of cast iron, and Fig. 14 a flat casting with the edges rounded. The lines of crystallization are still perpendicular to the cool- ing surface, but there are less lines of weakness, the change being gradual. Fig. 1 5 is the same figure as Fig. 9, but all angles are rounded, and all lines of weakness are prevented. If the hydraulic cylinder of Fig. 1 1 had been round as in Fig. 1 6, the cylinder would not have given way and less metal would have been used. In the construction of all castings all angles should be avoided, and all curves should be made as large as possible. Molten cast iron should be poured as cold as it can be and fill the mold, especially in large castings, to allow of as little decrease of volume as possible. In the case of the same iron poured in a casting that has both heavy and light portions, Fig. 17, the light portion cools first and shrinks most, v/hile the heavy part cools last and shrinks least. This causes a tension all along the line ab, each particle in the thick part trying to break away from the thin part. This may not cause fracture, for the thin part will probably buckle. If the corner at ab had been made a large round, the change in cooling would have been more gradual. This explains the cracking of the arms of pulleys where they join the rim, either in the mold or after being taken out. Flanges or thickening on any part of a casting act in the same way. A runner often keeps one part of a casting hot while that around it has become solid. The cooling of the runner will afterwards often pull away and crack the casting. CHAPTER V. CARBON IN CAST IRON. CARBON is found in nature in various forms. The purest form of carbon is the diamond with a specific gravity of 3.33 to 3.55. Soot from flame, as lamp-black, ivory-black, etc., is amorphous carbon. Graphite is nearly pure carbon, sp. gr. 2.15 to 2.35. The greater part of anthracite and bituminous coal and of char- coal is carbon. Iron has great affinity for carbon, and if com- mercially pure iron is placed in contact with carbon while at a high heat, it will absorb it and become steel. With heat carbon unites with oxygen and forms carbonic acid gas. In this state the carbon is invisible, but, as it cannot be made to take a gaseous form, the carbon in carbonic acid is prob- ably still a solid hid by the oxygen. This is indicated by the fact that the carbon can be made to separate in the form of soot. The coloring power of carbon is very great. In the blackest smoke there is not more than one grain weight of solid carbon in a single cubic foot, and this one grain would color black one gallon of water. The carbon in iron is derived from the fuel with which the iron was smelted, and the carbon from wood-charcoal, from coke, from soft coal or anthracite coal, each gives a different color and quality to the iron. Carbon in Cast Iron. Carbon is the most important element in cast iron. Without it iron could not be melted readily and made into castings. The percentage of total carbon determines the melting-point of the iron. Without carbon the degree of hardness and strength needed for various uses could not be given 25 26 CAST IRON. to pure iron. Carbon is present in ordinary cast iron in larger proportion than any other constituent. The only practical method of obtaining iron from iron ore is in the blast-furnace, and by this process carbon is necessarily absorbed by the iron. Origin of Carbon. Iron ore is an oxide of iron. In the blast-furnace oxygen is removed and the remaining iron is melted. The process consists of charging into the furnace fuel and ore which descend together. At the lower end of the furnace hot air is blown in, which burns the fuel near the point where it enters, producing carbonic acid, which, ascending through the hot fuel higher up, becomes carbonic oxide by taking the extra carbon from the fuel. This carbonic oxide takes the oxygen from the ore, thus again becoming carbonic acid. After the ore is relieved of its oxygen it is iron in the form of a sponge exposing a very large surface to the action of gases. As it passes downward toward the hottest part of the furnace, through the portion where the gas is carbonic oxide and carbonic acid, the iron sponge absorbs carbon from the gas. The amount that will be absorbed is inversely as the speed at which the ore descends. The greater the heat, the more carbon will be absorbed. For an iron to be high in carbon the furnace must be hot and the iron sponge must remain in contact with the gases a sufficient length of time. The sponge does not absorb carbon by being in contact with fuel, but by contact with the gas produced by the combustion of fuel. The more surface, the greater the contact; therefore iron made from dense, hard ore in large lumps will not contain as much carbon as when the same ore is broken very fine or when a more open ore is used. After the sponge reaches a point where the heat is sufficient to melt it there will be no increase of carbon. The iron that runs out of the furnace has absorbed as much carbon as was possible under the existing conditions in the furnace. Iron is at such a high temperature in the furnace that it is often able to hold more carbon in combination than it can when CARBON IN CAST IRON. 27 the temperature is lowered. As soon, therefore, as it leaves the furnace the surface of the flowing metal is covered with floating graphite ( <4 kish "). More or less will rise on the surface of the pigs. After the pig becomes solid the carbon which separates is caught as graphite between the grains and gives the fracture its gray color. From the time the sponge iron was melted in the furnace until the pig is cold there is a constant loss of carbon. A cupola can never heat iron much above its melting-point, and it can therefore never absorb more carbon than it already has ; and as a general thing it cannot reabsorb quite all of the graphite that has lodged among its own crystals. In the cupola the iron is not in a finely divided state when compared with the sponge iron of the furnace. In the cupola the melted drops are not in contact with the fuel or carbonic acid gas to any great extent and never in contact with carbonic oxide gas, and the passage into the hearth is very rapid. Quantity of Carbon Present. This is due to the character of the ore and to other conditions, and is from about 2$ to 4.25$. If by any means the iron has absorbed in the furnace more carbon than it can hold after it becomes solid, the excess will separate the same as soot separates from a clear flame. Saturation of Carbon. By saturation is generally understood the largest percentage that it is possible for any iron to hold when solid. For charcoal-iron it may be 4$, while for coke and anthracite-irons carbon does not generally exceed 3.50^ or 3.75$. Iron made in a cold furnace or with all the ore that the fuel can possibly take care of will contain much less carbon than if made under contrary conditions. Condition of Carbon in Cast Iron. When cast iron is melted all of its carbon is supposed to be combined with or dissolved in the iron. It is supposed that if it were in any other form it would rise on account of its lightness and float on the surface. When the iron is solidified its carbon will remain in the combined state unless some influence is present to change it. If any such influ- 28 CAST IRON. ence should change the carbon and the influence is removed, the carbon will resume the combined state. The combined is the natural form for carbon to assume which has been absorbed by pig iron ; therefore, owing to the conditions present in the furnace, if the iron has not absorbed more carbon than it can hold when cold, it will be white pig iron. If the casting contains more carbon than it can hold, the darkest casting will be that which contains most carbon to begin with; and secondly, that casting will be darkest which is cooled most slowly. The larger the mass, under ordinary conditions, the slower the cooling; therefore large castings are darker than small castings from the same iron. Diffusion of Carbon in Cast Iron. Probably a variation of .05$ is as near as can be expected in the different parts of a single casting, and the extremes of variation will be o$ to .12%. By a very interesting series of experiments described in Chapter VIII on Keep's "Cooling Curves," it is shown that graphite was not formed until after the casting had become solid, and the graphitic scales exude into the spaces between the crystals. After this takes place the casting must consist of a network or sponge more or less crystalline, with the spaces filled with closely packed graphite. The network must be a lower carbide of iron than the fluid iron. Bulk of the Carbon in Cast Iron. It is evident that the more carbon the less will be the weight of a casting of a given size. Castings containing a large percentage of graphite are coarse-grained, and the coarse crystals have large spaces between them, which causes the casting to be larger. A cubic foot of pure iron weighs 489 Ibs. , with sp. gr. about 7.84. A cubic foot of white cast iron weighs 474 Ibs., with sp. gr. about 7.60. A cubic foot of mottled iron weighs 458 Ibs., with sp. gr. about 7.35. CARBON IN CAST IRON. 29 A cubic foot of light gray iron weighs 450 Ibs., with sp. gr. about 7.20. A cubic foot of dark gray iron weighs 425 Ibs., with sp. gr. about 6.80. Thus in iron with 3$ of carbon the carbon forms about 12% of the entire bulk, and in pig iron containing 4$ the carbon forms about 15$ of the bulk (see Fig. 5). Carbon and Chill. Iron with its carbon combined is white. Slow cooling of any carbonized iron causes some carbon to take the graphite form. If the metal is cooled so suddenly that the carbon cannot have time to become graphite, the casting will be white. This sudden cooling is accomplished by making a portion of the mold out -of iron, which, as the fluid iron runs against it, will suddenly draw the heat from the metal. The white portion of the casting is chilled. Carbon and Hardness. The purest iron of commerce has a hardness of about 40 (Turner's test). I melted a sample which had in the casting 0.29 of carbon and had a hardness of 44. Ingot-iron test-bars with carbon 0.39 had a hardness of 45. Carbon and Fusibility. Truran says ' * fusibility is directly dependent on the proportion of the volumes of the respective ingredients forming the pig iron.'' Pure iron cannot be melted by ordinary furnace heat. Steel containing from 0.25$ to 1.50^ of carbon can be melted with difficulty. The cast iron which contains the most carbon melts the most readily. At a bright red heat the graphite changes into combined carbon and the iron at once melts, beginning at the outside of the pig. Thus pure iron and graphite, neither of which can be melted alone, at a high heat form a fusible alloy. Any dilution of carbon makes the iron harder to melt. At each remelting of cast iron the metal became more infusible. In the series Table IV when carbon reached 2.25 it was very difficult to fill the molds. Fluidity. Carbon is the principal agent for imparting fluidity. The presence of carbon, by lowering the temperature of fusion, increases the fluidity of cast iron. 30 CAST IRON. Gray iron is more fluid than white iron, but requires a much higher temperature for fusion. White iron, slightly cooled, is thickly liquid and passes through a pasty stage between the solid and fluid states which does not allow gases to escape freely, which partly accounts for the blow-holes in white-iron castings. As soon as the graphite in gray cast iron becomes combined the iron melts and becomes very fluid in proportion as the percentage oi graphite was great or small. The author has made a number of experiments to show the influence of different percentages of carbon, but in nearly all of them the physical quality was also influenced by variations of some other elements or by other conditions. Carbon maybe lowered by repeated remeltings. In ordinary foundry practice some of the metal melted each day is returned to the cupola the next, and is remelted with the pig iron. In Chapter IX, Fig. 29 and Table XVII are derived from remelts of Gaylord white pig iron, and Fig. 32 and Table XX from remelts of F L M (Method 12). The fluidity grew less as carbon decreased. The decrease in strength shown in the tests of white iron was largely due to the blow-holes. A very small quantity of graphite will prevent blow-holes, and it need not be sufficient to make the casting gray. Another method by which the carbon may be lowered in test- bars is to introduce into the melted cast iron wrought or steel scrap. This adds strength so long as the carbon is not so much reduced as to make unsound castings. The following tests illus- trate this. The first (Fig. 33 and Table XXI) is that of F L M remelted (Method 13) with additions of iron wire to cause suc- cessive dilutions of the carbon. Fig. 30 and Table XVIII are " Gaylord " white pig iron with iron wire. In the series thus far considered not only has carbon varied, but silicon also. In fact, the cause of blow-holes was not wholly due^ to a lack of carbon, but to too low silicon. In the series Fig. 1 8 and Table IV (Method 6) silicon is kept uniform and the CARBON IN CAST IRON. carbon is therefore the only variable. The lessening of carbon turns the iron white, and bars 564 and 568 would not run full. 568 C. 0.31 FIG. 18. Dilution of Carbon. TABLE IV. Strength. ij Shrinkage. Test .2S No. Mixture. o\ 20 parts of wrought increased the strength 32$; 30 parts, 60^; but 40 parts only 33$. The maximum result was therefore pro- duced with about 30$ of wrought scrap.* * Thomas Turner, Journal Soc. of Chemical Industry, May 29, 1886. 3 2 CAST IRON. Table V is introduced in this place to show how the results under the column ' ' Impact ' ' of Table IV and of other tables to be introduced were obtained. TABLE V. Inches Fall of 2 5 -lb. Hammer. Inch-pounds Developed. ArbitraryValue for each Blow in Ibs. Avoirdupois. Inches Fall of 25-lb. Hammer. Inch-pounds Developed. ArbitraryValue for each Blow in Ibs. Avoirdupois. .12 3-12 16.94 3-12 78.12 423-53 25 6.25 33-88 3.25 81.25 440.47 37 9-37 50.82 3-37 84-37 457-41 50 12.55 67.76 3.50 87-55 474-35 .62 15.63 84.70 3.62 90.63 491.29 -75 18.70 101.65 3-75 93-70 508.24 .87 21.87 118.59 3.87 96.87 525.18 .00 25.00 135.53 4.00 100.00 542.12 .12 28.15 151.46 4.12 103.12 559-06 25 31-25 169.41 4-25 106.25 576.00 37 34.38 186.35 4-37 109.37 592.94 -50 37-41 203.29 4-50 112.55 609.88 .62 40.61 220. 23 4.62 115.63 626.82 75 43-72 237-18 4-75 118.70 647.77 .87 46.87 254.12 4.87 121.87 660 . 7 i 2.OO 50.00 271.06 5.00 125.00 677.65 2.12 53.13 288.00 5.12 128.12 694.59 2.25 56.25 304 . 94 5.25 131.25 7H-53 2-37 59-37 321.88 5-37 134-37 728.47 2.50 62.55 338.82 4-50 137.55 745.41 2.62 65-63 355.76 5.62 140.63 762.35 2-75 68.70 372.71 5-75 143.70 779.30 2.8 7 71.87 389.65 5-87 146.87 796.24 3.00 75.00 406.59 . 6.00 150.00 813.18 Table V shows the number of inch-pounds developed by each blow. Also gives an arbi- trary value to each blow in pounds avoirdupois, to allow comparison and tabulating with records from Keep's dead-load machine. The value for an inch-pound was obtained by testing a good sample of Swedish gray pig iron which broke in the dead-load machine with 288 pounds and by impact with a z^-inch fall. The table was constructed from this data. Table VI shows the influence of annealing ordinary gray cast iron. The test-bars were all measured for shrinkage, then one of each pair was tested, while the companion bars were packed in wood shavings and heated to a white heat in 12 hours and then cooled in another 12 hours. The only chemical change is in graphite and combined carbon. The iron is more than 30$ softer, has lost one-half of its chill, CARBON IN CAST IRON. 33 33$ of its shrinkage, and its deflection is increased nearly 50$. The hardness in this table is obtained by the Keep machine, Fig. 9 6. This does not prove, however, that the same proportions of graphite and combined carbon in the original mixture would produce castings with these physical properties. Probably the carbons in the annealed castings have peculiari- ties which the chemist did not determine. TABLE VI - Shrinkage. Chill. St'gth Dead Load. 435 4*5 Def. Hardness. Analysis by Dickman and Mackenzie. *"Q i X |"D 14 .14 One' End *"D Other End *"D 2 8. 5 29.0 U H 3-77 3-77 cJ 3-32 3-68 U u 45 5 t/5 1.78 1.78 eu 505 .51 t/j .041 043 G .568 .568 No. i bar not an- nealed. No 3 bar not an- '55 55 .169 173 .28 .27 30.8 31.0 Average No. 2 bar not an- 155 1-57 ^55 .171 .167 ..69 4 14 .14 425 400 400 .27" 36 34 30 ig : i 9 I9 l 19 No. 4 bar not an- nealed Average No. 2 bar annealed No.4 " Average .156 2J .096 .168 .069 .070 14 .06 .08 .090 .069" .07 400 35 19 Influence of Remelting on Carbon. For carbons in the A. S. M. E. tests see Tables LXI, LXII, LXIV; also see analy- sis of original irons in Chapter III. Also see Table LXIII. CHAPTER VI. SILICON IN CAST IRON. Silicon. This metalloid is reduced from its oxide, silica, of which quartz is an example. Silica is contained in all the iron ores, and in the ash of all fuel ; therefore more or less silicon is always found in cast iron. Metallic silicon is said to be of a dark iron-gray color, and has such an affinity for oxygen that it must be kept away from air. The specific gravity of silicon is 2.49. It alloys with iron in all proportions up to 10$, and by special treatment to 20$ or 30$. Silicon can be present to a certain per cent in iron that is otherwise comparatively pure and still leave it malleable, such inetal being called silicon steel. The presence of silicon is neces- sary if cast iron is to be used for ordinary casting purposes. According to Truran, silicon, being as bulky as carbon, should increase the fusibility of cast iron. Silicon in Cast Iron. During the year 1885 Professor Turner published the results of his researches showing that addi- tions of silicon to a specially made 'white iron would change it to gray, and that by varying the silicon the softness and grayness could be controlled at will. This discovery was of such practical importance to founders that the question was raised whether the influence of silicon would be the same in ordinary pig iron made In a blast-furnace. To determine this question, during the years 1886 to 1888 the author made a large number of series of tests using Professor Turner's methods. Only a few series are given. The following irons (Table VII) were used to determine the 34 SILICON IN CAST IRON. 35 influence of silicon in cast iron in the first six series of tests. Professor Turner's white and silicon irons are inserted for com- parison. TABLE VII. IRONS USED IN TESTS. Test No. Kind of Iron. T.C G.C. C.C. Si. P. s. Mn. How made. 376 441 396 397 401 White (Gaylord), " (Turner) Gray (F^M) 2.98 1.98 3-55 i.Bi 0.95 0.38 3.22 1. 12 2.03 i. 60 o-33 0.6y 0.186 0.190 1.249 9.80 0.26 0.32 0.084 0.210 0.03 0.05 0.04 0.04 o.cg 0.14 0.187 1-95 Blast furnace Cementation Blast furnace Blast furnace Silicon iron (Turner).. " " (imported) " " (Pencost) . 1.99 1.92 0.07 K>-34 0.45 0.52 Dead Load. Impact. Shrinkage. Chill. Fluidity. Hard ness. Str'n. Def. Str'n. Def. I" D rV'X." 376 441 396 397 401 White (Gaylord) " (Turner) 379 .14 237 16 .248 .246 all 5.60 100 72 74 57 64 132 Gray (Fi-M). 362 157 144 478 231 .27 339 .29 .168 .186 .40 4-5 Silicon iron (Turner) . . " " (Imported) " " (Pencost).. .08 25 .12 68 425 125 .06 .27 .11 39 139 .199 29? .146 .201 . 12 .10 6.62 376 441 Si. 0.25 Si. 1.20 397 401 Si. 4.36 Si. 10.34 FIG. 19. The fractures of four of these irons are shown in Fig. 19 (Method 4). Table VIII shows the percentages of silicon, the hardness, and the calculated transverse strength of a i''Q test- bar of the original series by Professor Turner. 36 CAST IRON. TABLE VIII. TURNER'S TESTS. Silicon in test bars 0.19 0.45 0.96 1.37 1.96 2.51 2.96 3.92 4.74 7.33 9.80 Calculated dead- load transverse strength of test bars \^" C] X 12" long 337 410 421 437 431 442 356 318 293 188 157 Hardness (Turner's test) 72 52 42 ... 22 22 22 27 32 42 57 Silicon added to white iron changes it to gray iron. The upper series, Table IX (Method n), using Gaylord white pig iron 376 and the 16$ ferrosilicon 396 shows this. The fractures are shown in Fig. 20. TABLE IX. No. Test. Calcu- lated Si. Dead Load. Impact Strength Shrinkage. Chill Turner's Hardness Strength i Def. .j" n iV"xi" 37 6 0.25 379 14 237 .248 .246 White 108 IQ7 o . 40 406 . 1-7 216 2S9 257 08 + - 1 VJ IQ4. \j f\s o. 71 i r vyvy 4^2 J 13 271 * ^ 7 .256 -258 y 06 C C o o i y-f 195 ** / j 1.25 *T J ^* 468 J .16 314 .232 .251 .90 v u 93 .~ y I 9 6 1.71 412 .20 322 .18 7 .65 9i 'In 197 2.22 445 .21 374 .200 .232 .90 76 Jc, 198 2.90 420 .19 348 .l6o .230 .80 65 > 200 3-24 422 .22 374 .149 .201 35 61 f^ 201 3-59 436 23 373 .147 .204 25 54 2O2 3.84 430 .20 433 ISO 213 45 52 _1_ T 0/l I 2^ A 9n ^2 4=^8 172 . 20 T^ C 104 185 i . ^5 1-50 ?*v 348 .> 25 TO U 381 1 / .167 .182 -35 o 186 2.00 332 23 305 .160 .I6 7 .27 "* C >-. o 187 2.50 356 .29 339 .156 . 164 .10 - 188 3-00 362 30 382 .142 .162 .07 bf/^ 189 3-50 343 25 322 .140 .158 15 s 190 4.00 378 .29 330 137 -158 17 191 4-50 424 .28 389 .146 .165 15 Silicon added to gray iron lo^v in silicon will make it more gray. The lower series, Table IX (Method n), with F L M gray iron and the i6# FeSi, proves this. The fractures of this series are shown by 184 to 191 of Fig. 20. It is evident that while additions of silicon change combined SILICON IN CAST IRON. 37 carbon into graphite, yet a given percentage of silicon does not always produce a given physical quality. White Pig and 16.32 FeSi. 194 195 196 197 19 s 20 201 202 Si. 0.73 Si. 1.25 Si. 1.71 Si. 2.82 Si. 2.90 Si. 3.24 Si. 3.59 Si. 3.84 F L M Gray Pig and 10.62 FeSi. 184 185 186 187 188 189 190 191 Si. 1.25 Si. 1.50 Si.' 2.00 Si. 2.50 Si, 3.00 Si. 3.50 Si. 4.00 Si. 4.50 ji iiiiiiiiiiiiHiiiiiiL iiiiiiiiiiriiiniiiiiiiiiiiiitiiiiiiiiiiiiiii niiiiiiiiiiiiiiiinn fiiiiiiniiniiiifiiii FIG. 20. // is the influence of silicon, not the percentage, that produces the physical quality, as is shown more clearly by Table X and Fig. 2 1 (Method 4). Fig. 19 shows the appearance of grain of the irons used to make the mixtures, and Table VII the chemical composition. Examine in Fig. 21 the fractures of the bars of each series CAST IRON. containing 1.50$ silicon, or the bars of each containing 2%, or those with 2.50$ or 3.00$: though the silicon is the same in each, the influence of silicon in each bar containing the same per- centage depends upon the total carbon and its condition in the original irons, and upon other unknown conditions. To more fully illustrate this see Fig. 22 and Table XI (Method 4). The TABLE X. No. Test. 344 345 346 347 340 34i 342 343 Calculated Percentage Silicon. White + Silicon. \\ per cent silicon 2 " " White + Silicon, g per cent silicon. . F L M -j- Silicon. if, per cent silicon. . 349' 2 " .. 350 2^ " .. 351 3 " " F L M -f Silicon. 352 if per cent silicon 353, 2 354 2! 355 3 Sir'n 466 486 416 421 501 464 498 452 362 363 380 370 365 348 318 348 Def. S 23 23 .28 .26 1 Impact. Shrinkage. Str'n 299 356 362 451 Det. Sq. Bar. Thin Bar. 243 339 .220 .167 15 .21 .21 .27 .227 1.163 -163 155 486 .27 .179 .243 390 466 390 .20 .29 .22 !i56 .147 .217 .161 .160 345 334 350 385 25 .22 .28 27 .162 .156 143 134 175 .158 .149 153 373 228 .29 23 .163 .161 .181 .162 277 401 .24 32 .141 .148 .149 . I4 B Hard ness. 58 45 42 28 54 50 39 42 Chill All 1.25 .60 .22 1-25 .60 .18 .40 .03 , 4 .07 03 Act- ual per cent Sili- con. 1.48 1.92 2.42 1.49 1.97 .202.55 2-95 1.50 22 2.01 10 2.42 2.86 i-53 20 2 . 1 1 2.4T 2.86 same F L M (441) was used for the bulk of each mixture. The silicon of each mixture was 2.50$. The 1.25 silicon added in each case was obtained by using a different silicon iron, the silicons of which are given in the table. The influence of silicon is indirect, acting through the carbon which the iron contains. Silicon lowers the saturation point of carbon, that is, an addi- SILICON IN CAST IRON. 39 tion of silicon to iron containing combined carbon expels carbon in graphitic form, which is caught between the grains of the iron, giving it a grayer color. The influence of silicon is modified by the various conditions 344 345 346 347 Si. 1.48 Si. 1.92 Si. 2.42 Si. 2.87 340 34i 342 343 Si. 1.49 Si. 1.97 Si. 2.55 Si. 2.95 348 349 Si. 1.50 Si. 2. or 350 Si. 2.42 35i Si. 2.86 352 353 354 355 Si. 1.53 Si. 2.H Si. 2.41 Si. 2.86 FIG. 21. attending remelting and cooling of the iron. The more total carbon or the less combined carbon in the iron the less silicon will be required to produce a given effect. Less silicon acting a long time, as in slow cooling of large castings, or more silicon acting in a short time, as in rapid cooling of small castings, will produce similar effects. CAST IRON. TABLE XI. 2.5 PER CENT SILICON IN EACH TEST-BAR. Silicon Irons used in Mixture with FLM. No. Test. Dead Load. Impact. Shrinkage. Chill. Fluidity. Hardness. No. 'Test. Name. Per cent of Si. Str. Def. Str. Def. i" n iV'x i" 36 397 403 140 & 213 S2I 35 $ _^- Wellston Pencost Dayton Govan 5.66 4-36 12.08 4-35 9.85 6.76 11.99 10.62 6 99 S-5 16.32 10.34 447 35 448 449 45 451 45 2 453 454 455 456 354 384 380 375 375 367 357 349 342 34i 337 327 2*1 .26 .28 .20 .26 .28 23 .22 .21 :3 .20 .24 362 35o 322 333 384 345 277 3" 93 339 282 277 .26 .26 .24 .26 30 .27 .22 .24 .29 .27 23 23 .146 143 .144 i45 .147 .141 .138 .140 .148 .150 .140 .141 155 .149 .208 .158 '59 J 55 .182 .171 157 *57 .164 .149 .08 .10 35 .04 .04 .08 35 23 -13 05 23 .08 4.12 IO.OO 837 12. OO 7-03 .07 8.62 11.00 10.62 11 .50 7.84 9 25 58 5 2 .60 .64 .48 .48 .40 .40 58 .80 .48 .48 Pencost foreign Wellston foreign , Oncost . . . 449 450 FIG. 22. Silicon 2.5$ in each. SILICON IN CAST IRON. 41 It is the influence of silicon upon the physical quality in a casting that is of value to a founder. By changing silicon in an iron mixture we can control the state of carbon and also the chill. Silicon, acting through carbon, is therefore the controlling element in cast iron. White iron contains less than \% of silicon. No. 3 Foundry should contain about 1.50$ of silicon. ' ' 2 ' ' ' ' ' ' ' ' 2 1 ' I ' ' ' ' ' ' ' ' 2 < 2 Soft " " " 3-00$ " " < < j < < < < < < 3 2 ^^ ' ' ' ' Silvery iron " " " 4 to 6% " " High-priced silicon iron contains from 6$ to 10$. By mixing these irons we can get any desired percentage of silicon in the casting. Diffusion of Silicon in Cast Iron. Table XII gives the sili- cons of nineteen series of tests made by the author for the Ameri- can Society of Mechanical Engineers. It has been supposed that if the product of a blast-furnace could be caught in a mixing-ladle before it was run into the pig bed, an iron could be produced which would make castings of homogeneous chemical composi- tion. If in Table XII (Method i) we subtract the smallest record in a series from each other record, the remainders will show the excess of silicon in each test-bar poured from one ladle from iron which was supposed to contain the same percentage of silicon. Series 19 was from iron taken from a i2OO-lb. mixing-ladle, and the mixture was made up of pig irons of very nearly uniform composition, and the average silicon in all test-bars was .77$. The variation was . 19 with an average of .09$. Series 18 was made of pig irons of nearly uniform composition and good scrap mixed in a i2OO-lb. ladle. The average silicon in the test-bars was 2.96, with variation .42 and an average of Series 15 is an example of the practice of imparting silicon by a low silicon silvery iron. The iron was melted at the rate of CAST IRON. eleven tons an hour, and was caught as fast as it was melted, and no effort was made to obtain even diffusion. The variations of .26 and average of .22 show that as perfect diffusion may be obtained without as with the mixing-ladle. Series 2 to 6 are made from the same pig as series I, with increasing additions of an I \% ferrosilicon. The diffusion of silicon is very irregular, but the average variations increase as the silicon increases, showing, in connection with the other series, that the diffusion is less com- plete as silicon increases in quantity. TABLE XII. A. S. M. E. TESTS. | I c** u c c til Per Cent of Silicon. > < c/5 |" l" D i" X 2" 2"D 3 "D 4 "o I I.OO 8^ ,7Q 78 .82 72 .88 .81 2 1.50 **j 1.09 / V 1.14 . / u 1.70 1-33 / ^ I.IO .88 1. 2O 3 2.00 1-73 1-73 1.70 1.50 2.17 2.50 1.88 4 2.50 2.13 1.69 1.60 i. 80 2.17 2.07 2.01 5 3.00 2.42 265 2.40 3.36 3.67 4.67 3-19 6 3.50 2.74 2.69 2.70 2.62 4-30 3-22 3-04 Hinkle 7 I.OO .01 .qa .86 .QO .85 1. 12 Q-l / 8 1.50 * V* 1.16 V J 1.29 I.IO . y\j 1.22 * W J 1.25 1.03 y.> 1.17 9 2.00 93 1.40 1.05 I.OO 2.15 3-50 1.67 10 2.50 2.84 2 55 2.70 2.OO 1-75 1-57 2.23 ii 3.00 2.56 2-75 2.97 2.49 2.64 284 2.71 12 3-50 2.77 3-7D 34i 2.91 2.89 2-95 3-05 2. 7O 2.80 2. Si 2.7Q 2.04 2.81 2.81 13 ~* / ** 3-13 3-22 3-17 ** 1 y 3-19 *" VI 3-20 3.15 3^18 15 3-29 3-50 3-52 3.48 3-75 3-42 3-51 C.G. Bretting &Co. 16 I QO 1.86 1.68 1.61 i 8^ I 7O ' 76 Mich Mall.IronCo. 1 7 j..y<~ 81 .67 i . OJ .86 j.. f\j I 2J. *.. /u O2 Bement, Miles&Co. * i 18 2 2Q 2 OQ 2 24. v> j 1.82 2.06 i . ~ -4 1.88 V? 2 O^ A. Whitney & Sons 19 ^V .87 ^>. wv^ .72 * ^"T .78 .81 .68 73 ^O .76 Loss of Silicon in Remelting. An analysis of drillings from three pigs of Iroquois iron by Mr. C. D. Chamberlain gave .88$, and an analysis of drillings from twenty-five pigs of the same lot of iron by Messrs. Dickman and Mackenzie gave 98$ silicon. The analysis of each of the six sizes of test-bars of SILICON IN CAST IRON. 43 series I showed a variation of silicon between .72 and .83, c r the same variation as the two analyses of the pig iron. The average silicon in the six castings was .80$, showing the loss during the remelting of .08$ to .18$ of silicon, and indicates that there is a loss of as much as 10$ of the silicon in the iron during remelting in a cupola. See also Table LXIII. There could not have been a more careful mixture of iron to produce given percentages of silicon than in the series I to 6, and a slight excess of silicon was added (Method i), for it was expected that there would be a loss in remelting. The calculated and average silicons, by analysis, are shown in the upper part of Table XIII. TABLE XIII. LOSS OF SILICON IN MELTING IROQUOIS PIG IRON. No. Series. i a 3 4 5 6 I] 3 i % u 1. 00 .80 I.OO 1.23 1-50 I. 21 1.50 1.64 2.00 1.88 2.OO 2.04 2.50 2.OI 2.50 2.25 3-00 3-19 3.OO 3.10 3-50 3-04 3-50 3-49 Si by Analysis . . . Calculated Si To show how impossible it is to produce an exact percentage of silicon in castings, even by an analysis of pig iron, the lower part of Table XIII gives the calculated and exact silicons in six crucible heats of the same mixtures of iron. The Iroquois pig iron from the analysis was supposed to contain .88$ of silicon. The additional silicon to make even percentages was obtained by additions of n# ferrosilicon (no excess was added to make up for loss). The variations in silicon, as shown by the analysis of Mr. E. E . Mains in Table XIII (crucible), indicates that there must have been an unequal diffusion of silicon in the material used. These examples seem to prove that it is only possible to 44 CAST IRON. obtain an approximation of the silicon in a pile of pig iron, and that it is much less possible to obtain the percentage of silicon in a pile of scrap, and that, having obtained an approximation by analysis, the loss of silicon during remelting, and the uneven diffusion of silicon in the casting, make it possible to obtain, by an analysis of the original materials, only an approximation of the percentage of silicon in the casting. An analysis of the casting itself gives the only reliable infor- mation, and this is only an approximation from the uneven diffu- sion of silicon. Silicon and Hardness. All records of hardness (except those in the chapter on Hardness) are given in terms of grams placed on a diamond to cause it to make a scratch on a polished surface with Professor Turner's machine. Silicon is a softener when added to hard cast iron. From 2$ to 3.50$ of silicon will change into graphite all of the combined carbon that can be changed. This change, we know from experience, generally reduces hardness. The iron that remains is softer, because iron of itself is soft and was hard- ened by the carbon that was combined with it. The more the combined carbon is changed into graphite the softer the remain- ing iron. Probably silicon itself, however small the quantity present, hardens cast iron ; but the decrease of hardness from the change of the combined carbon to graphite is so much more rapid that the total effect is to decrease hardness, until the silicon reaches from 2% to 3$, which is as much as is required in practical foundry work. Silicon and Chill. Silicon removes the chilling tendency of an iron. This influence of silicon is greatly modified by the state and quantity of carbon in the iron used. Silicon and Fluidity. While the records are only an indica- tion of the fluidity of each cast due to the heat of that cast, yet when silicon is mixed with irons previously low in silicon the fluidity is increased. SILICON IN CAST IRON. 45 Silicon and the Surface of a Casting. One of the most use- ful influences of silicon is the color and freedom from sand that it imparts to the surface of a casting. The liberation of graphite at the surface interposes between the hot iron and the sand mold a layer of graphite, which in thin castings prevents the sand from burning on. In heavy castings the graphite is deposited under the scale and causes it to peel off and leave a smooth surface. The Relation of Carbon to Silicon in Pig Iron. In Table XIV, with silicon at from 2% to 2j$ or even 3$, the maximum percentage of carbon may be present. Above this point silicon replaces carbon. At one end of the series, a very low silicon and low carbon, with about 2\fi silicon we have the highest carbon, and at the otfier extreme of the series we find the highest silicon and no carbon. TABLE XIV. RELATION SILICON TO CARBON. o.S* 0^ 1 be 1 c/> >, frl ! cfc fc d Calumet No. i Foundry. B o c c.H S J M 14 m * i >, *> c75 4-39 3-44 3.40 .04 1.42 ..':. C >, 5| i/5 ^ 02 u U. i! E Pence Ferro Si oil I 1 ! Si 0.18 2-53 .49 2.12 .26 03 .09 1-25 3-55 3-22 33 .08 .04 .18 2.03 3-75 3-12 63 1.65 .01 .87 3.15 3-50 2.46 I. 10 1. 06 .02 1-35 5.89 3-15 2.85 30 I. 10 .02 1. 00 9. 10 2.58 .90 1.68 .09 03 2.20 10.34 1.99 1.92 .07 45 tr 57 12.08 I. 5 8 1.52 .06 .48 tr .76 16.27 75 .01 .01 .60 20.00 .00 T C G C C C p s Mn CHAPTER VII. SHRINKAGE OF CAST IRON. THE general understanding is that the shrinkage of a casting is the difference in length (or any other linear dimension) between the casting and the pattern from which it was made, or rather between it and the mold in which it was cast. It has generally been understood that the shrinkage of cast iron is one-eighth of an inch per foot. The pattern-maker, in taking measurements for the different dimensions of the pattern, uses a " shrink-rule, " which is one-eighth inch longer than the standard foot-rule which is used to measure the casting. In the investigation of shrinkage it must be understood that one size of test-bar must be used when the influence of variations in chemical composition is being observed. Any one size of test-bar could be used as a measure of shrink- age, but the author has adopted a test-bar \" G X 12" long because it is convenient to make and handle, is more sensitive to any change in composition of the iron mixture, and shows a wider range of shrinkage than any other size. In the examination of the variation in shrinkage caused by a variation in size of test-bar , iron of unvarying composition should be put into all the sizes of test-bars. Carbon and Shrinkage. When iron becomes crystalline it will occupy more space than before it became so. Carbon causes iron to become crystalline, therefore it reduces shrinkage. Pure iron shrinks about .300 of an inch per foot. Iron in melting expands for each degree of heat that has entered it. It is poured into the mold at this high temperature, therefore in cooling it must shrink as much as it has been expanded. 46 SHRINKAGE OF CAST IRON. 47 Iron containing carbon melts at a lower temperature than pure iron, and consequently shows less shrinkage. " Dayton " white iron had a shrinkage of .221". " mottled " n tt .179". close gray " " " " " .167". 11 dark " " " -145". Silicon reduces shrinkage by changing combined carbon into graphite. Its direct influence is overcome, and its influence as a reducer of shrinkage is seen, through its action on carbon. See Tables IX to XIV. Anything that decreases silicon increases shrinkage. For- tunately any increase of shrinkage produced by any chemical element in cast iron can be largely overcome by increasing silicon. In June, 1894, from the data then at hand, the author pre- pared the following proposition: 4 ' To produce a given uniform grain and a sound casting, and with one-eighth of an inch shrinkage to the foot, the silicon must vary with each variation in the size of the casting. And such a variation in silicon will cause a variation in the shrinkage of a half-inch test-bar." Table XV shows the shrinkage of a -in. test-bar that will indicate the correct amount of silicon for each size of casting to produce a shrinkage of -J in. per foot according to this proposition. TABLE XV. AN APPROXIMATE KEY FOR REGULATING FOUNDRY MIXTURES. Size of the Casting. Silicon required in the Casting. Shrinkage of the Casting. Shrinkage of a \" Q Test Bar from the same Ladle. |-inch square 3.25 per cent .125 ins. per foot .125 ins. per foot I. " 2.75 " " .125 " " " .135 " " " 2- " 2.25 " " .125 " " " .145 " " " 3. M ,. 1-75 " " .125 " " " .155 ' 4- " 1.25 " " .125 " " " .165 " " " 4 8 CAST IRON. The author had just been made a member of the A. S. M. E. committee on Standard Tests and Methods of Testing Mate- rials, and he agreed to make a series of tests with these sizes of test-bars sufficient to prove the proposition. Table XVI gives the shrinkage records. TABLE XVI. SHRINKAGE OF THE SERIES OF A. S. M. E. TESTS. Kind of Iron Average Per cent Silicon. No. of Series. Test Bars *"O i"D 2"X l" 2"D 3"D 4"D f o So I .183 .160 .148 .131 .116 . IO2 " Iroquois " with I. 21 2 .172 .150 .138 .125 .110 .10^ silicon added by J 1.88 3 .166 145 .130 .109 .069 3 ( . Pencost ferro ', 2.01 4 .162 143 .123 .099 .066 .128 silicon. 3-19 5 157 .105 .094 .075 .067 057 I 3-04 6 .169 .130 .086 .077 .085 033 r 0-93 7 .176 .149 .144 139 115 .072 I. 17 8 .160 145 .126 .122 093 .092 "Hinkle" and 1 1.6 7 9 .156 .141 134 .128 .083 .036 Pencost. ] 2.23 10 .154 .124 .092 .094 075 .067 1 2-71 ii 157 . IO2 .O9O .062 053 .023 I 3-50 12 .144 .098 .092 .068 043 .023 Mich Stove Co . .. 2.82 Id. Id8 .098 O8^ O72 06^ 07 C Do 3-i8 *-*T 13 A*f ^ .130 095 <-"> O .091 f\tj4 .079 . filled, increased until 3^ minutes, then decreased until 7 minutes. This is named the 1st Expansion. The expansion then increased until 8 minutes, and decreased again until 10 minutes. This I called the 2d Expansion. A very great expansion then takes place, reaching its maximum between 12^ and 14 minutes, ami decreasing until 16 minutes, or a little later. This is the $d Expansion. When these expansions are completed the regular shrinkage curve from the loss of heat is formed, the same as in the simple metals. This shrinkage had been acting from the beginning, for the metal had been parting with its heat all the time, but the expan- sions were great enough to overcome all this shrinkage during the first 16 minutes. Another proof of this is that the shrinkage curve of all i-in. D cast-iron bars takes substantially the same direction after the 3d Expansion is completed. This is beautifully shown in Curve No. 35 of Fig. 28, where the dotted line shows the location of the shrinkage curve if no expansion had occurred. Solidifying of Cast Iron. To get an explanation of Curve No. u, eighteen test-bars I in. Q and I ft. long were poured at the same time as the 2 -ft. test-bar from which the diagram was taken. As the bars were made in a snap flask, there was nothing around the bar but sand. The first bar was numbered 19. At the end of I minute the iron in the gate was still fluid. At i| minutes the sand was cut away and the bar taken out, but it broke by its own weight, though it was not fluid. One-half of this bar was dropped into a barrel of ice-water and the other half was allowed to cool in the air. At the end of each minute there- after the sand was cut away from a bar, which was broken, and half of it dropped into the ice-water. From the fact that the cooling of a i-in. D bar in water cannot be instantaneous, and that anything short of that would allow a change in crystalliza- KEEP'S COOLfNG CURVES. 55 tion, the quenched bars give only a faint idea of the condition of the iron at the time it was taken from the mold. Each bar was a little stronger than the preceding one, and as soon as it could not be broken with a pair of pincers alone, one- half of the bar was placed in a hole in a heavy block of iron, when a wrench of the pincers would break it. Then a light blow of a hammer, and toward the end quite a sharp blow from a 5-lb. hammer was needed. Curve No. 1 1 , Fig. 24, was then divided according to the times of breaking the 18 bars, to see which belonged to the different parts of the curve. Previous to making the bars described, and while Curve No. 13 (Fig. 25) was being made, a similar number of bars numbered from I to 18 had been cast, cooled, and broken. Hard or Soft Cast Iron. An examination of the fracture of these two series of quenched bars shows a great change in the crystalline structure before and after the 3d Expansion, but these fractures do not at all show what the iron really was, because quenching cannot entirely prevent the crystals assuming their natural form. The whole change from melted iron to a soft gray crystalline casting, shown by Curve No. n, can take place in a thin casting in less than a minute (see Curve No. 17, Fig. 26). If a non-chilling iron, like that from which the Curve No. n is made, is poured against a chill, only a very thin portion will be chilled, and behind this, toward the molten mass, will be formed a dense black soft grain, probably at the same instant with the chilled portion. This instantaneous passage of cast iron through all of the stages of crystallization, from fluidity through the 3d Expansion, makes it impossible to fix the iron at any instant. To get an approximate idea of the state of the iron, the bars numbered 19, 25, and 30 were selected for analysis; No. 19 before the iron was solid, No. 25 during the 2d Expansion, and No. 30 just as the 3d Expansion had reached its maximum. The combined carbon in bar No. 19 was 0.60, in bar No. 25, 0.45, and in bar No. 30 it was 0.06. From the location'of the 56 CAST IRON. curves, bars 26, 27, and 28 were probably as hard and contained as much combined carbon as No. 25. As bar No. 28 probably contained 0.40 combined carbon, and as bar No. 30 contained only 0.06 combined carbon, which is the same percentage as was contained by the portion of the bars which were allowed to cool in the air, it appears that the change of combined carbon into graphite takes place in less than I minute in a casting I in. D cooled in its own mold, and that this is the time when hard iron changes to soft iron. After the 3d Expansion no further change in the crystalline structure took place, and the shrinkage curve was that ordinarily made by the loss of heat. The bars Nos. 19 and 25 were so hard that they could not be touched with a drill, and it was very difficult to break off enough for analysis. It would seem that the bars were much harder than could be accounted for by the 0.60$ of combined carbon, while bar No. 30 was very soft. The final arrangement of crystals took place during the 3d Expansion, and at that time the iron became soft. Calling an iron by the number of its curve, No. 1 1 was intensely hard for the first 10 minutes and became soft during the 3d Expansion. In Nos. 12 and 13 the 3d Expansion was almost lacking and the iron was left in its hard state. The facts were that the castings made from the mixture of No. n were all soft, while those made from Nos. 12 and 13 were difficult to drill. Much depends upon the character of the original irons. No. 10 would almost scratch glass, but Nos. 6, 7, 8, and 9, made from pig iron only, and melted in a crucible, though the 3d Expansion is not great, were soft. An investigation may show that the 1st Expansion, being so large, had a softening influence, or that the entire absence of the 2d Expansion may account for it. Silicon is a Softener and a Lessener of Shrinkage. Curve No. 4, Fig. 25, shows an immense 3d Expansion, and the iron is so soft and open as to be very weak, and the silicon is 3.49^. Curve No. 5 is one of iron containing 3.10$ of silicon. The 3d Expansion is not as great, and the iron is not quite as soft, as in No. 4. Each lessening of silicon lessens the 3d Expansion, KEEP'S COOLING CURVES. 57 FIG. 25. 58 CAST IRON. and the iron is harder each time. The silicon of No. 1 1 is higher than that of No. 4, being 3.85$, but No. u is from a regular cupola mixture of close-grained low-carbon irons, and 40$ of the mixture is the sprues made the previous day, and the latter have been melted over each day. In irons producing curves Nos. 4 to 9, and 14, 15, and 16, the total carbon was nearly 4.00, and all are open-grain pig iron, and melted without scrap in a cruci- ble. In Nos. n, 12, and 13 the carbon was about 3.10$, phos- phorus was i. 00$, and sulphur 0.10$, while in the crucible irons Nos. 4 to 9 P was only 0.20 and S 0.04. In the practical application of cooling curves to foundry work the mold can be made in 20 minutes, and as soon as the iron is running the bar can be poured. It takes 15 minutes to find the 3d Expansion. It is at once apparent whether the mixture needs more or less silicon, and the charges of iron can be changed at once if necessary. Phosphorus, Sulphur, and Manganese in Cast Iron. In Curve No. 14 (Fig. 25) phosphorus is 1.14$ and the silicon is 2.44$. The 1st Expansion continued longer than in Curve No. 6, the 3d Expansion was greater, and the casting, therefore, is softer. The final shrinkage begins higher up or from a greater initial expansion, and the total shrinkage is therefore less than in No. 6. In Curve No. 15 manganese was increased to 0.83$, while the silicon is substantially the same as in No. 6, which was about 2.50. The iron was hotter, and for this reason it remained fluid for 2 minutes. The 1st Expansion was of shorter duration. A 2d Expansion is almost apparent, and the 3d Expansion occurred later, and was greater than in No. 6, therefore the iron was no harder. In Curve No. 16 the sulphur was 0.169$. This has greatly lessened the duration of the 1st Expansion, and has both shortened and reduced the 3d Expansion, and has therefore caused the iron to be harder than that of No. 6. Size of Casting, and Expansion. Figs. 24 and 26 show Curves Nos. 17, 18, II, 19, 20, 21, and 22, from test-bars -J" X I", i" D, i" D, i" X 2", 2" D, 3" D, and 4'' D, which are KEEP'S COOLING CURVES. 59 the sizes that were made for the Committee's strength tests. In No. 17 the casting became solid in 20 seconds, with a very slight 1st Expansion, and the 3d Expansion probably occurred in i minutes. In No. 18 the 1st Expansion began as soon as the bar was poured, and the curve shows the 2d and 3d Expansions. In No. 19 the thickness of the bar was the same as in No. 1 1, but the width was twice as great, and the ratio of cooling was slower, and therefore all three expansions are retarded. In Nos. 20 and 2 i the size of the bar was so great that it was not congealed in the center for some time after pouring, and the early beginning of the ist Expansion must have been on account of the pins of the test-bar being located on the edge of the mold. As soon as the shell became rigid enough it expanded, the same as any solid casting, and the slowness of cooling prolonged the period of each expansion. The rate of cooling causes the location of the expan- sion curves to be formed either earlier or later. Effect of Hot or Dull Iron on Shrinkage. Fig. 27 gives four examples of hot- and cold-poured test-bars. The apparatus was arranged to make two curves at one time, and the test- bars were half as long as those already examined. The enlargement of the diagrams showing shrinkage in this chart is therefore only ten times, but the time measure is the same as before. In each of the four examples presented, iron was caught in a -ladle and emptied out several times in succession so as to heat it very hot, and then 35 Ibs. of iron was caught and a bar i ft. long and i in. D was poured immediately. The ladle was then allowed to stand until a shell had formed on the top of the remaining iron. A hole was broken through this shell, and the iron under it poured into another test-bar of the same size. This iron was as dull as would fill the mold, and to insure a full test- bar the gates had been cut nearly as large as the mould for the bar. The iron put into the first mold was white-hot and flowed like water. The last was red and sluggish. The hot bar, No. 26, became solid in a little more than a minute, when the 6o CAST IRON. FTG 26. KEEP'S COOLING CURVES. 61 expansion began. The 2cl Expansion had begun when the dull bar was poured, yet the dull bar went through the expansions so much more rapidly that the temperature that produced the 3d Expansion was reached in both the hot- and dull-poured bars at nearly the same time. The final shrinkage of the t\vo did not vary much, though the hot bars shrank a little the most. The dull-poured bar went through the changes more rapidly, because it entered a cold mold, and was nearer the temperature at which the 3d Expansion would occur, to begin with. Temperatures at which the Three Expansions Take Place. The diagrams show that each expansion occurs at a definite temperature. In Fig. 27 the hot-poured bars had a greater amount of heat to impart to the mold than the cold-poured bars, FIG. 27. and the temperatures necessary for the formation of the curves were reached after a longer interval of time. The No. 10 bar in Fig. 25 was poured very hot, and the 3d Expansion occurs after a greater interval of time. Nos. 7 and 9 were dull, and the 3d Expansion occurs earlier than in the others. If the rate of cooling is slower it will take a longer time to reach the temperature at which each expansion takes place. For example, in No. 1 1, Fig. 26, the 3d Expansion took place in 12 minutes; in No. 19 it was 20 minutes; in No. 20 it was 40 minutes; in No. 21 it was 85 minutes; and in No. 22 it was 140 62 C^ST IRON. minutes, which corresponds with the rates of cooling. It is important to prove that each expansion occurs at a definite tem- perature, and it would be a great satisfaction to know the exact degree of heat. The cast-iron test-bar, as shown by the 1 8 bars that were broken, was at quite a red heat at the 3d Expansion. It may be found that a change in chemical composition may hasten or retard the formation of the curves, irrespective of tem- perature. For example, in the curves of iron and steel, Fig. 28, the bars had just a reddish tinge in the sunlight, while the expan- sion was taking place, and were a dull red, if shaded; and this curve must correspond with the 3d Expansion in cast iron, which takes place at a bright-red heat. When does Carbon Combine when Heated towards Fusion ? The cast-iron test-bar from which Curve No. 1 1 was taken was heated as much as it was thought it would stand without breaking, and was placed at a bright-red heat on the pins of the machine. The result was a curve, I \a, Fig. 25. As this bar was cooled in the open air the change was very rapid, and the proportions of the diagram are different from the original. The diagram begins just before the 3d Expansion. This shows that the crystalline structure which produced the 3d Expansion had been changed, during the latter part of the heating, to the struc- ture which preceded the 3d Expansion. At that time most of the carbon was combined, and the iron was extremely hard. This experiment shows that in melting graphitic cast iron the graphite changes to combined carbon when the temperature of the 3d Expansion is reached, instead of at the temperature of fusion. Unlike white cast iron, the iron is in an expanded state from the 3d Expansion to the point of fusion; i.e., the atoms are not as close together. In white iron, with the carbon combined in the cold casting, there is no change in the crystalline struc- ture during the heating, and the iron does not reach the expansion which causes it to fuse until just before fusion. Gray cast iron reaches its greatest expansion much sooner than white iron, with the result that it melts from the outside of the casting, KEEP'S COOLING CURVES. 63 and does not become plastic to the extent that white cast iron does. The bar which produced Curve No. n was again heated, to determine if a lower point on the curve could be reached, but it fell apart in handling. Practically, the 3d Expansion is all that can be reached by reheating. It was found that the bar was too long to go in the machine after the second reheating, showing that two heatings above the 3d Expansion had increased the size of the crystals the same as ordinary annealing. The temperature for annealing should, therefore, be that of the 3d Expansion. To illustrate the expanded condition of cast iron of the quality of No. 11, two of the gates from the 18 bars that were broken were cleaned, and one of them was polished. Two ladles of melted iron from the same heat were placed on the floor; one of the 14-ounce gates was placed in each. They were plunged into the fluid metal at first to cause the melted iron to come in contact with the surface. Both gates (about I in. round X 4 ins. long) lay on top of the melted metal until they were melted, about one- fourth being above the surface. This took 2 minutes. The fact that gray cast iron just before fusion is more expanded than when cold or when fluid explains the phenomenon of the floating of gray iron on the surface of fluid iron while it is melting. Drop a piece of cold iron into molten iron. At first it sinks, then rises and floats with about one-half its bulk above the surface until it is all melted. This seems to the author the first correct explanation for this phenomenon. Curves from Heated Rolled Steel (Fig. 28). The first bar treated was a bar of merchant iron I in. Q by 26 ins. long, with the holes for the pins 23f ins. apart. The expansion was so great that when white-hot it was 24^ ins. long. As these bars were cooled in the open air the shrinkage was very rapid. The curve of No. 31 changed slightly after I minute, but it would need other tests to show whether the metal became at all crystalline. The next tested was a bar I in. Q of Jessops tool-steel, Curve CAST IRON. No. 32. This was then heated again, to see if it would become more coarsely crystalline, Curve No. 33. The expansion (which is the 3d) at the first heating was blended into the curve of shrinkage, and was of shorter duration than that of the second heating, showing that it became more coarsely crystalline by reheating past the 3d Expansion. This was on account of its JESSOP'S TOOL STEEL mil FIG. 28. high carbon. (The pins of the machine, which were of Stubbs steel, became enlarged by repeated heatings.) The next tested was a bar of ij-in. Q mild steel, with carbon 0.45$, which was expected to behave more like No. 31. The expansion curve, 34, was so great, however, that while the 2 -ft. bar was shrinking at the rate of ^/inr in. in 4 minutes, the 3d Expansion overcame this shrinkage and carried the pencil backwards --- oi an inch. KEEP'S COOLING CURl'ES. 65 The second heating gave Curve No. 35. These curves show that the shrinkage is going on at the same time with the expan- sion, for the direction of the shrinkage curve after the expansion is the same as it would have been if no expansion had taken place, as shown in each case by the dotted line. The total shrinkage of any iron or steel is therefore decreased by the amount of the expansion. At the second heating of the 0.45 C. steel, when the expan- sion began, the color in sunlight was dark, with a faint red tinge; by shading it from the light the side of the bar away from the light was red. When the expansion was over, the bar on the side away from the light was a dull red. The foreman said that if the steel was red-short it would break if forged at such a color as existed during the expansion. This remark, and the difference between the expansions of Jf.ssops high-carbon steel and the 0.45% carbon mild steel, suggest the possibility of determining this property in such metal by the use of these expansion curves. In the practical application of these cooling curves any bar of iron can have two J-in. holes drilled, 23! ins. apart, in 10 minutes; it can be heated in 10 minutes, and the record is made in 5 minutes. Relation of these Expansions to the Critical Points of Iron and Steel. This cannot be ascertained until the temperature at which each expansion takes place is determined. If these expan- sions should occur at the temperatures 850, 750, and 650 C., which correspond to the critical .points Ar. 3, Ar. 2, and Ar. I, these expansions are caused by a change in the length of the test-bar; in other words, it is purely a physical change and not at all caused by any increase in temperature. If the expansion was caused by a rise in temperature, then, in diagram No. 11, during the 3d Expansion the temperature must have been higher than when the iron was melted, which idea is absurd. The expansion curves are caused by a rearrangement of crystals, and is purely a physical process. CHAPTER IX. PHOSPHORUS IN CAST IRON. PHOSPHORUS is reduced from phosphoric acid, which is found in all fuel and in iron ores. Phosphorus is produced in two forms, the ordinary white variety, with specific gravity 1.82, and red phosphorus, with a specific gravity of 2.14. The latter is. produced by heating the first variety, and will unite with pure iron to form iron phosphide. The author could not make ordinary white stick phosphorus exert any influence when introduced into cast iron (Method 7). By melting wrought-iron drillings with red phosphorus Mr. H. S. Fleming produced in his laboratory 10 Ibs. of phosphide of iron which contained 10.22$ of phos- phorus. Metallic phosphorus is said to have a specific gravity of 2.34. If Truran's opinion is correct, the bulk of phosphorus being greater than carbon should make iron more fusible. Ten per cent iron phosphide melts as easily as cast iron. Very small per- centages of phosphorus alloyed with commercially pure iron are said to make it brittle when cold, but not when hot, and for this reason such brittleness is called cold-shortness. This term can- not be correctly applied to cast iron. As iron has an affinity for phosphorus, it is found in all cast iron. In the Journal of the Iron and Steel Institute, No. i, for 1886, Professor Turner says: "What is badly wanted at the presenl moment is a series of experiments in which various proportions of phosphorus should be added to a specimen of cast iron in which all other constituents should be kept as nearly as possible con- stant." < 66 PHOSPHORUS IN CAST IRON. 67 With the 10.22$ phosphide were made two series of test-bars. The first with Gaylord white iron, Table XIX and Fig. 31, and the second with F L M gray iron, Table XXII and Fig. 34 (Method 11). 376 377 378 379 38o 381 G. C. 1.62 G. C. 2.14 G. C. 1.53 G. C. 1.83 G. C. 1.57 G. C. 1.44 C. C. 1.51 C. C. .83 C. C. 1.59 C. C. 1.59 C. C. - C. C. 1.08 Si. .25 Si. .23 Si. .20 Si. .20 Si. .15 Si. .12 P. .26 P. .26 P. .26 P. .26 P. .26 P. .26 FIG. 29. White Iron (Remelts). Runners remelted with new iron. Carbons in all, and Si. in 376 and 381, and P. of 376 from actual analysis. Other percentages are estimated. 436 437 438 439 440 ale ;.W FIG. 30. Steel Wire added to White Iron, remelted as in Fig. 29, and all ele- ments diluted as in Fig. 31. Analysis of 436 same as 376. To 440 was added one per cent of aluminum, which made 440^, a solid casting. 68 CAST IRON. TABLE XVII. WHITE IRON REMELTED. Strength. Shrinkage. Dead Test No. Mixture. Phos Added. Load Def. Chill. Dead Load. Impact. i sq. in 5xA in. 370 White iron alone. none 379 237 .14 2 4 8 246 377 t J-7 1 178 . 13 2C-I 5 11 " 399 365 .22 .183 .244 X -25 J6 " 420 331 .22 .186 .2 4 8 1-75 57 479 340 23 .208 .250 white TABLE XXI. F'-M -{- STEEL WIRE. Strength. Shrinkage. Test No. Mixture. Phos. Added. Dead Load Chill. Dead Load. Impact. Def. isq. in. i X A in. 441 F L M and wire. none 362 339 .27 .168 .186 .40 442 it " 409 404 23 175 .216 .70 443 *' 4 ' 444 272 .19 .203 .225 2.00 444 " Cl 493 330 .15 239 .258 white 445 " " 533 437 .16 .258 .264 " 446 549 526 .17 .258 .263 TABLE XXII. GRAY F L M -\- PHOSPHORUS. Strength. Shrinkage. Test No. Mixture. Phos. Added. Dead Load Chill. Dead Load. Impact. Def. \ sq. in in. 417 FLM 4- phos. of iron. o 08 409 491 .29 .174 .I8 7 25 418 0.50 461 458 .25 .166 .184 55 419 ii I.OO 382 280 .15 .168 .I8 7 .5 42O < < 1.50 384 220 14 .I 7 8 .191 .90 421 < 2.00 2 7 6 137 .09 .192 .190 white 422 2.50 223 IOI .08 .192 .186 412 P. 0.26 4^3 4H P. 0.50 P. i. oo CAST IRON. P. 1.50 416 4 i 6 2 P. 2 00 P. 2.50 FIG. 31. 10.22$ Iron Phosphide was added to White Iron, which diluted all elements same as Fig. 30, except P. The analysis of 412 was T. C. 3.00, Si. 0.25, P. 0.26. All other percentages are estimated. 382 383 384 385 386 387 T. C. 3.97 G*C. 2.81 C. C. 1,16 Si. 1.16 P. .08 T. C. 2.61 G.C.2.37 C. C. .24 Si. 1. 12 P. .08 T-C.3.37 G. C. 2.32 C. C. 1.05 Si. 1.04 P. .08 .3.57 cc. 2 ' 77 Si. i. 06 P. .08 FIG. 32. FLM Gray Iron (Remelts). Runners returned with new F L M. The carbons of all with Si. and P. of 382 are by actual analysis. All other percentages are estimated. NOTE. The shadme of the cross-fmctures of all figures in this chapter should be carried to the boundary of each square The following questions presented themselves: First, what effect would this manipulation of returning sprues to be remelted and the long-continued heat exert on the test-bars ? Second, as 441 PHOSPHORUS IN CAST IRON. 442 443 444 445 446 "it FIG. 33. Wire added to F^M to dilute all elements as in Fig. 34. The per- centages of 4*1 by analysis were T. C. 3.55, Si. 1.20, P. 0.08. All other elements were diluted as in Fig. 34. 417 P. 0.08 FIG. 34. 10.22$ Iron Phoshide was added to F L M, which diluted all elements, same as last, except P. By analysis 417 was T. C. 3.55, Si. 1.30. All others estimated. the phosphide consists of 90$ pure iron, what would be the effect due to the introduction of this iron without the phosphorus ? CAST IRON. To find the influence of the manipulation a series of test-bars was made of the pig iron alone (Method 12), Table XVII, Fig. 29, and Table XX, Fig. 32. To find the influence of the iron of the phosphide there were made series Table XVIII, Fig. 30, and Table XXI, Fig. 33. Figs. 29 and 32 show that by the remelting of the sprues that were returned silicon was decreased at each cast, and that the increased size of blow-holes was due to this cause. Figs. 30 and 33 show that the iron additions (from the phos- phide) further decreased the silicon, which still farther increased the number and size of blow-holes. While in Figs. 29 and 32 the carbons remained substantially uniform, Figs. 30 and 33, in which wire was added, decreased in both carbon and silicon. In 440 the iron became so sluggish that the author could not carry the series farther. In the phosphide series, Figs. 3 1 and 34, which are exactly like the series Figs. 30 and 33 except that the phosphorus con- tents steadily increase, the blow-holes do not increase and are less in number even than in Figs. 29 and 32. This proves that phosphorus lessens the tendency to form blow-holes, because the iron remains fluid longer. For a further proof of this some samples of ' ' Durham ' ' white pig iron were received which had been made for the basic steel process. One had P. 3.57 (white), Si. 0.06; another (white) TABLE XXIII. WHITE -f- DURHAM PIG IRON. Test No. Mixture. Calcu- lated % Phos. Strength. Dead- load Def. Shrinkage. Chill. Dead Load. Impact. \" ."XaV 37 6 484 483 482 481 481* 0.25 0.50 1.50 2.50 3-50 3-57 379 314 296 257 205 175 237 153 I6 9 101 101 .14 . IO .09 .08 .06 05 .2 4 8 237 .211 195 .188 .182 .246 .224 .205 193 .184 .182 li \* \< White -f- Durham Do. Do. Do PHOSPHORUS IN CAST IRON. 73 P. 3.52, Si. 0.32. A test-bar was made from the sample with only 0.06 silicon and 3.57 phosphorus and got a perfectly solid casting. Each number of the Durham series, Table XXIII (Method 10), was perfectly solid, but white, and would have made a picture like the one next to 440. If it had not been for the phosphorus, the low silicon of the Durham series would have caused the castings to be very porous. During the summer of 1888 a man at Moxahala Furnace dug up in an old stock-yard two samples with 4.71$ of phosphorus and 3.45$ of silicon. About the same time the author procured a 2-lb. sample of gray Hafnden charcoal-iron made in 1854. Professor Locke's analysis of this sample from Vol. V. of the Ohio Geol. Report gave: Gr. 1.80, Cd. C. 0.50, P. 4.22, Si. I.93- The author procured I ton of pig made from Hamden ore (6.90$ phosphoric acid) with P. 4.59, Si. 3.16. From Hamden iron, with F L M, was made the Hamden series, Table XXIV, Fig. 35 (Method 4). TABLE XXIV. FLM -f HAMDEN 4.59 P. Test No. 479 457 458 459 460 461 462 463 464 465 480 467 Mixture. Phos. Added. Strength. Dead- load Def. Shrinkage. Chill. Dead Load. Impact. i" D i"xA" FLM + Hamden 25 .50 1. 00 1.50 2.OO 2.50 3-00 3-50 4.00 4.50 4.59 4.71 361 372 362 349 347 313 264 280 260 212 194 174 441 33 6 300 271 231 219 I 9 8 169 130 H3 IOI 101 32 .22 .18 .17 .16 15 *3 13 .11 .10 .10 .09 .154 .142 .137 137 .144 .143 .144 .145 .147 .141 .142 .146 .155 .156 .154 .150 .152 153 .1:3 .155 .161 .155 .163 .156 .07 .07 .12 15 .15 15 15 ib .15 Do Do Do Do. Do. Do Do Do .... Do 74 CAST IRON. 479 P. 0.25 i IPtQ. 35. F^M and Hamden Series. F^M pig had Si. 1.25, P. 0.08. Hamden pig had Si. 3.16, P. 4.59 per cent. Enough " Pencost " with Si. n.oo per cent was added to make 3.16 Si. in each. 479 is F L M with Si. 3.16. 480 is Hamden. 467 is Moxahala. For a more delicate test of phosphorus the author obtained 50 Ibs. of "Norway " gray iron with 1.65$ of phosphorus (Si. 2.03, S. o.oi % Mn. 0.87, Gr.C. 3.12, Cd.C. 0.63). To dilute the PHOSPHORUS IN CAST IRON. 75 phosphorus of this iron he used " Stewart " pig iron (P. 0.092, Si. 1.99, S. 0.015, Mn. 0.50) (Method 9). 123 P. O.OI 124 125 126 p. 0.25 p. 0.50 p. 0.75 127 128 P. i.oo P. 1.25 FIG. 36. Stewart and Norway Series. 123 is Stewart, 118 is Norway. Of each iron the pig used was the one which was analyzed, and each iron was of nearly the same constitution, ex- cept phosphorus, and" of this element " Stewart " has less, and Norway more, than would be found in ordinary cast- ings. Therefore a mixture of these irons gives us a series corresponding very nearly with the metal used in ordinary foundry practice. This series, Figs. 36 and 36^, is free from all objections. As there are thus five series from which to formulate conclusions regard- FIG. 36^. ing the influence of phosphorus in cast iron, and as series is made on an independent plan, we may place each con- CAST IRON. fidence in conclusions based upon them. In any test, consid- eration is to be given, not to individual variations, but to the general tendency of a whole series. TABLE XXV. STEWART + NORWAY PIG IRON. Test No. Mixture. Phos. Added. Strength. Dead- load Def. Shrinkage. Chill. Dead Load. Impact. *"D i"x T y 123 124 125 126 127 128 I2 9 118 0.01 0.25 0.50 0-75 I.OO 1.25 1.50 1.65 385 392 407 362 388 408 345 406 466 432 401 262 339 433 246 356 35 3i .26 .18 .22 .29 17 .21 .156 151 .149 .152 .149 .151 153 153 .161 .158 154 158 157 159 .159 .158 .04 17 .10 02 .02 .20 Stewart -f- Norway. . . . Do Do Do Do Do. TABLE XXVI. PHOSPHORUS IN THE 19 SERIES OF A. S. M. E. Series. Per Cent of Phosphorus. i"ti i"D i"Xa" 2"C1 3"D 4"D I .211 .213 .214 215 .216 .216 2 273 .269 .270 .271 .272 .270 3 4 .270 .284 .267 .283 .268 .281 .267 .280 .266 .283 .267 .281 5 333 331 330 327 .325 329 . 6 .300 299 .296 .298 .297 .299 r 7 .201 .199 .197 .199 . 198 .200 8 . 164 .161 -I6 3 .163 .160 .l6l 9 10 .258 .211 .260 .218 253 .218 .251 .220 .250 .222 .261 .219 ii .264 275 275 300 255 .283 i 12 301 .296 -.295 .300 .299 303 14 T-J .809 .826 .801 .828 795 .830 .800 8l7 .80 4 goQ 797 2 e * J 15 .980 975 .972 . 01 / 972 O j\J .971 0^3 973 C. O. Bretting & Co 16 ^OQ aan 3OT . ive. 198 iin Mich. Mall. I. Co 17 J'-'V jj-~ j~ i j-j j~ j j-" Average .222 per cent Bement, Miles & Co . * / 18 " 342 " A. Whitney & Sons 19 jf~ 352 PHOSPHORUS IN CAST IRON. 77 Table XXVI gives the analysis of phosphorus in each size of test-bar in each mixture that was made for the Testing Committee of the A. S. M. E. (Method i), and when compared with the tables of physical properties of these same test-bars, especially as compared with the Southern series, it will be seen that if c of phosphorus does no harm. Influence of Remelting on Phosphorus. There is very little change. See analysis of Original Irons in Chapter III, and com- pare with Table XXVI; also see Table LXIII. The Influence of Phosphorus upon the Grain of Cast Iron. Phosphorus produces a most peculiar grain, and of such a char- acter that when once observed it is always readily recognized. The 10.22$ phosphide, which is, as nearly as possible, iron and phosphorus, has a flat peculiar fracture, with each grain standing alone, and appearing as though it could be separated easily from those next to it. It is no doubt this peculiarity in the crystallization of iron and phosphorus which causes the weak- ness of high phosphorus irons. The color is very different from that of the fracture of wrought iron or steel. It is almost white, with a tendency to straw color. If broken when hot, the fracture was beautifully iridescent, shading from brilliant blues to bright gold. In many places there are cavities filled with fine needles of rich colors, the sides of which are brilliantly smooth and straw- colored. White and straw seem to be the colors imparted by phos- phorus. In all further examinations of grain this peculiarity of phosphorus as to color must be remembered. In the Hamden iron the groundwork of the face of the fracture is light-colored, with a very slight yellow tint, and apparently partly imbedded in the surface are what appear to be round grains, either intensely black and shining like minute glass beads, or else reflecting light so as to have such an appearance. The fracture of the gray F L M iron exhibited a dark spongy grain. The grain of each individual member is such that the grain of a series gradually changed from one of these extremes to the other. ?8 CAST IRON. By rubbing the finger across the fracture of the last member of the series, the grains do not cling to each other or to the finger, as in a strong tenacious iron, but to some extent feel as though they were rubbed off. In an iron deficient in graphite, or containing practically all of its carbon in a combined state, phosphorus has but little effect on the grain. Iron so deficient in silicon that the carbon is almost wholly combined tends to a lamellar fracture. Judging from the Durham samples of white iron, we are led to think that phosphorus rather increases this tendency. Incidentally it may be mentioned in this connection that in irons even as low as from i to 2% of phosphorus. We must therefore ascribe the reputation of some of them largely to the phosphorus and not wholly to the silicon which they contain. It is new to consider phosphorus in this light in foundry prac- tice, but it must be remembered that from \% to \% will do all 8o CAST IRON. that can be done in a beneficial way, and that all above that amount weakens the iron without corresponding benefit. The Influence of Phosphorus upon the Strength of Cast- iron. All of these tests show that while phosphorus of itself, in whatever quantity present, weakens cast iron, yet in quantities less than 1.5$ its influence in this direction is not sufficiently great to overbalance other beneficial effects which are exerted before the percentage reaches \%. Probably no element, of itself, weakens cast iron as much as phosphorus, especially when present in large quantities. The Influence of Phosphorus upon the Chill of Cast Iron. In the F L M series, Table XXII and Fig. 34, silicon decreases with each addition, and as a consequence the depth of chill increases. In the F L M and Hamden series in cast 479, which is of F L M, with its silicon brought to 3.16$, the chill is only a little over o. 10 of an inch deep. The chill is only slightly more in the case of 480, which is of Hamden alone, with 3.16$ of silicon. The chill of each cast of the series does not materially vary from that of the others ; 467 is a cast of Moxahala pig iron which was made from a different ore, and where the furnace was running regularly. It shows less chill than Hamden iron, while its phos- phorus is higher. In the Stewart-Norway series the chill remains the same, in fact it decreases as phosphorus increases. Phosphorus and Hardness. In the chapter on Hardness and that on Silicon Irons I have shown that the ordinary cause of hardness in cast iron is combined carbon, and that under ordinary conditions silicon may exert a softening influence, but that exces- sive silicon causes hardness. The tests reported in this chapter show that phosphorus does not ordinarily harden cast iron, probably because it does not increase combined carbon. Further experiments might show exceptions, for the 10.22$ phosphide is hard, while wrought iron cast alone is soft. Hamden is hard, as also is Moxahala. PHOSPHORUS IN CAST IRON. 81 The mixture of the hard Hamden with soft 479 shows hard- ness to increase as Hamden increases in quantity. The hardness records as a whole indicate that phosphorus is neither a hardening nor softening agent; but if it is either, the tendency is on the side of hardening. Phosphorus and Fluidity. The conclusion reached from these tests is that the fluidity of the metal is slightly increased by phosphorus, but not to any such extent as has been ascribed to it. By watching the cooling of the metal in the mold after being poured, it was found that the time required for freezing varies directly with the percentage of phosphorus, the phosphide cast holding its heat longer than any other. This property of remaining long in the fluid state must not be confounded with fluidity, for it is not the measure of its ability to make sharp cast- ings or to run into the very thin parts of a mold. Generally speaking, however, the statement is justified that, to some extent, phosphorus prolongs the period of fluidity of the iron while it is filling the mold. General Remarks. The endeavor has been to exhibit the action of phosphorus separated from all other influences. But we must not expect that a given percentage of phosphorus will behave at all times as it has done in these tests, for other elements may be present in such a way as to modify results. Are the favorite irons high in phosphorus ? The old Scotch irons contained about i$. The foundry irons which make the best thin castings in the Eastern States contain, as a general rule, over \% of phosphorus. From an extensive examination of English foundry irons Professor Turner concludes that the best foundry iron should contain about \%. It has always been notice- able that the irons which are rejected by the rolling-mill people on account of phosphorus are most acceptable to the founder. American pig iron will rarely impart to castings more than this percentage. CHAPTER X. SULPHUR IN CAST IRON. SULPHUR exists in nature as a brittle solid, of lemon-yellow color, and has a specific gravity of 2.05. It is of volcanic origin and alloys with pure iron in all proportions. Very small percen- tages are said to cause wrought iron and steel to become brittle at a red heat, though strong at a white heat and when cold, which peculiarity has given to this kind of brittleness the name of red-shortness. All mineral fuel and many ores of iron contain sulphur, therefore most cast iron contains a small percentage of sulphur, which in gray cast iron is very small indeed, as carbon seems to prevent its absorption. The term red-short cannot be correctly applied to cast iron. In 1886 Professor Turner said: "We are still in need of exact information as to the influence of sulphur in cast iron." Almost without exception, writers on the subject say that sulphur in cast iron will cause it to be white and is in every way injurious. All founders believe that a small amount of sulphur in the fuel will work great damage, and that if any castings crack, or if anything out of the general run occurs, i maybe charged to sulphur in the fuel. Fuel rarely contains mrr than iti of sulphur, and if this can produce so marked au effect, one would suppose it an easy matter to introduce sulphui into cast iron, so as to produce castings containing different desired percentages of sulphur. As carbon is increased toward saturation, as in white cast iron, about 0.05$ of sulphur can be retained; and as silicon 83 SULPHUR IN CAST IRON. increases, changing carbon into graphite, less sulphur is found in the iron. Writers say that the influence of sulphur in all cast iron -is to drive out carbon and to increase chill, to increase shrinkage, and as a general thing to decrease strength. By melting iron of uniform composition in a crucible, and adding brimstone, the following results were obtained (Table XXVII, Method 4): TABLE Test No. Mixture. K^ Shrinka ^ 41 With no sulphur. 411 .I2O 48 With sulphur 432 .160 . 47 " " 358 .176 46 .... 263 .207 By three tests, melting the same kind of iron as above in a. crucible, and adding fluor-spar free from galena to one crucible, and to other crucibles adding almost pure galena (containing over 14$ sulphur), Table XXVIII (Method 4) was obtained. TABLE XXVIII. Test No. Mixture Dead-load Strength. Shrinkage. 51 c e With good fluorspar W^ith galena. 367 260 137 ifis ^6 jfiT . ID;, I7O To soft iron from a cupola were added, in one ladle, iron pyrites (containing 39.88$ sulphur); in another "blue billy" (with S. 9.15$); in another were placed selected brown spots from coke supposed to be rich in sulphur; in another, the purest coke that could be selected ; and in still another, iron ore of practically the 8 4 CAST IRON. same composition as that in the pyrites and "blue billy," only with no sulphur. Each of these materials was ground to a fine powder, and small foundry ladles were lined with a pound of each, and into each ladle 15 Ibs. of iron was poured from the cupola. The result was Table XXIX (Method 2). TABLE XXIX. Test No. Mixture. Strength. Dead-load Deflection. Shrinkage. Chill. Deafc Load. Impact. 60S 603 604 60 5 606 60 7 342 332 346 380 274 255 271 322 271 345 119 169 .18 .18 .18 .19 .08 .07 151 154 153 159 .208 .194 15 .10 . 10 13 White Sulphury coke Iron ore " Blue Billy " This test shows that iron pyrites imparts its sulphur to gray iron and turns it white. In this series of tests the iron alone showed an unusually high shrinkage and chill, because of the ladles being freshly lined and not having had iron caught in them before the test. To find whether this would influence the iron, three ladles were freshly lined and dried. In one of these the iron was caught, and the ladle was then allowed to become cold. Another was used contin- uously, and the third was not used prior to the test. A set of bars was then poured from each ladle with the following results, Table XXX (Method 2): TABLE XXX. Test No. Dead-load Strength. Dead-load Deflection. Shrinkage. Chill. 624 Hot dried ladle.. . 335 .18 143 .07 623 Cold " " ... 386 .20 .151 .04 622 Green fresh ladle. 369 .19 .156 .06 SULPHUR IN CAST IRON. This series shows that conditions often overlooked may influ- ence castings more than chemical constitution. The iron in all the ladles was of exactly the same composition, for it was caught in a large ladle and at once poured into the small ladles. In the endeavor to make a high-sulphur pig iron (Method 4), 25 Ibs. of F L M gray iron was melted in a crucible, and fed slowly with 8 Ibs. of brimstone, keeping it as closely covered as possible. The operation occupied about two hours, and when completed there was at the bottom about 5 Ibs. of silver-white iron, covered with a very rich sulphide of iron. This white iron showed, by analysis, 0.58 sulphur, which was conclusive proof that a carbonized iron rich in sulphur could not be made artificially. Messrs. Weston and Smith, in Dr. Percy's laboratory, found that it was difficult in any case to cause cast iron to take up sul- phur, and they did not say that it turned the iron white. With our. knowledge of the action of silicon gained from the preceding pages, their experiments prove rather the reverse. The author made a series of tests with F L M gray pig iron, Table XXXI, adding brimstone to the melted metal (Method 7). This series was remelted and the results of the original are given in heavy figures, and of the remelt series in light figures, side by side. TABLE XXXI. Test No. Mixture, Crucible. Strength Dead- load De- flection. Shrinkage. Chill. -SO .60 50 .70 .65 .70 .70 .70 .55 -60 .50 .70 .70 .40 Dead Load. Impact. Square Sar. Flat Bar. 253 265 266 267 268 269 270 F^M and o.ooS Do. o.ioS Do. 0.308 397 -413 .378 .396 395 -397 081 -3$c .305 -333 .322 .367 .384 -368 333 .384 .209 .330 .362 .322 351 .271 .23 .21 .23 .23 .26 .25 ,-22 .25 .17 .23 .22 .21 .26 .18 .166 168 .169 178 .172 181 .179 186 193 193 .192 193 193 194 .221 .226 225 225 .231 .231 .219 212 .242 .223 .228 .222 .237 .220 Do. o.SoS Do i ooS. ... 3" 369 345 -336 374 -34 Do. 2.ooS The amount of brimstone necessary to produce so great a percentage of sulphur required, towards the last, two or three separate additions, and consequently the burning of two or three 86 CAST IRON. hollow plugs in the metal. The same metal (F L M) was melted, and the insertion of two empty plugs gave the results shown by the heavy figures of Table XXXII. The remelt is shown by the light figures. The grain of the remelt of 254 is shown in the last number of Fig. 40. TABLE XXXII. Test No. Iron. Strength. Dead- load De- flection. Shrinkage. Chill. Dead Load. Impact. Square Bar. Flat Bar. 254 F^M o.o S. with 2 plugs .425 -408 339 -369 .33 .26 .180 .166 .179 .250 0.40 .60 Fig. 37 shows the fracture and chill of each of the remelted bars in Table XXXI. 253 265 266 267 268 269 270 S. o.oo S. o.io 8.0.30 8.0.50 S. 0.80 S. i.oo S. 2.00 FIG. 37. F L M + Brimstone. While the strength decreases slightly with each addition of brimstone, the shrinkage shows a decided increase, and this increase remains constant in the remelts, while in Table XXXII the shrinkage of the remelt returns to its original figure. The remarkable thing is that the chill is not affected, and the grain, to say the least, is as even after the addition of sulphur as before. The following series of tests are on a different plan. Sulphide of iron was introduced into molten cast iron from a cupola. SULPHUR IN CAST IRON. As the sulphide would not give up its sulphur when remelted, it formed a compound of carbonized iron and sulphide of iron. The sulphide was only in the iron a very few moments before it was put in the molds, and therefore most of the sulphur is prob- ably in the test-bars. A number of series were made in this way. In ordinary foundry practice this union would not occur, but this is the only way to cause a large percentage of sulphur to .stay in carbonized iron. The first sulphide that was used contained 25$ of sulphur and 75$ of pure iron. It had a shrinkage, when cast alone, of square bar .107 and thin bar .125. Enough sul- phide was placed in six foundry ladles to give the desired per- centage of sulphur to the castings. Enough 11$ ferrosilicon was added to keep the silicon the same as in the cupola iron, which was about 2.50$. Enough scrap of the same composition was also added to make the cooling effect the same in each ladle. The iron was caught from the cupola in a large ladle and at once divided among the small ladles. The materials added were cemented to the bottom of the ladle with fire-clay. As the added portions melted, they came to the top through the molten iron and remained there. The bars poured last from a ladle showed that more and more of the sulphide was absorbed as more time was given. The test gave the results of Talkie XXXIII (Method 2). TABLE XXXIII. Test No. Mixture. Strength. Dead-load Deflection. Shrinkage. Chill. Dead Load. Impact. Square Bar. 625 626 Cupola iron, o.ooS. . . Do. .108... 397 422 .21 .21 r .i5i .152 .04 -03 299 627 Do. .308... 410 352 .21 .152 .06 628 Do. .508... 400 339 .21 .150 03 629 Do. .8oS. .. 385 307 .19 157 05 630 Do. i.ooS... 340 271 .19 .165 .08 88 CAST IRON. In the 0.80$ set of bars the first bar had a shrinkage of .153,, and the last .165. In the \% bars the shrinkage of the bars was .153, .169, .168, .175, and .174, showing that sulphide was absorbed as time was given, or else that the iron highest in sul- phide did not run out first. The influence of cooling in the ladle with no sulphide present produces a contrary effect, as is shown, for example, in the following five sets, Table XXXIV. The first bar was poured when the iron was first caught; the second when the iron was half cooled down ; and the third when the iron would just fill the mold. TABLE XXXIV. No SULPHUR. Test No. Dead-load Strength. Shrinkage. Very hot. Partly cooled. Would just run. Very hot. Partly cooled. Would just run. 22 23 24 25 26 366 427 392 441 413 424 456 427 450 418 478 450 473 427 .156 .120 .151 154 .157 157 .128 .148 .154 .162 .142 .121 135 .151 .170 Table XXXV is exactly like the last sulphide series (Table XXXIII) except that the additions were pounded fine and were thrown into the ladle loose, just before the iron was poured in (Method 2). In this series the sulphide was melted more quickly and formed a more perfect alloy, though a considerable amount floated on top. The grain and chill are shown in Fig. 38. TABLE XXXV. Strength. Shrinkage Dead-load Test No. Mixture. Deflection Chill. Dead Load. Impact Square Bar. 614 Cupola iron and o.oo S. 392 299 .21 .144 .01 609 .ioS. 395 282 .19 .144 .04 610 .308. 365 254 .20 .I6l 03 611 .. .508. 382 237 .19 173 05 612 .8oS. 348 212 17 .178 .10 613 i.ooS. 312 225 .16 177 .10 SULPHUR IN CAST IRON. 89 614 S. o.oo FIG. 38. Cupola Iron + S. A series of test-bars was made in a crucible using F L M gray pig iron (1.25 silicon and 3.50 carbon), adding a sulphide of iron containing 22% of sulphur. This, unlike Table XXXIII, had no silicon added to keep the silicon uniform, and therefore by the addition of the iron sulphide both silicon and carbon were diluted. The tests gave the following record, Table XXXVI (Method 1 1): TABLE XXXVI. Strength. Shrinkage. Test No. Mixture, Crucible. Dead-load Deflection Chill. Dead Load. Impact. Square Bar. Flat Bar 429 F L M. with o.oo S. 335 330 23 .166 .200 .60 430 .loS. 344 322 .20 193 234 65 431 30S. 364 263 .14 .208 .240 1-25 432 .8oS. 389 212 .12 245 .244 all 433 i. oo S. 422 305 .14 .241 .236 all 434 1.50 S. 410 2 4 8 .14 245 .244 all 435 2.00 S. 366 212 .18 .242 .241 all Fig. 39 shows the grain of this series. The decrease of silicon and carbon along with the increase of sulphur introduces a complication ; and to find out which produced CAST IRON. the change in strength, shrinkage, chill, etc., examine the results shown in Fig. 33 and Table XXI (Method 13) in Chapter IX. 429 430 431 432 433 434 435 T. 0.3.55 T. .3.46 T. 0.3.37 T. C. 3.28 T.C. 3.19 T. C. 3.11 T. C. 3.02 Si. 1.20 Si. 1.17 Si. 1.14 Si. i.n Si. 1.08 Si. 1.05 Si. 1.02 P. 0.08 P. 0.08 P. 0.07 P. 0.07 P. 0.07 P. 0.06 P. 0.05 S. 0.04 S. o.io S. 0.30 S. 0.80 S. i. oo T. 1.50 S. 2.00 \\\ FIG. 39. F L M + S. If we subtract each member of Table XXI from the corre- sponding member of Table XXXVI, we shall find approximately the change due to sulphur. A series was made (Method 11) with white pig iron in a crucible using the 22% sulphide. The white iron contained o. 186 silicon and 2.98 carbon, and the additions diluted the silicon and carbon as sulphur increased. This is shown in Table XXXVII. TABLE XXXVII. Test No. Mixture. Strength. 11 Shrinkage. Chill. Dead Load. Impact. a silicon iron. TABLE XXXVIII. Strength. -a c o Shiinkage. Test No Mixture. i U Chill. Per Cent. Dead Load. Impact. C/3 .056 3 "D 4" D r I 054 .050 .046 .049 .041 i 2 .046 .040 .040 039 039 039 I roo uois ^ 3 4 032 045 .030 .046 .030 .047 033 .040 .036 .044 .O3O .044 5 .017 .021 .027 .031 .030 .030 I 6 -034 033 034 033 .028 .028 r 7 .029 .030 .031 033 .030 .030 I 8 .015 .Oil .010 .on .010 .009 Hinkle \ 9 .015 .on .009 .010 .010 .007 10 .021 .019 .017 .019 .020 .022 ii .030 .027 .025 .030 .022 .027 ( 12 .031 .030 033 .026 .029 -025 ( 14 .093 .096 .100 .092 .094 .090 Southern < I ^ OQI QQC .091 QQ'l .OQI .O9O * J 15 . VJ^ L .088 ^V J .093 .088 ^yj .089 . \j\j i .087 .089 Bretting & Co 16 .025 .030 .030 .029 .030 .029 Mich. Mall. I. Co.. 17 Average 0.031 per cent. Bement,Miles&Co. 18 0.052 " A. Whitney & Sons. 19 O.IOI " There is, however, not the slightest indication that sulphur is in any way beneficial. A small percentage of sulphur will get into a casting from the fuel, and chemists are accustomed to lay any unexplainable peculiarity to sulphur. In the laboratory tests that have been described, in the effort to prove that sulphur had a tendency to increase shrinkage and chill and to turn iron white, it was necessary to resort to unusual methods, and percentages of sulphur were added which can never be found in commercial pig iron or fuel. The foregoing tests were made to prove the correctness of the general opinion that sulphur was never anything but injurious, or at least to show how such an opinion originated. Having failed in this effort, if these records shall suggest that the opinion is partly a superstition, and that the gray-iron founder must look 9 8 CAST IRON. to some other cause for defects in castings, the work will do a great deal of good. If iron pyrites is put into a cupola, it will harden cast iron at once. At one time a car of limestone which we used as a flux contained pyrites. The first day the shrinkage was .133, but complaints came in that the iron was hard. The second day the shrinkage ran up to 176, and much of the iron could not be drilled. The pyrites were then discovered and the limestone changed (leaving the iron mixture unchanged all the time), and that day the shrinkage was back to 135, and the iron was soft. Sulphur in coke is not in a condition to be so easily taken up by gray iron. NOTE. The sulphur percentages given in all tables and figures of this chapter, except Table XXXIX, are amounts added. CHAPTER XL MANGANESE IN CAST IRON. THE ferromanganese used in these experiments contains Mn. 81.62, Si. 0.256, and Cd.C. 6. 153$. This cracks in small pieces when cooling. Manganese combines with iron in almost any proportion ; but if an iron containing manganese is remelted, more or less of the manganese will escape by volatilization, and with oxidation with other elements present in the iron, especially sulphur. Owing to this escape of manganese the amount of manganese given in the tables is no doubt greater than would be found by analysis. The following three casts were made, Table XL (Method 2), with three hot ladles : in the first was placed five ounces of gray scrap; in the second four ounces of white iron and one ounce of 1 1 % ferrosilicon ; in the third four ounces of ferromanganese and one ounce of the ferrosilicon. TABLE XL. 5 li> c . bo ll 11 "!* O L II .. c u 2 - rt *j o> M rt j 1 fe 'o N li a ij$ V II 1* c J3 u a |u o c/>^ Q CQ M Q 1/5 E ^0 I ;O 2.50 380 .190 .029 365 .225 .022 .121 .075 70 124 2 O 2.51 398 195 .029 323 .225 .050 123 .060 70 I2| 3 1-08 2.51 380 .165 .017 271 .015 -155 .065 99 1 60 Into each of these ladles was poured the same amount of cupola melted iron. As the cupola iron contained 2.50$ silicon, 99 IOO CAST IRON. the ounce of ferrosilicon kept the percentage of silicon practically uniform. To prove still further the influence of silicon in manganese irons, and that much of the influence exerted by manganese can in most cases be overcome by silicon, Table XLI and Fig. 42 (Method 10) are presented. TABLE XLI. ujj o c 1 d 11 u |s || |(j) c 15 5j c o i 8 a QQ Q ^w Q .;S'' j 'V^'-j 104 CAST IRON. Manganese and Shrinkage. Nearly all of these tests show- that manganese increases shrinkage. In the tests of Table XL, an increase of \% has raised the shrinkage 26%. So long as any carbon remains in the combined form an increase of silicon will drive such carbon into graphite, and this in some degree decreases shrinkage, but a high shrinkage caused by manganese seems to- be independent of carbon and cannot be taken out without removing the manganese. Much depends, however, upon how the manganese got inco the iron. If present in small quantities in the pig it may not raise shrinkage so much as if introduced by a high-grade ferro- manganese. For a soft iron without shrinkage manganese should be absent, yet it does not seem that where manganese in the pig is below .75$ or even \% its presence will ever be noticed. Strength. A general glance at the records conveys an im- pression that manganese does not influence the ability to resist a dead load, though adding ferromanganese to molten iron gen- erally reduces strength. The small percentages of manganese found in commercial foundry-irons will have little if any influence on strength. It will be seen from this discussion that it is almost impossible to determine whether manganese is a benefit or an injury. It is only with the closest calculation and care that we have been able to determine its influence at all when present in cast iron. Much that is present in a pig iron may escape during remelting and it may aid in removing sulphur which has been brought in with the fuel. Chill. Judging from these records manganese does not materially influence chill. Manganese steel is not hardened by sudden cooling, and we might therefore expect that manganese would not add to the chilling quality of cast iron. A decrease of silicon often increases chill, and this may account for the small increase in the tests. If manganese has any chilling tendency, the 20% spiegel tests and MANGANESE IN CAST IRON. 105 the Swedish irons would show it, and yet each cast of. that test shows absolutely no chill, except the very slight chill with 7.30 silicon. It even seems to act in a contrary manner, as in irons without manganese we find it difficult to remove all chill by addi- tions of silicon. The presence of manganese under certain con- ditions may possibly aid in removing chill. TABLE XLVI. Series. Per Cent of Manganese. l a i" D l"X2" 2 " a 3 " a 4 " a f I 35 36 37 35 34 36 1 2 31 30 30 3i 30 32 3 4 50 35 51 30 49 32 .46 33 5i 34 .48 35 5 .36 39 38 37 37 38 I 6 43 45 4i .40 44 43 7 47 44 .46 47 45 .48 8 37 34 37 36 35 37 Hinkle 9 10 .48 63 47 71 44 43 49 43 47 47 5i 44 ii 58 74 7i .58 .64 54 12 59 .56 53 56 54 .61 ( 14 59 .60 .62 .60 59 .64 Southern 4 13 43 .48 43 47 .40 .41 ( 15 50 49 50 50 49 51 16 . 7 .38 . 17 38 .36 38 Mich. Mall. Iron Co... j / Av j^ J i i/ v .if >i ner re nt Bement, Miles & Co . . *7 T Q " 1C4 " A. Whitney & Son 1 19 *' 1% T o Hardness. (Turner's Test.) In the tests of Table XL an increase of ifo of manganese has increased the hardness 40$. In the first two casts the hardness of the chill is 44$ greater than that of the unchilled fracture, but in the cast with manganese the difference is only 38$, again showing that manganese does not increase chill. If, however, a hard chill is required manganese gives it by adding hardness to the whole casting. This hardness is probably due to the hardness of the manganese itself and not because more of the carbon has taken the combined form. It 106 CAST IRON. seems that in trying to make soft castings with low shrinkage, manganese should be avoided. The hardness of the spiegel series is remarkable. The cast- ings with 7.30 silicon look like the softest of open gray iron, but can hardly be touched with a file, and the piece that was left in the crucible could not be cut with a cold-chisel. It was necessary to bring it to a white heat before the chisel could be driven into it with a sledge. Influence of Remelting on Manganese. Table XLVI (Method i) gives the percentage of manganese in each bar of each series made for the Testing Committee of the A. S. M. E., and in Chapter III is given the analysis of the original irons. See also Table LXIII. CHAPTER XII. SEGREGATION. Segregation. Where any element collects in greater quantity within a casting than the average throughout the mass, such element is said to segregate. The separation of graphite into the spaces between the crystals and its even distribution throughout the mass is not generally spoken of as segregation, though the aggregation of graphite into patches, as in mottled iron, might be considered as such. It is known that sulphur is never evenly distributed, but is always greatest at the point which cooled last. Excellent examples of segregation of compounds of silicon, carbon, and iron were shown at the Columbian Exposition by makers of Swedish pig iron. This iron is the purest iron that can be found, on account of the purity of the ores from which it is made. All Swedish pig iron is said to be run into an iron pig- bed, and the surface of each pig is liable to show the influence of this chilling. The ordinary Swedish pig-iron slab is 8 or 9 ins. wide by about 16 ins. long. The samples shown in the exhibit were fractured across the slab at its center and the drawings, Figs. 46 and 47, of these fractures are on a scale off of an inch to I inch. Fig. 48 is a very small drawing of a fracture of a pig of American charcoal iron from Lake Superior ores which could not be sold because buyers were afraid of the white spot at the center. F L M (Fig. 49) is a greatly reduced drawing of a pig of Swedish gray iron made by Laxa Iron Works (Ltd.), Carlsdal, near 107 io8 CAST IRON. S9E For Lancashire Process Steel. Degafors For Open- hearth Steel. Phos. below 0.024. For Crucible Cast Steel. OeNI Phos. below 0.02 For Siemens- Martin Open- hearth Stetl. For Lancashire Firing Process. IAOBI Dannemora. IAOBI Dannemora. IAOBI Dannemora. SEGREGATION. 109 Kortfors. The pigs are cast with deep notches so that they can be broken into pieces about 5 ins. long, 3 ins. wide, and 2 ins. thick. This iron was intended for crucible steel. A most remarkable example of segregation is shown in Fig. 50. The drawing is the natural thickness of the casting. This iron is from a regular cupola mixture. At the left-hand RWn FIG. 47. FIG. 49. FIG. 52. FIG. 53. FIG. 51. FIG. 50. portion of the casting there are gray surfaces, then next under each is a white section of about the same thickness, then two gray sections, and at the center another white section, all of equal thickness, making seven distinct strata in a casting fa .of an inch thick. At the right the central white section is again divided making nine strata, five gray and four white. Considering the short time for such a casting to become solid the change of conditions to produce each separate strata must have been nearly instantaneous. no CAST IRON. Fig. 5 1 is a section from the edge of a stove-cover, circular in shape and about 9 ins. in diameter. The white parts are therefore white rings entirely surrounded by a perfectly soft gray exterior. In the upper part of Table XLVII is given the analysis. TABLE XLVII. T. C. G. C. C. C. Si. P. s. Mn. Gray, exterior . 3.628 1.874 1-754 2.846 1. 00 0.04 0.501 White, interior 3.860 1-307 2-554 2.742 1. 00 0.04 0.501 Gray, part 3.010 2.480 0.95 0.35 0.90 Gray -f- white.. 2.640 2.470 1. 00 0.42 0.91 The analyses below the line are from a segregation (Fig. 52), where white beads were driven down into the mold below the casting. The chemist did not separate all the gray from the white, but made his determination from the gray alone, and then from the gray and white together. Fig. 53 shows the flat side and edge of a casting which did not run full. Small drops of white iron exuded from the rounded surface of the edge of the casting. It merely suggests a way that the white beads of Fig. 52 may have been formed. Figs. 54 and 55 show a number of sections of castings, all of true size, showing stratification or segregation in various forms. Figures marked a show the fracture at the center and those marked b show the end of a test-bar \ in. G and I ft. long. Those marked c are cross-fractures at the center of a bar I ft. long with the section shown, the stratification extending nearly or entirely the length of the bar. Those marked */ are cross-fractures of a bar \" X TTF" X 12", run from two gates on one side. No. 208 d shows a fracture through the gate, and it would be the same for 209 e, and between the gates of 210 e and 211 e. The drawings 210 e, 211 e, and 212^ show the horizontal structure of these thin test-bars. 196 . 208 e. SEGREGATION. 197 a. 209 c. 209 e ill 217 c. 2IO-II-I2 C. } Gate. 210 &~~. f . M 405 450 A, 447'. 453'- FIG. 55- mixture of Gaylord white, and 16.32$ ferrosilicon ; gray pockets extended into the white casting. Test 197 a is the same mixture with 2.82$ Si. 4.37$ Pencost (Test 397 c) shows no tendency to segregate, "'Hie the remelt of this same mixture (Test 405 c) shows very SEGREGATION. 113 marked segregation of a remarkable kind, and exactly the same occurred in the remelt of 6.54$ Pencost (Test 406 c). The tests 208 to 212 are a mixture of F L M with the 16.27$ ferrosilicon (Method No. 8). It is evident from an examination of these analyses and draw- ings that the occurrence of white spots in castings is not always due to a decrease of silicon, for the silicon is substantially the same in both the gray and the white part of the casting; neither is it due to any increase or decrease in carbon or to the presence of any other element. For many years such peculiar castings would occasionally appear among our regular castings. All we could find out was that they were made very early in the heat. One day we wanted some test-bars from the first iron and got a white core, but the molder did not know that he did anything out of the regular order. He could not produce the same the next day. The third day, however, we got the same core, and by watching found that the iron that he used was the first that lay on the sand bottom of the cupola and that it very likely boiled before tapping. It was caught in a fresh ladle and boiled in that, and was changed to another fresh ladle and boiled again. Pro- ceeding in this way we could produce the core when the iron has such a tendency. Ordinary Commercial Impurities a Benefit. The irons from which the foregoing examples were made were exceptionally pure, which is an indication that the less the .impurities the greater the tendency to segregate. In ordinary foundry practice such pure iron would not be used and the commercial foundry pig irons do not ordinarily form segregations. At the right of Fig. 36 a (Chapter IX) is a fracture from a test-bar (118) from "Norway" pig which has P. 1.65, S. o.oi, Si. 2.03, Mn. 0.87; which is about as large a percentage of these elements as will often be found, and the grain is absolutely uniform. H4 CAST IRON. In the Swedish exhibit were examples of iron quite high in manganese (Figs. 43 to 45 a) but otherwise as pure as those already referred to, but these do not segregate. Chill as Explained by Segregation. Nearly all who have expressed an opinion on the subject of chill ascribe this phenome- non to a union of silicon and carbon. It is quite certain that the best chilling irons are not those with the lowest silicon. The best chilling irons are charcoal-irons, that is, pure irons with low phosphorus and sulphur. CHAPTER XIII. STRENGTH OF CAST IRON. WHILE a member of the Committee on Testing of the A. S. M. E., the author made nineteen series of tests with silicon varying between \% and 3.5$- Records of Tests for Committee A. S. M. E. with all details appear in the Transactions of the American Society of Mechanical Engineers, vol. XVI., 1895, pp. 542 to 568 and 1066 to 1141. also vol. XVII., 1896, pp. 675 to 729. The intended size of pairs of test-bars of each series were " D, i" D> 2 " X i" and 12" long, 2" X i', i'', 2", 3", and 4" square, and 24" long, and round bars Y 9 /'' i" an d ii" diameter. The average strength of each pair of test-bars is given in Tables XLVIII and XLIX. These test-bars were made in green-sand molds lying flat, and all melting and pouring were exactly as for ordinary work. When these tests were presented to the society they were so much of a surprise that the committee wished that a parallel set of tests should be made by tension and by compression. This has been done by a committee of the American Foundrymen's Association, with Dr. R. G. Moldenke as chairman. Record of Tests of Committee A. F. A. For a series of test- bars the iron was caught in one ladle, then poured into a pool in the foundry floor. At a signal a gate was raised and all molds bedded in the foundry floor at a lower level were filled simul- taneously with iron of supposed uniform composition and tem- perature. This would insure test-bars of the most uniform structure that it was possible to make. All bars were cast on end. 116 CAST IRON. TABLE XLVIII. AVERAGE MAXIMUM DEAD LOAD (TRANSVERSE) FOR NOMINAL SIZE OF TEST- BARS, A. S. M. E. TESTS. 5 "N Iron. j u c/> d X a M X a X X "CM M X u * X a "TO X D "V "ej X a 00 X a X a X X r i 282 9 .8 I,68 7 5,962 24,112 52,555 2292 4 68 439 3,975 2 326 914 1.883 6.661 21,556 49,201 i'.8 94 520 430 3,759 Iroquois (Coke). . . . -{ 3 4 365 422 1,046 2-173 1.823 6,853 6.427 *7,459 18,015 40.632 46,744 2,369 2,329 "s'8" 481 483 4,216 I 5 6 423 454 1,113 1.962 7,838 5,826 22,136 17,202 40,247 40,422 1,982 555 504 503 400 4> J 47 [ 7 8 321 386 972 1,046 1,614 1,925 6,299 6,419 20,791 20,929 40,964 51,628 2,018 2,146 43 * 469 414 436 3,793 3,956 Hinkle (Charcoal).^ 9 10 316 432 985 936 1,665 6.440 6,274 17,823 19,363 44,114 4?,468 2,158 1,981 474 461 428 3,563 ii 424 1,036 2,104 6,091 20,406 42,142 2,047 523 468 4,066 I 12 426 1,198 2,012 7,299 26,728 42,994 !,944 446 490 3 1 6 34 Southern < 13 359 379 1,181 i, 080 2,13- 7.936 7,910 24,409 22,048 50,649 50,904 2,424 611 506 3,858 ( 15 416 1,015 8,273 21,661 53,357 Foundries < 3 358 363 1,183 1,024 2,73 2,066 8,941 6,789 30,221 20,882 65.094 44,2i7 1 18 44 1 1,287 2,342 7,919 24,526 56,079 Malleable t 1 2 2,651 12 36 41,968 QA 771 ' ,33 VT-I / 3* TABLE XLIX. AVERAGE MAXIMUM LOAD FOR NOMINAL SIZE BAR IN TERMS OF SECTIONS OF TEST-BAR " D X 12" A. S. M. E. TESTS. Iron. X D X a of X X i" n x 24" X n X D X D 00 X D :" D X 54" X X * i 282 220 2IT 186 223 205 286 234 247 248 2 326 228 235 208 200 192 237 260 242 235 3 4 365 422 285 262 2 7 2 228 214 20 1 162 159 183 296 291 259 270 271 261 264 423 2 7 8 245 245 205 277 283 454 182 I ^8 248 252 225 259 r 7 321 243 202 197 192 160 252 216 233 235 i 8 262 241 201 194 20 1 268 235 246 247 Hinkle (Charcoal) .4 9 TO II 19 316 432 424 2 4 6 234 259 299 208 "263" 251 201 196 190 228 165 179 189 247 172 93 165 168 270 248 256 243 237 ' 26l' 223 259 241 263 278 223 229 237 Southern < 14 13 359 379 270 235 267 248 247 226 204 198 109 33 305 284 241 1 15 254 258 201 208 Foundries... .,. a 358 364 2 9 6 2 5 6 338 258 276 212 280 250 '73 1 18 441 3 22 293 247 227 219 Malleable .... 17 46s 33* 008 ?88 770 *r 4 V J jy JWW STRENGTH OF CAST IRON. 117 Square bars were intended to be , I, i, 2, 2j, 3, 3J, and 4 ins. square. Round bars were to be of the same areas, and all 12 ins. between supports. Two lengths and the same shape and sizes of cross-section both D and O were made for tensile tests, using no bars larger than 2 ins. D or 2.15 ins. diameter. Other bars of the same sizes were machined down to the next smaller size by cutting \ in. from each surface, thus a I -in. D bar was machined to a bar \ in. D- In tabulating, the in. D machined section will be placed in the column of the size that it was cast (i in. Q) and the same for other sizes, for the reason that the grain is varied by the size as cast, and therefore comparison can only be made between bars originally cast the same size. TABLE L. SERIES E. TRANSVERSE STRENGTH PER SECTION \" D X 12" LONG. A. F. A. TESTS. /\ \ n 4" tf xi" n \v / 2 3 g 238 Drv " ' " " 410 31-1 247 f j J Diameter ^^ 56" 1.13" 1.69" 2.15" 2.82" 338" 3.95" 4-51" Green sand not machined 480 290 267 2 86 280 Drv " " " 287 SERIES D. TRANSVERSE STRENGTH PER SECTION \" O X 12" LONG. /\ \ 1 1 i /r i /x ij" ff 2 i'/ _// ,!// ff ^/ / Green sand not machined 45 3.51 359 348 08 n 98? Drv " " 2.91 ,68 "j ^ Diameter Q .56" 1.13" 1.69" 2.15" 2.82" 3.38" J-95" 4.51" 280 _0_ 0- Hrv 4< ** fc * 180 39 39 308 jy* 322 4 2 5 352 295 340 363 CAST IRON. SERIES A. TRANSVERSE STRENGTH PER SECTION V' 3 X 12" LONG. < r \. \ 1 1 > > LJ i" l" if ai 3" 3i 4" 08 218 180 3 216 187 1 68 262 186 176 2 g 5 ^ ^ J Diameter Q .56" I.I3'' 1.69" 2.15" 2.82" 3-3S'' 3.95" 4.51" 238 218 218 176 1 60 1 68 SERIES B. TRANSVERSE STRENGTH PER SECTION " D X 12" LONG. /\ \ II \2 7 206 212 26O 280 2J.8 2O2 221 1 80 44 O ) Diameter Q .56" I.I3" 1.69" 2.15" 2.82" 3.38" 3-95" 4-5 1 " 288 262 264 192 288 196 " " machined... 22^ 227 220 172 IQ2 181 181 SERIES C. TRANSVERSE STRENGTH PER SECTION \" D X 12" LONG. r /\ "s r~i \y 7 i" 272 705 290 660 i" 2 2i" 3" 34" 4 Green sand not machined 3 6o 380 252 279 290 270 263 257 2,2 246 223 234 228 272 216 231 214 227 211 224 208 211 2O2 193 1 99 177 " " machined . \^J J Diameter Q .56" I-!3" 1.69" 2.15'' 2.82" 3.38"' 3-95" 4-15" Green sand not machined. 365 257 248 232 308 T08 STRENGTH OF CAST IRON. 119 Instead of comparing the measured strengths of each test-bar and to assist in making comparisons, Tables XLIX and L give the average measured strength of a section of each test-bar Wb h H 4" D X 12" long. (Formula W l = , , 1 2 . 1 -, where the plain letters are for the measured test-bar, the letters sub. one are for the other size.) TABLE LI. CHEMICAL ANALYSES OF TEST-BARS l" D CAST IN DRY SAND. A. F. A. Series. Total Carbon. Graphitic Carbon. Combined Carbon. Silicon. Phosphorus Sulphur. Manganese. E 3-04 3-04 0.72 0-45 0.07 0.17 D 2.36 0.06 2.30 0.85 0.48 0.07 0.15 A 3-87 3-44 0-43 1.6 7 0-95 0.03 0.29 B 3-82 3-23 0-59 1-95 0.41 0.04 o-39 C 3-84 3-52 0.32 2.04 0.58 O.O4 0-39 It is unfortunate that only one test-bar in each series of A. F. A. tests was analyzed, and that no analysis of the pig irons entering into the composition was made. Fig. 56 is the average measured strength of a section \" D X 12" long of each size of test-bar from 4" D X 4" D A. S. M. E. Series I to 6, as given in Table XLIX. Fig. 57 is the record of Fig. 56 with the curves made regular, i.e. the influences which caused the variations in the different sizes of bars have been eliminated, but the general conditions which influenced each series as a whole are left unchanged. Referring to the records given in this chapter and especially Fig. 82, Chapter XV: Strength generally increases with each increase of silicon (up to j $), in the bars which were cast 4 in. Q. An increase of silicon diminishes the combined carbon and removes brittleness and thus increases strength. Strength does not follow a variation in silicon, but it is foundry experience that the lower, the silicon the weaker the small fastings, and the stronger the large castings and the higher the I2O CAST IRON. silicon (up to j#), the stronger the small castings and the weaker the larger castings. For large castings therefore it is the practice to use the least silicon that will produce the requisite softness. Referring to Fig. 57 (and more especially Fig. 82), by follow- ing the curve which represents the strength of the different sizes 01 1 LU01 I I 1201 II UOI I I 1401 I I (.501 i ' 1601 | I L70L I I 1801 I I 1901 U'.OO iRATIO^CUBIC INCHES CONTENTS 4- SQ IN. COOLING SURFACE 460 410 420 ' 400 3! ml 340 t 320*5 -in- s 300 o 280 * 200 160 I UOI 1 ! ].20| I I U 2'xl" ^ 2' DQUQIS" SERIES 1\ 0| I I UOI I I '.SOI I ! I.60I I ! L \ lYrTTT"K 420 400 _380 340 320 300 280 260 240 220 -- 200 -Xi prnsj 6 FIG. 56. A. S. M. E. Tests as Measured. of test-bars made from one mixture of iron, we see that with castings poured from the same iron : Strength of a $ in. Q section of each test-bar decreases as the size of the casting increases. This is because under ordinary conditions large castings cool STRENGTH OF CAST IRON. 121 more slowly than small castings. Slow cooling gives time for the grains to become larger than when the casting cools rapidly. Referring again to any one of the curves in Fig. 82 we see that, Strength of a % in. Q section of each test- bar decreases more 01 II 11 460 01 [ 1 -12 ,RAT0 K" Oj I I U -rCC'BIC jo- 01 I ! \.i INCHES 2Vl" o| 1 1 U CONTEN 2 oi I | ifl rs-^-so. a" 01 1 1 L' IN. CO 01 1 1 IS 3LINGSI .30- 01 I 1 U RFACE 0) [1.00 4" 6 1 , 460 440 o 420'C4 5 "IR DQU( )IS" 420 X 400* 1 400 1 So 4 380 1- 360 8 1 300 *?" | 340 ! 320 \1 \ 320 i 3005 \ 1 \\ 300 3 fc 280 1 \ \v 280 S *,! \ ^ S 260 u. u 240 v: ^ \VN Si \ 240 I 220 \ ^ f^ 1 ^ \ ^RJES i 200 < \ s. ^ ^ -4's '? 220 200 180 x^ 4 a t^ ^<>^ > ^ 180 180 1 U ' on f u 1' 01 1 ! 1.3 2*xl" 01 I 1 U 1 PJJJJJ 0" 01 1 1 t< ** v -y^r oi 1 1 i7oi r i LI 0! I I U 01 inooT FIG. 57. Curves of Fig. 56 made uniform. rapidly with each increase in size near the \ in. Q end of the series than near the 4. in. Q end. Also, Decrease in strength due to each increase in size of a casting is greater and more rapid with each increase of silicon. In Fig. 57 the curve from iron containing \% of silicon begins 122 CAST IRON. lowest and ends highest. With increase of silicon the curve begins higher and ends lower than curves from iron with lower silicon. The slower cooling of each larger casting causes each curve to drop quickly at first and less rapidly as the castings increase in size. An increase in silicon causes a more rapid drop through- out. It is in accordance with shop experience and general opinion that an increase in silicon weakens large castings. The fact that small castings grow stronger with each increase in silicon (at least up towards 3$) does not seem to have been noticed, probably because, until the introduction of a J-in. Q test-bar, a I -in. D test-bar was the smallest in general use. TABLE LII. TENSILE TEST. Position of Test-bar in a Casting 9" D X 18" Long. Load per a" Corner of casti Side Middle " ng 16,450 15,406 13,750 Corner o (i Side < < Middle f cast 24,600 25,000 25,400 25,400 19,700 19,850 2O.2OO 15,730 Total carbon 2.84 per cent Graphite " .60 Combined" 2.24 Silicon i.io Phosphorus. .34 Sulphur 09 Manganese.. .49 The center of a casting has a coarser grain and is weaker than the surface of a casting. This is shown by Table LII, giving the strength of two sets of eight bars which were cut from different portions of two castings, also by the compression tests of Table LVII of J-in. cubes cut from the surface and from the center of square test-bars, therefore, Removing the surface of a test-bar or casting diminishes its strength per unit of cross- section. Table LIII proves the following proposition : STRENGTH OF CAST IRON. 123 Casting test-bars in liorizontal molds gives more even strength and more uniform size than in vertical molds. (See also Table LXIV.) The A. F. A. bars were all cast on end, while the A. S. M. E. bars were cast flat. These two series of records can only be compared by finding the ratio of variation in size and strength. The bars were supposed to be the size of the pattern. The ratio of variation would therefore be the difference between the size as cast and the intended size, divided by the intended size. TABLE LIII. PERCENTAGE OF VARIATION IN SIZE OF D BARS IN GREEN SAND. Size of Test-bars Averaged. 1"D i"D 2"D 3"D 4"D Average. A. S. M.E. Number of bars averaged I 2O 2.62 38 38 3 ? 271 4.16 A F A Percentage of variation ... 7 .8e 7.^6 8 g g f. PERCENTAGE IN VARIATION IN STRENGTH OF D BARS IN GREEN SAND. Size of Test-bars Averaged. i"D i"Q 2" D 3"D 4"D Average. A. S. M. E. Number of bars averaged 120 3 Ss 38 1o6 38 268 A. F. A.- Percentage of variation Number of bars averaged 15-99 12 ,3.96 4.06 6 ^ J' 55 8-57 32 The ratio of variation in strength is found by finding the differ- ence between the breaking load of the companion bars of each size and dividing by their average strength. One reason for such a large percentage of variation in strength of the A. F. A. tests of J-in. and i-in. test-bars was that they used a ioo,ooo-lb. tensile testing-machine, while the committee of A. S. M. E. used very sensitive transverse machines. Tables LIV and LV give the average percentage of variation in strength of each size of test-bars of each size of A. F. A. bars. These include unmachined and machined bars tested trans- versely and by tension. The recorded strength of all unmachined bars tested trans- versely is the strength as actually measured, and the size of test-bar was in nearly every case slightly larger than intended. 124 CAST IRON. All machined bars were the exact intended size. The strengths of unmachined tension-bars were reduced to strength per square inch. The average per cent variation of these bars is therefore not due to variation in size, but to the inherent quality of the material. The variation in strength is as great in test-bars machined to exact size, or in tensile bars per square inch, as that of unmachined bars of the measured sizes. AVERAGE PERCENTAGE TABLE LIV. OF VARIATION IN TRANSVERSE STRENGTH, A. F. A. TESTS. J"D i"D J"D 2"Q 2|"D 3"D 4-32 2.87 3 .a6 5-5 I3-58 3*"D 4"D Gen. Av. SERIES A, B, C. Green sand bars, not machined ** " " machined 15.99 13.96 7-15 8.90 5-44 3-42 4-65 "43 3-65 i-95 8.05 4.06 2-73 1.07 3.98 20.39 .5.68 4-77 5-5 3.62 3.48 3-72 5.98 2.IO 4.27 3-53 4-55 4-95 85 1.81 4.02 7.12 5-84 4-93 3-79 7.06 SERIES A, B, C. Dry sand bars, not machined 14 ii " " " machined SERIES D, E. Dry sand bars, not machined 4.00 % c5 O O & p N O fe O <> r"> 4- SERIES A, B, C. Green sand bars, not machined machined SERIES A, B, C. Dry sand bars, not machined 11.31 9-47 14.24 3.12 9.90 5-65 10.99 7.22 4-47 5-8. 4.68 l 3- 6 3 4.36 3-34 2.49 6.17 19.23 4.00 1.02 2.6 4 1.89 5-99 2.70 3-24 1.88 2.09 4-47 7-59 1.63 4.69 1.26 3 48 3.29 4-05 2.04 2.05 8.18 6.84 2.98 4.86 3-40 8.25 SERIES D, E. Dry sand bars, not machined ... Average of all Q test-bars o " " it. 37 9.63 7-77 8.78 6.15 7.16 6-45 7.12 4.61 3-" 5.91 2.88 3 92 3-73 3-24 3-92 5-75 5-27 The variation in strength to be provided for in ordinary cast- ings is at least 50$. The average of all the bars of the A. S. M. E. series is 5.48$, and the widest variation is 20^; the average of the five A. F. A. series was 8.64$, and the widest is 26*. In all of these castings the greatest care was taken to have the castings uniform, and each test-bar was tested, while the STRENGTH OF C4ST IRON. castings of commerce often contain concealed flaws that only a test would reveal. TABLE LV. AVERAGE PERCENTAGE OF VARIATION IN TENSILE STRENGTH, A. F. A. TESTS. i : V 1? V'O i"D ro 3 "D Av. ,"D ."D 2"D Av. SERIES A, B, C. Green sand, not machined machined 7.27 5.08 o.S5 6.30 5-48 6.6! 16.42 6.31 10.82 5.99 5-42 9-22 5-^6 4.19 ,5:3 5-75 11.24 SI :RIES A. B, C. Dry sand, not ma machir chined, ed. 6. 7 8 2.18 8.14 8.69 4.29 4.17 6.56 5-45 6-33 4-97 4.II 5-46 12.47 10.78 2.6 3 9.6. 6.04 8.28 5-74 7-3 1 O O | O O 4 I? ^ v IP m > ^ M g ? u - N M < ~ M < "^ SERIES A, B, c. Green sand, not machined " " machined 6.56 3-37 4-73 5'45 6.03 5.40 5-32 5-19 5-3 9-33 "34 6.66 7.12 7 .l6 I3.8 9 6. 5 6 II. 12 8.07 SERIES A, B, c. Dry sand, not machined.. ' " machined Average of all D bars. . ' Qbars.. 7-31 7.02 6.88 4 !o6 4-25 10. 08 3 26 5-75 5-28 6.63 7.61 4.10 13.00 16.03 12. IO 3.80 12.26 7.88 12.46 6.08 9-54 7 ~6 7-39 6.49 4.64 6.19 6-45 8.44 4-93 7-23 5-53 8-47 6.05 S-'3 10 14 .0.48 7 .50 9.28 7-79 9-5 It is quite surprising that the A. S. M. E. test-bars, molded and cast flat and poured from various ladles with iron caught from the cupola in the usual way, should not vary as much in size and strength as with the A. F. A. test-bars molded and poured with every precaution to insure uniformity. The variation in strength is due to the natural uneven structure of cast iron, and not in any great degree to varying temperature of iron entering a mold, or to varying chemical constitution, or to the character of the mold. The strength of any size of casting cannot be calculated by any mathematical formula from data obtained from testing a test-bar of another size. (The ordinary formula is given just before Table LI, page 119.) A record obtained by a mathematical formula from the test of 126 CAST IRON. a bar 01 any single size would give the same strength for a given unit of section for a large as for a small bar, and the graphic record would be a straight horizontal line. The formula wouhl give for small castings too small a strength, and for large castings too great a strength. TABLE LVI. TENSILE STRENGTH PER D INCH OF AREA A. F. A. TESTS. i 1 J d. SERIES A. Side of square i"D i"D i*"n 2"D rn i"D i*" D *" D Green sand, not machined.. 15,857 13,93 12,140 12,940 10,650 15,575 13,245 ",525 10,225 Dry " not machined.. 14,840 ",950 13,840 12,285 13,420 9,795 12,124 15,872 13,075 12,000 10,55 Diam. 0.56" O 1.13" 1.69" O 2.15" C 0.56" O 1.13" O 1.69" O 2.15" C Green sand, not machined.. " " machined Dry " not machined. 16,015 14,277 13,770 13,760 13,725 12,520 13,49 11,685 13,225 1 1 ,065 12,230 10,530 14,962 15,622 12,785 13,980 13,220 II, 860 !2,45 11,485 10.675 10.715 10,140 SERIES B. Side. *"D i"D i*"n 2"D rn i" a i*"D 2"C Green sand, not machined.. 4> " machined Dry " not machined.. " " machined.. ... 17,100 16,310 i5. T 55 17,620 15.090 18,420 12,870 15.005 13,290 14,935 11,460 12,780 ",135 12,105 16,295 15,7^5 14,970 17,780 15,123 16,380 13,315 14,97 11,740 I 4, I 35 10,515 12,150 n37 Diam. 0.56" o 1.13" C 1.69" 2-15" O 0.56" O '.13" 1.69" O 2.15" O Green sand, not machined.. 4< machined.. Dry " not machined.. 16,537 16,730 15,865 17,000 16, 160 16 860 13.115 15,375 13,160 11,405 12,535 II,OIO 16,210 16,830 14,81-5 18,040 16,260 13.705 13,870 14,170 ",335 13,75 11,090 SERIES C. Side. J"D i"D ii"D 2"D i"D i"D i*" a 2" a Green sand, not machined . " " machined...... Dry " not machined.. 17,702 16,352 16,020 18,460 16,020 17,080 12,520 15,13 12,170 14,125 11,055 11,670 ",315 9,77 16,430 16,772 15,970 17,640 15,785 17,100 11,665 14,030 12,760 13,390 10,765 io,595 10,890 12 885 Diam. Green sand, not machined.. 0.56" O 17,830 1.13" 15,865 1.69" 14,170 2.15" O J2,O3O 0.56" D 17,275 1-13" 16.535 1.69" O 14,410 2.15" 11,810 Dry " not machined.. " " machined 16,402 15,93 17,720 14,045 15,850 11,590 10,430 17,850 lO^OQO 18,460 12,705 14.925 10,970 10,230 A glance at the graphic records, Fig. 57, shows how far from the actual strength such a record would be. When these results were made known in 1895, they were so at variance with general opinion that the question arose whether the STRENGTH OF CAST IRON. 127 same results would be obtained by tensile tests. It was ques- tioned Avhether the peculiar-shaped curves of Fig. 82 were not due to applying the ordinary formula to transverse tests to obtain the strength of a i-in. D section of each size of test-bar. Lira SQ. IN. 18000 JIT ijjj Mi l^ IH 2 i '- llJ !'56i J TENSILE STRENGTH, GRE i.i3 ; -|i | i. l 6U J 2.y EN SAND BARS. 17000 s (A.F.A.|)\C TESTS \ 16000 "\\ A \\ s X 15000 \\ \ \; s 14000 S \ fe '- \\ \\ 13000 v % k \ \ \ \ \ 12000 AV *\N \ s. \ N \\ 11000 %1 5 loooo 36000 B\ \ 32000 \ 28000 2 \ . . . COMPRESSION LBS. ON &"cUBES 24000 \ V FROM MIDDLE OF SQUARE DRY SAND BARS. (A.F.A. TESTS) 20000 v< X 16000 "Vy \ X X N ^-\ 12000 X ^ N ^ 8000| Til, !' (Ml .iff infl 'MM s^____ ? ( M' TfiT' """STV l?^ 1 " |JP FIG. 58. A. F. A. Tensile and Compression Tests make the same Character of Curves as A. S. M. E. Tests. The committee of 'A. F. A. have in Table LVI supplied exactly what was wanted in their complete set of tensile tests of G and O bars varying in size from J sq. in. area to 4 sq. ins. area of cross-section, and they have used two forms of tensile test- bars. Plotting the results we have the upper part of Fig. 58, 128 CAST IRON. exactly the same kind of diagrams as produced from transverse tests. Calculating the strength of one size of tensile bar from the strength of another size is by a simple proportion, but observing the decrease in the tensile strength per square inch of bars of greater area on account of slow cooling, it is seen that the proposition is proven that the strength of a casting cast one size cannot be calculated from another casting cast another size by any formula. Compression Tests, Table LVII, by the committee of A. F. A. prove the same thing. Cubes with -J in. Q sides were cut from the square bars cast in dry sand of Series A, B, and C. One cube was taken from the surface of each bar equidistant from the corners, another cube was taken \ in. nearer the center of each of the bars, and so on. The results are given in Table LVII. One cube was always taken from the exact center. At the bottom of Fig. 58 these plotted results show the same-shaped diagrams as in Fig. 57. TABLE LVII. COMPRESSION TESTS A. F. A. OF CUBES FROM EACH OF DRY SAND D BARS. Distance from Surface. j"n i"D i*"D 2 "D IPQ 3"D 3J"D 4"D c First . 12,040 II,2OO 10,770 10,340 St 4> J Second Third 17,180 13,880 11,430 10,950 10.270 10,430 9,830 9,540 9,950 9,570 11 9,360 4 f Center First . . 29.57 ^3 360 2O,OIO 17,180 24,820 13,810 21,640 10.950 18,270 9,830 17,000 9,35 15,970 9,100 16,140 Second .... 38,740 14,410 15,200 13,950 sJ Third 13.760 13,160 12,830 $ Center. 38,360 33,COO 20,980 18,130 15,060 13,790 13,160 12,430 . - f First ng r O O 19 800 18.170 16,410 u Second 20,750 19,340 18,050 16,850 16,510 15,250 gj Third 17,840 16,080 14,880 fl 15 880 14,200 II Center 38,50 24,890 20,750 18,010 i 7 ,8 4 o 15,950 15,880 14,220 Stronger castings are made in green- sand than in dry- sand molds, because castings cool slower in a dry mold. See Table LIV. Square test-bars are stronger than round test-bars with equal areas of cross-section, shown by Tables LVIII and LIX. STRENGTH OF CAST IRON. 129 TABLE LVIII. AVERAGE TRANSVERSE STRENGTH OF D AND O TEST-BARS, A. S. M. E. TESTS. t" D X 12" .56"OXw" i" D X 24" reduced to i" D x 12" ... 3 ;px 2107 A verage of 38 bars ; 2 each of the 19 series 401 362 2361 TABLE LIX. AVERAGE TRANSVERSE STRENGTH OF d" SECTIONS OF ALL D AND O BARS, A. F. A. i"D i"D *"D 2"Q 2j"D 3"D 3i"D 4"D Averages of all Q bars 363 336 280 258 2 7 2 242 242 239 56" O "3" O ,.6 9 " 2-15" O 2.82" O 3-38" O 3.95" O 4-5i"O Averages of all Q bars ... 3i3 2 6 7 281 283 261 247 222 220 T'7~ Diff. -f in favor of [J bars ... + 50 + 69 i + 25 + 11 - 5 + 20 AVERAGES OF TENSILE STRENGTHS OF D AND O BARS, A. F. A. TESTS. r rn =c ' . ^C^=^ 1 rq i"Ll *'D 2"D i"D ,D ii"D a"D Averages of all n bars . 16,360 15,834 13,403 ",234 16,109 15.531 13,003 11,219 56" .3" 15,803 i.6 9 "O 13,894 - 491 2. 15" O 56" O 1.13" O 1.69" O 2.15" O Averages of all Q bars 16,173 + i8 7 11,510 -2 7 6 16,410 291 15,697 - 166 13,573 - 570 11.427 - 208 Diff. + in favor of D bars The difference is not great, but the average shows in favor of the square bar. This shows that practically the grain of one is as uniform as the other, and it is fortunate that it is so, because the rectangular shape is more common than the round for ordinary castings, but it is well known that all corners of a pattern should be rounded and all reentrant angles should be as round as possible. For further remarks on Tensile-bars see Chapter XIX. A committee appointed to report to the Western Foundry Association, Nov. 21, 1894,* as to whether a round test-bar cast * The Iron Trade Review, Nov. 29, 1894. 1 3 CAST IRON. on end was better than a square bar cast flat, reported that in one group of tests, all square bars cast flat were perfect, while 43 # of the round bars cast on end were defective. In another group of tests, 1 8$ of the square bars cast flat were defective, and 54$ of the round bars cast on end were defective. The committee reported that they could not endorse the round bar cast on end as against the square bar cast flat. During 1894 Mr. West made a large number of tests with round bars cast on end and square bars cast flat.* The results exhibited the same differences in favor of the square bar as above stated. Test-bars 2" X i" X 36" cast flat and tested flat, and also on edge. The test-bar 2" X i", tested with the wide side down > TABLE LX. A. S. M. E. TESTS. Size of Test-bar, 2 // Xi // X36 // . No Shear- Kind of Iron. Test- bar. Max. Deflec- Stress per ing Stress Modulus of Elasticity. Resili- ence. Breadth Height. Load in Lbs. tion in Inches. D" in Outer Fibre. per D" oj TESTED NARROW SIDE I )OWN. at 359 .02 2.OO 2,800 .295 37,057 686 I5,OOO,OOO 413 366 .OI 1.98 3,000 355 40,915 750 14,900,000 533 JO, 367 .OI 1.98 3,020 350 4M73 ' 754 I4,9OO,OOO 431 T3 358 .00 1.99 3,028 386 41,300 767 16,500,000 637 364 99 1.96 3,054 .416 43,000 786 l6,OOO,OOO 635 43 361 .00 2.00 3,056 433 4L254 764 14 900,000 66> CU ^60 .00 I QQ 3,100 A QOO 6 J^\7 372 .04 2 00 3,950 .380 51,400 968 17,304,000 750 Av'ge 7 2O 077 42 2QQ 782 TC fi.17 /IOR c8r> en Oi J * 31 1 il* t +\f\f IV" * J ) Vfc tv7 ,*f ~ J~ v JO TESTED WIDE SIDE DOWN. i 363 1-97 99 ,342 623 37,600 345 16,000 ooo 4 I8 357 1.98 .01 ,404 .604 38,400 352 15,200,000 422 c 368 2.00 .00 ,438 555 38,842 359 14,100,000 399 4) 365 1.98 .02 ,480 750 38,800 282 T4,5OO,OOO 560 3 . 06 7 08^ Bement, Miles & Co *- 1 18 3-35 3-35 3-42 j VJ *-' 3-30 3-23 3-3i j) *-"- 3 3.326 A. Whitney & Sons 19 3-74 3-85 3-79 3.8l 3-89 3-86 | 3.823 i 1 i Total Carbon. The series made for the A. S. M. E., Table LXI, do not present enough data to form any conclusions. The only way to make comparisons is to compare series containing exactly the same silicon, and otherwise substantially having, the same chemical composition. Then it would take a large number of tests to prove anything, for any influence that would cause the grain to be close would increase strength independently of chemical composition, and vice versa. It is very difficult to make experiments on carbon in cast iron and preserve uniformity in the rest of the composition. It will not answer to add wrought scrap, for this will not only decrease carbon, but at the same time would decrease the percentage of every other element; and also STRENGTH OF CAST IRON. because such scrap will close the grain and increase strength, independently of the lessening of carbon. TABLE LXII. A. S. M. E. TESTS. Series. Per Cent of Combined Carbon. 1 "D i"D i" X 2" 2 " D 3"Q 4"D r I 1 .46 1-25 1.05 .80 .76 .70 2 .70 54 59 56 54 .60 i 3 4 .48 45 45 .48 .42 43 37 .36 34 . ii .13 50 i 5 35 .16 .20 . ii . IO i 6 37 38 .30 .15 .11 .08 r 7 1.24 .88 .72 53 52 .46 i 8 .67 44 .50 49 .46 .42 Hinkle { 9 10 53 .29 .42 .36 50 43 .46 37 .15 44 .11 45 \ ii 32 .12 .09 .09 .08 .08 ( 12 27 .09 .09 .09 .09 .09 ( M .26 15 .14 .09 .09 .08 Southern < I ** . 1 1 . IO .OQ .08 O7 O7 X J 15 .10 .09 \j\j .09 .09 **/ .11 . \j / .08 C. G. Bretting & Co 16 49 78 73 .49 58 44 Mich. Mall. Iron Co.. . . 17 2 8s 2 78 1 . 2O i . 20 Bement, Miles & Co .. . . A / 18 ^ " <-> J 45 52 50 * /o 24 .12 .11 A. Whitney & Sons 19 2-95 99 .81 .81 87 .89 Combined Carbon. This must have been uniformly diffused in the molten metal, to have produced such a uniform variation in the test-bars. It is the universal opinion that strength is mainly due to the combined carbon which the castings contain, and that weakness is caused by changing it into graphite, which is supposed to mechanically separate the grains. This opinion originated with the makers of heavy castings, who in making strong castings invariably used irons with high combined carbon, which is always an accompaniment of low silicon, and produces a close grain, and for this last reason gives great strength in a large casting. For example, an 8-ton anvil- 134 CAST IRON. block was made from white pig iron which contained about one half of i% of silicon, with the carbon nearly all combined. This made a very strong, fine-grained gray casting. Series 17 made white castings in the J in. O, I in. Q, and \" X 2" test-bars, but the 2-in. Q, 3-in. D, and 4-in. Q bars were very close- grained gray castings, and of extraordinary strength. <4 Iro- quois," with combined carbon 1.46$ in the ^-in. Q bar, produced a stronger 4-in. Q bar than any other of the six ."Iroquois " mixtures, with less combined carbon. Viewing the subject of strength and of combined carbon in the light of chemical analyses alone, no other conclusion could be drawn. But if the whole nineteen series of test-bars are examined, we shall see that com- bined carbon weakens castings, and never strengthens them. We shall proceed to prove, from these same series from which we have shown how the accepted opinion was obtained, that the decrease in strength of large castings is wholly due to loosely united crystals, and not to any change in the proportion of com- bined or graphitic carbon. Combined Carbon weakens Cast Iron. In each of the charts we see in the ^--in. D test-bars, that with each increase of silicon the combined carbon is decreased, and that the strength is increased in the same proportion. In the ^-in. Q test-bars of each series containing about \ of silicon, the combined carbon was about 1.50$, and the iron was weak because it was brittle. As combined carbon decreased in the J-in. [U bars with each addition of silicon, the brittleness decreased. This is shown strikingly in Series 14, 13, and 15. Combined carbon may decrease as castings are larger, but the strength always decreases. This decrease of combined carbon and of strength are botJi caused by the slow cooling, and the decrease of combined carbon has not king to do with the decrease of strength. "Iroquois," Series I, had 1.46$ of Cd. C. in the -in. D bar, which was about one half white, and Cd. C. decreased in the other sizes to 1.25, 1.05, 0.80, 0.76, and 0.70. In this case strength decreased exactly as Cd. carbon decreased (silicon and STRENGTH OF CAST IRON, 135 other chemical elements were practically uniform in each size), and as a chemist would look at it, it would appear that there could be no other reason for decrease in strength than the decrease in combined carbon, for this is the only chemical variable. The fact is, however, the lessening of the combined carbon made the I -in. G test-bar gray, and each successive decrease of Cd.C. darkened the color and made the casting more ductile; in other words, slow cooling has done for the larger sizes of test-bars of the series just what the increases in silicon did for the -in. G bars of the six series, and this should therefore have increased the strength, and it did. But the increase in the looseness of the grains on account of the slow cooling decreased the strength more rapidly than this increase of strength. Whatever the decrease in strength on account of loose crystallization was, it was lessened in Series I by the increase in strength due to the decrease in combined carbon, with the result that the larger bars were stronger than any others of the six series. A further proof is found in the various series in which com- bined carbon is the same in each size of test-bar; for example, Series 19, which was from a car-wheel mixture in which the iron was mixed in a large ladle and therefore of uniform composition ; Cd.C. remained the same in all sizes of test-bars, but the decrease in strength follows the general law. Another example proving the same thing is Series 2, "Iro- quois." The silicon has been increased about two tenths of i$, and in all but the J-in. G bars the combined carbon is uniform at about .54$, but slow cooling decreases strength in the large test-bars, exactly the same as in Series I. The increase in sili- con has, in the J-in. G bar, taken out brittleness, by diminishing combined carbon, and has thereby increased the strength 45 Ibs. This increase in silicon causes the grain to become coarse, in the larger bars, more rapidly than in Series I. The large bars grow weak faster in Series 2 than in Series I, in spite of the combined carbon not decreasing in the larger bars. In Series 1 5 the 136 CAST IRON. J-in. n bar begins with o. 10 of \% combined carbon, and there is not enough decrease in this element, in the larger bars, to make any difference in any respect, but slow cooling causes the same proportional weakening of the larger bars. A most interesting test was made by Mr. A. L. Colby, chemist of the Bethlehem Steel Co. Half of a furnace cast was run in sand and half in the iron molds of their casting machine. The former made an open-grained pig and the latter a pig of very close grain as shown in Fig. 59- A portion of each kind was melted separately in a cupola under as uniform conditions as possible, and test-bars 3^ ins. square were cast in horizontal and in vertical molds. The castings from each were exactly alike in grain. The fractures are shown in Fig. 59 and the composition and strength of each. is given in Table LXIII. Mr. Colby ascribed the great strength of the machine-cast pig to the combined carbon, which he considered the only variable. Whereas the sudden cooling in the iron molds caused the grain to be very uniform and close, while the slow cooling of the sand- cast pig produced a very coarse and irregular grain. If the same close grain could have been produced without any increase of combined carbon the strength would not have been any less. TABLE LXIII. SAND OR IRON MOLDS FOR PIG IRON. Sand-cast Pig. Machine- cast Pig. Test-bars 3*" D X 18" Long. From Sand-cast Pigs. From Machine cas- Pigs Cast Hor- izontally. Cast Ver- tically. Cast Hor- izontally. Cast V - deal y Total Carbon 3.460 3.210 .250 3.000 .770 .041 950 15.000 3.380 2.460 .920 2.990 773 .041 950 41 .000 3.400 2.930 47 2 930 .766 .071 .840 18.000 3-300 3.022 -368 2.QIO .769 .064 .850 16.300 3-364 3.028 .336 2.960 .772 .077 .840 17.000 3-35- 3. 100 257 2 . Q ^ .-> ^4 .071 .840 i 7 . ooo Graphitic Carbon Phrsphorus Manganese Tensile strength Ibs. per sq. in. ... STRENGTH OF CAST IRON. 137 FIG. 59. Showing Effect of Grain on Strength. CAST IRON. To make it possible to judge of the quality of castings from the fracture of machine-cast pig iron it would be necessary that the conditions in all cases, at all furnaces, should be uniform, and there seems no chance of this. At present there seems to be more variation in conditions than in sand-cast pig. TABLE LXIV. A. S. M. E. TESTS. Per Cent of Graphitic Carbon. Series. r n i'"D i" X 2" a"Q 3"D 4"D r I 2.36 2.60 2.8 3 3.08 3-05 3-13 2 3.20 3-32 3-24 3-33 3-32 3-31 3 4 3.21 3-28 3-24 3-32 3-27 3-40 3-41 3.56 3.60 3-55 3-25 i 5 3-19 3 40 3-24 3-47 3-38 3-42 I 6 3-oi 3.08 3.08 3-24 3-19 3-23 c 7 2.78 3-13 3- 2 3-42 3-48 3-55 i 8 3.17 3-44 3- 3 3-29 3-37 3-42 Hinkle \ 9 10 3-28 2.91 3-47 2.87 3-36 2.86 2.84 3-67 2.80 3-72 2.82 i 1 it 3-oo 3-22 3.22 3-25 3.28 3-23 I 12 3-07 3-28 3-24 3.26 3.28 3.22 ( 14 2.89 3.08 3-13 3.18 3-14 3.20 / *-xrt 3.06 "3 O*" "7 O*7 3 04 i oft ( 15 3.03 J VVP 3-oi 3-07 3.06 3-03 3-H j C. G. Bretting & Co 16 3.30 3-10 3.08 3-32 3-22 3-3i Mich. Mall. Iron Co.. . . 17 .26 . 24 I . 90 1.86 Bement, Miles & Co / 18 2.90 2. ,8 3 2.92 3.06 3-n 3-20 A. Whitney & Sons 19 .79 2.86 2.98 3. co 3.02 2.97 Graphitic Carbon. The general opinion is that it causes weakness. If C. C. decreases, the G. C. must increase ; there- fore, if it was thought that combined carbon produced strength, the same facts that seemed to warrant this conclusion implied that graphitic carbon produced weakness. Again, in graphitic iron, the grain was coarse, and the flakes of graphite lay between the grains, and it seemed self-evident that these graphitic flakes must of necessity separate the grains of iron and cut the casting up. The same proof that has been produced in the case of com- bined carbon will apply regarding graphitic carbon. STRENGTH OF CAST IRON. 139 From an examination of these series, strength or weakness seem to be absolutely independent of this element. The loose- ness of the grain, produced by slow cooling, so separates the grains that there seems to be more than enough room for the flakes of graphite to lie in the open spaces. It may be even doubted if the graphite ever gets between the grains to make their union less perfect. The graphitic scales seem to have formed in the spaces after the openings have been formed, and either act as a cushion, or the scales lie loosely in the cavities. This latter supposition seems plausible, from the fact that when pig iron, or a casting as large as a pig of iron, is broken, scales of graphite fall out in great abundance. For the influence of sulphur and manganese on strength con- sult the chapters on these elements. Annealing Castings. To produce very soft castings with very low shrinkage some founders melt only the softest No. I pig iron, and do not even use the scrap made from such iron ; while others use cheaper irons for the castings, and afterwards place the castings in an annealing oven until most of the combined carbon which they contain is changed into graphite. Instead of increasing silicon in their mixtures to cause a decrease in com- bined carbon, they prefer to anneal the castings. It would be impossible to get as low a shrinkage or as soft iron in the cupola as by this process. Table VI, Chapter V, is an example of this. All bars were poured from one ladle. The chill in the annealed castings is very dull and only half as deep as before annealing. The grain is much darker and is filled with glistening points. The unannealed thin bars broke without taking set, while the annealed thin bars took a set of over four tenths of an inch at the center before breaking. The unannealed square bars took a set of . 10 of an inch at 300 Ibs., and after annealing took a set of .20 of an inch. The annealing changed three fourths of the combined carbon into graphite and the annealing temperature was high enough to enlarge the grain, thereby weakening the test-bars. CHAPTER XIV. IMPACT. THE object of this chapter is to show the influence of impact upon test-bars of various sizes. The Test-bars are I in. X i in., i in. X i in-, I in. X I in., and j- in. X i in. in section. One set is 24 ins. long and another set is 12 ins. The bars are of tool-steel having a uniform spring-temper produced in a gas-muffle. Each bar was then ground to the exact standard size on a surface-grinder. The Recording Apparatus (Fig. 99) holds the ends of the bars in clamps which rest on bearings exactly 24 ins. or 12 ins. apart. The impulse is received by a cage clamped to the center of the bar, and the motion is multiplied five times by an arm which carries a pencil at its end, which makes an autographic diagram of the movement of the center of the test-bar. Dead Load. A single-lever machine, shown in Figs. 92 and 93. TABLE LXV. Dead Load. 25 Ibs. 50 Ibs. 75 Ibs. TOO Ibs. 200 Ibs. 300 Ibs. 400 Ibs. 500 Ibs. i in. Xi in. X 24 in. .0028 .0056 .0084 .0112 .O224 .0336 .0448 .0560 i X 1 X 24 .0056 .0112 .0168 .0224 .0448 .0672 .0896 .1120 D C XI X 24 .0224 .0448 .0672 .0896 .1792 .2688 .3584 .4480 C i x i x 24 .0448 .0896 1344 .1792 .3584 .5376 .7168 .8960 G 2 I XI X 12 .0003 .COO7 .0011 .0014 .0028 .0042 .0056 .0070 I Xi XI2 .0007 .OOI4 .0021 .0028 .0056 .0084 .OII2 .0140 1 Xi Xi2 .0028 .0056 .0084 .OII2 .0224 .0336 .0448 .0560 Q i X| Xia .0056 .0112 .0168 .O224 .0448 .0672 .0896 .1120 140 IMPACT. 141 DEAD LOAD BARS 24" LONG DEAD LOAD BARS 12" LONG Keep FIG. 60. 142 CAST IRON. Ball bearing Dead-load Diagrams (Fig. 60). It must be remembered that all of these test-bars are perfectly elastic for the loads applied to them, and for this reason dead-load deflections are propor- tioned to loads. The diagrams show that: Deflection is inversely as the breadth of the test-bar. ' " " " " cube of the height of the test-bar. " "directly " " " " " length of the test-bar. Impact Testing Machine (Figs. 61 and 98) with its Hammer Swinging on a wooden vertical arm 6 ft. long. The weight of the hammer can be varied be- tween 25 and 100 Ibs. With a swinging hammer the bar receives the impulse in a horizontal direc- tion and the bar bends until the motion of the hammer is stopped. The bar then springs back and throws the hammer away, ^ when it is caught by the hand and clasped -L to the trip for the drop from the next higher point. The dead load of the hammer is carried by the vertical arm to which it is attached. Impact Diagrams made with a Swinging Hammer. To obtain a complete record of each motion of the center of the test-bar, a shaft J in. in diameter was made to revolve 1250 revolutions per minute. The paper-holder was connected to this shaft by a cord. Just as the blow was given the shaft was caused to revolve, and the winding of the cord caused the paper to move its whole 24-ins. length in half a second. The length of the diagram of a single impulse, and of the vibrations until the bar comes to rest, is from I to 4 ins., and the pencil makes this record in from T ^ to T V of a second. The diagrams (Figs. 62 to 68) are reduced to one third actual size. Figs. 62 and 64 are swinging blows from 25 and loo-lb. hammers, both i-m. drop, on the same test-bar. The bar Anvil FIG. 61. IMPACT. TABLE LXVI. DEFLECTIONS WITH SWINGING HAMMAR. 143 Weight of Hammer. Height of Drop in Inches. o in. i inch. i inch. i^ inches. 2 inches. 2$ inches. 3 in. 25 Ibs. .0 .060 .087 .105 . 122 .130 .144 50 " .O 075 .107 .130 .150 .162 .176 75 ' .O .088 .126 .152 175 193 .2IO 100 " .O .099 .141 .17-2 .200 . 22O .240 25 " .O .077 .110 139 -158 175 .194 50 " .0 .100 M5 .174 .201 .226 .250 75 ' .0 .122 175 .2IO .241 .271 300 100 " .0 .141 .200 .241 .280 314 347 25 " .0 .110 .165 .205 237 .270 .300 50 ' .0 .168 .242 300 350 .400 435 75 " .0 .210 314 .385 .448 .508 550 100 " .0 .260 370 .460 532 .605 .665 25 " .0 .150 .225 .280 .322 -3f>5 -410 50 " .0 .230 332 .410 .470 .530 .580 75 " .0 295 .425 .525 .600 .670 .732 100 ' .0 342 493 .6lO .705 .791 .870 25 " .0 035 .048 -057 .069 .078 .084 50 " .0 .044 .065 .080 093 .107 .119 75 ' .0 -053 .080 .099 .118 .133 .149 100 " .0 .062 093 .H5 .138 .155 174 25 " .0 .046 .062 .026 .090 . IOO .no 50 " .0 .058 .084 .100 .118 .133 .147 75 ' .0 .069 .100 .120 . 140 .160 .J77 100 4 .0 .080 US .138 .162 .186 .203 25 " .0 .060 .090 . no 125 .139 .150 50 " .0 .080 US .140 .161 .182 .200 75 " .0 .100 .120 175 .2OO .223 .246 100 " .0 .119 .166 .200 234 .265 .293 25 " .0 .070 .100 .126 .148 .155 -181 50 " .0 .091 .131 .I6 4 . igi .218 .239 75 ' .0 .117 .165 .207 .238 .270 .295 100 " .0 .140 .197 .241 .280 315 -345 vibrated once and a trifle more with 25 Ibs., and twice as many times with 100 Ibs. The hammer is caught by the hand and allowed to strike the test-bar but once. By letting the hammer swing against the bar until it came to rest, the 25-lb. hammer 144 CAST IRON. IMPACT. 145 made the series of records shown in Fig. 63 and the loo-lb. hammer the record of Fig. 65. The regular diagram is made by moving the paper slightly for a base-line when the test-bar is at rest (see Fig. 63). The hammer is then raised to a catch hung in the hole of the graduated arc for -J-in. drop and let drop upon the test-bar. The record is a vertical line five times as long as the actual deflection of the center of the test-bar. The paper is then moved T \ in. and the hammer let drop J in. and so on, each drop being -J in. higher than the next preceding. Each record is continued above the base-line, because when the hammer swings back the center of the test-bar goes past its original position, and the bar vibrates several times before it comes to rest. By drawing a line through the lower ends of the record line we get a curve showing the total deflections for drops from o to the highest drop. By connecting the upper ends of the record-lines we have a curve showing the vibrations of the test- bar. Diagrams from Test-bars which are not Perfectly Elastic. Fig. 69 shows a dead-load diagram from a bar of ingot iron \ in. by in. by 12 ins. When the metal began to flow rapidly the load was removed, and the pencil rested a distance below its original position equal to the set taken by the test-bar. When the load was again applied, a new diagram was made which joins the former diagram at the point where the load was released. A diagram from a J-in. square steel bar is added, which is shown by the dotted line and coincides with the spring-line of the ingot iron. Fig. 70 shows an impact diagram from a test-bar cut from the same bar of ingot iron. With dead load the bar took no set until it gave way. With impact, set began almost at once. The upper line shows the vibration of the bar. The dotted lines show impact diagrams from the tempered-steel bar. As it took no set the line of sets coincides with the base-line. Figs. 71 and 72 are diagrams from bars of rolled puddled i 4 6 CAST IRON. Mild Steel or Ingot Iron and Steel Bar FIG. 70. IMPACT Ingot Iron and Tempered FIG. 71. DEAD LOAD Wrought Iron MM N -^ \ ! J X^ t ui 0" \ i? F,G. 72. IMPACT Wrought Iron ^~ N - > - > f ill ---^ -- s \ \ f M 'f M FIG. 75. Swinging Base Line Vibration of Bar All Test Bars on this page are M '~x. % * 12, 'all Impacts are with a 25 Ib. Swinging Hammer, all diagrams Full Size. Deflections multiplied by 5. Keep Am.Bk.Nott CO.N.T. IMPACT. 148 CAST IRON. IMPACT. 149 iron. Figs. 73 and 74 are diagrams from gray cast iron, which takes set with the smallest load, but after having been subjected to a given dead load, in this case 300 Ibs., it is perfectly elastic for less loads. In every case the deflection is greater with impact than with dead load. Figs. 76 and 78 show the deflections with drops from o to 3 ins. of a 75-lb. hammer on each of the test-bars, and Fig. 77 shows deflections from all the hammers on a i-in. G bar. Impact Diagrams made with a Hammer having a Direct Drop. Fig. 79 shows the same hammers as with swinging blows, hung on a horizontal wooden arm 8 ft. long, which allows the hammer to drop practically in a vertical line without guides. 8 feet Elevation IMPACT Direct Fall Plan FIG. 79. The hammer is fastened by a cord to a timber overhead, and is then raised or lowered by a thumb-screw until it is exactly the required height above the test-bar. The cord is cut to let the hammer drop. With a direct drop the hammer acts on the test-bar as a dead load, and also by impact. When the motion of the hammer has been stopped by the elasticity of the test-bar, in springing back the bar must lift CAST IRON. the hammer as a dead load, and toss it upward when it reaches its normal position. The test-bar when the hammer leaves it vibrates until it comes to rest. The hammer drops again on the bar, is tossed a second time, the bar vibrating until at rest, when it receives the hammer again, which finally rests on the bar, bending it as a dead load. If a loo-lb. hammer is suspended as in Fig. 79, so that it just rests on the test-bar, say J in. by J in. by 24 ins., but does not bend it, and is lowered slowly, the bar will bend .1792 in., and come to rest supporting the loo-lb. weight; but if the cord was cut and the hammer allowed to drop, it is the same as a . 1792 in. drop on the bar, less the resistance of the bar during this drop, and the bar bends nearly .1792 from the drop -|- .1792 for dead load. Fig. 66 on page 144 is a diagram from o drop of loo-lb. hammer. TABLE LXVII. DEFLECTIONS WITH DIRECT DROP HAMMER. tfl If Weight of Height of Drop in Inches. rt ^ Hammer. o in. } inch. i inch. T| inches. 2 inches. s$ inches. 3 'n- 25 Ibs. 005 .065 .092 .119 .130 145 .150 50 " .OIO .080 US .140 .159 174 .187 75 ' .015 .100 135 .162 .I8l .201 .219 100 " .020 .116 .160 .I8 7 .209 .230 .248 25 " .Oil .087 .130 155 .174 .191 .205 50 " 023 .112 .I6 5 .196 .221 .241 .265 75 ' 034 134 .191 .229 .262 .289 .316 100 ' .047 .161 .221 .266 .305 339 370 25 " .046 145 .205 .251 .285 .315 340 50 " .084 .230 .317 376 .420 454 .483 75 " .126 .300 .415 .485 531 571 .611 IOO ' .171 359 .482 064 .620 .670 715 i 25 ' .070 .206 300 .365 .418 .456 . 4 S2 50 " .141 332 453 525 578 .618 6 5 2 75 ' .230 453 .586 .663 .720 .766 .800 IOO " 333 .566 .702 .790 .8 5 I .900 .941 *" 1 ! IMPACT. 1 152 CAST IRON. Fig. 67 is a diagram showing the influence of a single direct i -in. drop of 25 Ibs. as compared with the influence of a single j-in. drop of a 25-lb. swinging hammer (Fig. 62). The diagram (Fig. 80) was constructed from actual records of direct drop. Impact Tests. To determine the resilience of a material, support a test-bar at the ends and deliver blows at the center. After a test-bar has been tested in this way to find its resistance to impact without any distortion as a test of brittleness, a portion of the same bar should be clamped on the anvil of the testing- machine so that one end shall project (see Fig. 75, p. 146). Blows should be delivered on the projecting end as far from the clamp as i j- times the depth of the bar. An inch bar would receive blows if inches from the clamp and a J-in. bar f in. from the clamp. Size of Test-bars for Impact. Some one ^Softest-bar must be selected for comparisons. The size for cast iron which would seem to give the best results is a bar I in. by I in. by 24 ins. struck with a 5O-lb. hammer. This has the same relative pro- portions as a bar j- in. by f in. by 12 ins., and if a 25-lb. hammer is used for the latter, the record is the same as with a i-in. by I -in. by 24-in. bar with a 5O-lb. hammer; but this does not take into account the change in grain due to size of casting. Impact with a Swinging Hammer. On account of this giving simple impact unmixed with dead load and other modifications which accompany the direct drop, the swinging hammer appears to be the best mode of application of impact for ordinary test-bars. Its convenience is greatly in its favor. It does not affect the surface of the test-bar. The height of drop is exact to in. It can be operated by hand rapidly. Blows should be begun with the same drop at all times, which should be less than the lowest possible breaking drop, and then each drop should be increased by \ in. until fracture takes place. Influence of Shock on Cast Iron. In Trans. A. S. M. E., IMPACT. 153 vol. XIX. pp. 351 to 386, a large number of tables prove the fol- lowing : Striking test-bars on the side or in the direction of their length decreases their length slightly, probably by the grains readjusting themselves so as to lie closer together. This explains the frequent cracking of castings while breaking off the gates. Tumbling, in contact with other castings, test-bars which are covered by sheet-iron cases which Jit perfectly, slightly shortens them. Tumbling test-bars in contact with other castings lengthens them. The amount that they are lengthened is proportioned to their malleability. A tempered-steel bar and a bar of white cast iron were not lengthened. A bar of soft wrought iron was lengthened. It was proved that at first, all soft cast-iron bars were slightly shortened, and then the peening action of the blows on their sides lengthened them. Shipping test-bars 1000 miles, loose in a box and in contact with each other, with the box lying on the floor of a box-car, did not produce any difference in length that could be measured. Influence of Shock on Strength. The test-bars shipped by rail were apparently not influenced. Blows delivered on the side or end of a test-bar (even 500,000 blows) did not alter the strength, at least very slightly. Test-bars tumbled in a tumbling -barrel are always stronger than companion bars not tumbled. This fact seems to have been discovered by Mr. A. E. Outerbridge. He described this in a letter to the author in the last part of 1894, and he published a description in Transac- tions of the American Institute of Mining Engineers, vol. XXVI., 1896, p. 176. He explains the gain in strength by the 4 mobility of molecules, " relieving an overcrowded condition of the grains at the surface of a casting. The chapter on Crystal- lization shows that in a test-bar there can be no crowding, but 154 CAST IRON. that each individual grain tends to pull away from those next to it. A shock therefore allows the grains to settle more closely together, making the test-bar slightly shorter. The author endeavored to find the true reason for the increase in strength and discovered the following: Test- bars -J in. square increase in strength until they have been tumbled two or three hours, but not materially by longer tumbling. Of tumbled test-bars, the weakest bars are strengthened most ; , and the strongest bars are strengthened very little. The removal of the surface weakens a test-bar . Smoothing the surface of a test- bar without removing the sur- face strengthens it. Smoothing the surface of a test- bar by pounding with a hammer increases its strength. Pounding the surface of a test-bar strengthens it by condensing the grain. Therefore the strength gained by tumbling is due to making the surface of the test-bar smooth and to condensing the surface by peening. Test-bars of gray iron containing least silicon gain most by the process of tumbling. This therefore proves that tumbling test-bars does not strengthen to any great extent, if any, on account of the grains moving on each other and readjusting themselves ; but the great increase in strength is on account of the condensing the grain by pounding, and by smoothing, thus removing notches which would induce fracture. CHAPTER XV. GRAPHIC METHOD FOR CLOSELY APPROXIMATING THE PERCENTAGE OF SILICON, THE SHRINKAGE, AND STRENGTH OF ANY OTHER SIZE OF CASTING THAN THE ONE TESTED. WE may take ' ' Iroquois ' ' Series I as a fair representation of irons suitable for a moderately heavy casting a little more than i in. thick, having a considerable surface and with a shrinkage of one eighth of an inch to the foot, and containing about \% of silicon. We may take Series 15 as a fair representation of iron suitable for the lightest castings, say from one quarter of an inch down to one sixteenth of an inch thick, and of considerable surface, and with a shrinkage of one eighth of an inch per foot. Shrinkage. Table LXVIII is constructed with these series as extremes, and it will be found that " Iroquois," Series I to 6, with Series 15, nearly correspond to the shrinkages of this table. TABLE LXVIII. APPROXIMATE RELATION OF SHRINKAGE TO SIZE AND PERCENTAGE OF SILICON *"G ,-a 2" X i" ,"D 3 " a 4" a Percentage Silicon. o S 2 c . PS rt o .183 .158 .146 .130 113 .102 I. 00 4) (U 4> > ^ Ratio = Cubic Inche s Contents quare Inches c ling Surface ^ ^ ! i 1 JO - i. ._. ~-v. 1 - d 4-c ,180 _l \ ^ IS, 1 -, 1.3-2 ~\ n n .180 1 ^ ---, ^ B ^^ \ ^ - ^ .170 si 1 M -- \ 't w ~~-^_^ 7 TO \ ^ * -^ \ s 2 i -I r f ~> \ ^ ^ .IK) - s ^ .IK) i 2 (X) n \ s ^ ~^_ \ -~^ ~ 3 \! \ ~> ^ "^^ .150 \ s ** 2 i T \ 5 \ ] ..'. - \ \ i ~N - _ i \ V \ \ -- N .140 \ .140 \ \ \ ^ V V j Si 3 i -j \ \ K \ _. i \ s \ .1!*) \ \ V j \ .130 \ s \ \ x x \ ' \ \ ! \ 31 5 M) i \ - V ' V, x .130 \ XI s t II! \ s 5 s^ \ > ^ \ \ \ s ^ X \ \ \ x .1 10 \ ^ s. \ x X, .110 \ s k - - ~ s \ x. \ x "N, ^ 5 \ 5 \ X 00 \ \ s ^ s^ _3fc IX) \ \ \ X J |_ 000 s J \ X - x, v s , \ x X ~ \ I \ - - X s \ \ N X, ^ f, 2m .090 \ ^ S, X s^ \ x $ s s L \ s s x. r ^ .0 s^ x, ** x .080 ^ - x at A/ * v, \ s, X, ^^ \ x .(TO ~ \ \ \ .070 \ V x. | \ - S, x\, - X x^ I t- t \ .go - .06L s X ^jsy s ^ * - x s x - c X t. x x; .05C C/i X A "".050 s x kW^ \ ^>^ x (M) ^ .040 i V C 1 A. ^L"~ u 2 T tt M L J 1. L T ' u - ^ ^ 4*0 1 ' ) j L . t fjo ( ) .1 ) .'J _js5 FIG. 81. Keep's Shrinkage Chart. Approximate Relation of Shrinl age to GRAPHIC METHOD FOR SHRINKAGE AND STRENGTH. 157 In Fig. Si, the shrinkages of this table are shown graphically. The figures on each side of this denote the shrinkage in inches per foot. The numbers at the top and bottom show the ratio of cooling. Each curved line shows the variation in shrinkage in varying sizes of castings, containing a given percentage of silicon. The percentage of silicon to produce these curves is marked at each end of each curve. To find the approximate percentage of silicon in any iron mixture, locate on the ^-in. d vertical line (.12) the shrinkage of a ^-in. square test-bar from that mixture, and this will show the approximate percentage of silicon that should produce this shrinkage. From this chart a founder can at a glance see the difference in shrinkage between different parts of a casting on account of size. He can tell the shrinkage of any casting larger or smaller, made from the same mixture, from the shrinkage of any size of test-bar which he may use. If he knows the size of a casting and the shrinkage that is desired, he can calculate the ratio of the cast- ing and can locate the shrinkage on Fig. 81 and, by following the curved line either way, can find approximately the percentage of silicon which the iron mixture should contain to produce the desired shrinkage. The following examples illustrate some of the uses of Fig. 81. Example I . Wanted, to make a cylinder with walls 3 ins. thick and so long that we may neglect the end cooling-surface. The shrinkage of a J-in. test-bar from the iron mixture is .153. What percentage of silicon does it contain, and what will be the shrinkage of the casting ? Imagine a strip of the 3-in. casting of any size, say 10" X i"; this contains 30 cu. ins. and 20 sq. ins. of cooling-surface; 30 divided by 20 equals a ratio of 1.50. In Fig. 81 find shrinkage .153 on the left-hand margin. A horizontal line will cut the silicon scale at 2.25, which is the approximate silicon. Follow between the curves until the perpendicular for the ratio 1.50 is 158 CAST IRON. reached (in this case outside the chart), and it will be found that the approximate shrinkage of the casting will be .062. If it had been required to make a casting of these dimensions with a shrinkage of .062 per foot, which had been found satisfac- tory for hydraulic cylinders, or for water-pipe, and it was required to find the silicon in a mixture to produce such a casting, follow down the ratio 1.50 until .062 is reached, then run along the curve to a silicon scale, and we will find the silicon to be 2.25. If we had wished a shrinkage of one eighth of an inch per foot, we would have lessened the percentage of silicon in the mixture to about \. O co O n O N r -. CO in be 5 .2 "5 1 D C/J - vo O O r^oo Tf o d r^. ^ M Tt" CM to \O O Tf" u*> O *"1" CM to TWO in invO CO CO CO c*^ 1 S CO M "* ti S o. -jj oo M ir>o co O O *^vC to O sl g2c7l C1 M ino^comoco ^1-r^M - . o ooo : 0* o 2 : ' "rt -Z t/5 vO w coo m r^ co d f^ o^oo Tj-Tt-COCOCOCOCOCOCOCOCO w in rf co co w o N CO CO CO S * s >* > \ Hf 1 wCK^^-o -^^ ^"^ C : * 6 d ^-^1 '^ 5^ "s M . M o m H t3 H O O ^l : - * E rt J3 ^11 ll 1 " u o S G^r^como M MCO ^-OC'O ss oo v ^^ r ^co S 0> i86 CAST IRON. autographic record of the behavior of the test-bar at each instant of the test. With a tester applying the load by a hand-wheel and screw, DEAD LOAD FIG. 93. Outline of No. 40 Testing Machine. and moving a jockey- weight to measure the load, the test-bar often breaks before the reading can be taken, and in any event the time taken would be from half an hour to an hour. 1000 900 i 500 200 100 iioo _-- I1190 Ibs. FIG. 94. Diagram from Dead-load Machine, No. 40. Fig. 95 shows extension pieces to allow test-bars to be used in ordinary testing machines. (Cut is with Fig. 85.) Keep's Hardness Testing Machine (Fig. 96). (Chas. A. Bauer's Drill-test.) A f straight-fluted drill of standard hard- ness at 200 revolutions per minute makes an autographic record of the workability of a pig or test-bar to any depth, and shows tendency to sponginess or blow-holes. A tray that slides under the table, on which the test-piece is clamped, catches all of the drillings for chemical analysis. The drill enters the test-piece on the under side, and each particle falls into the tray. To the table, by four rods, is suspended a load near the floor so that, including the test-bar, the load on the drill-point is 150 Ibs. The table is raised, or lowered, or held by a hand- wheel at the top of the machine. The machine is started by a clutch, and KEEP'S TEST APPARATUS. 187 thrown out at any desired point by a trip. The downward motion of the table is transferred through steel ribbons, through a ball- bearing arm, to the pencil, which makes a record on a curved FIG. 96. Keep's Hardness-testing Machine. paper-holder which moves at right angles to the path of the pencil by means of a screw. For diagrams see Fig. 84, Chapter XVI. Wet Grinder for f Drills, making both lips to cut exactly alike with standard angles (see Fig. 97). Enter the drill in the holder i88 CAST IRON. FIG. 97. Keep's Wet Drill-grinder. FIG. 08. Keep's Impact-testing Machine, No. 2. KEEP'S TEST APPARATUS. 189 with the cutting edges horizontal, while the holder is clear up to- the stone. Tighten the set-screw to hold the drill and press the drill against the stone until it will cut no more. Turn the holder over and press the other lip against the stone as before. Keep's Impact-testing Machine, No. 2 (Fig. 98). The anvil weighs 1000 Ibs., and the hammer is varied from 25 to 100 Ibs. It tests up to \" D X 24" bars. Recorder for No. 2 Impact, or No. 4.0 Dead-load Machine (Fig. 99). It multiplies the movement of the center of the test- bar five times, and a parallel motion gives the pencil a vertical motion. Recorder for No. 40 Dead-load and No. 2 Impact Machines. Pencil K | ^ ' _--Wf -- 3 Paper holder RECORDER g J ^ Keep FIG. 99. Outline of Recorder. Operation. The test-bar is clamped in position. The loca- tion of the pencil is marked on the paper. The hammer is hung on the graduated arc for a -in. drop. A trip releases the hammer, which swings against the test-bar. On its rebound it is caught by the operator's left hand and returned to the trip, which has been moved to -in. drop with his right hand. The left hand depresses a lever on the bed-plate, which moves the paper T \ in., and so on. The highest drop is 6 ins. After the test a base-line is drawn parallel to the edge of the paper. The autographic 19 CAST IRON. diagrams, Figs. 69 to 74, Chapter XIV, show deflection and set for each blow. Use a 25~lb. hammer for a J-in. D bar, and a loo-lb. hammer for a i-in. Q bar. Professor Thomas Turner's Hardness-test Machine (Fig. 100). The number of grams on a diamond that are required to FIG. 100. Professor Turner's Hardness-test Machine. make a scratch on a polished surface is used as the degree of hardness. This is especially adapted to white and chilled iron and to tempered steel. Keep's Cooling-curve Machine (Fig. 101). This makes an FIG. 101. Keep's Cooling-curve Recording Machine. autographic record of the behavior of a test-bar i" D X 24" long, while becoming solid, while graphite is forming, and during cool- ing. The recording drum is moved by a clock. It also makes a diagram of the critical points of iron and steel. One-half-inch Test-bar Patterns with Yokes. Fig. 102 shows an iron follow-board with two brass J-in. D test-bar KEEP'S TEST APPARATUS. 191 patterns, and Fig. 103 shows one J-in. [J test-bar with one T V X i", all i ft. long. Yokes with the taper-scale are used for measuring shrinkage. FIG. 102. FIG. 103. Record-paper. One-hundred-pound manila paper makes the best diagram. Strips 31 ins. wide and 14 ins. long fit No. 10, and 6J ins. wide and 18 ins. long fit No. 40 recorder. CHAPTER XXI. PIG IRONS AND SILICON IRONS. Pig Iron. Cast iron from the blast-furnace is run into a pig- bed, the runner which feeds the pigs being called a sow. In America the pig-bed for all the irons to be used in the foundry is generally made in sand, and sand is often thrown on the surface of the pigs to cause the iron to cool more slowly, that the fracture may show an open grain and dark color. Pig iron is graded into different classes according to the color or the size of the grain, or to the percentage of silicon indicated by analysis. On account of silicon changing combined carbon into graphite, hereby softening iron, silicon irons have come to be called "softeners." Ferrosilicon is the name applied to irons which carry 10$ and over of silicon. The fracture of the pig is coarser than No. I Silvery in Fig. 104. Flaky Silvery, named from the appearance of its fracture, is made in a very hot furnace, and generally contains from 7% to 10$ silicon. No. i Silvery (Fig. 104). An open-grained, light-colored iron, with more than 6% silicon. No. 2 Silvery (Fig. 104) is of closer grain than No. I. No. 3 or Close Silvery (Fig. 105). Quite often silvery iron is offered for sale that has a very close uniform grain and which has the ring of white iron. It is not generally very high in silicon, averaging perhaps 4.50$ to 6%. The Tennessee Coal, Iron and R. R. Co. selected sample pigs from the eleven grades of iron made by them, and after 192 PIG IRONS AND SILICON IRONS. 193 No. 1. Silvery. FIG. 104. I 9 4 CAST IRON Southern Silvery Grey. FIG. 105. PIG IRONS AND SILICON IRONS. making a very careful chemical analysis sent the pigs for use in this description. Table LXXVII gives for ready reference both the composition and physical qualities of these samples. TABLE LXXVII. ANALYSIS OF PIG IRONS AND SI. IRONS. CHEMICAL ANALYSIS. No. of Test. Southern Pig Iron. i! Graphitic Carbon. Combined Carbon. Silicon. II | Sulphur. 1 si c S T\ I No 2 silvery i 51 0.58 o cn 4QI o 58 o 08 O 2^ 7^O " i " o 17 1. 60 O. 57 47O O 5Q o 06 O 27 728 " j soft 2 QJ. 211 0.8-} } 65 o 60 o 06 O 27 2 " 2 8l 2 OO o 81 3 24. o 76 Z V 721 2 88 2J2 o 46 2e T o 60 72J. 2 2Q 2 28 o.oi 2 l6 o 62 O IO U.4J 722 No 2 foundry 2 16 I QO o 26 2 OO O 7Q o 08 O 21 72"? < _ < i 2 14. 2 O4. O IO i 8^ O 77 O IO O-I T 725 Grey forge I 7Q 1 . 51 o 28 I 7J. o 70 O T7 o 38 726 Mottled 2 74. I OO I 74. 1.^6 o 76 o 36 O ^1 727 White 2 oi c 67 I 34. PHYSICAL ANALYSIS. 1 1 731 730 728 729 -721 724 722 723 725 726 727 Southern Pig Iron. Dead Load. Impact. Shrinkage. Chill. .01 .02 .04 .02 .01 .15 .01 .30 .0 white white Strength. Def. .24 .27 .27 25 25 .20 .27 .19 23 .11 13 Strength. Def. .27 32 32 -32 -30 .21 30 .20 .27 .14 15 Square .140 .131 .149 .130 .156 .164 159 .161 .160 .226 .240 Flat. .136 .138 157 .154 .145 .164 .161 .148 .163 .218 No 2 silvery ... 295 360 375 293 384 362 365 354 365 372 435 305 407 407 322 407 305 373 254 356 186 237 " i " .... " i soft ' ' 2 " . ( I foundry .... Foundry forge. . . . No. 2 foundry. . . . " 3 " Mottled White Fig. 1 06 shows the appearance of the fracture of these first samples of Southern irons and of various silicon irons in a J-in. n test-bar. The drawings of fractures of silicon pigs in this chapter 196 CAST IRON. are from a set of Ashland pigs with silicons determined by analysis. Ferro- Flaky No. I No. 2 No. 3 No. T No. 2 silicon Silvery Silvery Silvery Silvery Soft Soft No. T No. 2 No. 3 Foundry Gray .. . , Foundry Foundry Foundry Forge Forge Mo FIG. 106. Fractures of Test-bars The fractures of Southern pigs in this chapter are from a second set sent by the Tennessee Coal, Iron and R. R. Co. without silicon analysis. The fractures represent their ordinary grading. Southern Silvery (Fig. 105) is made when the furnace is very hot and is not a regular product. It is as good a softener as regular silvery iron. The silicon ranges between 4$ and 5.^- No. i Soft and No. 2 Soft (Fig. 107) are grades peculiar to a PIG IRONS AND SILICON IRONS. I 97 No. 1 Soft. FIG. 107. I9 8 CAST IRON. Southern furnace. They contain too low silicon to be classed as silvery iron and too much to be classed as foundry iron. If the fracture is close it is No. 2 soft. Both are expected to contain over 3.25$ silicon. These pigs have a smooth face. There is a temptation to mix close-grained light-colored low silicon irons with No. 2 soft before shipping. No. i Foundry (Fig. 108) is a choice grade, having coarse grain and dark color. It makes fine-appearing and accurate thin and intricate castings. The surface of the molten metal is dark and sluggish looking, and does not give off sparks. Under the surface of the melted iron there are splashes of light. The pig has a blue velvet face v/here the surface is smooth. No. 2 Foundry (Fig. 108). The fracture is lighter in color than No. I, and usually the surface of the pig is smoother. The grain is closer and there is often a closeness around the edges of the pig. It is generally a little harder and stronger than No. I and it is not quite as fluid, as its carbon and silicon are generally less. The surface of the melted iron is a clearer red, and throws off some sparks, splashes only a little as it cools, and its surface exhibits a series of lines or figures ever varying as though the surface were in circulation, such appearance continuing until the iron becomes pasty. The closeness of No. 2 may often arise from the way it is handled in the pig-bed, and it often has as high carbon and silicon as No. I. It is generally as good an iron as is needed for the best foundry castings. No. 3 Foundry (Fig. 109). The fracture is still lighter in color, the crystals are much smaller, and the fracture is smoother but contains some pits. It is stronger than the preceding grades. It contains less silicon and carbon than Nos. I and 2, and has not the same fluidity. The molten metal throws off sparks abundantly as it runs from the cupola. The surface figuring is less apparent than with No. 2. It is used in the foundry for heavy work, but it will not take much scrap on account of its low silicon. No. 4. Foundry or Foundry Forge (Fig. 109) is a grade that PIG IRONS AND SILICON IRONS. 199 No. 1 Foundry FIG. 108. 200 CAST IRON. No. 3 Foundry FIG. 109. PIG IRONS AND SILICON IRONS. 201 is too close and with silicon too low for No. 3* Foundry and too open for Gray Forge. The surface of the pig has larger and more pits than No. 3. Gray Forge is known by its larger pits. It has too close a grain to be classed as Foundry Forge. It may be used to advan- tage for heavy castings. In Chapter XVI, Table LXX are given analyses and records of tests of several Gray Forge Irons. Mottled is the highest pitted iron that is made. The name does not indicate any grade, but a mottled appearance of grain formation. The iron may be quite gray, or it may be white with only a gray tinge. The latter would be high mottled. White is made in a cold furnace and carries more sulphur than any other grade. One of the advantages of making the pigs of foundry iron in iron molds or in one of the modern casting machines would be that the fracture of the pig would indicate the kind of casting lhac the iron would make. A non-chilling iron would show a gray even fracture against the mold as in Fig. 59, while the slightest tendency to chill would be shown in chilled surfaces of the pig. The close-grained foundry iron which is cast in iron molds will make as soft castings as the same iron pigged in sand. (See Table LXIII.) REMARKS ON SILICON IRONS. Weakness of Silicon Irons. Softeners are invariably weak in the pig and would make very brittle and weak castings if used alone or to excess. Silicon irons often run from \% to 2% phos- phorus. Mill cinder is quite extensively used in the ore mixture for silicon iron, and imparts a silvery whiteness to the pig. History of the Use of Silicon Iron. Until within the last fifty years Scotch pig iron which contained about 3$ silicon was imported and used as a softener. It was found that No. i American pig was often as good a 'softener as Scotch pig. About thirty years ago it was found that the iron ores found in Ohio 202 CAST IRON. would make a peculiar light-colored iron which imparted great fluidity and softness to the irons made from the refractory Lake Superior ores. At once these irons took the name of Ohio Softeners. In 1888 the author made a large number of tests to determine the influence of remelting on silicon. The tables of records and remarks can be found in Trans. A. I. M. E., vol. xvil. pp. 252- 261 . In remelting in a crucible not in contact with blast, the loss of silicon in irons containing less than 10$ of silicon was 0.55$, and of the irons containing over 10$ the loss was 2.80$ of the silicon contained. In a second remelting the proportion of the loss was reversed, but the loss was very small in either case. It was proved that in the use of silicon iron to impart silicon to low-silicon irons the mixture contained within o. 10$ of what was calculated. The percentage loss was, however, not half as much when silicon irons with less than 6% silicon were used. CHAPTER XXII. TESTING SMALL SAMPLES OF PIG IRON. A CRUCIBLE furnace may be made by lining a sheet-iron drum 26 ins. diameter and 30 ins. deep with fire-brick set on end. It should be set in a pit with its top even with the floor. Cross- bars are fixed across the brick ashpit to allow i-in. square grate^ bars to slide under the furnace. The upper part should be connected to a chimney 12 ins. square inside and 30 ft. high to give a natural draft. The furnace-top is closed with a round cover. Seventy-two-hour coke is the best fuel. A No. 16 brass crucible is the most convenient size and can be handled with long-handled tongs taking hold of the edge. The crucible will settle down as the fuel burns away. Fifteen pounds of iron will melt in from 30 to 45 minutes. When the pot and iron cannot be seen in the furnace the iron is fit to pour. By keeping the furnace ready for use, after lunch a fire can be put in, and when hot the iron can be melted before four o'clock, when the molder would have put up his day's work and can put up a flask of test-bars. A Small Cupola. A sheet-iron cylinder 16 ins. diameter and 4 ft. high, with a cast-iron bottom bolted on, is lined 2 ins, thick with pounded fire-brick and fire-clay. This is set on a block of stone about 2 ft. from the floor. Just above the bottom lining is fixed a spout lined like the cupola, and a foot above is a 3-m pipe fastened to the side for a tuyere. Bring wind from a small 203 204 CAST IRON. fan by a portable pipe. When a heat is to be made, put in enough /2-hour coke the size of a hen's egg to heat the cupola red hot, and to leave the incandescent coke 14 ins. above the tuyere. Then charge the iron and cover with coke. Such a test is almost exactly like using iron from an ordinary cupola. CHAPTER XXIII. ALUMINUM IN CAST IRON. A NUMBER of series of tests were made in 1887 with both F L M and with " Gaylord " white pig iron, introducing aluminum by ferroaluminum (containing 10$ Al. and 3$ Si.). After the results of these tests were published it was claimed that the 3$ of silicon produced the results. TABLE LXXVIII. Calculated. Actual Percentage foundjin Test-bars.; C O I , 1 00 g U ^_J a s . ( j u H 3 C < 3 C . rt U t! o H . c3 a a o a 3 O 1 _u 3 s U a rt 1 "o o | 1 f a * < c/l "* ^ i H O U fc * H o c3 WHITE BASE IRON AND ALUMINUM. WHITE REMELTS. Pig. 363 .00 .00 .186 .18 .18 .18 .00 .00 186 08 2.98 2.98 . -95 7* 2.63 1.71 P,g. 376. . 186 25 2.98 3-13 2.03 25 .26 .27 25 37 377 2.97 X 4 83 36s .50 35 54 .. ... 2.84 35 1.49 378 3.12 53 1.59 366 75 -43 47 .89 3.10 .81 1.29 379 2-75 83 .92 ,67 .62 i 28 62 380 12 2.52 44 1. 08 FLM GRAY IRON AND ALUMINUM. FLM REMELTS. Pig. .00 25 25 .00 i 249 3-55 3-22 33 Pig. 1.249 3-55 3.22 33 368 .00 2 5 25 .00 1 245 .26 382 1.160 3-97 .81 1.16 3t>9 25 33 .28 .10 4.09 .72 1-37 383 2.61 37 .24 5 .41 .29 .14 . 3-55 .66 .89 384 3-86 .98 .88 37 1 75 50 37 32 3-53 78 75 3'57 .60 97 372 I. 00 .58 50 75 3-45 .89 386 . . 3.65 77 .88 373 2.OO 92 75 1.50 . 50 i. 08 . . 3-37 32 1.05 374 3.00 .26 99 2.23 3-34 .76 58 388 3-72 49 1.23 375 4.00 .70 54 3-84 2 280 3.10 44 .66 389 978 3-53 .82 1.71 205 2O6 CAST IRON. WHITE IRON AND ALUMINUM, COWLES' AND PURE METAL. Al. .89 Gr. i. Si CC. 1.29 Si. .48 Al. 1.28 Gr. 1.71 CC. 1.19 Si. .62 g3 m tr~ / 06 ^ iiii ALUMINUM IN CAST IRON. 207 The author then repeated the tests, using pure aluminum. Fractures of the test-bars are shown in Figs. 1 10 and 1 1 r. The chemical analyses are given in Table LXXVIII and in Figs. 112, 113, and 1 14. The test numbers of bars accompany each record for reference. Al. 1.50 Gr. 2.50 CC. i. 08 Si. 1.66 FIG. in. F L M Gray Iron and Aluminum. Influence of Aluminum on the Grain of Cast Iron. In tests with white iron, Fig. 1 10, in bar 364 one quarter of \% of aluminum has prevented blow-holes. Test-bar 471 made with silicon alone 208 CAST IRON. CALCULATED & ACTUAL ALUMINUM & SILICON IN "FLM"-AL.i WHITE-AL. SERIES FIG. 112. FIG. 113. FIG. 114. ALUMINUM IN CAST IRON. 209 NO. OF TEST GREY IRON OASE .30 -vl ill. ight FIG. 115. FIG. 116. 2io CAST IRON. is free from blow-holes. In 365 one half of \% of aluminum has made the test-bar a light gray. In 366 with 0.75$ of alumi- num, the casting is decidedly gray, and in 367 with 1.28$ of aluminum the chill is reduced to \ in. Test-bar 473 was made to prove 367, and when compared with 366 and 367 it is just what a \% bar should be. So far the aluminum was introduced by 10$ ferroaluminum. There was added to 471 exactly the same silicon and carbon as there was in 473. The difference in the appearance is due to the aluminum in 473. One quarter of \ of aluminum in 364 produced the same effect as 0.62 silicon in 471. Pure Al. added to 471 made 472 the exact duplicate of 473. The grain is the same, and these tests prove that 0.50$ of Si. and ifo of Al. will change a white porous iron into a solid gray casting. The author added \% pure aluminum to white iron to make 469, and 0.75$ Al. for 468. (See also 440 of Fig. 30, Chapter IX.) This shows that Al. alone will make solid gray castings out of porous white iron. It had been claimed that aluminum would not stay in a cast- ing. 468 and 469 were melted together and produced 470, which was still gray, showing that the Al. was still there. Fig. 1 1 1 shows the influence of Al. from o to 3.82^. The formation of an intensely black grain directly back of the chill was first shown in Fig. ill in 1887, and it was soon found to be more or less apparent back of any chill. Analysis of similar castings shows more graphite just behind the chill in this dark portion than farther back where the casting cooled more slowly. The author's discovery of the influence of aluminum on cast iron was made in 1887 and published August, 1888 (see Trans. American Association for the Advancement of Science, 1888). Aluminum in Steel Castings. Very full records are in Trans. A. I. M. E., vol. xvm. pp. 835-850. Aluminum in Wrought-iron Castings, ibid., pp. 851-858. These were the first series of records made to show the influence ALUMINUM IN CAST IRON. 211 of Al. in steel and wrought iron. Castings were made of both steel and wrought iron with 0.25$, i$, 2$, and 3$ aluminum. The Mitis Co. had used Fe.Al. for some time in castings of wrought iron, but never used more than 0.25$, and they claimed that none of the aluminum remained in the casting, and that its only effect was to make a sound casting. Dr. Mabery's analyses proved that aluminum remained in the casting. The author discovered that iron with 50$ of Al. would in a short time fall to a powder. These experiments for the first time determined the shrinkage of Al. and the influence of Al. on the shrinkage and strength of steel and cast iron. CHAPTER XXIV. INFLUENCE OF VARIOUS METALS IN CAST IRON. THE object of this chapter is to show the effect of a chance introduction to F L M (Method 3) of the more common metals into cast iron. Instead of using cupola iron the F L M was melted in a crucible. TABLE LXXIX. No. Test. Melted in a Crucible. Dead Load. Impact. Shrinkage. Chill. Str. Def. Str. Def. i"D &"Xi" 755 756 757 758 75i 752 753 754 759 760 761 762 763 764 765 766 767 768 769 770 776 777 778 779 500 375 327 295 330 60 379 375 370 383 50 363 340 34i 330 20 370 357 357 333 TO 387 377 390 390 403 463 481 483 .18 .28 23 .19 .24 .11 .27 25 23 .23 05 .28 .22 .22 .18 .09 25 23 .21 .20 .04 .24 .22 .22 .21 .21 .26 .26 .26 400 356 356 246 356 390 348 337 432 .21 30 .29 .19 .26 30 .26 25 .28 .198 .169 .I 7 8 .184 .190 .2 4 8 .172 .172 173 .173 .148 .168 .169 .170 1 66 195 .190 .200 .205 .192 .I8 7 195 .198 .198 .182 .190 195 .90 .70 .70 .70 .70 .60 .60 50 .60 .60 50 .40 .70 .70 .60 50 03 .02 03 .OI .02 .02 .02 .01 F L M 0% Nickel 1 t4 " 8 F L M o% Copper f I t 382 373 38i 31 .28 I < .053 .168 .178 .183 .187 113 .125 .128 .129 .129 .127 .126 .127 .127 .I8 5 195 .200 .199 .141 137 .138 .138 .140 137 135 F L M o$ Tin 322 399 350 305 350 350 407 365 296 313 423 304 25 .28 " i " i I -:::::.. 23 .22 FLM 0% Lead 4 4 " . . F .. ^ . S .22 , t o 8 = 2nd -s o o ^"" ,, o.a ard . . 5 v- b .24 .20 JS " l 212 INFLUENCE OF VARIOUS METALS IN CAST IRON. 213 In all cases the iron was hot enough to melt all of the added metal, although the last metal barely filled the molds. The effect of the stirring alone is shown by tests 776-9. TABLE LXXX. No. Test. Chromium. Dead Load. Shrinkage. Chill. Str. Def. J"D &"xi" 429 594 595 596 597 598 599 620 6i5 616 617 618 619 335 332 345 475 345 355 365 374 362 366 373 363 352 23 .20 .24 .24 .17 .21 .12 .21 .21 .20 15 .19 .18 .167 .171 .167 .187 .186 .172 .227 .132 .136 .142 .154 .150 .156 .200 195 .194 .222 .220 .199 .60 65 75 .60 75 75 All 03 .01 03 .07 .05 .08 I OO . . . Gaylord White 1.00$ Chromium 1 o IO . 4 0.50 i oo ' (i *o) 4 (2 00) 429 0% 594 0.10$ 595 0.50$ 596 597 1.50* 598 FIG. 117. Chromium in Cast Iron. Chromium in Cast Iron. Mr. R. A. Hadfield in his paper before the Iron and Steel Institute, "Chromium in Steel," pro- posed that the author should determine the influence of chromium 214 CAST IRON. in cast iron, and ne furnished the ferrochromium to make the necessary tests. The ferrochromium was added after the iron was melted (Method 4), and the pot was returned until all was melted, which gave time for the chromium to become thoroughly incorporated and to exert its influence on the metal. Chromium does not seem to be of any benefit in cast iron, and it exerts little or no influence except to slightly increase shrinkage when present in quantities less than i#. INDEX. Action of fluid iron in a foundry ladle, 21 ; of iron floating in, 20, 63. Aluminum, and solid castings, 67, 207; changes combined carbon into graphite, 205; in cast iron, 205; in wrought iron and in steel castings, 210; with 50% iron, crumbles to powder, 211. American Foundrymen's Association, analysis of test-bars, 119; description of tests, 115, 117; shrinkages, 48; strength tests, 117, 118, 123, 124, 125, 126, 128, 129. American Society of Mechanical Engineers, Testing Committee, 115; strength tests, 116, 123, 129, 130. Analysis, A. F. A. test-bars, 119; aluminum mixtures, etc., by Dr. C. F. Mabery, 205; annealed test-bars, 33, 139; Ashland pig, 176; A. S. M. E. test-bars, Si. 42, P. 76, S. 97, Mn. 105, T.C. 132, Cd.C. 133, G.C. 138, by Dickman and Mackenzie; cube, 122; cupola iron, 177; Durham pig, 72; eleven pigs (Pencost by Dr. Lord), 45; F L M pig and test-bars, li, 35, 70, 176, 205; ferromanganese, 99; Gaylord pig iron and test-bars, n, 35, 205; Hamden pig iron by Dr. Locke, 73; Hinkle pig iron by C. D. Chamberlain, 12; Iroquois pig iron by D. & M., n, 12, 42, 43; machine and sand cast pig and test-bars by A. L. Colby, 136; Norway pig, 74; Pencost pig, 12, 45; segregation, no; silicon iron test-bars, 35, 38, 40; six pigs, 167; Sloss pig, 176; Star pig, 78; Stewart pig, 74; strong and weak test-bars, 174; Swedish Mn. irons, 103; Tenn. C., I. & R.R. Co. pigs, by Dr. W. B. Phillips, 195; test-bars of cooling curves by E. E. Mains, 53, 55; three pigs, 176; twenty pigs (Dayton by H. S. Fleming), 178. Annealed castings, analysis, 33, 139; physical quality, 33, 139; temperature for, 63. Ashland Iron and Steel Co., 12. Axes of crystals of cast iron, 20, 22. Blast-furnace, reduction of ore in, 26. Blowholes, 30, 72; and aluminum, 207; and phosphorus, 72; and silicon, 72, 210; and sulphur, 91, 96; prevented by graphite, 30. Blue-billy added to gray iron, 84. Borings of cast iron, close grain, 131. 215 21 6 INDEX. Breaking or ultimate strength, 7. Brimstone added to gray iron, 83, 85. Brittleness by combined carbon removed by silicon, 119; caused by Cd.C., 134, 135; of phosphorus irons, 66. Britton, J. Blodget, analysis of Star iron, 78. Bulk of carbon, 10; of other elements, 10, 28, 29. Carbon, 25; and chill, and fusibility, 29; and fluidity, 29, 30; and hardness 29, 167; and phosphorus, 78; and silicon, 28, 38, 45, 176; and strengtn. 132, 133, 134, 136 138; and sulphur, 82 to 98; bulk of, 10, 28, 29; coloring power of, 25; combined, 27, 133; combines at red heat, 62; combined, the natural condition, 27, 28; combined, produces hardness, 166; diffusion of, 28, 133; diluted by adding steel, 30, 71; diluted and Si. kept uniform, 31; dissolved in fluid iron, 27; driven out by sulphur, 83, 93, 94; if de- creased, gray iron turns white, 30, 31, 94, 95; in cast iron, 25; influence of remelting, 27, 30, 33; invisible in carbonic acid, 25; not increased in cupola, 27; of the three expansions, 55; origin of carbon, 26; quantity of, 27; reduces shrinkage, 46; relation to silicon, 45; replaced by silicon, 38, 45; saturation of, 26, 27; segregation of, 107; silicon acts through, 38, 47, 176; specific gravity, 25; total, and strength, 132; total, determines in- fluence of silicon, 38, 39. Carbonic acid, 25, 26. Carbonic oxide, 26. Cast iron, 3; crystals of, 19; does it expand in freezing, 20; fusibility of, 29; with low silicon tends to lamellar fracture, 78; various metals in, 212. Casting machine, pig-iron, 136; conditions should be uniform, 138. Castings, classes of, 159; cracking of, 24, 153; green sand stronger than dry, 28; with close-grained pig, 162; spongy, 21, 22, 162; with heavy and light parts, 24. Chemical analysis, i; drawbacks of, 171; drillings for, 165; will not account for all physical properties, 38, 39, 40, 173; what it does, 170. Chemical composition, and hardness, 166; a shrinkage test for, 46; does not always determine strength, 163, 168, 177. Chemists assume that chemical composition insures physical quality, 169, 171; must measure physical quality to prove prediction, 169; not agreed on methods of sampling, 171. Chill, 7; and carbon, 29; and manganese, 100, log, 104, 105, 106; and phos- phorus, 80; and silicon, 44, 101, 102; and sulphur, 83, 86, 96; explained by segregation, 114; measure of, 183; of annealed castings, 33, 139; pro- cess instantaneous, 51. Chromium, in cast iron, 213; in steel, R. A. Hadfield, 213. Coal, 4. Coke, 4. Colby, A. L., machine and sand cast pig iron, 136. Cold-shortness of phosphorus iron, 66. Cold-shut in test-bars hastens fracture, 181. Color of grain of phosphorus irons, 77. INDEX. 217 Coloring power of carbon, 25. Combined carbon, 27, 133; accompanied by low silicon, 133; causes brittle- ness, 134, 135; changed to graphite by aluminum, 205, by annealing, 139; by silicon, 36, 119; during 3d expansion, 56; influence of increase in size on strength, 133, 135; produces hardness, 166; the natural con- dition, 27, 28; weakens, never strengthens, 133, 135. Composition of irons may be alike and yet make unlike castings, 171. Compression test, 2, 127, 128. Conditions influence castings, 85, 168, 176; unavoidable, 173. Cooling curves, 50; begin when casting is solid, 51, 54; carbons of three ex- pansions, 55; for hard and soft iron, 55; for rolled steel, 63; influence of size of casting, 58, 59, 61; of ist, 2d, and 3d expansions, 54; of hot and dull iron, 59; of phosphorus, manganese, and sulphur, 58; of test-bars in yokes, 51; of various metals, 51; of yokes, 51; practical application, 58; relation of size to expansion, 58; temperature of expansions, 61; test of quality at each minute, 54. Cooling in ladle, 88. Cooling ratio, 157. Copper in cast iron, 212; cooling curve of, 52. Cracking of castings, 24, 153. Critical points of cast iron, 65. Crucible, furnace, 203; iron for tests, 14. Crushing test, 2, 127, 128. Crystallization, 19; carbon causes, 46; castings larger after, 46; no change after 3d expansion, 56; of a mill roll, 21; very rapid, 55, 109. Crystals, arrangement during 3d expansion, 56; axes of, 20, 22; do they expand, 20; formation of, 19, 20, 21; octahedral, 19; of cast iron, 9. Cubic foot of cast iron, weight of, 28. Cupola, iron for tests, 13; small, for testing samples, 203; temperature, 175. Dead load, 7; diagram, 186; diagrams of cast iron, 5, 146, 161; diagrams of steel bars, 140, 141; Keep's testing machine, 183, 184; records of deflection, 140, 141. Deflection, 5; dead-load records, 140, 141; elastic, 6; greater for impact, 149; impact records, 143, 150; relation of, to dimensions of bar, 142; set, 6. Dickman and Mackenzie, analyses, u, 12, 42, 76, 97, 105, 132, 133, 138. Diffusion, of carbon, 28, 133; of silicon, 41. Direct test, i. Drills, angles for, 164; for hardness test, 164. Drillings for analysis, 165. Dry-sand castings weaker than green-sand, 128. Dull iron, cooling curves of, 59, 61, 62. Durham pig iron, 72. Elasticity, 6; measure of, 6. Elements in cast iron, bulk of, 10, 28, 29. 2i8 INDEX. Expansion, and shrinkage, 54; carbons of, 55; combined carbon changes to graphite during 3d, 56; crystals take final form during 3d, 56; ist, 2d, and 3d, 54; is the cause of less shrinkage, 64; silicon increases the 3d, 57; takes place at definite temperatures, 61, 62, 65; 3d indicates red-shortness, 65; 3d is when carbon combines before fusion, 62; time it takes place, 54. Factor of safety for cast iron, 124, 163. Feeding castings, 21, 22. Ferromanganese, 99; cracks when cooling, 99. Ferrosilicon, 192. Ferrostatic pressure in mold, 22. Flask for test-bars, 182. Fleming, H. S., analysis of Dayton pig iron, 178; phosphide of iron, 66. . FLM pig iron, analysis of, n, 35. Floating of solid on liquid iron, 20, 63. Fluid iron, its action in foundry ladle, 21; less dense than solid iron, 21; phosphorus irons remain so, 81 ; should be poured dull, 24. Fluidity, and carbon, 29, 30; and phosphorus, 81; and silicon, 44. Fluor-spar, melted with gray iron, 83. Formula for strength, 119, 163; strength of cast iron cannot be calculated by, 125. Foundry irons, silicon in, 41. Fracture, 5; from a notch or cold-shut, 181. Fractures, illustrations of, of aluminum mixtures, 206, 207; of castings, 23; of chromium in cast iron, 213; of dilutions of carbon, 31, 67, 71; of F L M remelted, 70; of irons used in tests, 35; of machine and sand cast pig, 132; of phosphorus and FLM, 71; of phosphorus and white iron, 70; of segregations of American pig, 109; of segregation of stove castings. 109; of segregation of Swedish irons, 108, 109; of segregations of test- bars, in, 112; of silicon and F L M, 37; of silicon and white iron, 37; of silicon i l /2% to 3%, 39; of silicon 2 l / 2 %, 40; of silicon pig iron, 193, 194; of spiegel and silicon, 100; of sulphide of iron and F L M, 90; of sulphide and silicon and F L M, 91; of sulphur and cupola iron, 89; of sulphur (brimstone) and F L M, 86; of sulphur and white iron, 91; of Swedish manganese irons, 103; of Tenn. C., I. & R.R. Co. pig irons, 169, J97) !Q9r 200; of white iron remelted, 67. Fuel, 4; sulphur in, 5, 82. Fusibility, and carbon, 29; and phosphorus, 66; Truran's rule for, 29. Galena melted with gray iron, 83. Gaylord pig-iron analysis, n, 35. Grain, 7; and aluminum, 207; and phosphorus, 77, 78; darkened and made coarser by annealing, 139; of phosphide of iron, 77. (See Fractures.) Graphic method for approximating silicon and shrinkage, 155, 156; strength, 160; to find the silicon in any iron mixture, 157; to find the shrinkage of any size of casting, 157. INDEX. . 219 Graphic records, 8; advantages, 8. Graphite, 25; and strength, 133; changes to combined carbon before melting, 62; does it separate the grains, 138; flakes of, 19, 138, 139; in cast iron, 27; in spaces between the grains, 139; segregation of, 107; silicon changes combined carbon to, 36, 119; strength and weakness independent of, 139- Gray forge pig iron, 167, 195, 201. Gray iron, melting of, 30; more fluid than white, 30. Green-sand castings stronger than dry-sand, 128. Grinder for Y& drills, 164, 187. Hadfield, R. A., chromium in steel, 213; Spiegel and silicon, 100. Hamden pig iron, 73; chill, 80; grain, 77, 78; hardness, 80. Hardness, and carbon, 29, 167; and chemical composition, 166; and man- ganese, 104, 105, 106; and phosphorus, 58, 80; and silicon, 44; and sul- phur, 93, 98; diagrams, 165; follows combined carbon, 44, 167; Keep's testing machine, 164, 186; or workability, 166; record of, 165; records of A. S. M. E. and A. F. A. tests, 166; test does not distinguish between hardness and tenacity, 164; Turner's definition of, 164; Turner's test for, 29, 190; Turner's testing machine, 190. Hard spots in castings, 107 to 114, 165, 167. Heavy and light portions of the same casting, 24. Hinkle pig iron, analysis, 12. Horizontal stronger than vertically cast test-bars, 123. Hot and dull iron, cooling curves of, 59, 61, 62. Impact, 140; advantages of swinging hammer, 152; arbitrary values for, 32; diagrams from test-bars not perfectly elastic, 145; diagrams of cast iron, 146; influence of shock on cast iron, 152; Keep's testing machine, 142, 189; recording apparatus, 140, 189; records of steel test-bars with direct drop, 144, 149, 150, 151; records of test-bars with swinging hammer, 142, 143, 144, 146, 147, 148; tests, i, 152; striking bars shortens them, 153; to determine resilience, 152. Impurities, a benefit, 113; prevent segregation, 113. Iroquois pig iron analysis, 12. Iron, cast, 3; ingot, 3; to find if low-priced is economical, 169. " Keep's Test " name suggested, 49. Kish, graphite floating on liquid iron, 27. Ladle, action of fluid iron in, 21; hot, cold, and green, 84, 175; iron standing or stirred in, 59, 88, 173, 212. Lamellar fracture caused by low silicon, 78, 100. Large castings proportionally weaker than small, 22, 120. Lead, cooling curve of, 52; in cast iron, 212. Lime, 4. Lines of weakness in cast iron, 22, 23. 24. 220 INDEX. Machine cast pig, conditions should be uniform, 138; tests and analyses, 136, 137; advantages of, 201. Machines, see Testing machines. Manganese, analysis, A. S. M. E. tests, 105; and carbon, 101, 102, 103, 104; and chill, 100, 103, 104, 105, 106; and hardness, 58, 104, 105, 106; and remelting, 99, 102, 104, 106; and shrinkage, 100 to 104; and silicon, 100, 102, 103, 104; and strength, 104; does not make gray iron white, 102, 103; gray Spiegel, Ward's, 102; Pourcel's, 103; influence on cooling curves, 58; nothing to do with grayness or chill, 103; removes chill, 105; removes sulphur, 99, 104; Swedish irons, 103; whiteness of spiegel not from Mn., 102, 103. Measure, of elasticity, 6; of rigidity, 6; of usefulness, 7. Mechanical analysis, 168; advantages of, 169; can be used alone by any- one, 169; it is chemistry for the uneducated founder, 170; Keep's ap- paratus for, 182 to 190; measures physical properties, 168; simple method benefits ordinary founder most, 170; shows combined influence of chemi- cal elements and of all conditions, 168; tells whether more or less silicon is needed, 168. Mechanics, 5. Metis Co.'s use of aluminum, 211. Methods of investigation, 12; methods, I to 9, pp. 13 to 16; 10 to 14, pp. 17 and 18. Mill cinder in silvery iron, 201. Moldenke, Dr. Richard, A. F, A. tests, 115. Molds, wet and dry, 175. Mottled pig iron, 93, 94. Moxahala pig iron, 73, 78; grain, and color of fracture, 78; chill and hard- ness, 80. Nickel in cast iron, 212. Norway pig iron, 74, 113. Outerbridge, A. E., tumbled test-bars, 153. Paper for records, 191. Patterns for test-bars, 182, 184, 191. Pencost pig iron, 12, 35, 45. Phosphide of iron, 66; grain of, 77; shrinkage of, 79; hardness, 80. Phosphorus, 66; analysis of A. S. M. E. test-bars, 76; and blowholes, 72; and brittleness, 66; and carbon, 78; and chill, 78, 80; and cold-shortness, 66; and cooling curves, 58; and fluidity, 81; and fusibility, 66; and grain, 77, 78; and hardness, 58, 80; and remelting, 77; and shrinkage, 79: and strength, 80; color of casting sometimes straw-colored, 78; in favorite foundry irons, 81; irons low in silicon have lamellar fracture, 78; spe- cific gravity, 66; Turner on, 66, 81; irons, weakness of, 77; white and red, 66. Physical tests, I. INDEX. 221 Pig irons, 192; used in tests, n, 12; ferrosilicon and silvery iron, 192; foundry Nos. i, 2, 3, 4, 198; gray, 28; gray forge, 201; mottled, 201; not homogeneous, 171; physical tests of, 195; soft, Nos. i and 2, 196; Southern, analysis of, 178, 195, 196; testing small samples of, 203; to find if low-priced is desirable, 169; white, 28, 201. Pyrites added to gray iron, 84, 96, 98. Quenching cast iron in ice-water, 55. Ratio of cooling, 157. Red-short, 82; at time of 3d expansion, 65. Relative tests, i. Remelting, influence on carbon, 27, 30, 33; on manganese, 99, 102, 104, 106; on phosphorus, 77; on silicon, 42, 202; loss in, 12; on sulphur, 85, 93, 96; reduces carbon, 30. Resilience, test for, 152. Rigidity, 6; diagram of, 6; measure of, 6; no, 6; perfect, 6. Round and square test-bars, of Mr. West, 130; of Western Founders' Asso- ciation, 129; strength of, 128. Samples of pig iron, to test, 203. Sampling, chemists not agreed on method of, 171. Saturation of carbon, 26, 27, 38. Scale, of casting peels, 45; tapr for measuring shrinkage, 183. Scotch pig iron, 81, 201. Scrap, appearance a practical guide, 172; no correct analysis can be made of, 173- Scrap-carrier, the pig that has the lowest shrinkage, 161. Segregation, 107; analysis of, no; explains chill, 114; in a stove cover, no; of American pig, 107; of graphite, 107; of sulphur, 107; of Swedish iron, 107; prevented by impurities, 113; remarkable examples, 108; to produce a white core, 113; white beads, no. Set, 6; increased by annealing, 139. Shock, influence on cast iron, 152. Shrinkage, 46; A. F. A. tests, 48; a safer guide than silicon analysis, 176; A. S. M. E. tests, 48; a test for chemical composition, 46, 168; and carbon, 38, 47, 176; and expansion go on at the same time, 65; and manganese, 100 to 104; and phosphorus, 79; and silicon, 47, 57; and sulphur, 83, 86 to 95; autographic curves of, 50; carbon reduces, 46 caused by other elements, 47; chart showing percentage of silicon, 156; decrease in cooling the same as increase in volume by melting, 46, 51; decreases as silicon increases, 48; decreases as size increases, 49; graphic method, 155; if lower, the pig iron is a better scrap-carrier, 169; is lessened by expansion, 64; least with greatest graphite, silicon, and phos- phorus, 79; less in casting than in pig, 96; measure of, is equivalent to chemical analysis of silicon, 168; measure of, gives relative influence of silicon. 158. 159; measure of, tells whether more or less silicon is needed, 222 INDEX. 159, 168; measuring, 183; of annealed castings, 138; of different grades of iron, 47; of fluid iron, 51; of heavy and light parts of the same casting, 24; of large and small castings, 22; of pure iron, 46; one size of test-bar to study variations in chemical composition, 46; y%", for any size of casting, 47, 158, 159; relation to percentage of silicon, 44, 155, 168; silicon lessens, 47, 56; soft castings with low, 139; standard, 159, 168; taper scale for measuring, 183; there is no definite, for a definite percentage, of silicon, 176; to find, for different sizes of castings, 157; to study relation of, to variation in size of test-bar, use iron of uniform compo- sition, 46; varies with size of casting, 46, 155. (See Cooling curves.) Shrink-holes, 25, 165. Shrink-rule, 46. Silicon, 34; acts through carbon, 38, 47, 176; a definite percentage will not always produce a definite shrinkage, 37, 176; a greater quantity acting a shorter time or vice versa, 40; a softener, 44, 56, 79; alloys with iron, 34; amount needed told by shrinkage, 49, 168; analyses, A. S. M. E. tests, 42; analysis not as safe a guide as shrinkage, 176; an approximation of, all that can be obtained by analysis, 44; and carbon, 28, 38, 45, 176; and chill, 44; and fluidity, 44; and hardness, 44; and manganese. 100 to 104; and Spiegel, 100; and strength, 119, 121; an increase counteracts most evil influences, 168; approximated by shrinkage chart, 156; as i increases shrinkage decreases, 48, 100; carriers of, 79; changes white iron to gray, 36, 37; controls mixture, 41; decreased, increases chill, 101, 102; decreasing, increases shrinkage, 47; decreasing, turns gray iron white, 94, 95; diffusion of, 41, 43; exact percentage, even by analysis, impossible, 43; expels carbon, 39; graphic method for determination of, 155; history of its use, 201; in foundry irons, 41; in some irons more effective than in others, 176; increases 3d expansion, 57; influence de- pends on total carbon, 38; influence indirect through carbon, 38; in- fluence of remelting, 42, 202; influence modified by various conditions, 39; its influence, not its percentage, 37, 38; low for large castings, 131* lowers saturation of carbon, 38; makes gray iron more gray, 36; the more total or less combined carbon the less silicon required, 39; overcomes sulphur, 93; percentage increased or decreased according to shrinkage, 159; pig iron, 192, 201; reduces combined carbon, thereby removes brittleness and increases strength, 119; specific gravity, 34; reduces shrinkage, 47, 57; relation to carbon, 45; the percentage in pig iron must be known, 172; the percentage is but an approximation, 44; to find percentage from a test-bar, 157; to produce y&" shrinkage per foot, 47, 159; Turner on, 34, 36; varies shrinkage, 168; weakness of silicon irons, 201. Size, influence on cooling curve, 58, 59, 61; of a casting varies shrinkage, 46, 49; of test-bars due to poor molding, 180; one size of test-bars should be used for comparison, 180; of test-bars only approximate, 180; varia- tion of, in test-bars, 123. Slag, 4. Slow cooling increases size of grain, 121. Small castings proportionately stronger than large, 22, 120. INDEX. 22 3 Smith and Weston's tests with sulphur, 85, 93. Smoothing surface of a test-bar strengthens it, 154. Soft, Nos. i and 2 pig iron, 96; castings, to produce, 139, Softeners, 79; Ohio, 202; Scotch, 201; silicon, 44, 56, 192; weak, 201. Solidifying, of cast iron, 19 to 22, 54; test to find quality at each minute, 54. Southern pig irons, analysis of, 167, 176, 178, 195. Specific gravity, of elements of cast iron, 10; of fluid and solid cast iron, 21; of graphite, 25; of phosphorus red and white, 66; of pig irons, 28, 29; of pure iron, 2, 3; of silicon, 34; of sulphur, 82. Spiegeleisen, and silicon, 100; carbon of, 102; whiteness of, caused by low silicon, 102; hardness, 106; PourceFs gray, 103; Ward's gray, 102. Spongy spots in castings, 21, 22, 165; to prevent, 131. Spring-line, 6. Square test-bars stronger than round bars, 128, 129, 130. Standard shrinkage, 2, 159, 168. Star pig-iron analysis, 78. Steel, 3; additions of, dilute all elements, 30, 71; and aluminum, 210; man- ganese, 104, scrap increases strength, 30, 31, 131, 132. Stewart pig iron, 75. Stiffness, 6. Stirling's toughened cast iron, 31. Strain, 5. Strength, 7; and carbon, 132, 133, 134, 136, 138; and manganese, 104; and phosphorus, 80; and silicon, 119, 121; and sulphur 86, 96; breaking, 7; by pounding surface of test-bar, 154; by smoothing the surface of test- bar, 154; by steel or wrought-iron scrap, 30, 31, 131, 132; cannot be cal- culated by formula, 125, 128, 163; chemical composition and, 163, 168, 171; close approximation of, 163; comparisons of, must be made of bars that were cast the same size, 117; curves, 127; decrease of, due to coarse grain, 134; decreases as size increases, 22, 120; diagrams of A. S M. E. tests, 120, 121 ; formula for, 119; graphic method for, 155, 159; increased by tumbling test-bars, 153; increases with each increase of silicon in small castings, 119; influence of shock on, 153; maximum breaking, 7; more dependent on size of grain than on chemical com- position, 171; of A. S. M. E. tests, 115, 116, 119, 130; of annealed cast- ings, 33, 139; of castings made in green and in dry sand, 128; of center oi a casting less than the outside, 122, 128; of horizontally and vertically cast test-bars, 123; of machined and unmachined test-bars, 117, 118, 124; of square and round test-bars, 128, 129, 130; of test-bars 2" X i" tested flat or on edge, 130; proportionate, decreases as size of casting increases, 22, 120; proportional decrease in, due to size, is more rapid and greater with each increase of silicon, 121 ; proportionate, decreases more rapidly with increase in size of small than of large test-bars, 121; removing surface of test-bar decreases, 122, 154; sulphide of iron reduces it, 92; tests from various parts of a cube, 122; to find the actual for any size of test-bar, 162; to find, for a y 2 " D section of a large cast- ing, 162; the lower the silicon the greater, for large test-bars, and vice 224 INDEX. versa, 119; ultimate, 7; variations in, due to character of cast iron, 125; variations of, in test-bars, 123; variations in ordinary castings, 124; with combined carbon and low silicon, 132. Stress, 5. Striking test-bars shortens them, 153. Strong castings, by using an iron low in silicon, 131 ; by using cast iron bor- ings, 131; by using close-grained pig iron, 131; by using wrought iron or steel scrap, 30, 31, 131, 132; to produce, 131. Sulphide of iron, analysis of, 87, 89; compound of, with carbonized iron, 8;; its own iron reduces silicon and carbon in test-bars, 94; reduces strength. 92; shrinkage, 87. Sulphur, 82; and carbon, 82, 83; and chill, 83, 86, 96; and cooling curves, 50; and F L M, 85; and hardness, 58, 93, 98; and shrinkage, 83, 86 to 97; and strength, 86 to 97; added to gray iron by brimstone, 83, 85; by blue-billy, 84; fluor-spar, 83; galena, 83; pyrites, 84, 98; sulphide, 87 to 92; castings contain more than the pig, 96; drives out carbon, 83, 93, 94; in A. S. M. E tests, 96, 97; in cast iron injurious, 82, 95, 97; influence of, in com- mercial pig iron, 95, 96; not beneficial, 97; loss in remelting, 85, 93, 96; overcome by silicon, 93; percentage in pig iron, 82, 95; segregation of, 107; turns gray iron white, 92, 95; Turner's experiments with, 93, 95; Weston and Smith's experiments, 85, 93, 94. Surface, and silicon, 45; removing, weakens, 122, 154; smoothing, strength ns, 154- Swedish irons, and pyrites, 96; segregation of, 107, 108; manganese, 103. Taper scale for measuring shrinkage, 183. Tennessee Coal, Iron and R.R. Co., various grades of iron, 178, 195 to 201. Tensile test, 2, 122, 127, 129, 136. Test-bars, 12, 179; a cold-shut will hasten fracture, 181; advantages and disadvantages of i" n, 180; analysis of (see Analysis); as a coupon not satisfactory, 181; best size, to use, 179; flask for, 182; in yokes, cooling curves, 51; machined or unmachined, 117; patterns for, 182, 184, 191; same size, should be used by all, 179; size of, 12, 179; size, best that shows greatest variation for the least variation in chemical composition, 169; smallest that will run gray is best, 180; strength of square and round, 128; that give most uniform results not best, 180; 2" X i" tested flat or on edge, 130; used in impact tests, 140; variation in size and strength of, 123; Waterworks Association, 131; with yokes, 190; why i" D. gives uniform results, 180. Test, chemical, i; compression or crushing, 2; definition of, i; direct, i: impact, 2; physical, i; preparing for, 12; relative, i; tensile, transverse, 2. Tests, consider general tendencies not individual variations, 76; of aluminum in cast iron, 210; of silicon, Prof. Turner's, 36. Testing machines, advantages, 184; cooling curve, 50, 190; dead-load, 183, 184; drill-grinder, 188; hardness. 186; impact, 189; patterns, 182, 184, 190; recorder, 189; Turner's hardness, 190. Testing small samples of pig iron, 203. I ' ' INDEX. 225 Tin, in cast iron, 212; cooling curve of, 52. Transverse test, 2. Truran's rule for fusibility, 29. Tumbling strengthens test-bars, 153; discovery of A. E. Outerbridge, 153. Turner, Professor Thomas, definition of hardness, 164; experiments with sul- phur, 93, 95; hardness machine, 190; hardness test, 29, 190; on phos- phorus, 66; on silicon, 34, 35, 36; on strength, 180. Usefulness, measure of, 7. Variation, of carbon, 28; in size of test-bars, 123; in strength due to character of cast iron, 125; in strength of machined and unmachined test-bars, 124; in strength of ordinary castings, 124; in test-bars due to poor moulding, 180. Vertically cast test-bars, strength of, 123. Waterworks Association, size of test-bars, 131. Weakness, lines of, 22; of phosphorus irons, 77; of silicon irons, 20. Western Founders' Association report, 129. Weston and Smith's experiments with sulphur, 85 to 93. Wet and dry molds, 175. White pig iron, 28, 201; melts more easily than gray, but is not as fluid, 30; silicon changes to gray, 36; thick and pasty as it cools, 30. White spots in castings, 113, 107 to 114. Workability or hardness, 164. Wrought iron, 3; and aluminum, 210; added to cast iron, 77, 132; shrinkage, 79- Yokes, cooling curves from, 51; of test-bars in, 51; for test-bars, 190. Zinc in cast iron, 212; cooling curves of, 51. SHORT-TITLE CATALOGUE OF THE PUBLICATIONS OF 'JOHN WILEY & SONS, NEW YORK. LOMX>X: CHAPMAN & HALL, LIMITED. ARRANGED UNDER SUBJECTS. Descriptive circulars sent on application. Books marked with an asterisk are sold at net prices only. All books are bound in cloth unless otherwise stated. AGRICULTURE. Armsby's Manual of Cattle-feeding 12mo, $1 75- Downing's Fruits and Fruit-trees of America 8vo, 5 OO 1 Orotenfelt's Principles of Modern Dairy Practice. ( Woll.) . . 12mo, 2 00* Kemp's Landscape Gardening 12mo, 2 SO 7 Maynard's Landscape Gardening as Applied to Home Decora- tion 12mo, 1 50 Stockbridge's Rocks and Soils 8vo, 2 50 Woll's Handbook for Farmers and Dairymen 16nio, 1 50 ARCHITECTURE, Baldwin's Steam Heating for Buildings 12mo, 2 50* Berg's Buildings and Structures of American Railroads. .. .4to, 5 00 Birkm ire's Planning and Construction of American Theatres.Svo, 3 00' Architectural Iron and Steel 8vo, 3 50 Compound Riveted Girders as Applied in Build- ings 8vo, 2 00 " Planning and Construction of High Office Build- ings 8vo, 3 50 Skeleton Construction in Buildings 8vo, 3 00 Briggs's Modern American School Buildings 8vo, 4 00 Carpenter's Heating and Ventilating of Buildings 8vo, 3 00 Freitag's Architectural Engineering 8vo, 3 50 " Fireproofing of Steel Buildings 8vo, 2 50 Gerhard's Guide to Sanitary House-inspection 16mo, 1 00 " Theatre Fires and Panics 12mo, 1 50 Hatfield's American House Carpenter 8vo, 5 00 Holly's Carpenters' and Joiners' Handbook 18mo, 75 Kidder's Architect's and Builder's Pocket-book.. 16mo, morocco, 4 00 Merrill's Stones for Building and Decoration 8vo, 5 00 1 Monckton's Stair-building _... 4to, 4 00 Pattern's Practical Treatise on Foundations 8vo, 5 00 Siebert and Biggin's Modern Stone-cutting and Masonry .. 8vo, 1 50 Wait's Engineering and Architectural Jurisprudence 8vo, 6 00 Sheep, 6 50 " Law of Operations Preliminary to Construction in En- gineering and Architecture v*y 8vo > 5 M Sheep, 5 50 " Law of Contracts 8vo, 3 00 Woodbury's Fire Protection of Mills 8vo, 2 50 Worcester and Atkinson's Small Hospitals, Establishment and Maintenance, and Suggestions for Hospital Architecture, with Plans for a Small Hospital 12mo, 1 25 The World's Columbian Exposition of 1893 Large 4to, 1 00 AEMY AND NAVY. Bernadou's Smokeless Powder, Nitro-cellulose, and the Theory of the Cellulose Molecule 12mo, 2 50 * Bruff's Text-book of Ordnance and Gunnery 8vo, 6 00 Chase's Screw Propellers and Marine Propulsion 8vo, 3 00 Craig's Azimuth. 4to, 3 50 Crehore and Squire's Polarizing Photo-chronograph 8vo, 3 00 Cronkhite's Gunnery for Non-commissioned Officers..24mo, mor., 2 00 * Da vis's Elements of Law 8vo, 2 50 * " Treatise on the Military Law of United States ... 8vo, 7 00 Sheep, 7 50 De Brack's Cavalry Outpost Duties. (Carr.) 24mo, morocco, 2 00 Dietz's Soldier's First Aid Handbook 16mo, morocco, 1 25 * Dredge's Modern French Artillery 4to, half morocco, 15 00 Durand's Resistance and Propulsion of Ships 8vo, 5 00 * Dyer's Handbook of Light Artillery 12mo, 3 00 Eissler's Modern High Explosives 8vo, 4 00 * Fiebeger's Text-book on Field Fortification Small 8vo, 2 00 * Hoff's Elementary Naval Tactics 8vo, 1 50 Ingalls's Handbook of Problems in Direct Fire 8vo, 4 00 * " Ballistic Tables 8vo, 1 50 Lyons's Treatise on Electromagnetic Phenomena 8vo, 600 * Mahan's Permanent Fortifications. (Mercur's.).8vo, half mor. 7 50 Manual for Courts-martial 16mo, morocco, 1 50 * Mercur's Attack of Fortified Places 12mo, 2 00 * " Elements of the Art of War 8vo, 400 Metcalfe's Cost of Manufactures And the Administration of Workshops, Public and Private 8vo, 5 00 " Ordnance and Gunnery 12mo, 5 00 Murray's Infantry Drill Regulations 18mo, paper, 10 * Phelps's Practical Marine Surveying 8vo, 2 50 Powell's Army Officer's Examiner 12mo, 4 00 Sharpe's Art of Subsisting Armies in War . 18mo, morocco, 1 50 Walke's Lectures on Explosives 8vo, 4 00 * Wheeler's Siege Operations and Military Mining 8vo, 2 00 Winthrop's Abridgment of Military Law 12mo, 2 50 WoodhiiU's Notes on Military Hygiene 16mo, 1 50 Young's Simple Elements of Navigation 16mo, morocco, 1 00 Second Edition, Enlarged and Revised 16mo, mor., 2 00 ASSAYING. Fletcher's Practical Instructions in Quantitative Assaying with the Blowpipe 12mo, morocco, 1 50 Furman's Manual of Practical Assaying 8vo, 3 00 Miller's Manual of Assaying i*. nt .> .... 12mo, 1 00 O'DriscolPs Notes on the Treatment of Gold Ores 8vo, 2 00 Kicketts and Miller's Notes on Assaying :> j>'ie4-8vo, 3 00 Wilson's Cyanide Processes 12mo, 1 50 , Chlorination Process 12mo, 1 50 ASTRONOMY. Craig's Azimuth 4to, 3 50 Doolittle's Treatise on Practical Astronomy 8vo, 4 00 Gore's Elements of Geodesy 8vo, 2 50 Hayford's Text-book of Geodetic Astronomy 8vo, 3 00 Merriman's Elements of Precise Surveying and Geodesy .... 8vo, 2 50 * Michie and Harlow's Practical Astronomy 8vo, 3 00 * White's Elements of Theoretical and Descriptive Astronomy. 12mo, 2 00 BOTANY. Baldwin's Orchids of New England Small 8vo, 1 50 Davenport's Statistical Methods, with Special Reference to Bio- logical Variation 16mo, morocco, 1 25 Thome and Bennett's Structural and Physiological Botany. 16mo, 2 25 Westermaier's Compendium of General Botany. ( Schneider. )8vo, 2 00 CHEMISTRY. Adriance's Laboratory Calculations and Specific Gravity Tables, 12mo, 1 25 Allen's Tables for Iron Analysis 8vo, 3 00 Arnold's Compendium of Chemistry. (Mandel.) (In preparation.) Austen's Notes for Chemical Students 12mo, 1 50 Bernadou's Smokeless Powder. Nitro-cellulose, and Theory of the Cellulose Molecule 12mo, 2 50 Bolton's Quantitative Analysis 8vo, 1 50 Brush and Penfield's Manual of Determinative Mineralogy..8vo, 4 00 Classen's Quantitative Chemical Analysis by Electrolysis. (Her- rick Boltwood.) 8vo, 300 Cohn's Indicators and Test-papers 12mo, 2 00* Craft's Short Course in Qualitative Chemical Analysis. (Schaef- fer.) 12mo, 2 00 DrechsePs Chemical Reactions. (Merrill.) 12mo, 1 25 Eissler's Modern High Explosives 8vo, 4 00 EfTront's Enzymes and their Applications. ( Prescott. ) ( In preparation. ) Erdmann's Introduction to Chemical Preparations. (Dunlap.) 12mo, 1 25 Fletcher's Practical Instructions in Quantitative Assaying with the Blowpipe 12mo, morocco, 1 50" Fresenius's Manual of Qualitative Chemical Analysis. (Wells.) 8vo, 5 00 System of Instruction in Quantitative Chemical Analysis. (Allen.) 8vo, 600 Fuertes's Water and Public Health 12mo, 1 50 Furman's Manual of Practical Assaying 8vo, 3 00 Gill's Gas and Fuel Analysis for Engineers 12mo, 1 25 Grotenfelt's Principles of Modern Dairy Practice. (Woll.) . . 12mo, 2 00 Hammarsten's Text-book of Physiological Chemistry. (Mandel.) 8vo, 4 00 Helm's Principles of Mathematical Chemistry. (Morgan.) . 12mo, 1 50 Holleman's Text-book of Inorganic Chemistry. (Cooper.) (In preparation.) Hopkins's Oil-chemists' Handbook 8vo, 3 00 Keep's Cast Iron. (In preparation.) Ladd's Manual of Quantitative Chemical Analysis 12mo, 1 00 Landauer's Spectrum Analysis. (Tingle.) 8vo, 3 00 Lassar-Cohn's Practical Urinary Analysis. (Lorenz.) (In preparation.) Lob's Electrolysis and Electrosynthesis of Organic Compounds. (Lorenz.) . .' 12mo, 1 00 Mandel's Handbook for Bio-chemical Laboratory 12mo, 1 5Q Mason's Water-supply. (Considered Principally from a Sani- tary Standpoint.) 8vo, 5 00 " Examination of Water. (Chemical and Bacterio- logical.) 12mo, 1 25 Meyer's Determination of Radicles in Carbon Compounds. (Tingle.) 12mo, 1 00 Miller's Manual of Assaying 12mo, 1 00 Mixter's Elementary Text-book of Chemistry 12mo, 1 50 Morgan's Outline of Theory of Solution and its Results . . . 12ino, 1 Oft " Elements of Physical Chemistry 12mo, 2 00 Nichols's Water-supply. (Considered mainly from a Chemical and Sanitary Standpoint, 1883.) 8vo, 2 50 O'Brine's Laboratory Guide in Chemical Analysis 8vo, 2 00 O'Driscoll's Notes on the Treatment of Gold Ores 8vo, 2 00- Ost and Kolbeck's Text-book of Chemical Technology. (Lor- enz Bozart.) (In. preparation.) 4 * Penfield's Notes on Determinative Mineralogy and Record of Mineral Tests 8vo, paper, 50 Pinner's Introduction to Organic Chemistry. (Austen.) . . . 12mo, 1 50 Poole's Calorific Power of Fuels 8vo, 3 00 "* Reisig's Guide to Piece-dyeing 8vo, 25 00 Richards and Woodman's Air, Water, and Food from a Sanitary Standpoint 8vo, 2 00 Richards's Cost of Living as Modified by Sanitary Science . 12mo, 1 00 Cost of Food, a Study in Dietaries 12ma, 1 00 Ricketts and Russell's Skeleton Notes upon Inorganic Chem- istry. (Part I. Non-metallic Elements.) . .8vo, morocco, 75 Ricketts and Miller's Notes on Assaying 8vo, 3 00 Rideal's Sewage and, the Bacterial Purification of Sewage. .8vo, 3 50 Ruddiman's Incompatibilities in Prescriptions 8vo, 2 00 Schimpf a Text-book of Volumetric Analysis 12mo, 2 50 Spencer's Handbook for Chemists of Beet-sugar Houses. 16mo, morocco, 3 00 " Handbook for Sugar Manufacturers and their Chem- ists 16mo, morocco, 2 00 Stockbridge's Rocks and Soils , 8vo, 2 50 * Tillman's Elementary Lessons in Heat 8vo, 1 50 Descriptive General Chemistry 8vo, 3 00 Turneaure and Russell's Public Water-supplies 8vo, 5 00 Van Deventer's Physical Chemistry for Beginners. (Boltwood.) 12mo, 1 50 Walke's Lectures on Explosives 8vo, 4 00 Wells's Laboratory Guide in Qualitative Chemical Analysis. ..8vo, 1 50 " Short Course in Inorganic Qualitative Chemical Analy- sis for Engineering Students 12mo, 1 50 Whipple's Microscopy of Drinking-water 8vo, 3 SO Wiechmann's Sugar Analysis Small 8vo, 2 50 " Lecture-notes on Theoretical Chemistry .... 12mo, 3 00 Wilson's Cyanide Processes 12mo, 1 50 " Chlorination Process 12mo, 1 50 Wulling's Elementary Course in Inorganic Pharmaceutical and Medical Chemistry 12mo, 2 00 CIVIL ENGINEERING. BRIDGES AND ROOFS. HYDRAULICS. MATERIALS OF ENGINEERING. RAILWAY ENGINEERING. Baker's Engineers* Surveying Instruments. 12mo, 3 00 Bixby's Graphical Computing Table Paper, 19 x 24 inches. 25 Davis's Elevation and Stadia Tables 8vo, 1 00 Folwell's Sewerage. (Designing and Maintenance.) 8vo, 3 00 Freitag's Architectural Engineering 8vo, 3 50 5 Goodhue's Municipal Improvements 12mo, 1 75 Goodrich's Economic Disposal of Towns' Refuse 8vo, 3 50 Gore's Elements of Geodesy 8vo, 2 50 Hayford's Text-book of Geodetic Astronomy 8vo, 3 00 Howe's Retaining-walls for Earth 12mo, 1 25 Johnson's Theory and Practice of Surveying Small 8vo, 4 00 " Stadia and Earth-work Tables 8vo, 1 25 Kiersted's Sewage Disposal 12mo, 1 25 Mahan's Treatise on Civil Engineering. (1873.) (Wood.) . .8vo, 5 00 * Mahan's Descriptive Geometry 8vo, 1 50 Merriman's Elements of Precise Surveying and Geodesy. . . .8vo, 2 50 Merriman and Brooks's Handbook for Surveyors. . . .16mo, mor., 2 00 Merriman's Elements of Sanitary Engineering. . . 8vo, 2 00 Nugent's Plane Surveying. (In preparation.) Ogden's Sewer Design 12mo, 2 00 Patton's Treatise on Civil Engineering. ..... .8vo, half leather, 7 50 Reed's Topographical Drawing and Sketching 4to, 5 00 Rideal's Sewage and the Bacterial Purification of Sewage . . 8vo, 3 50 Siebert and Biggin's Modern Stone-cutting and Masonry. .8vo, 1 50 Smith's Manual of Topographical Drawing. (McMillan.) . .8vo, 2 50 * Trautwine's Civil Engineer's Pocket-book. .. .16mo, morocco, 5 00 Wait's Engineering and Architectural Jurisprudence 8vo, 6 00 Sheep, 6 50 " Law of Operations Preliminary to Construction in En- gineering and Architecture 8vo, 5 00 Sheep, 5 50 " Law of Contracts 8vo, 3 00 Warren's Stereotomy Problems in Stone-cutting 8vo, 2 50 Webb's Problems in the Use and Adjustment of Engineering Instruments 16mo, morocco, 1 25 * Wheeler's Elementary Course of Civil Engineering 8vo, 4 00 Wilson's Topographic Surveying 8vo, 3 50 BRIDGES AND ROOFS. Boiler's Practical Treatise on the Construction of Iron Highway Bridges 8vo, 2 00 * Boiler's Thames River Bridge 4to, paper, 5 00 Burr's Course on the Stresses in Bridges and Roof Trusses, Arched Ribs, and Suspension Bridges 8vo, 3 50 Du Bois's Stresses in Framed Structures Small 4to, 10 00 Foster's Treatise on Wooden Trestle Bridges 4to, 5 00 Fowler's Coffer-dam Process for Piers 8vo, 2 50 Greene's Roof Trusses 8vo, 1 25 " Bridge Trusses 8vo, 2 50 Arches in Wood. Iron, and Stone 8vo, 2 50- Howe's Treatise on Arches. . 8vo, 4 00 Johnson, Bryan and Tumeaure's Tlieory and Practice in the Designing of Modern Framed Structures Small 4to, 10 00 Merriman and Jaeoby's Text-book on Roofs and Bridges: Part I. Stresses in Simple Trusses 8vo, 2 50 Part II. Graphic Statics 8vo, 2 00 Part IIL Bridge Design. Fourth Ed. (In preparation.) . .&vo, 250 Part IV. Higher Structures 8vo, 2 50 Morison's Memphis Bridge 4to, 10 00 Waddell's De Pontibus, a Pocket Book for Bridge Engineers. 16mo, mor., 3 00 Specifications for Steel Bridges 12mo, 1 25 Wood's Treatise on the Theory of the Construction of Bridges and Roofs Svo, 2 00 Wright's Designing of Draw-spans: Part I. Plate-girder Draws .8vo, 2 50 Part II. Riveted-truss and Pin-connected Long-span Draws. 8vo, 2 50 Two parts in one volume 8vo, 3 50 HYDRAULICS. Bazin's Experiments upon the Contraction of the Liquid Vein Issuing from an Orifice. (Trautwine.) Svo, 2 00 Bovey's Treatise on Hydraulics 8vo, 5 00 Church's Mechanics of Engineering Svo, 6 00 Coffin's Graphical Solution of Hydraulic Problems. . 16mo, mor., 2 50 Flather's Dynamometers, and the Measurement of Power. 12mo, 3 00 Fehvell's Water-supply Engineering Svo, 4 00 Frizell's Water-power Svo, 5 00 Fuertes's Water and Public Health 12mo, 1 50 Water-filtration Works 12mo, 2 50 Ganguillet and Kutter's General Formula for the Uniform Flow of Water in Rivers and Other Channels. (Her- ing and Trautwine.) Svo, 4 00 Hazen's Filtration of Public Water-supply Svo, 3 00 Hazleurst's Towers and Tanks for Water-works Svo, 2 50 HerschePs 115 Experiments on the Carrying Capacity of Large, Riveted, Metal Conduits Svo, 2 00 Mason's Water-supply. (Considered Principally from a Sani- tary Standpoint.) Svo, 5 00 Merriman's Treatise on Hydraulics Svo, 4 00 * Michie's Elements of Analytical Mechanics Svo, 4 00 Schuyler's Reservoirs for Irrigation, Water-power, and Domestic Water-supply Large Svo, 5 00 Turneaure and Russell. Public Water-supplies Svo, 5 00 Wegmann's Design and. Construction of Dams 4to, 5 00 Water-supply of the City of New York from 1C58 to 1895 .-I'.-; 4to, 10 00 Weisbach's Hydraulics and Hydraulic Motors. (Du Bois.) . .8vo, 5 00 Wilson's Manual of Irrigation Engineering Small Svo, 4 00 Wolff's Windmill as a Prime Mover Svo, 3 00 Wood's Turbines Svo, 2 50 " Elements of Analytical Mechanics Svo, 3 00 MATERIALS OF ENGINEERING. Baker's Treatise on Masonry Construction Svo, 500 Black's United States Public Works Oblong 4to, 5 00 Bovey's Strength of Materials and Theory of Structures .... Svo, 7 50 7 Burr's Elasticity and Resistance of the Materials of Engineer- ing , . . 8vo, 5 00 Byrne's Highway Construction 8vo, 5 00 " Inspection of the Materials and Workmanship Em- ployed in Construction 16mo, 3 00 Church's Mechanics of Engineering 8vo, 6 00 Du Bois's Mechanics of Engineering. Vol. I Small 4to, 10 00 Johnson's Materials of Construction Large 8vo, 6 00 Keep's Cast Iron. (In preparation.) Lanza's Applied Mechanics 8vo, 7 50 Martens's Handbook on Testing Materials. (Henning.) 2 vols., 8vo, 7 50 Merrill's Stones for Building and Decoration 8vo, 5 00 Merriman's Text-book on the Mechanics of Materials 8vo, 4 00 Merriman's Strength of Materials 12mo, 1 00 Metcalf s Steel. A Manual for Steel-users 12mo, 2 00 Patton's Practical Treatise on Foundations 8vo, 5 00 Rockwell's Roads and Pavements in France 12mo, 1 25 Smith's Wire: Its Use and Manufacture Small 4to, 3 00 Spalding's Hydraulic Cement 12mo, 2 00 Text-book on Roads and Pavements 12mo, 2 00 Thurston's Materials of Engineering 3 Parts, 8vo, 8 00 Part I. Non-metallic Materials of Engineering and Metal- lurgy 8vo, 200 Part II. Iron and Steel 8vo, 3 50 Part III. A Treatise on Brasses, Bronzes and Other Alloys and Their Constituents 8vo, 2 50 Thurston's Text-book of the Materials of Construction 8vo, 5 00 Tillson's Street Pavements and Paving Materials 8vo, 4 00 Waddell's De Pontibus. (A Pocket-book for Bridge Engineers.) 16mo, morocco, 3 00 Specifications for Steel Bridges 12mo, 1 25 Wood's Treatise on the Resistance of Materials, and an Ap- pendix on the Preservation of Timber .8vo, 2 00 " Elements of Analytical Mechanics 8vo, 3 00 RAILWAY ENGINEERING. Berg's Buildings and Structures of American Railroads . . 4to, 5 00 Brooks's Handbook of Street Railroad Location. . 16mo, morocco, 1 50 Butts's Civil Engineer's Field-book 16mo, morocco, 2 50 CrandalFs Transition Curve 16mo, morocco, 1 50 Railway and Other Earthwork Tables 8vo, 1 50 Dawson's Electric Railways and Tramways . Small 4to, half mor., 12 50 " " Engineering " and Electric Traction Pocket-book. 16mo, morocco, 4 00 Dredge's History of the Pennsylvania Railroad: (1879.) .Paper, 5 00 * Drinker's Tunneling, Explosive Compounds, and Rock Drills. 4to, half morocco, 25 00 Fisher's Table of Cubic Yards Cardboard, 25 Godwin's Railroad Engineers' Field-book and Explorers' .Guide. 16mo, morocco, 2 50 Howard's Transition Curve Field-book 16mo, morocco, 1 50 Hudson's Tables for Calculating the Cubic Contents of Exca- vations and Embankments .._. 8vo, 1 00 Nagle's Field Manual for Railroad Engineers. . . . 16mo, morocco, 3 00 Philbrick's Field Manual for Engineers 16mo, morocco, 3 00 Pratt and Alden's Street-railway Road-bed 8vo, 2 00 Searles's Field Engineering 16rno, morocco, 3 00 " Railroad Spiral 16mo, morocco, 1 50 Taylor's Prismoidal Formulae and Earthwork 8vo, 1 50 * Trautwine's Method of Calculating the Cubic Contents of Ex- cavations and Embankments by the Aid of Dia- grams 8vo, 2 00 * The Field Practice of Laying Out Circular Curves for Railroads 12mo, morocco, 2 50 * " Cross-section Sheet Paper, 25 Webb's Railroad Construction 8vo, 4 00 Wellington's Economic Theory of the Location of Railways. . Small Svo, 5 00 DRAWING. Barr's Kinematics of Machinery 8vo, 2 50 * Bartlett's Mechanical Drawing 8vo, 3 00 Durley's Elementary Text-book of the Kinematics of Machines. (In preparation.) Hill's Text-book on Shades and Shadows, and Perspective. . 8vo, 2 00 Jones's Machine Design : Part I. Kinematics of Machinery 8vo, 1 50 Part II. Form, Strength and Proportions of Parts 8vo, 3 00 MacCord's Elements of Descriptive Geometry Svo, 3 00 Kinematics; or, Practical Mechanism Svo, 500 Mechanical Drawing 4to, 4 00 Velocity Diagrams Svo, I 50 * Mahan's Descriptive Geometry and Stone-cutting Svo, 1 50 Mahan's Industrial Drawing. (Thompson.) Svo, 3 50 Reed's Topographical Drawing and Sketching 4to, 5 00 Reid's Course in Mechanical Drawing Svo, 2 00 " Text-book of Mechanical Drawing and Elementary Ma- chine Design Svo, 3 00 Robinson's Principles of Mechanism Svo, 3 00 Smith's Manual of Topographical Drawing. (McMillan.) .Svo, 2 50 Warren's Elements of Plane and Solid Free-hand Geometrical Drawing 12mo, 1 00 Drafting Instruments and Operations 12mo, 1 25 Manual of Elementary Projection Drawing. ... 12mo, 1 50 " Manual of Elementary Problems in the Linear Per- spective of Form and Shadow 12mo, 1 00 " Plane Problems in Elementary Geometry 12mo, 1 25 Primary Geometry 12mo, 75 Elements of Descriptive Geometry, Shadows, and Per- spective Svo, 3 50 General Problems of Shades and Shadows Svo, 3 00 Elements of Machine Construction and Drawing. .Svo, 7 50 Problems, Theorems, and Examples in Descriptive Geometry Svo, 2 50 Weisbach's Kinematics and the Power of Transmission. (Herr- mann and Klein.) Svo, 5 00 Whelpley's Practical Instruction in the Art of Letter En- graving 12mo, 2 00 Wilson's Topographic Surveying Svo, 3 50 Wilson's Free-hand Perspective Svo, 2 56 Woolf's Elementary Course in Descriptive Geometry. .Large Svo, 3 00 ELECTRICITY AND PHYSICS. Anthony and Brackett's Text-book of Physics. (Magie.) Small 8vo, 3 00- Anthony's Lecture-notes on the Theory of Electrical Measur- ments 12mo. 1 00 Benjamin's History of Electricity. 8vo. 3 00 Benjamin's Voltaic Cell 8vo. 3 00 Classen's Qantitative Chemical Analysis by Electrolysis. Her- rick and Boltvvood.) Svo. 3 00 Crehore and Squier's Polarizing Photo-chronograph Svo, 3 00 Dawson's Electric Railways and Tramways.. Small 4to, half mor.. 12 50 Dawson's " Engineering " and Electric Traction Pocket-book. 16mo, morocco, 4 00 Flather's Dynamometers, and the Measurement of Power. . 12mo, 3 00 Gilbert's De Magnete. (Mottelay.) Svo, 2 50 Holman's Precision of Measurements Svo, 2 00 Telescopic Mirror-scale Method, Adjustments, and Tests Large Svo, 75 Landauer's Spectrum Analysis. (Tingle.) Svo. 3 00 Le Chatelier's High-temperature Measurements. (Boudouard Burgess.) 12mo. 3 00 Lob's Electrolysis and Electrosynthesis of Organic Compounds. (LorenzO 12mo, 1 00 Lyons's Treatise on Electromagnetic Phenomena Svo, 6 00 * Michie. Elements of Wave Motion Relating to Sound and Light Svo. 4 00 Niaudet's Elementary Treatise on Electric Batteries (Fish- back.) " 12mo. 2 50 * Parshall and Hobart's Electric Generators..Small 4to, half mor., 10 00 Thurston's Stationary Steam-engines Svo. 2 50 * Tillman. Elementary Lessons in Heat Svo. 1 50 Tory and Pitcher. Manual of Laboratory Physics. .Small Svo. 2 00 LAW. * Davis. Elements of Law Svo. 2 50 * " Treatise on the Military Law of United States.. Svo, 7 00 Sheep, 7 50 Manual for Courts-martial IGmo, morocco, 1 50 Wait's Engineering and Architectural Jurisprudence Svo, 6 00 Sheep', 6 50 " Law of Operations Preliminary to Construction in En- gineering and Architecture Svo, 5 00 Sheep, 5 50 " Law of Contracts Svo, 3 00 Winthrop's Abridgment of Military Law 12mo, 2 50 MANUFACTURES. Beaumont's Woollen and Worsted Cloth Manufacture. . . . 12mo, 1 50 Bernadou's Smokeless Powder Nitro-cellulose and Theory of the Cellulose Molecule *2mo, 2 50 Bolland's Iron Founder 12mo, cloth, 2 50 '" The Iron Founder " Supplement 12mo, 2 50 " Encyclopedia of Founding and Dictionary of Foundry Terms Used in the Practice of Moulding. .. .12mo, 3 00 Eissler's Modern High Explosives Svo, 4 00 Effront's Enzymes and their Applications. (Prescott.) (In preparation.) Fitzgerald's Boston Machinist ISmo, 1 00 10 Ford's Boiler flaking for Boiler Makers 18mo, 1 00 Hopkins' is Oil-chemists' Handbook 8vo, 3 00 Keep's Cast Iron. (In preparation.) Metcalf's Steel. A Manual for Steel-users 12mo, 2 00 Metcalfs Cost of Manufactures And the Administration of Workshops, Public and Private 8vo, 5 00 Meyer's Modern Locomotive Construction 4to, 10 00 * Reisig's Guide to Piece-dyeing 8vo, 25 00 Smith's Press-working of Metals 8vo, 3 00 " Wire: Its Use and Manufacture Small 4to, 3 00 Spalding's Hydraulic Cement 12mo, 2 00 Spencer's Handbook for Chemists of Beet-sugar Houses. 16mo, morocco, 3 00 Handbook for Sugar Manufacturers and their Chem- ists 16mo, morocco, 2. 00 Thurston's Manual of Steam-boilers, their Designs, Construc- tion and Operation 8vo, 5 00 Walke's Lectures on Explosives 8vo, 4 00 West's American Foundry Practice 12mo, 2 50 " Moulder's Text-book 12mo, 2 50 Wiechmann's Sugar Analysis Small 8vo, 2 50 Wolff's Windmill as a Prime Mover 8vo, 3 00 Woodbury's Fire Protection of Mills 8vo. 2 50 MATHEMATICS. Baker's Elliptic Functions 8vo, 1 50 * Bass's Elements of Differential Calculus 12mo. 4 00 Briggs's Elements of Plane Analytic Geometry 12mo. 1 00 Chapman's Elementary Course in Theory of Equations. . .12mo, 1 50 Compton's Manual of Logarithmic Computations 12mo, 1 .~>