CITY OF NEW YORK THE AQUEDUCT COMMISSIONERS. REPORT OF THE RESULTS EXPERIMENTS AT JEROME PARK RESERVOIR TO DETERMINE THE LAWS OF PROPORTIONING CONCRETE. MADE BY AUTHORITY OF THE AQUEDUCT COMMISSIONERS, UNDER THE DIRECTION OF THE CHIEF ENGINEER. BY WM. B. FULLER, MEMBER AMERICAN SOCIETY OF CIVIL ENGINEERS. Evening lpost 3ob (printing ©ffice, mew H?orft. TABLE OF CONTENTS- PAGE Conclusions from Tests. 4 Further Tests Suggested and Enumerated. 8 Description of Tests. 10 Mechanical Analysis. 13 Sieves and Other Apparatus. 13 Plotting Analysis Curves. 15 Preparation of Materials. 16 Cement. 16 Purity Test of Cement. 17 Aggregate . 17 Character and Composition of Jerome Park Rock. 18 Uselessness of Determination of Voids in Sand. 20 Screening Aggregate in Laboratory. 20 Diameters of Aggregate. 22 Methods and Apparatus for Determining Density. 23 Weighing. 23 Measuring. 23 Mixing... 25 Ramming. 26 Removing Surplus Water. 26 Final weighing. 26 Measurement.. . 28 Recording and Computing Data. 28 Proportioning the Ingredients for Maximum Density. 30 Necessity for Using Cement in Density Tests. 31 Density Tests with 2 14 -Inch Jerome Park Stone. 32 Density Tests with 1-Inch and %-Inch Stone. 33 Cement vs. Fine Sand. 35 Density Tests with Cowe Bay Material. 35 Ideal Sand. 36 Equations of Ideal Mechanical Analysis Curves. 36 Directions for Plotting Ellipses. 37 Diagram of Ideal Curves. 37 Methods and Apparatus for Beams. 38 Weighing. 38 Mixing. 38 Consistency. 38 Marks for Specimens. 39 Placing in Molds. 39 Recording and Computing Data on Beams. 40 Beam Testing Machine. 41 Compression Pieces. 42 Compressive Strength of True Prisms vs. Capped Pieces of Beams .... 43 1904 Beam Tests. 44 1905 Beam Tests. 44 Methods of Preparing Materials for Natural Proportions. 47 Mechanical Analysis Curves Used in Beam Tests. 48 Equations of Mechanical Analysis Curves for Artificial Propor¬ tioning of Aggregates and Cement... 57 Comparative Tests Forming Basis of Conclusions. 57 Aggregates of Different Maximum Size, Comparative Density and Strength. 57 I 0 IV PAGE Jerome Park vs. Cowe Bay vs. Mixed Aggregates, Comparative Density and Strength. 59 Graded Aggregate vs. Mixtures in Ordinary Proportions, Compara¬ tive Density and Strength. 63 Graded Coarse Aggregate vs. Uniform Size, Comparative Density and Strength. 63 Influence of the Analysis of the Coarse Aggregate upon the Strength of the Concrete. 66 Effect of Fineness of Sand upon Density and Strength. 67 Relation of Ideal Curves of Different Size Stone. 67 Effect on Density and Strength of Concrete by Increasing Percent¬ age of Sand and Decreasing Stone. 68 Proportioning Sand and Stone in Practice. 69 Concrete with Different Percentages of Cement, Comparative Den¬ sity and Strength. 69 Effect upon Density of Substituting Cement for Fine Sand. 70 Complete Schedule of Transverse and Compressive Tests, 1905. 70 Beams Made with Different Brands of Cement, Comparative Density and Strength. 73 Permeability Tests. 74 Method of Making Permeability Tests. 74 Apparatus for Testing Permeability. . .. 76 Recording the Data. 76 Adhesion of Cement Casing to Concrete Block. 78 Results of Permeability Tests. 78 Effect of Per Cent, of Cement upon Permeability. 78 Effect of Size of Stone upon Permeability. 78 Decrease of Permeability with Age. 79 Increase of Permeability with Pressure. 80 Effect of Thickness of Concrete upon Permeability.. 80 Rate of Flow During A Four-Hour Period. 80 Comparative Strength of Mortars and Permeability of Concretes with Several Available Sands. 81 Volumetric Tests of Density of Neat Cement and Mortar. 83 Effect of Different Percentages of Water upon Volume and Density. 83 Tensile Tests of Neat Cement Used in Concrete Beams, 1905. 84 Mechanical Analysis of Average Jerome Park and Cowe Bay Mate¬ rials Used in Beam Tests. 85 Elasticity Tests. 86 Effect of Size of Stone on Modulus of Elasticity. 92 Effect of Percentage of Cement on Modulus of Elasticity. 92 Effect of Character of Material on Modulus of Elasticity. 92 CITY OF NEW YORK THE AQUEDUCT COMMISSIONERS Hon. GEORGE B. McCLELLAN, Mayor; HERMAN A, METZ, Comptroller; JOHN F. COWAN; JOHN P. WINDOLPH; WILLIAM H. TEN EYCK; JOHN J. RYAN. WALTER H. SEARS, Chief Engineer. Printed in Accordance with a Resolution of THE AQUEDUCT COMMISSIONERS March 16, 1906. Digitized by the Internet Archive in 2017 with funding from University of Illinois Urbana-Champaign Alternates https://archive.org/details/reportofresultexOOnewy 170 Broadway, New York City, Oct. 11, 1905. To the Aqueduct Commissioners , New York City: Gentlemen : In accordance with your resolution and acting under the in¬ structions of your Chief Engineer, I herewith present a report on the results of the concrete experiments made at Jerome Park Reser¬ voir during the past two years. The object of the tests has been the comparative study of the laws governing the mixtures of cement and aggregates for con¬ crete with a view to determine the most economical mixtures and the mixtures giving the greatest strength and impermeability with a minimum percentage of cement, with special reference to the ad¬ visability and propriety of using materials available at Jerome Park Reservoir in that work. Previous tests and studies of the writer indicated that with the same percentage of cement, but different arrangement of the ag¬ gregates, the strongest concrete is that in which the aggregate is proportioned so as to give a concrete of the greatest density, that is, with the smallest percentage of voids. With this law established, it would be possible to compare the value of different aggregates and various proportions of the same aggregates by volumetric tests, the best mixture in general giving the smallest volume of concrete. The tests were accordingly arranged with the view to fix the limitations of this theory, to ascertain the effect of different ag¬ gregates upon the density, strength and permeability of concrete, and particularly to determine the exact sizes of aggregate which, mixed together with a given proportion of cement, would form the best concrete. The density and strength of concrete made up of an aggregate of ideal mechanical analysis and of different maximum size was compared with the density and strength of concrete made of average sand and stone or gravel in ordinary nominal mixtures. 4 The relative permeability of concrete with different aggregates and different proportions of cement was also studied. In this connection it became necessary to study many related subjects, such as the physical characteristics of different cements, broken stone, broken stone screenings, gravel and sand, especially with reference to specific gravity, weight, voids, density, etc. Each material was tested by itself and when mixed in various proportions with others. The methods and results of all these studies are given in detail in the accompanying appendix by Mr. Sanford E. Thompson, Mem. Am. Soc. C. E., who had direct supervision of the details of the tests and arrangement of their records. The practical side of the question was not lost sight of; in fact, all of the preliminary studies were made to throw light on the best attainable concrete in the actual construction of the reservoir, and from the results of these studies several changes were made from time to time in the sizes, proportions and methods of mixing and placing, which it is believed have resulted in a grade of concrete construction at the reservoir fully in accordance with the best engineering practice of the day and of a higher grade than was at first thought practicable under the contract. As shown by the tests, an ideal mixture of Co we Bay sand and gravel with a given proportion of cement gave a denser concrete than an ideal mixture of Jerome Park screenings and broken stone with the same proportion of cement, but on the other hand, there was under equal conditions a stronger adhesion of the cement to the broken stone than to the gravel. As the proportion of cement specified for the reservoir construction was such as to produce im¬ permeability in either case, it was advised that screenings and broken stone be used on account of the additional adhesion. The following conclusions have been drawn from the tests. Many of these introduce laws which we believe have not hitherto been recognized, and which are opposed in certain cases to current ideas, so that some of the conclusions are of a tentative character and require further confirmation. 1.—The largest size stone makes the strongest concrete under both compression and transverse loading, i. e., an aggregate whose maximum size stone is 2i in. diameter gives stronger concrete than 5 an aggregate with 1 in. maximum size, and the 1-in. stone gives a stronger concrete than a |-in. stone. A concrete whose aggregate runs to 1 in. maximum size will require for equal strength about & more cement, and with aggregate running to i in. maximum size about i more cement than concrete with aggregate whose maximum size is 2\ in. 2. —The largest stone makes the densest concrete. Concrete made with stone having a maximum diameter of 2£ in. is notice¬ ably denser than that with 1-in. stone, and this is denser than that with i-in. stone. 3. —Round material like gravel gives under similar conditions a denser concrete than broken stone. 4. —Sand produces a denser concrete than screenings when used with the same proportions of stone and cement. 5. —Cement, sand and gravel concrete is stronger than concrete •of cement, screenings and broken stone, probably because of this greater density. Concrete of cement, sand and broken stone, how¬ ever, is found to be stronger than concrete of cement, sand and gravel, although the latter mix is denser, thus indicating a stronger adhesion of cement to broken stone than to gravel. 6. —Aggregates whose mechanical analysis (including the cement), has been formed so as to give, when water is added, an artificial mixture of greatest density produce concrete of higher strength than mixtures of cement and natural materials in similar propor¬ tions. The average improvement in strength by artificial grading under the conditions of the tests was about 14 per cent. Comparing the tests of strength of concrete having different percentages of ce¬ ment, it is found that for similar strength the best artificially graded aggregate would require about 12% less cement than like mixtures of natural materials. Artificial grading to obtain strength is evidently economical only up to the point where its additional cost is equal to the cost of cement saved by the process. 7. —The ideal mechanical analysis curve, i. e., the best curve, is slightly different for different materials. Cowe Bay sand and gravel, for example, pack closer than Jerome Park stone and screen¬ ings, and therefore require less of the size of grain which we desig¬ nate as sand. 8. —The tests indicate that the best mixture of cement and ag- e gregate has a mechanical analysis curve resembling a parabola, which is a combination of a curve approaching an ellipse for the sand portion and a tangent straight line for the stone portion. The ellipse runs to a diameter to the diameter of the maximum size stone, and the stone from this point is uniformly graded. 9. —The strength and density of concrete appears to be but slightly, if at all, affected by decreasing the quantity of the medium size stone of the aggregate and increasing the quantity of the coarsest stone. An excess of stone of medium size, on the other hand, appreciably decreases the density and strength of the con¬ crete. 10. —The density of concrete is affected by the variation in di¬ ameter of the particles of sand more than by variation in the diame¬ ters of the stone particles, provided the maximum diameter of the stone remains the same and there is not an excess of medium-size stone. 11. —An excess of fine or of medium sand decreases the density and also the strength of the concrete, while, on the other hand, a deficiency of fine grains of sand in a lean concerte may result in un¬ filled voids. 12. —The form of the best analysis curve for any given material is nearly the same for all sizes of stone, that is, the curve for l*in., 1-in., and 21-in. maximum stone may be described by an equation with the maximum diameter as the only variable. In other words, suppose a diagram in which the left ordinate is zero, and the ex¬ treme right ordinate corresponds to 21-in. stone, with the best curve for this stone drawn upon it. If, now, on this diagram the vertical scale remains the same, but the horizontal scale is increased two and a quarter times, so that the diameter of 1-in. stone corresponds to the extreme right-hand ordinate, the best curve for the 1-in. stone will be very nearly the one already drawn for the 21-in. stone. The chief difference between the two appears to be that the larger size stone requires a slightly higher curve in the fine sand portion. If the two curves are drawn to the same scale, the 21-in. curve is of course lengthened out so that much less fine material is actually required than for the 1-in. curve. 13. —It follows from this last conclusion that from a scientific standpoint the term sand is a relative one. With 21-in. stone, the 7 best sand would range in size from 0 to 0.22 in. diameter, while the best sand for ^-in. stone would range in size from 0 to 0.05 in. diameter. 14. —In ordinary proportioning with two given kinds of aggre¬ gate and a given percentage of cement, the densest and strongest mixture is attained when the volume of the mixture of sand, ce¬ ment and water is so small as to just fill the voids in the stone. In ether words, in practical construction use as small a proportion of sand and as large a proportion of stone as is possible without pro¬ ducing visible voids in the concrete. 15. —The substitution of cement for fine sand of the same size grains does not affect the density of the mixture, but increases the strength, although in a slightly smaller ratio than the increase in the percentage of cement. 16. —Correct proportioning of concrete consists in finding with any percentage of cement a concrete mixture of maximum density, and increasing or decreasing the cement by substituting it for sand having the same size grains or vice versa. This very important law requires further tests for complete demonstration, since certain of the results are inconclusive. 17. —Permeability or rate of flow through concrete is less as the per cent, of cement is increased, and in very much larger in¬ verse ratio. 18. —Rate of flow is less as the maximum size of the stone is greater. Concrete with maximum size stone of 2^-in. diameter is, in general, less permeable than one with 1-in. diameter maximum stone, and this is less permeable than one with i-in. stone. 19. —Concrete of cement, sand and gravel is less permeable— that is, the rate of flow is less—than concrete of cement, screenings and broken stone. The difference, however, is a relative one, and indicates that for equal permeability a slightly less amount of ce¬ ment is required • with rounded aggregates like gravel than with sharp aggregates like broken stone. 20. —Concrete of mixed broken stone and sand is more permeable than concrete of gravel and sand, and less permeable than con¬ crete of broken stone and screenings, which indicates that for water-tightness less cement is required with rounded sand and gravel than with broken stone and screenings. 8 21. —The rate of flow decreases materially with age. 22. —Rate of flow increases nearly uniformly with the increase in pressure. 23. —Rate of flow increases as the thickness of the concrete de¬ creases, but in a much larger inverse ratio. The use of concrete in water works construction is yet in its infancy, and there are yet many unsolved questions relating to its properties which are not touched upon in this report, or need fur¬ ther investigation for confirming the tentative conclusions now ar¬ rived at, and I would recommend that these tests be continued along the following general lines: A. Further tests to confirm the conclusions on relation of density to strength, in order to more fully adapt the laws to prac¬ tical proportioning in the field. B. Further permeability tests to establish laws for proportion¬ ing water-tight concrete and fixing required thickness for different heads. C. Tests to determine relative economy of plain vs. reinforced structures in hydraulic designs. More specifically, tests in the directions just mentioned may be enumerated: 1. —Specific tests in connection with the design of structures about to be built. 2. —Comparison of water-tightness of different thicknesses of concrete and mortar to determine proper thicknesses of concrete of different proportions for given heads. 3. —Determination of relative permeability of mortar and of ^ concrete made with this same mortar. 4. —Determination of relative permeability of mortars of coarse and fine sand and various mixtures of sand. 5. —Further tests upon relative permeability of concrete with different sizes of coarse aggregate. 6. —Determination of effect upon impermeability of adding and substituting Puzzolan cement. Y.—Determination of effect upon permeability of adding and substituting hydrated lime. 8.—Study of effect of troweling and other treatment of the sur¬ face upon permeability. 9. —Continuation of tests relating to density vs. strength of concrete to confirm the laws established tentatively. 10. —Extension of the series of tests just completed with richer and leaner mixtures to formulate a law and evolve formulas by which the strength of a specimen made with a certain brand of cement may be calculated from its density. 11-—Comparison of effect of different classes of coarse stone, such as limestone, sandstone, and trap, upon the strength and per¬ meability of concrete. 12. —Determination of effect of the surface of the stones upon the strength of the'concrete. 13. —Tests to determine under what conditions the ultimate strength of a concrete is dependent solely upon the character of the coarse aggregate. When such is the case, a lean mixture on a long time test will catch up to a rich mixture, and a mixture in which a portion of the cement is replaced by an inert fine material will be equivalent in strength and water-tightness to a richer mixture for structures which will not be used immediately. 14. —Evolution of formulas by which, having given the mechan¬ ical analysis of an aggregate, and its percentage of a certain brand of cement, the density, strength and water-tightness may be ap¬ proximately calculated. 15. —Study of effect of the fineness of the cement upon the density and strength of a mortar. The necessity for such study is shown by the fact that while an increase in the fineness of a ce¬ ment appears to increase its cementitious value, there may be an economical limit of fineness because the very fine grains tend to be chemically acted upon by an excess of water so as to form laitance, and also tend to increase the bulk of the paste. It is thus impossible to reduce the density of a mortar below a certain point, and this limits the strength which the mortar can attain. 16. —Tests of the resistance to erosion of concretes made up of various aggregates, sand and cement. In conclusion I desire to express to my assistants in this work my hearty appreciation of their interest and zeal, without which co¬ operation the results accomplished with the time and funds available would have been impossible. Especial thanks are due to Mr. 10 Thomas H. Wiggin, for his valued advice and assistance in start¬ ing the experiments; to Mr. William Hauck, for his careful atten¬ tion to the large amount of calculations involved in the tables of this report, and to Messrs. James L. Davis, Charles M. Mont¬ gomery, and William E. King, for their faithful labors in con¬ nection with the tests. Kespectfully submitted, Wm. B. Euller, Expert APPENDIX. DETAILS OF METHODS AND RESULTS OF CON¬ CRETE TESTS FOR THE AQUEDUCT COMMISSIONERS, MADE AT JEROME PARK RESERVOIR, NEW YORK CITY, 1904-5. By Sanford E. Thompson. Description of Tests. Experiments made by Mr. Puller at Little Falls, N. J., in 1901, upon the density and transverse strength of concrete beams mixed in various proportions, by weight, ranging in proportions from 1: 0 to 1: 6:10, indicated that the strength of concrete varies with the percentage of cement contained in a unit volume of the set con¬ crete, and also with the density of the specimen. With the same per¬ centage of cement in a given volume of concrete, the densest mix¬ ture, irrespective of the relative proportions of the sand and stone, was in general the strongest. The Little Falls tests further indicated that for the materials used there was a certain mixture of sizes of grains of the aggre¬ gate which, with a given percentage, by weight, of cement to the total aggregate, gave the highest breaking strength. In practice, also, it was found that the concrete made with this mixture worked most smoothly in placing. The mixture of sizes of particles of aggregate which appeared to give the best results gave for its me¬ chanical analysis a curve approaching a parabola, with its beginning at zero co-ordinates, and passing through the intersection of the curve of the coarsest stone with the 100% line, that is, passing through the upper end of the coarsest stone curve. Graded stone of the same maximum diameter and character was used in all the Little Falls experiments, and the laws men- 12 tioned were discovered by a comparison of the final results from the tests rather than by logical investigation. Therefore, before the laws could be fully established, further experiments were essen¬ tial with other materials and mixtures, together with a careful study into the limitations of the laws. The purpose of the experiments at Jerome Park Reservoir was not only to obtain results which might prove of practical value in the construction of the reservoir, but to go further and make a more general study with the object of assisting in the design of other structures to be subsequently built for the water supply of New York. The study, therefore, was chiefly of the laws of proportion¬ ing plain concrete, to determine the mixtures which would give maximum strength and water-tightness at the least cost. The experiments were begun with a series of tests of the density of different mixtures of aggregates and cement to study the laws of proportioning for maximum density with different materials, and these density experiments were followed with the manufacture of concrete beams 6 in. by 6 in. by 72 in. for comparing the laws of strength with the laws of density, and determining the relation be¬ tween these two causes. As the compressive strength of concrete is a truer measure of its quality than the transverse strength, which in the case of concrete is really one form of tension test, two pieces of each beam after being broken in the beam machine were capped with neat cement so as to form prisms about 6 in. square and 18 in. long, and these were tested for compressive strength and a num¬ ber of them were also tested for the compressive modulus of elas¬ ticity. A selected number of pieces of the broken beams were also tested for permeability. Other secondary experiments upon the density of mortars and the quantity of water required for different sizes of sand were be¬ gun, and although not extended far enough to reach definite con¬ clusions, the results are of interest as indicating the possibilities of another important line of investigation. It was intended to use these results further as preliminary to a series of permeability tests of mortars composed of cement and sand of different size grains, and containing admixtures of Puzzolan cement and of hydrated lime. One of the valuable results of the present series of concrete tests 13 was the discovery of a method for testing the permeability of speci¬ mens of different composition and thickness by simply coating the sides with neat cement and forming a dome over the top, which was readily connected with a tank of water under various pressures. The methods of making all of the tests and the apparatus employed are briefly described in subsequent paragraphs. Mechanical Analysis.* Mechanical analysis consists in separating the particles or grains of a sample of any material—such as broken stone, gravel, sand or cement—into the various sizes of which it is composed, so that the material may be represented by a curve (see Fig. 2, p. 15) each of whose ordinates is the percentage of the weight of the total sample which passes a sieve having holes of a diameter represented by the distance of this ordinate from the origin in the diagram. The objects of mechanical analysis curves as applied to con¬ crete aggregates are (1) to show graphically the sizes and relative sizes of the particles; (2) to indicate what sized particles are needed to make the aggregate more nearly perfect and so enable the engineer to improve it by the addition or substitution of an¬ other material; and (3) to afford means for determining best pro¬ portions of different aggregates. To determine the relative sizes of the particles or grains of which a given sample of stone or sand is composed, the different sizes are separated from each other by screening the material through successive sieves of increasing fineness. After sieving, the residue on each sieve is carefully weighed, and beginning with that which has passed the finest sieve, the weights are successively added, so that each sum will represent the total weight of the particles which have passed through a certain sieve. The sums thus obtained are expressed as percentages of the total weight of the sample and plotted upon a diagram with diameters of the particles as abscissas and percentages as ordinates. The method of plotting and the uses of the curves thus ob¬ tained are more fully described in the pages which follow. Sieves and Other Apparatus. —Fig. 1 illustrates a convenient outfit for such a mechanical analysis as above described, consisting of a set of sieves, an apparatus for shaking the sieves, and scales for weighing.! A standard size of sieve is 8 in. in diameter and .. paragraphs and diagrams under this heading are quoted by permission from the chapter by Mr. Fuller on Proportioning Concrete, in Taylor and Thompson’s “Con¬ crete, Plain and Reinforced,” 1905. t This apparatus is used in the laboratory at Jerome Park. 14 2i in. high. Sieves with openings exceeding 0.10 in. are preferably made of spun hard brass with circular openings drilled to the exact dimensions required. Sieves with openings of 0.10 in. and less are preferably of woven brass wire set into a hard brass frame. Woven brass sieves are made for many purposes, and are sold by numbers which approximately coincide with the number of meshes to the linear inch. As the actual diameter of the hole varies with the gage of wire used by different manufacturers, every set of sieves must be separately calibrated. An approximate idea of the diameters of holes which may be expected in commercial sizes of sieves is presented in the follow¬ ing table, which is sufficiently exact to serve as a guide to the pur¬ chase of the sieves: Commercial No. of sieve. Diameter of hole in inches. Commercial No. of sieve. Diameter of hole in inches. 10 0.073 60 0.009 15 0.047 74 0.0078 16 0.042 100 0.0045 18 0.037 140 0.003625 20 0.034 150 0.00325 30 0.022 170 0.0031 35 0.017 180 0.00306 40 0.015 190 0.0028 50 0.011 200 0.00275 For separating particles smaller than those passing through a No. 200 sieve, recourse must be had to processes of elutrition which have been developed to great precision by soil analysis chemists. In selecting the right series of sieves to purchase, first decide on the limiting diameters, say, from 3.00 in. to No. 200 = 0.00275 in. Then decide on the total number of sieves, say, twenty. Look up the logarithm of 3.00 and of 0.00275 and by proportion find eighteen other logarithms between these having equal difference between each. Look for the number corresponding and take the nearest commercial sieve giving this diameter. The diameters of holes exceeding 0.10 in. can be made as required. A convenient set of twenty sieves—ten for stone, which give the diameter of the holes in inches, and ten for sand, giving the commercial number —is as follows:* * These sizes nearly, but not quite, correspond to those adopted at Jerome Park laboratory, which are tabulated on page 22. I2E jvrnm •"LLOllD- ■f "'r w T*' f I : a jr » tv if. i • , i 1 i -*W \b# / i si W'a $=s Fig. 1.—Outfit for Mechanical Analysis of Aggregates. PEftCENf, By WElGHt, SMALLER Than GIVEN BlAMETEft 15 Stone sieves, inches. Sand sieves, Commercial No. Stone sieves, inches. Sand sieves, Commercial No. 3 00 10 0.45 60 2.25 15 0.30 74 1.50 20 0.20 100 1.00 30 0.15 150 0.67 40 0.10 200 After the sieves are obtained it is necessary that they should be very carefully calibrated to ascertain the average diameter of the mesh. This should be done by averaging the diameters of the open¬ ings measured in two positions at right angles to each other, as the meshes of commercial sieving are not exactly square. Sieves havi .ig meshes exceeding 0.10 inch are most conveniently calibrated by ordinary outside calipers; those having meshes of less diameter, by a micrometer microscope. Plotting Analysis Curves .—Tor those who are unfamiliar with mechanical analysis a detailed explanation of the method of locating the curve is here given. The method can best be understood by referring to the diagrams of typical materials which are also of practical interest as illustrating the curves which may be expected in special cases. Tig. 2 represents a typical mechanical analysis of crusher- run micaceous quartz stone which has been run through a \ -in. revolving screen so as to separate particles finer than \ in., that is the dust, for use with sand. Fig. 2.—Typical Mechanical Analysis of Micaceous Quartz Crusher Run Stone. For a sample of stone, which may be taken by the method of quartering, 1 000 grams is a convenient quantity for 8-in. diameter sieves 2J in. in depth, and also permits of easy reduction from weights to percentages. To obtain the analysis shown in Fig. 2 the sample of stone is placed in the upper (coarsest) sieve of the nest of stone sieves given on page 15 and after 1 000 shakes the nest is taken apart, and the quantity caught on each sieve is weighed.. The results obtained in the particular case under consideration are illus¬ trated in the following table, which shows the method of finding the percentages: Results of Screening Samples of Stone of Fig . 2. Size sieve, inches. Retained in each sieve*, grams. Amount finer than each sieve, grams. Percent, finer than each sieve. 0.10 8 0 0.15 11 8 1 0.20 8 19 2 0.30 72 27 3 0.45 123 99 10 0.67 235 222 22 1.00 344 457 46 1.50 199 801 80 Total. 100 . -I- I - * In practise this column is not required, the weights in the next tained directly by placing each successive residue on the scale pan with that already weighed. The various percentages are plotted on the diagram and the curve drawn through the points. The vertical distance from tne bottom of the diagram to the curve, that is, the ordinate at any point, represents the percentage of the material which passed through a single sieve having holes of the diameter represented hy this particular ordinate. Since the percentage of material passing any sieve is always the complement of the percentage of grains coarser than that sieve, the vertical distances from the top of the diagram down to the curve represents the percentages which would be retained upon each sieve if employed alone. For example, tak¬ ing* 1.25, 62%, the distance from the bottom of the diagram, repre¬ sents the percentage of material finer than li in. diameter, and 38%, the distance down from the top of diagram, represents the per¬ centage coarser than li in. Preparation of Materials. Cement .—Portland cement from the regular shipments to the reservoir was used in all except a few comparative tests of different IT brands. This cement had been tested in the regular fashion for reservoir work, and, in addition, every bag which was used in the later experiments was subjected to the purity test. It was found that the cement used in the 1904 beams which did not set up satis¬ factorily, failed to pass this test, while all the cement passing this test appeared satisfactory from a chemical standpoint. Purity Test .—The purity test is as follows: Provide a glass-stoppered bottle of muriatic acid, two shallow white bowls or two i-inch by 6-inch test tubes, a glass rod, and a pair of rubber gloves. Put in a bowl or a tube as much cement as can be taken on a nickel 5-cent piece; moisten it with half a tea¬ spoonful of water; cover with clear muriatic acid poured slowly upon the cement while stirring it with the glass rod. Pure Port¬ land cement will effervesce slightly, give off some pungent gas, and gradually form a bright yellow jelly without any sediment. Pow¬ dered limestone or powdered cement-rock mixed with the pure ce¬ ment will cause a violent effervescence, the acid boiling and giving off strong fumes until all the carbonate of lime has been consumed, when the bright yellow jelly will form. Powdered sand or quartz or silica mixed with cement will produce no other effect than to re¬ main undissolved as a sediment at the bottom of the yellow jelly. Eeject cement which has either of these adulterants.* Aggregate .—Two classes of materials were used for the aggre¬ gate; broken stone and screenings from the crushers at the reservoir, and Cowe Bay gravel and sand dredged from the river. The stone as it comes from the crushers at Jerome Park is run through revolving screens to remove the stones greater than 2-in. diameter, which are not permitted in the reservoir lining, and to separate the screenings, which are measured separately from the broken stone in proportioning the concrete. The stone and screenings were brought to the laboratory, and the screenings dried in large pans on the stove. To hasten the dry¬ ing, they were continually stirred, and when dry fine dust freely arose from them, tests indicated that they contained no appreciable moisture. Tests made of the material which had been stored in the laboratory for some time during the winter showed that it did not collect enough moisture to affect the weights used in making the tests. The 1904 beams were made with material as it came from the * Judson’s City Roads and Pavements, 1902. 18 crusher screens, this series being for approximate comparison of the strength of different proportions, while the density experiments were in progress. The density tests were a necessary preliminary to the more scientific mixtures made in 1905. For the density tests and the specimens made in 1905, the stone and screenings were separated in the laboratory by twenty-one sieves ranging in size from 3-in. openings to No. 200 mesh the latter corresponding to an opening of .0027 in. By employing dif¬ ferent mixtures of the sizes separated by these sieves, it was pos¬ sible to obtain an infinite variety of proportions. Character of Jerome Parh Rock. —The rock at Jerome Park is technically a mica schist, although the mica is in such a state that it does not form in concrete or mortar planes which seriously affect the strength. This has been investigated, and discussed in a pre¬ vious report. The variation in the quality of the rock at different parts of the reservoir gave some trouble in the laboratory and led to less accuracy in the results of the density experiments, and in forming the proportions of the mixtures, than if the specific gravity had been uniform. The most noticeable difference in different ledges was in the predominance of quartz, some of the lots being lighter in color than the others. Since the specific gravity of the minerals in the native rock ranged from 2.6 to 3.2, the specific gravity of different lots of screenings varied appreciably, and the specific gravity of different diameters of the same rock also varied because some of the minerals of which the rock was composed crushed more readily than others, and therefore certain sizes con¬ tained predominating minerals which determined their specific gravities. Specific gravities of different sizes are given in a subse¬ quent portion of this report. Composition of Rock.— The following report was made upon the mica in the screenings by Charles P. Berkey: Report of Tests Made for Quantity of Mica in Two Samples of Crushed Bock. No. 7Jj-. —(Screenings at 74 mesh), and General Mixture— (Whole rock crushed to less than 40). Tests were made first by Heavy Solution (Thoulet’s Solution), and 3 separations made in each case, obtaining: 19 1st. The white minerals (lower specific gravity constituents). 2d. The lighter weight portion of the dark constituents, that is, the most flaky micas and less basic ones, or, otherwise, the por¬ tion that floats in the densest solution. 3d. The heaviest minerals, that is, all that sink to the bottom in the densest solution, or otherwise, mostly the perfectly fresh solid micas, and a little magnetite, a little zircon, and some hornblende. No. 74 with this method gives out of sample of 5 g. 1st. White minerals .2.977 g. = 59.54% 2d. Middlings .982 g. = 19.64% 3d. Heaviest part .1.041 g. = 20.82% Of No. 1—there is no mica. Of No. 2—all of it is mica = .982 g. Of No. 3—83% is mica . . = .864 g. In No. 3 the chief impurity is hornblende. General Sample —with this method gives in 5 g. 1st. White constituents .2.848 g. = 56.96% 2d. Middlings .913 g. 3d. Heaviest constituents = 1.239 g. In these No. 1 has no mica. No. 2 has 90% mica.= .821 g. No. 3 has 90% mica.= 1.115 g. Total mica . = 1.846 g. = 36.92% The chief impurity in No. 2 is fine grains of (No. 1). The chief impurity in No. 3 is hornblende. Check tests were tried for mechanical separation by sliding over an incline and agitating the plate. The results are not satisfactory in every case. For No. 74— The separation is fairly good. The flaky portion—mica = 35%. The general sample— This method did not succeed. Too much fine material, i. e., less than 74 mesh. It will neither roll nor slide—making the residue much too high in value. Conclusions : 1st. About 74 mesh is the best size to work with. If all grains were about this size a fairly good separation could be obtained. 20 2d. In the average crushed sample, unless controlled carefully to avoid over-fine grains, a mechanical separation is a failure. 3d. Thoulet’s heavy solution will separate the light from the dark minerals with great precision. (K I -j- Hgl 2 , proportion 1:2.24, dissolved in distilled water and used as a settling hath.) 4th. About 90% of the dark constituents of the rock prove to be mica. The grains are prominently flaky, brown, yellow, black color, with very perfect clevage. 5th. The average amount of mica seems certainly to exceed 35% according to these samples. (Signed) Charles P. Berkey. New York City, Apr. 27, 1904. Uselessness of Determination of Voids in Sand. —The percen¬ tage of moisture and the degree of compactness of sand so affects its weight per cubic foot and its per cent, of voids that weight and void determinations are of no practical value in distinguishing sands. An illustration of the variation in weight and volume of the same sand occurred at Jerome Park this summer. A lot of Cowe Bay sand which had been stored in the laboratory was weighed in a measure and found to average 102.7 lb. per cu. ft., the weight of three determinations being respectively 103.6 lb. 101.8 lb., and 102.8 lb. As the sand was very dry, it contained 38% voids. The same sand was taken out of doors and exposed to the weather for several days. A storm occurred during this period. Three days after the rain it was shoveled into a measure and again weighed, and found to average 83.0 lb. per cu. ft., three weighings giving 82.9, 83.4 and 82.6 lb. per cu. ft., respectively. About 5% of this weight was moisture, hence the air voids were 45.7% and the air plus water voids 52.2 per cent. Screening Aggregate in Laboratory. Barge sieves about 2 ft. square, were used for screening the aggregate into the twenty-one sixes. The sieves rested in a frame, one above another, seven to a frame, so that any one of them could be pulled out like a drawer without disturbing the others. The frames rested upon a pair of rockers consisting of 2-in. plank, sawed with angles instead of rounding, so that the motion of the sieves when rocked by hand was a com¬ bined slide and jar. Spikes were driven into the rockers at the 21 angles to make them more durable. A frame of sieves and a single sieve is shown in the background of the photograph, Fig. 8, page 38 and sketched in Fig. 3 below. This apparatus worked fairly well, although difficulty was ex¬ perienced in getting uniformity of screenings. It is suggested that Rear View. Section of Sieve Fig. 3.—Detail of Frame for Sieves for Screening Aggregates in Laboratory. for future work some form of pedometer be placed upon the frames to record the number of shakes, which number should be definitely fixed for each series of sizes. Each lot of unscreened material placed in the top sieve of each series should be measured (as this is suffi¬ ciently accurate and requires less labor than weighing). 22 Diameters of Aggregate .—Experiments have shown that the diameter of the largest stone passing through any sieve corresponds almost exactly to the width of the opening. The method of rating is referred to in subsequent paragraphs, hut it may be said in gen¬ eral that the diameters of the stones can be designated by the width of openings in the wire cloth. The usual method of designating sieves below one-tenth inch diameter is by a commercial number corresponding to the number of meshes in a linear inch. This hears no exact relation to the size of the openings, nor, therefore, to the diameter of the particles passing through them, because of the variation in the size of the wire used in making the sieve. Accord¬ ingly, it is necessary to carefully calibrate wire cloth in every lot unless it comes from a maker who has adopted standard sizes of wire for the different sieves. The Howard & Morse Company of Brook¬ lyn now manufacture wire cloth upon a standard scheme, so that the sizes are very regular. The diameters of sand grains in the diagrams of this report are designated by the size of the openings in the sieve, that is, by the diameter of the largest particles passing through it, instead of by the commercial number of the sieve. The diameters of par¬ ticles with the corresponding commercial sieve numbers which have been adopted in this series of tests are as follows: TABLE I. — Sizes of Sieves. Normal sizes of sieves, inches. Diameters passing sieves, inches. Normal sizes of sieves, commercial number. Diameters passing sieves, inches. 2.25 2.25 10 0.075 1.50 1.50 15 0.046 1.00 1.00 20 0.034 0.75 0.75 30 0.020 0.60 0.60 40 0.016 0.45 0.48 50 0.014 0.35 0.36 74 0.0071 0.27 0.29 100 0.0058 0.20 0.20 150 0.0036 0.15 0.16 200 0.0027 0.10 0.10 Fig. 4.—Apparatus Used in Volumetric Tests. 23 The number of sieves adopted is so large as to require consid¬ erable time and labor for making each test. It may be possible in future experiments to omit a few of the sizes, if tests of the ma¬ terials indicate that this can be done without appreciable effect upon the result. Methods and Apparatus for Determining Density. Trial mixes of aggregates composed of various sized particles following a logical plan, and containing first 10% of cement to the total weight of the dry materials and then other percentages of cement, were made and mixed with water to the same medium wet consistency, and the resulting volumes thus obtained from exactly the saifie total weight of dry materials (corrected for specific grav¬ ity) were compared. The tests of density are termed volumetric tests in this report. The general procedure (described below) in making them is that adopted by the French Commission, in 1894, and the volumes of material per cubic foot were calculated by methods used by Mr. R. Feret, of Boulogne-sur-Mer, France, in his determination of elementary and of absolute volumes. The apparatus used in the volumetric tests is shown in the photograph in Fig. 4, page 22, and the apparatus and tools are sketched in Figs. 5 and 6, on pages 25 and 27. Weighing .—All materials were proportioned by dry weight. The natural sand and screenings as they came from the pit or the crusher were dried in the laboratory, and experiments showed that after being sifted, they did not collect in the laboratory a sufficient amount of moisture to appreciably affect their weights. For weighing the materials a Fairbanks scale No. 1288, with compound beam having seven scales, was employed. The scales read to half pounds, and by means of a wire rider made in the laboratory it was possible to read to hundredths pounds, although the accuracy was not much greater than tenths. In the density tests the materials were weighed directly in the mixing pan placed upon the platform of the scales. The water was usually weighed in a 16-quart galvanized iron pail. Measuring. —For measuring the volume of concrete made in the volumetric tests an old cast-iron air brake cylinder and piston was found convenient. The cylinder was 8 in. in diameter inside meas¬ urement, flanged at both ends, with a blank flange bolted to it, thua 24 forming a vessel 8 in. in diameter and 9 in. deep. This vessel was carefully calibrated, so that the volume of the contents could be obtained by measuring down from the top. To measure the depth of concrete and thus calculate its volume, the piston was set in the cylinder on top of the concrete; the projecting iron rod formed a convenient handle and also provided means for measurement. A yoke of hard wood consisting of two vertical uprights, connected by a crosspiece at the top, and at a distance apart corresponding to the diameter of the pipe, was made to place upon the cylinder and straddle the piston, the cross-piece resting against the handle of the piston. These are all shown in Tig. 4 and are sketched in Figs. 5 and 6. A mark was scratched on the handle of the piston at such a point that for any position of the piston in the cylinder the distance from this mark to the top of the crosspiece of the wooden yoke represented the exact distance from the bottom of the piston to the average sur¬ face of the flange forming the bottom of the cylinder. After placing the concrete in the cylinder, the piston was pressed down upon it until its bottom surface exactly coincided with the surface of the concrete. The depth of the concrete could then be exactly measured by measuring with a boxwood scale, reading to hundredths of inches, the distance between the mark on the handle of the piston and the top of the crossbar of the wooden yoke. From this depth the volume of the concrete was readily calculated. The larger cylinder. Fig. 6, 12 in. in diameter and 18 in. deep, was made for volumetric tests of materials containing large size stone, but comparative tests with this and the smaller cylinder showed that the densities obtained by the two apparatus were so nearly identical that the larger one was not used to any great extent. The materials were proportioned by mechanical analysis curves, as described in preceding paragraphs, and the weights of each diame¬ ter were scheduled from these curves. The mixing pan, which was about 2 ft. by 2 ft. by 3 in. deep, was placed upon the scales, and any trowels and rammers or other apparatus to be used in the opera¬ tion, and therefore to he more or less coated with cement, were placed in the pan, and the weight of this tare recorded on one of the beams. The aggregates, beginning with the finest diameter, were then weighed. The schedule of weights of the aggregate were made 25 up so that each weight included the weight of all the finer aggre¬ gates. In this way the weighing poise was reset for each size of material, and the materials placed in the pan, one on top of the other. The cement was weighed last, so as to be on top of the other materials. Mixing .—The mixing was done by two men working on opposite sides with large trowels and turning the dry material until of uni¬ form color, in the same manner as concrete is turned by hand. The material was then formed in a ring, water poured into the center, Fig. 5.—Tools and Tray Used in Volumetric Tests. and the mass turned until the mixing was thorough. The mixing pan, trowels, shovel (in plan and elevation), cleaner (in plan and section), and rammer are sketched in Fig. 5, above. Because of the variation in the sizes of the grains of the aggre¬ gate in the different volumetric mixes, it was impossible to select a definite percentage of water to use in all the tests, or to select in advance definite percentages for each mix. The mixtures containing the largest quantities of very fine material required the largest per¬ centage of water. A series of volumetric tests of mortars made with cement and sands with grains of different size, which are referred to later in the report, was started in order to study the reasons for the variation in quantity of water required under different condi¬ tions and to formulate some rule for proportioning it. The water, therefore, was added by judgment to obtain a soft mushy mixture which would scarcely hold its form in the mixing pan, but which was not so fluid that the mortar would run away from the stones. A pail of water, with its dipper, was first weighed, as much water added as was required, and the weight of the remaining quantity deducted from the original weight to determine the net weight used. As the surplus water was removed from the specimen after placing in the cylinder, the quantity finally given in the den¬ sity sheets represents the water actually contained in the specimen as it began to set. Ramming .—The mixed concrete was introduced into the cylin¬ der, which had been previously weighed, and rammed in 2-in. layers. This frequent ramming was necessary because of the small size of the receptacle, the friction on the sides preventing the material from settling even with very wet mixtures if ramming was delayed until the entire amount, which averaged a little over 6 in. in thickness, was placed. The rammer which was found best was a cast iron disc about 4 in. in diameter, with an upright handle, sketched in Tig. 5. A mixture of the consistency described above gave the best results in the cylinder. If too dry, there was always a possibility of occasional large air voids, and it was more difficult to manipu¬ late, although the resulting rammed volume was about the same as the mushy mix which was used. A very wet mixture did not work quite so well as either the mushy or the dry, the mortar probably not being in a sufficiently plastic condition to lubricate the stones. Rather curiously, a very wet mixture required more fine material to fill the voids than the others. Removing Surplus Water .—As the concrete was rammed, the sur¬ plus mortar, if there was any, rose to the surface, and the water separated forming a layer from | to l in. in depth. After this had become clear, it was removed by a small suction pump, and the weight deducted from the weight of the water used in the mix. Final Weighing .—The cylinder containing the concrete was weighed as a check upon the weights of the dry material and the water, the mixing tray, together with the tools, was weighed, and the quantity of material adhering to them was thus found, being 27 4 ' Level. Bridge used on smoill cylinder. Fra. 6 .—Apparatus Used in Volumetric Tests. 28 the difference between this weight and the weight of the clean tray and tools. The weight of the portion of the mix adhering was intro¬ duced into the final calculations as described below. Measurement .—After the surplus water was removed, care was taken to see that the surface of the concrete was level, with no pro¬ jecting stones, and the cylinder was leveled by placing it upon a tripod consisting of a board with three lag screws for legs. The > piston was introduced and pressed firmly down, but not with suffi- * cient force to disturb the mortar on top and thus force it up between the piston and the side of the cylinder. The wooden yoke wa& I placed in position, and the depth of the concrete in the cylinder measured, as described above, by measuring the height from the j top of the yoke to the mark on the piston. The cylinder, piston, j yoke, scale and magnifying glass are shown in Fig. 6. The test j was now complete and ready for computation, the records having ; been entered upon the blank form described below. The concrete ! was thrown away, and the tools cleaned ready for the next experi- j ment. Recording and Computing Data .—The form for recording the j data in the volumetric tests of concrete is given on the following page, in Table 2, with a typical test of Cowe Bay material recorded upon it. It is so arranged that all of the weights may be entered in the laboratory, and the printed column of items also gives the con¬ stants to be used in calculation and the method of combining the j various items so that the calculation of each sheet may be made by rote by an unskilled computer. The experiments in each class were numbered consecutively for convenience in reference. The items recorded in the laboratory were the weights, and the depths of I piston in cylinder, items 1 to IT, 24, 34 to 40. The weight of aggregate finer than .0071 in. diameter, item 6, is separated from the weight of aggregate coarser than this, item 7, because the former is separately used in subsequent items for the calculation of the material adhering to the tray and tools. Item 8 gives the weight of the vessel and water before any of it is used for mixing, and item 9 the weight remaining after mixing. From the difference, item 10, must also be deducted the free water drawn from the surface of the concrete, item 17, which is the difference between items 15 and 16, and also the weight of the water left on the tray, 29 TABLE 2. The Aqueduct Commissioners, N. Y. Subject: Concrete Experiments. Blank Form for Volumetric Tests. City. File No. Ac. No. Sheet No. iuted by Checked by. Date, Mar. 10, 1905. 377£ 3/10/05 2.65 .85 21.50 e. ninal Mix. d of Cement.j ^ 0 iant ight. . of Aggreg. finer than .0071' “ “ coarser “ “ “ Vessel + Water (1). 29.00 “ “ “ (2).27.40 “ Water Used. 1.60 al Wt. Mixed. 26.00 . of Tray + Tools (2). 13.15 “ “ “ (1;. 13.08 “ Mixed Adhering.07 Syringe + W ater. 2.62 Syringe. 2.60 Free Water.02 Mix Set = 11—14—17. 10 v 14 ter left on Tray 5 —|— 6 —i— 10 Water Set = 10 — 17 —19 • n * « - 5 X 14 Cement “ =5 26.51 .02 Aggreg. = 6 X finer 6 X 14 + 5 4- 6 + 10 than .0071 1 1.58 2.61 .84 5 + 10 23. Net Aggreg. coarser than .0071. 21.50 24. Depth of concrete in Cylinder, in. 5.99 25. Vol. of concrete in Cylinder in., .02934 X 24 — .001.... v .174 26. Net Water per cu. ft. as mixed „ . 9.2 set Cement Aggreg. 1 9.1 15.0 .128.4 27 Abs. Vol. Water per cu. ft. as set — Cement Aggreg. 28_ 194' 165" ‘ .146 .077 .778 1.001 “ “ Total 30 -j- 31 4- 32. Wt. of Form -f Concrete... 77.23 “ “ Form. 50,75 “ “ Concrete. 26.48 Temp, of Water. 40° Time of Mixing after Wetting. Remarks on Consistency. dARRs.— Looked little stony in mix. Top surface filled. No excess of mortar on top. Small No. 10, right under top surface. given in item 19, thus leaving the net water in the set concrete in item 20. Figures following many of the items refer to the numbers of other items; the fraction following item 19 represents, for exam¬ ple, the portion of the mix adhering to the tray and tools which is water. The assumption is made, the fact having been determined by experiment, that the mortar sticking to the tray and tools consists of cement, and particles of aggregate finer than diameter .0071, and water. The weight of the water in this mortar which adheres may be found from the proportion: Mix adhering: total fine mortar = water in mix adhering: total water. Expressed in item numbers, this becomes: item 14 Item 19 = item 5 + itenUT+TtemlO X item 10 - The net water contained in the concrete is thus item 20, which equals item 10—(item 17 + item 19). O' 30 The net weight of the concrete, item 18, should be the total weight mixed, item 11, minus the mix adhering to tray and tools, item 14, minus free surface water, item 17. This item 18 should coincide with item 36 obtained by finding the net weight of the concrete in the pipe. The net weight of the cement, item 21, is the weight originally used less the cement adhering to the tray and tools, the determina¬ tion of which is made in the same way as the determination of the water, item 19. Item 22 is similarly calculated to determine the actual amount of aggregate finer than .0071 left in the concrete. As no aggregate coarser than .0071 adheres to the tray and tools, item 23 is the same as item 7. The volume of concrete in cylinder in cubic inches, item 25, is calculated from the depth, item 24, by multiplying item 24 by a coefficient and deducting a constant. The weights of each of the materials per cubic foot, items 26 to 29, are, respectively, the quotients of the total net weight of each material divided by the volume of the concrete. The absolute vol¬ umes, items 30 to 32, which represent the total volume of the liquid, or the volumes of the grains of cement or aggregate, are the net volumes per cubic foot divided by the specific gravity of each of the materials. These absolute volumes represent simply the ratios of the actual volumes of each ingredient to the total volume of the con¬ crete. Proportioning the Ingredients for Maximum Density. The volumetric tests of concrete were begun in 1904. In the first place, it was the aim to determine for the various materials under consideration artificial mixtures of aggregates of diameters graded by methods of mechanical analysis already described, which would give concrete of maximum density. These ideal mixes having been determined, the concrete made from them could be compared with respect to density, strength and permeability with concrete made from simple mixtures of natural materials, and thus the best proportions to use with different materials upon actual construction work could be fixed. The larger part of the experiments were made upon the same proportions of total aggregate to cement. The tests with Jerome 31 Park stone and screenings for the most part were with 10%, by weight, of cement to the total dry materials. The tests with Cowe Bay sand and gravel were with similar proportions, except for a slight correction for the difference in specific gravity in order that the ratios of absolute volumes might be the same. In the earlier tests mechanical analysis mixtures were made to definite curves, such as parabolas and straight lines and curves intermediate between these, in order to select the curve giving the maximum density. In the later tests, in the winter of 1904-5, the methods were slightly changed and the best results were obtained by making mixtures on trial curves without reference to their mathematical equations, and then having found the best curves, equations were fitted to them, so that they could be more easily applied to different materials and more readily plotted. All the earlier density specimens were found to be inferior to the tests made at Little Falls. Subsequent results in the manufac¬ ture of the beams indicated that this was due to the quality of the cement, which has already been referred to on page 17. From the volumetric tests and tests of the composition of beams of neat cement, it appeared that the difference in the density was due to the cement taking so much water in gaging that the volume of the paste was increased, and the density lowered. In the winter of 1904-5 the experiments were continued with the cement which was then being furnished, and this gave satisfactory results. To avoid duplication of tests and reach conclusions as speedily as possible, the first series of experiments in the winter of 1904-5 was made with one class of material, broken stone and screenings excavated from Jerome Park Reservoir site, and with 10% of cement, by weight, of total dry material. This was followed by tests with Cowe Bay gravel and sand and with other percentages of cement. Necessity for Using Cement in Density Tests .—The necessity may be questioned for using cement in the tests for density, which were really for the purpose of determining the best proportions of the various sizes of particles of the aggregate. Why would it not have been simpler to use only the dry aggregate with no cement or water, and thus more readily obtain the mixtures which would give the least volume with the same weight? As a matter of fact, both theory and experiment prove that the 32 mixtures of aggregate which give the greatest density dry do not necessarily give the greatest density when mixed with the cement and water. The cement and water actually occupy space in the mass, as many of the voids-are too small for the grains of cement to fit into them without expanding the volume, and the water surrounds the grains of fine sand and of cement and actually increases the bulk. As an illustration of this, the weight per cubic foot loose of very fine sand, if weighed absolutely dry, is very nearly the same as the weight per cubic foot of a very coarse sand weighed dry. However, if the two sands are mixed with cement and water, the resulting mortar made with the fine sand will occupy a bulk perhaps 20% greater than the mortar of coarse sand, even when each of them is mixed with the cement in the same proportion by weight or by absolutely dry volume. The density of the mortar of fine sand will be correspondingly less than the mortar of coarse sand. Further¬ more, the proportion of cement to sand affects the relative bulk and density of the two mortars, 1:1 mixtures giving different compara¬ tive results from 1: 4 mortars. If fine aggregate having grains of the same size as cement par¬ ticles were used, the aggregate could have been mixed with water without using any cement, and the resulting density would probably have been the same as where real cement replaced the fine aggregate. However, this fine aggregate is more costly than cement because of the labor required to screen it. Moreover, the conditions would not have been so practical as where cement itself is used. Density Tests with 2 \-in. Jerome Varh Stone .—In all these tests the cement was included in the mechanical analysis curve. This was in accordance with the assumption, which was afterwards proved to be correct, that the grains of cement acted similarly to grains of sand of similar size so far as the density was concerned. A trial mechanical analysis curve was drawn based on previous tests, and a volumetric test was made to determine the density. The curve was then altered in various ways by raising and lowering it at different diameters of stone, and volumetric tests made with each experi¬ mental curve. By this means the general principles of the density of mixtures of Jerome Park broken stone and screenings with 10% Giant Portland cement were studied. The curve which gave the best result when using a graded coarse 33 aggregate with the cement was found to be one resembling a par¬ abola in appearance, but, more strictly, consisting of a curve having for the lower portion the form of an ellipse, and above this a straight line running to 100% on the maximum diameter of the stone, in this case 2i in. The tangent point of the curve and straight line was at about 0.2 in. diameter. The curve started below measurable diameters on the 7% line, indicating that at least 7% by weight of the very finest diameters of particles of cement or sand, or both, was required for a dense mixture. In studying these density curves, the attention of the writer was called to experiments by Mr. A. E. Schutte, for the Warren Brothers Company, Boston, on mixtures of aggregate to be used in their bituminous macadam pavement. Eor this class of work, which is really a scientifically graded concrete with bitumen for the matrix instead of cement, Mr. Schutte found the densest mixtures, and the best results in practice, to occur when a large percentage, about 50% in fact, of the aggregate consisted of the coarsest diameter of stone of uniform size. Density tests, using a mixture of this kind, were made at Jerome Park for comparison with the tests with graded coarse stone, as shown in Table 11, page 65. The resulting concrete was found to be slightly denser than the concrete with a graded stone. However, the mixture did not look so well in the mixing pan, and, while no doubt with a plastic substance like bitumen, a thor¬ ough mixture would produce excellent results, with the cement there seemed to be a tendency of the stones to separate from the mortar more than with the concrete containing graded coarse stone. Whether or not these conclusions would apply in practice has not yet been determined. Tests of beams made with the two kinds of mixtures, although somewhat erratic, indicated a scarcely appreci¬ able difference in the strength. Since under some conditions a stone of uniform size is as easy to obtain as a graded stone or “crusher run,” further experiments are desirable to compare these two methods of mixture, and to prove whether under some conditions the uniform stone may not be economical. This matter is further dis¬ cussed on page 63. The equations of the curve which was selected as the best are given in succeeding paragraphs. Density Tests with 1-in. and \-in. 8tone .—The best analysis curve for cement and aggregates whose maximum size was less than 34 2i in. was next studied in a similar fashion, using, as before, 10%, of cement to the weight of total dry materials. The curves for the smaller stone were found to resemble the 2^-in. curve, except that the tangent began at a smaller diameter. This diameter, where the curve ended and the straight line began, was found to be about one- tenth the diameter of the maximum size of stone used in the mix¬ ture; thus, for 1-in. stone, the tangent point was at about 0.1 in. diameter, and for i-in. stone the tangent point was at about 0.05 in. diameter. This suggested the possibility of a curve for all sizes of stone with the same equation, but with the diameter of the maximum size as a function of one of the terms. It was found that the shape of the curve was not exactly the same for the different sizes, but by introducing a small constant in the values of the axes of the ellipses, an equation was found which fitted all the diameters. Cement vs. Fine Sand .—The experiments just described assume that so far as density is concerned, cement acts in the same way as sand with grains of the same size. The cement was therefore in¬ cluded in making up the mechanical analysis curve. To prove this assumption, several other percentages, ranging from 8 to 15%, of cement to weight of total aggregate were tried, using the best curve already found for the mixtures with 10% of cement, and the result¬ ing densities were substantially identical with the density of the mixture by the same curve using 10% cement. It is evident from this that correct proportioning for concretes of maximum density, but of different strength, consists in not simply increasing the per¬ centage of cement, if a richer mixture is required, but substituting more cement for a like absolute volume of sand having grains of the same size as the cement. In other words, the larger the proportion of cement, the less very fine grains of sand are required, because the cement takes the place of them in increasing the density. Density Tests with Cowe Bay Material .—Tests of density of con¬ crete composed of cement and of gravel and sand from Cowe Bay were made in a similar manner to those of cement and Jerome Park stone and screenings. This investigation is of interest not only with reference to the work at Jerome Park, but to throw light on the mooted question of the relative value of broken stone and gravel, and form some basis for an economical comparison of the two in any given locality. Relative results are tabulated in Tables 8 and 9, pages 60 and 62. 35 Tests were made with concrete composed of straight Cowe Bay material, that is, gravel and sand and cement, and with concrete of Jerome Park broken stone, Cowe Bay sand and cement. The best curves for both of these combinations were found to be similar to the curves for concrete with straight Jerome Park material, except that, because of their rounded nature, the particles packed more closely together, so that similar mixtures gave with the Cowe Bay material a greater density, and for maximum density a smaller quantity of fine material was required, the curves in the diagram for Cowe Bay material being lower on the ordinate corresponding to one-tenth the maximum diameter of the stone. The combination of Jerome Park stone and Cowe Bay sand required more fine material than the Cowe Bay gravel and sand, but less than the Jerome Park stone and screenings. Ideal Sand .—The character of the best or ideal sand does not appear to depend upon the character of the coarse aggregate, the best sand for Jerome Park broken stone also being found best for the Cowe Bay gravel of the same sized grains, although less was re¬ quired with the latter. On the other hand, the curve for the Jerome Park screenings (plus cement) was slightly different from the Cowe Bay sand (plus cement) curve, showing a different arrangement of grains. Equations of Ideal Mechanical Analysis Curves. Having found by trial the curves for the best analyses of each size and class of material, mathematical curves were fitted to them for convenience in plotting. As already stated, all the curves for aggre¬ gate plus cement consist of ellipses with straight lines tangent to them. The curves all start upon and are tangent to the vertical zero axis of percentages at 7%—that is, at least 7% of the aggregate plus cement is finer than the Ho. 200 sieve—and run as ellipses with axes differing with the character of the materials, to a point which was found to be on a vertical ordinate or diameter whose value is about one-tenth the diameter of the maximum particles of stone, and thence by a tangent to the ordinate of maximum diameter, intersect¬ ing this on the 100% abscissa. The general equation of all the ellipses, using their own axes, is &2 _ y = ‘I" a2 V a 2 — x 2 - 36 Suisswa; s^uaoaaj Fig. 7.—Diagram of Ideal Curves for Mixing Concrete of Various Sized Aggregates. 37 This is the simplest equation to use for plotting, as it is in regular form, and only the values of a and b, that is, the major and minor axes, are required. Using zero coordinates on the mechanical analy¬ sis diagram, the equation becomes (:v — 0 2 = ( a x —* 2 )- The values of a and b for the different materials (plus cement) used are as follows: Materials. a. Jerome Park Stone and Screenings. . .035 -f- .14 D Cowe Bay Gravel and Sand.04 + .1'6 D \ Jerome Park Stone and Gowe Bay Sand .04 + .16 D b. 29.4 + 2.2 D 26.4 + 1.3 D 28.5 + 1.3 D In this table D = the maximum diameter of the stone in inches. The numerical values for these with different sizes of stone are given in Table 6 on page 50. Directions for Plotting Ellipses .—In practice the ellipses are plotted graphically by the trammel point method as follows: ^ Plot the major and minor axes on the diagram. The major or liorizontal axis in all cases is on a line 7%^ above the base. The niinor or vertical axis is at a distance, a, to the right of the vertical zero ordinate of the diagram. Lay a strip of paper or a thin straight-edge upon the major or horizontal axis, and mark upon it two points to represent the length of the semi-major axis, calling one of these points—the point on the zero ordinate— 0, and the other point A. Mark off on the strip or straight-edge, in the same direc¬ tion from 0, the length of the semi-minor axis, calling this point B. Mow, swing the strip of paper or straight-edge little by little so that the outline of the curve may be marked off by the point 0, while the points A and B are kept at all times upon the axes b and a re¬ spectively. The straight lines to continue the curves are drawn as tangents to them, or may be readily plotted from the data in Table 6. Diagram of Ideal Curves .—The ideal curves, that is, the best curve for each material and each size of material, are plotted in the diagram, Fig. 7, page 36, 38 Methods and Apparatus for Beams. The method of weighing the materials for the beams and mixing them was similar to the processes in the volumetric tests for density of concrete. The tools and various implements used in making the beams are shown in the photograph in Fig. 8, opposite. Weighing .—The weight of each diameter of aggregate and of cement was calculated by direct proportion from the percentage curve, and the approximate quantity of water to use estimated from the quantity employed in the volumetric tests made with materials of the same mechanical analysis. The total quantity of dry mate¬ rials for a beam 6 in. by 6 in. by 72 in. varied appreciably with the maximum size of stone and with the character of the material. The- larger the stone, the greater the weight of material required for a beam, because of the greater density. The Cowe Bay material has lower specific gravity than the Jerome Park stone and screenings, so that a less weight was required on this account, but the density of the resulting concrete was greater, so that this nearly balanced the other. The quantity for a mixture with 1: 9 proportions by weight averaged about 220 lb. of aggregate and 25 lb. of cement. The weights actually used for each beam may be directly figured from the data in Tables 14a to 14e, pp. 105 to 109, following page 72. The can for receiving the material, about 18 inches in diameter by 2 ft. deep, was placed on the scale and the tare recorded by the, weigher. Two other men scooped the different sizes of aggregates beginning with the finest, and also the cement, into the can, the poise being slid ahead for each size. Mixing .—The dry material was then dumped upon the mixing platform, which was 5 by 8 ft. and made chiefly of a plate of sheet iron surrounded by a strip of wood to prevent the soft material from overflowing. Four men turned the dry material with square pointed shovels, just as hand-mixed concrete is turned in practice. Three turnings were given, dry. The stuff was formed into a circle, when the approximate amount of water was weighed and turned into it. The material was then mixed wet by the four men, as in practice. Three turnings were given to it to insure a uniform mix throughout the beam. Consistency .—The required consistency was soft and mushy, but not wet enough for the mortar to run away from the stones, scoop Fig. 8.—Implements Used in Making Beams. 39 shovels being necessary to handle the wet concrete. If the calculated weight of water was not sufficient to give the required consistency, more was added and the weight of it recorded. Marks for Specimens .—Four marks were imbedded in each beam. The tags consisted of pieces of brass stamped by hand by a die with the number of the beam and followed by the letters A, B, C and D respectively. Each tag was about 1| in. long by i in. wide, and had projections bent up from it to run into the concrete. They were definitely located in the bottom of the mold by measurement, so as to lie nearly, but not quite, in the center of each of the four pieces into which the beam was finally broken. Each tag was placed in the bottom of the mold, and held with the blade of a shovel until the concrete was poured around and on top of it, thus holding it in *llron Casf/ng Plan. Section. Fig. 9.—Sketch of Moulds Used for Making Concrete Beams. place. The tags were often slightly covered by thin mortar which ran under them, but being exactly located, they were readily found by scraping the surface. Placing in Molds .—The mold for the beam, shown in Fig. 9, above was weighed and the concrete shoveled into it, and placed and slightly compacted with the aid of large trowels and shovels. The blade of a shovel was generally thrust down by the faces of the forms to insure a smooth surface, although this was not found abso¬ lutely necessary. The mold with the concrete was weighed to obtain the weight of concrete green for density determination, and placed one side to set. The quantity of material required had been approximately calcu¬ lated from the volumetric tests. If any was left over, after the mold 40 was filled, it was weighed and the proportion of this which was water was estimated. As the proportional amount of water may exert a difference of 2 or 3% upon the calculated density of the beam, it is suggested that in future tests the concrete left over be carefully separated from the excess water, so as to have about the same consistency as the beam, and then weighed, while the water remaining, together with the free water from the surface of the con¬ crete in the mold, be separately weighed and recorded. As this water contains some cement in suspension, its approximate specific gravity should be determined by a test, and this specific gravity used in the calculations of the density of the beams. As soon as the concrete was hard enough to handle easily, which was usually in about 7 days, it was buried in moist sand, where it I remained until the date of test. Recording and Computing Data on Beams .—The method of re¬ cording the data for mixing is given in Table 3. The form is filled TABLE 3.—Form for Mixing Beam Material. Wt. of Form No. 14 empty.. 131.25 •• “ “ filled. 364.00 “ “ Beam 217 .232.75 Total Wt. mixed.234.00 | Wt. of Mix left over. 0.00 Total Wt. of Beam.232.75 Wt. unaccounted for. 1.25 Wt. of Form No. 14 filled.359.75 Beams buried—day, hour, days be-| o/on/nA fore burying. \ 3/<5b/U4 Loss of wt. in setting . 7 Inspector, W. H. out with a typical test for illustration. When ready to break, after removing from the moist sand, measurements are taken for the pur¬ pose of figuring the density and calculating the modulus of rupture. The form for this, with typical measurements filled out, is given in Table 4, on page 41. As each beam was broken into four pieces, first in the middle, and then the two halves broken again, measure¬ ments were taken at half and quarter points, also at the two ends, making five sections in all. At each section, three dimensions were No. of Beam. 217 Approx. Vol. of Beam.1J4 cu. ft. Date.May 21, 1905. Hour.1.30 p. m. Temp, air in cellar.64° Kind. Lot No. Wt., lbs. Cement. .23.32 Kind. Analysis No.. Wt., lbs. Aggregate. .Ideal 36% ord. Temp. F. Total used.... Water. .54° ...14.0 Fig. 10.—Machine for Breaking Beams. 41 measured of depth and three of width. The beams were placed in the machine on their side, and therefore the depth as measured is really the width of the beam in the mold, and the width as measured is the height of the beam in the mold. As the quarters are labelled with the number of the beam followed by A, B, C and D, respectively, the measured sections are A, AB, BC, CD, and D. Upon each beam, thirty sectional measurements are thus made, each reading to hun- TABLE 4.—Eorm for Dimensions of Beams. Number of Beam. Section Letter. Depth (Inches). Width (Inches). Area. Av. Area. Length (Inches). Contents, cu. ft. Side. Mid. Side. Av. Side. Mid. Side. Av. f A 6.01 5.94 5.98 5.98 6.09 6.00 5.98 6.02 36.00 1 1 AB 6.01 6.03 6.07 6.04 6.05 6.00 6.00 6.02 36.36 166 ■{ BC 6.02 6.00 6.00 6.01 6.15 6.17 6.12 6.15 36.96 >■36.10 72 1.504 \ CD 5.97 5.95 5.98 5.97 6.02 6.02 5.85 5.96 35.58 | l D 6.03 6.00 6.01 6.01 5.89 5.95 1 5.91 5.92 35.58 J dredths of inches. From the beam mixing data and the calculated volume of the beam, the weights of the material in one cubic foot of the beam and the absolute unit volumes are calculated. The modu¬ lus of rupture for each break is calculated from the dimensions and the breaking weight. Beam Testing Machine .—The machine for breaking the beams was made by Riehle Brothers, Philadelphia, and the method of loading it improved in the laboratory. A photograph of the machine with a beam in place ready to break is shown in Fig. 10, page 40. In order to avoid the negative moment due to overhanging portions of the beam, the bearings for each specimen were 2 inches from each j end of the beam, thus giving a variable length between supports for i specimens of different length, but a overhang so short as to be neg- ! ligible in the calculation. Instead of using the poise for weighing \ the load upon the beam, an attachment was designed, as shown in the top of the photograph, so that the beam was loaded by dropping ! shot from a tin funnel into a vessel suspended from the scale beam. This avoided the irregularities incident to the machine as furnished by the manufacturers. I 42 Compression Pieces. As there is no constant relation between the transverse modulus of rupture of concrete and the compressive strength of the same mix¬ ture, a scheme was devised for obtaining the compressive strength of all of the mixtures in addition to the transverse modulus of rup¬ ture. After breaking the beams into four pieces, the two end pieces were capped with neat cement so as to form prisms about 6 inches square and 19 inches long. The method of capping these prisms is illustrated in the drawing in Pig. 11, below. Two pieces were ;/7gy/iote ELIevortion. Fig. 11. —Sketch Showing Method of Capping Test Prism for Compression. capped at the same operation. A piece of smooth, planed 2-in. plank was laid upon horses, and upright upon this, in a wooden frame, were set four pieces of i-in. plate glass, each 6 in. by 10 in. The lengths of the prisms were gaged by boards 1 by 6 by 19 in. placed length¬ wise and 6 in. apart, so that the broken piece of the beam fitted be¬ tween them with a space at each end between the rough ends of the beam and the plate glass. The pieces of beams were thoroughly soaked with water before beginning the operation. A plastic paste of neat cement as stiff as could be readily handled and molded 43 was next worked into the spaces between the ends of the specimen and the plate glass, and allowed to set over night. When removed from the mold, each prism was capped with neat cement with a smooth, glossy surface, and the two ends were parallel. These prisms were sent to Stevens Institute, at Hoboken, and broken in the compression machine. Certain ones were tested there also for elasticity. Comparative Compressive Strength of True Prisms vs. Capped Pieces of Beams. The question naturally arose whether the strength of prisms capped with neat cement in this way corresponds to the true com¬ pressive strength of the concrete. The neat cement capping, which was of different thickness in the different specimens, might affect the strength, and the specimens might be strained from the rupture in the transverse tests. In order to compare the strength of the capped pieces with the true prisms, one of the beam mixtures was used for making up four prisms 6 in. by 6 in. by 18 in., and these true prisms were broken at the same age as the capped pieces of beams mixed with the same ingredients in like proportions. The results of this comparison are shown in Table 5, below. It is noticeable that the TABLE 5.— Compressive Strength of True Prisms, 6 X 6 X 18 Inches, vs. Capped Pieces of Same Dimensions from Broken Beams, 1:9 Concrete, with Graded Jerome Park 2|-in. Aggregate. True Prisms. Capped Pieces of Beams.* Reference No’s. Compressive strength, lbs. per sq. in. Variations from mean. Reference No’s. Compressive strength, lbs. per sq. in. Variations from mean. 161 1 375 93 155 1 580 153 161 1 240 42 155 1 450 23 161 1 235 47 155 1 250 177 162 1 330 48 157 1 490 63 162 1 315 33 157 1 485 58 162 1 195 87 157 1 305 122 Average. 1 282 58 Average. 1 427 99 Per Cent. 4.5 14.4 - - | L. * The Compression Tests in the other Tables are made upon the capped pieces of 44 capped prisms give a higher average strength than the true prisms, and there is greater variation between the different specimens. The variation, however, is not so great but that the results from the capped pieces is of value, at least when used in connection with the values of transverse strength. In the tables which follow, the com¬ pressive strength of these capped prisms is therefore given side by side with the transverse strength and the density. The full results of the compressive tests are also given in Tables 14 a to e. 1904 Beam Tests. The first series of tests of beams were made in the winter of 1903-4. The materials were Portland cemenl crusher-run screen¬ ings brought directly from the crusher in the reservoir without laboratory screening, and crusher-run broken stone also brought directly from the crusher after there sifting out the screenings in the revolving screens. The results of the tests are somewhat erratic and the breaking strengths are low, because of the peculiarity of the cement, to which reference has already been made. The concrete set very slowly, and had to be stored in the mold for two or three weeks before the beams could be handled without breaking. In the diagram, Fig. 12, curves of equal modulus of rupture are drawn as contours interpolated between the plotted breaking- strengths as found. Notwithstanding the apparently poor results of the tests and the relatively low breaking strength, the curves fol¬ low the same general direction as those of the tests at Little Falls, N. J., made in 1901, with broken trap rock and sand, although the strength of similar mixtures is less at Jerome Park. Fig. 13, page 45, gives curves showing the weight of cement in pounds in 1 cu. ft. of the set concrete in the beam. These results also agree closely with those at Little Falls except that the latter contain more cement per cubic foot of set concrete than the same proportions at Jerome Park. 1905 Beam Tests. As a result of volumetric tests of density carried on in the spring and winter of 1904, the concrete beams in 1905 were made according to definite plans to compare the strength of concrete made with different materials and different proportions of the various diameters of particles. Cement and aggregate, graded by curves of mechanical 45 ttodu/us of Rupture of Concrete Beams,Gincbes 5quane-30and60 Span- tbundjperSquore inch, fbrts of Stone, by Weight. 0 I 2 3 A 5 6 y 8 9 /O II /2 /j ,q \ > \ Data from Tests at Jerome Park ReservoirN. Y, 1904. Aqueduct Commission Gian t Cement, Jerome ParRScrveninqs used assart? Jerome ParRS/oneiz"To 2/4 Concrete mixed very wet. Yr . 0 / Z 3 -f 5 6 7 Fig. 12 Weight of Cement, in Pounds,Required in One Cubic Foot of Concrete. Parts of Stone, by Weight. I I V. | <5 \ X / 9 '0 // /2 ' 3 / y ' X \ X \ \ \ data from tests at Jero/t/e Part Agueducf Comm 'Reserve 'ission vrNXRi m 0 / 2 : 3 \ ' 4 A r ^ y G/ant Ce/nent,Jero/ne fbrpScreenings usee/os sand. Jerome Park; Stone t" 1o 2.±" Concrete mixed very n/et. Fig. 13. 46 analysis which we're found by the volumetric tests to produce con¬ crete of maximum density for a given kind of material, were used as a basis for the tests. The strengths of these ideal specimens were then compared with the strengths of concrete mixtures proportioned by the artificial mechanical analysis curves which in the volumetric tests showed less density, to establish the general law that with the same percentage of cement the densest concrete is the strongest; the strengths of these ideal mixtures were compared with the strength of mixtures of natural material in ordinary proportions, and also in the best possible natural proportions; the relative strengths of con¬ cretes made with stone of different maximum size were compared; the relative strengths of concrete made with Jerome Park stone and screenings were compared with Cowe Bay gravel and sand and with Jerome Park stone and Cowe Bay sand; and the comparative strengths of neat cement and concrete beams made with different brands of cement were tested. The beams were broken in the beam machine into four pieces, and two of the end pieces, as already has been described, were capped with neat cement for compression tests, some of these specimens being also tested for elasticity. One piece each from twenty-seven beams, selected so as to show relative results, was tested for permea¬ bility, as described in subsequent pages. When a series of tests was commenced full instructions were typewritten, so that the experiments could proceed in the laboratory without direct supervision. It was intended that two beams should be made of each mixture, each beam being separately manufactured, however, in order that the exact weights of materials entering into it could be determined. In certain cases it was found expedient to make only one beam of a mixture, but the results with these speci¬ mens were not so good, three transverse breaks and two compressive breaks not being sufficient to give a fair average of the material. The full list of tests, of which there were 116 specimens in all, are briefly scheduled as follows: Straight Jerome Park material, using, respectively, 2f-in., 1-in. and i-in. stone, mixed with screenings and 10% cement, by weight, to total dry materials, and graded to the best ideal curves for each size, and to other artificial curves having more sand and less stone than the ideal. 47 Straight Jerome Park material, using, respectively, 2|-in., 1-in. and 1-in. stone, mixed with screenings and cement in three natural proportions—1: 21: 61, 1:3:6, and 1:31: 51. Straight Jerome Park material, using for the coarse aggregate, stone of uniform size, 21-in., 1-in. and 1-in. diameter, respectively, mixed with screenings and 10%, by weight, of cement, this fine ma¬ terial being graded to an ellipse. Straight Jerome Park material, 21-in. stone and screenings, mixed to the best ideal curve with 8, 121 and 15% cement to total weight of dry materials. Straight Cowe Bay material, using, respectively, 2-1-in., 1-in. and 1-in. gravel, mixed with Cowe Bay sand and 10% cement, by weight, and graded to best ideal curve for each size; also, in natural propor¬ tions, 1:3:6; and also to a curve identical with one of the straight Jerome Park curves. Straight Cowe Bay material, using 21-in. and 1-in. stone of uni¬ form size, mixed with Cowe Bay sand and 10%, by weight, and this fine material graded to an ellipse. Mixed Jerome Park broken stone and Cowe Bay sand, using, respectively, 21-in., 1-in. and 1-in. stone, mixed with Cowe Bay sand, and 10% cement, by weight, and graded to best ideal curve for each size; also, in natural proportions, 1:3:6; and also to a curve identical with one of the straight Jerome Park curves. Methods of Preparing Materials for Natural Proportions. A few mixtures were made in ordinary proportions, 1: 21: 61, 1:3:6, and 1: 31: 51, as scheduled above, for comparison with the ideal proportions. Instead of using for the aggregates the natural materials as they came from the crusher, which varied from day to day because of the difference in methods of handling and in the character of the stone in different parts of the ledge, an average mechanical analysis of the screenings as they came from the crusher was made by averaging a number of analyses of different lots of |screenings. Average analyses of Jerome Park broken stone, of Cowe Bay gravel, and of Cowe Bay sand, were similarly found. All of ( :hese average analyses are shown in Fig. 14, page 49, and in Tables 13 and 24, pages 85 and 86. The analyses of the two materials which were to be used together were plotted on a diagram and com- 48 bined in the required proportions by the ordinary method of com¬ bining mechanical analysis curves.* The weights of the various size particles for each of the beams which were graded to so-called natural mixtures were obtained from these combined curves. This eliminated the variation in the mate¬ rials coming from the crusher and from the dredge on different days, and gave truly average mixtures. Mechanical Analysis Curves Used in Beam Tests. All of the mechanical analyses of the aggregate and cement which were used in the concretes for the 1905 beam tests are drawn to scale in Figs. 15 to 20, inclusive. On two of the diagrams, Figs. 19 and 20, curves are drawn with the lower portions as parabolas instead of ellipses. These represent trial mixtures which were each used for a single beam, but, giving unsatisfactory results, were discarded. The combination of straight line and ellipse gives a curve which is really nearer a full parabola than does the combina¬ tion of a short parabola and straight line. When making the density tests, the analysis curves for obtaining the weights of the different size particles, instead of being drawn to a percentage scale, were drawn with the top line of the diagram representing the total weight of the aggregate plus the cement. In this way the weight could be read directly from the curve, thus avoiding the necessity of changing the percentages to weights by means of the slide rule. It was found by experiment that more sand was required in the best ideal curves used in the beam mixtures than was indicated by the best ideal curves in the volumetric tests for density. With the larger quantity of materials required for the beams, the mixing could not be so thorough, nor the placing and compacting so careful. Ac¬ cordingly, the ideal curves in Fig. 7 are all uniformly higher than would be expected from the density experiments. In some of the diagrams, a larger number of curves were drawn than would appear from the schedule to be needed. This was due to the fact just men¬ tioned. The first mixtures were made with curves corresponding to the best curves of the density tests, and it was found necessary to raise all of them about 2% at the tangent point to prevent a very *Full description of method of combining: mechanical analysis curves are given in a chapter by Mr. Fuller in Taylor and Thompson’s “Concrete, Plain and Reinforced,” 1905, pp. 194 to 209. ' 49 >ash “ “ Co we Bay 50 o a 5 its* EH ^ la 4- II OQ O 51 Fig. 16.—Mechanical Analyses of Cement, and Cowe Bay Sand and Stone of 2^-In. Maximum Diameter Used in 1906 Beam Tests. 53 Fig. 17.—Mechanical Analyses op Cement, and Cowe Bay Sand and Jerome Park Stone of \ and 1-In. Maximum Diameter Used 55 Fig. 19.—Mechanical Analyses of Cement, and Jerome Park Screenings and Stone of A and 1-In. Maximum Diameter Used in 1905 Beam Tests, 56 57 rough surface to the beam, which indicated that there was not suffi¬ cient mortar to fill the voids in the stone. In Tables 14 a to e, which gives the full data on the beam tests, these specimens are desig¬ nated in the remarks. The results from them, especially in the density tests, cannot be considered absolutely reliable. Equations of Mechanical Analysis Curves for Artificial Propor¬ tioning of Aggregates and Cement. Table 6 gives in full the equations of all the curves of artificial mixes which were used in manufacturing the beams. The curves are numbered in Table 6, and corresponding numbers placed on the curves in the diagrams, Fig. 7 (which contains all of the ideal curves) and Figs. 15 to 20. Comparative Tests Forming Basis of Conclusions. Before giving the full table (Tables 14 a to e) of beam tests, a number of short tables are presented, showing the comparative strength and density of the specimens which illustrate various laws of proportioning and form the basis of the conclusions presented. Comparative Density and Strength of 1: 9 (by Weight) Concrete with Aggregates of Different Maximum Size. In Table 7 are tabulated averages of density and of breaking | tests of concrete mixtures of cement and aggregate of different maxi¬ mum size. The figures represent averages of all specimens tested according to each character of mixing with the exception of four specimens omitted, where exceptionally low strength is due to voids not being filled. In every case but one the mixture with 2|-in. j stone for the maximum size is denser than that with 1-in. stone, ! and in every case but one the 1-in. is denser than the i-in-. The | modulus of rupture and the compressive strength of the specimens with the three sizes of stone follow the same general order, the con¬ crete with coarse stone being always stronger than the finer except in one case where the 1-in. is 1% stronger than the 2^-in. The general averages are calculated at the bottom of the table, and also the ratios of density and strength based on the 2^-in. material as unity. Comparing these ratios with the ratios of strength TxVBLE 7 .—Aggregates of Different Maximum Size. Comparative Density and Strength. 58 m ^ CO fn w S . (L w « 2 >> , O' oqOoq eo COO OC pOS H ►J Ou Sg IN P4 J H £ S g” C3 5 • > oPcc <1 S3 2 o g) .2 «« © C ’rt tn C8 tS tjo ►> s-* he . 0cq gw Sg*g n w 0-i •iqSiOM £q 'ijodoaj S3 »n tr'o f-1- 1 -H prf Qpioaocc ^ iO io co QOOOi 1— CM CM CM CM CL : a ^3 _c»_ 915 821 890 1 231 1185 1 008 .72 ® : a -o m 950 879 950 1402 1 585 1 153 .83 : ® -? § 02 +3 m 1 342 980 1 350 1 486 1 798 1 391 1.00 ® - a ^2 _co_ 208 162 180 189 207 ® *n GO J> ® u s ' a Q «o i-i at «d os C3t-OJTti£- OJ Ti Oi OJ o* 05 i-H (M 05 ! 04 OQ th t- in os 5* ag §S a§ oo oc §§ % - g .2 © > ce <1 P 3 59 with different percentages of cement as given in Table 13, page 71, it appears that an additional amount of cement to the exent of about i part with the maximum aggregate 1 in. and i part with the maxi¬ mum aggregate i inch will be required to produce a concrete equal in strength to a concrete having a maximum aggregate of 2\ in. I Comparative Strength of 1: 9 (by Weight) Concrete with Jerome Park vs. Cowe Bay vs. Mixed Aggregates. Tables 8 and 9 give the comparative density and strength of concrete made with Jerome Park stone and screenings vs. Cowe Bay gravel and sand vs. Jerome Park stone and Cowe Bay sand. The ! proportions are given as 1: 9 and 1:3:6 by weight. This is not | - strictly correct because in order to make a true comparison of the different materials a correction was made for the Cowe Bay mix¬ tures. Actually, the proportions for the straight Cowe Bay are \ 1: 8.43, and 1: 2.81: 5.62, and the actual proportions by weight for the mixed materials are 1: 8.80 and 1: 2.92: 5.88. By making this • correction, the proportions by absolute volume are exactly the same, 1 and as it is this which affects the composition, it is the proper method | of proportioning. The results, accordingly, show the true relation between the rounded gravel and the broken stone. The values in Table 8 are the results of tests of concrete com- 1 posed of the aggregates and 10%, by weight, of cement graded to the best elliptical curve for each material, which represents in general the best possible mixture of each material with the given percentage I of cement. In general, the straight Cowe Bay material produces the t greatest density. This is undoubtedly due to the rounded character j of the grains, and is in accord with the results of other experi- . menters. The mixed Jerome Park stone and Cowe Bay sand forms a concrete less dense than the straight Cowe Bay material but denser than the straight Jerome Park; in other words, it is intermediate I between the two. The breaking strength does not follow the same direction as the • density. The most noticeable difference is between the straight Cowe Bay and the mixed Jerome Park stone and Cowe Bay sand. The straight Cowe Bay material, although producing a denser mix¬ ture, is not so strong as the mixed material in two cases out of 60 3 W pH H O 9 ft rn & gg < a o % < 2 ® Q 5 > O M g ^ a © So x § 9 * S <1 < rn ^ >"H rq g H pP *< O h & O w 1—1 a w *< > Ch PS P g o O s >“5 CM Compressive Strength at 140 Days, Lbs. per Sq. In. J. Park stone. C. Bay sand. 1 888 1 770 1 243 1 634 3 C. Bay gravel and sand. 1 533 1 425 I 610 1 523 o J. Park stone and screenings. 1 425 1 052 888 1 119 C5 Modulus of Rupture at 90 Days, Lbs. per Sq. In. J. Park stone. C. Bay sand OJXN ©*©*©« 263 00 C. Bay gravel and sand. o ^nco co in th 235 - J. Park stone and screenings. MHO S ©j a? 221 CO Density. J. Park stone. C. Bay sand. .874 .820 .78? 8 o C. Bay gravel and sand. .874 .845 .827 a 00 J. Park stone and screenings. 40 C3 co 00 00 i- .814 CO •ouo^s 0 zis umuiixui^ ©?*"* HO * ©i •juSioav suox^jodojtj 05 05 OS j j ( - Reference Numbers 155, 157, 217, 220, 213, 214... 151, 152, 223, 226, 210. 164, 165, 259, 211, 212. Averages.1 O a --2 >H 61 three. An examination of the tests indicates that the one test with i-in. stone which does not follow this rule is the erratic one rather than the other two, since in the transverse tests the difference in the i-in. specimens is only 2%, whereas the difference in the tests of larger (2J in. and 1 in.) stone is much greater, while in the com¬ pressive tests the strength of the |-in. specimens, which averages 1 610 lb., appears abnormally large, being, in fact, greater than the tests above it with the coarser stone. It would appear from this that broken stone concrete of mica schist rock is stronger than gravel concrete under like conditions, and that this increase in strength is due to the difference in the surface of the coarse particles of stone. This, in general, is in accordance with the results of the best experi¬ ments, notably those of E. Candlot, in France, although it has been disputed in many quarters. In Table 9 the mixtures of the different materials are in the same proportions (corrected for different specific gravity), and not those demanded by the best analyses for each material. The results are, therefore, not apt to be so conclusive as those scheduled in Table 8, but it will be seen that they follow in general the same direction. The straight Cowe Bay gravel and sand mixture, while considerably denser than the mixed Jerome Park stone and Cowe Bay sand, is generally of lower strength than the latter. Both tables indicate that a mixture of Jerome Park stone and screenings gives a concrete of lower density and lower strength than either the straight Cowe Bay mixtures of the mixed Jerome Park stone and Cowe Bay sand. In the 1: 3: 6 mixtures. Table 9, this might be due in part to the fact that the materials are not graded to the proportions which are best for them, but this difference is eliminated in Table 8. Further tests are essential to determine whether the principles just stated I apply to other classes of broken stone. The age of the specimens also tends to affect the relative strength, because if the mixture is ! rich enough, as the concrete becomes older there is more and more tendency in compressive tests for the stones to shear, so that the actual strength of the particles of stone become more and more a function of the strength of the concrete. * 62 fc S h Eh ^ $ oq ^ W H H PS ^ H o m w ps A O 6 • w «c K. 5> £ H w H c* j 3 fe r. Park :one, 0. ly sand. 1 325 1 562 1 242 1 376 b u, 0* th • 05 m Sg §1 J. Park stone and screenings. t»lOO S (£§ a . o £> "d a®, ”|| o « 1 Sh to -T3 M ^ rt h «8.S So5w Note.— Weights of Cowe Bay material are corrected for difference in specific gravity, so as to have the same relative absolute volumes as the Jerome Park material. * Only two breaks. 63 Comparative Density and Strength of 1: 9 (by Weight) Concrete with Aggregate Graded by Ideal Mechanical Analysis Curves vs. Mixtures of Natural Materials in Ordinary Proportions. Methods of determining the best artificial mixtures of sizes of aggregates (plus cement) are discussed on pages 30 to 37, and the equations of the ideal mechanical analysis curves are given on page 35 and in Table 6. Table 10 compares density and strength of con¬ crete graded by ideal mechanical analysis curves with density and strength of natural proportions. The tests of graded mixtures in nearly every case are higher than the natural proportions. Greater differences would have appeared if the best ideal and best natural proportions had been selected instead of averaging all of them. Instead of using the average of all the specimens made by artifi¬ cial curves and the natural proportions, if a selection had been made of the specimens proportioned by the best ideal analysis curve with 2^-in. maximum stone—that is, by the lowest artificial curve with which the voids were filled, and by the best natural mixtures with the same size stone, which, on the whole, were the 1:3:6 propor¬ tions the average values of modulus of rupture would have been 241 lb. per square inch for the ideal and 211 lb. per square inch for the natural proportions, representing an increase in strength of 14% by using the ideal mixtures. The desirability of using an artificially graded mixture is therefore dependent upon economic conditions. The relative saving in cement is estimated with the aid of the data in Table 13. Comparative Density and Strength of 1: 9 (by Weight) Concrete with Aggregates Graded by Ideal Mechanical Analysis Curves Having Graded Coarse Aggregate vs. Aggregates Graded Similarly in the Sand Portion, but with Coarse Aggregate of Uniform Size. Trial mixtures of concrete made in the laboratory with uniform coarse stone gave greater density than could be obtained with uni¬ formly graded stone. However, in the former nearly 50% of the weight of total aggregate consisted of particles of the coarsest diam¬ eter of stone, and, in mixing, the concrete appeared very coarse, so TABLE 10. —Aggregates Graded by Ideal Mechanical Analysis Curves vs. Mixtures of Natural Materials in Ordinary Proportions. Comparative Density and Strength. 64 CO rH Oi f-H T-H CQ <3! v- rH <3M3* 212 05 Ideal. CO X) oooco CO 00(3* §? § e» 236 GO Average Density.* Natural. .821 .798 .768 .826 .844 .817 .834 .818 .787 .810 i> Ideal. .851 .810 .767 .863 .845 .827 .861 .820 .782 00 co Maxi¬ mum size of stone. 3? 1-1 Sf-r-l3? T-c --1 ” to Materials. Sand. M c£ b n Q? i i - - - - £ Hj O o T* Stone. M % M £ . . S _ Ph - - ® 3 s CL, G * a ^ CO Propor¬ tion by weight.* OS Os 05 Os Os Os os Os OS (N Reference No. it a.f (3? lO *> phNtKWXN •CO'^GO CO 1 I os (3? (31 CO «wOO ^aocr. tH^y-HtH(380^(3? • O* <3* <3* 1 COCOGG 1 1 1 1 1 *1 I 1 OrHHrHTHCDWHGCO • iO CO CO 1 1 *> os <3* (31 CO -OOO rH O CO T-H rH *> - Reference No. - 1 Bfi S 144- 146-148 149-150-155 156-157-158 145- 147-151 152- 159-160 153- 154-164 165-195-196 217-220 223-226 259 200-209 213-214 210 211-212 Jj «a 53 > < D 1 ABLE 11. Aggregates Graded by Ideal Mechanical Analysis Curves Having Graded Coarse Aggregate vs. Aggregates Graded Similarly in the Sand Portion, but with Coarse Aggregate of Uniform Size. Comparative Density and Strength. T 65 £5 % £ o H K3 GO S os < < £ O' gczj H £ p o > ◄ & 05 to H <1P © a s O ce .'ll! © ca ” OS ;ggss © T3 o © TJ W os o> o CO d o o © a •o o g? o OCMCSt-iOO lOrnrHMin 04C«O<(MC« 0Ji>O5C co os co c oo t- 1 - c © • CO © h a <3 o © 43 t> CO MO© III H iO O* CO i % % S3 cp. *= - r * a i l>»4J OX! .0 %a.<$ E o © P-T-£ is © g a.5© goofl n Cl'S © 2® 02 3 © » -d . S 3 . © ® 0-3 B ©£ go © bc w 02 a & £ M. Oh j s g- . P •-5 O o i-HC? 0* 66 that it was considered doubtful whether the concrete would work well in practice because of the apparent tendency of the mortar to run away from the stones, and leave pockets of stone and voids. Accordingly, in this 1905 series of experiments only a few beams were made with stone of uniform size in order to determine whether the results were sufficiently good to warrant the continuation of this line of tests at a future date if circumstances permitted. The results as indicated in Table 11 show no marked difference between the specimens with graded coarse stone and those with uni¬ form coarse stone. The former average slightly higher in strength, while the latter have slightly greater density. This is contrary to what would be expected, since the tests in general indicate that the densest mixtures produce the strongest concrete. The fact can be explained at the present time only by assuming that the uniform coarse stone does tend to separate, as suggested above, and thus pro¬ duces a less homogeneous concrete. Further tests are necessary, however, to determine whether this is true or whether these few results are abnormal. The analysis curves employed in the mixtures of uniform stone are indicated in Figs. 15, 16, 19 and 20. They were adopted as a result of the volumetric or density tests. The density tests indicated that a slightly better mixture would have been obtained by lowering the curve at the juncture of the ellipse and the first straight line, but since they also indicated that this style of curve required more very fine sand than a curve with graded coarse stone, it was thought best to adopt the curves shown, and then use for comparison with them the specimens with the graded stone, whose analysis curve in the sand portion was the same ellipse. Influence of the Analysis of the Coarse Aggregate upon the Strength of the Concrete. —One of the most important conclusions which may be drawn from the comparative tests shown in Table 11 is that there is comparatively little difference in the density and strength of concrete, whatever may be the analysis of the coarse aggregate (meaning by the coarse aggregate, all the particles whose sizes are greater than a diameter which is one-tenth the maximum diameter of the stone), provided this coarse aggregate does not con¬ tain a greater weight of medium size particles than can be repre¬ sented by a straight line. In other words, the coarse aggregate may be uniformly graded above the size corresponding to one-tenth the 67 diameter of the coarsest particles, or it may contain an excess of uniform size stone whose diameter is the same as the coarsest diam¬ eter of the graded stone. Expressed in another way, the effect upon the density and the strength is practically the same whether uniform stone or uniformly graded stone, or a mixture between these two, is used. However, tests should be made on a large scale with concrete mixed by hand and by machine before adopting uniform stone in regular construction work, in order to prove conclusively that the stones in the mixture of uniform stone do not separate from the mor¬ tar; in other words, that the concrete does not work harshly. While the tests of strength do not include tests to determine whether or not it might be possible to use an excess of medium size stone over and above what would be contained in a uniformly graded coarse aggregate, the tests of density indicate that when the curve of the coarse agg’regate is raised above a straight line, the density is immediately decreased, signifying that the uniformly graded coarse aggregate or an aggregate whose curve is below a straight line, is the t to use. Effect of Fineness of Sand upon Density and Strength. The relative effect of sand and stone upon the density and strength is evident from the preceding paragraphs. The density of I I • the concrete is affected very much more by the variation in diameter of the sand particles than of the stone particles, an excess of fine or medium sand decreases both the density and strength of the con¬ crete. The effect of the fineness of the sand, that is, the pro¬ portions of the different size grains, was carefully studied in the volumetric tests described on pages 30 to 37 inclusive, and the results ia were confirmed by laboratory and field tests. Relation of Ideal Curves of Different Size Stone. From the preceding paragraphs it is evident that the equations ' 1 of cury es given on page 35 for aggregate plus cement show that, for a given character of aggregate, a single equation whose only variable | is the diameter of the particles of maximum size will apply to aggre- j gates whose maximum diameter varies from \ in. to in. 68 Effect ox Density and Strength of Concrete by Increasing Per¬ centage of Sand and Decreasing Stone. The density experiments indicated very distinctly that the best ideal curve of mechanical analysis having been found, raising this curve in any portion of its length decreases the density of the mix¬ ture. Accordingly, to test this point in the beam tests, as shown in the diagrams, Figs. 15 to 20, specimens were also made with each material proportioned according to a curve which was raised above the best ideal at the tangent point of the ellipse and straight line, the lower portion of these higher curves being still in the form of an ellipse with the same horizontal or major axis but a longer minor axis. More simply expressed, this is equivalent to increasing the sand and decreasing the stone. It is evident from an examination of the full data in Tables 14 a to e that the results of this comparison were not so conclusive as might be desired. This is partly due to the slight difference between the various curves, to the comparatively few tests made with each curve, to the variation which must be expected in all tests of concrete, and to the fact that in some of the lowest curves the voids were not entirely filled and therefore the results less accurate. With the natural mixtures the same rather inconclusive results w T ere obtained, the proportions selected being so nearly alike that the difference is not marked. The general trend of the tests, however, indicates that the theory assumed is correct, and it is certainly substantiated by previous tests made by Mr. Fuller, which indicate quite conclusively that with the same proportions of cement to total aggregate, the strongest mixture is that with the largest possible proportion of stone and the smallest possible propor¬ tion of sand. TABLE 12.— Effect on Density and Strength of 1: 8 (by Weight) Concrete of Increasing Percentage of Sand and De¬ creasing Stone. (From Little Falls Experiments.) Proportion by Weight of Cement to Total Aggregate. Absolute Volume of Cement. Density, i. e. Total Av. Modulus of Cement Rupture at 33 Days, Sand + Stone. . Lbs. per Sq. In. 1:8 1:2:8 .091 .865 319 1:8 1:3:5 .086 .833 285 1:8 1:4:4 .082 .801 209 1:8 1:5:3 .080 .799 151 1:8 1:6:2 .076 .760 102 1:8 1:8:0 .073 .754 41 69 In Table 12 are given the results selected from tests at Little I Tails, N. J., made in 1901. All of the specimens presented have the i same proportions by weight, namely 1: 8. The ratio of sand to stone, however, is varied from 2: 6 to 8: 0. It is seen from an inspection of the table that in every case there is an increase in both density and strength as stone is substituted for an equal weight of sand. Proportioning Sand and Stone in Practice. Where two aggregates are used, such as natural sand and natural gravel, or crusher-run screenings and crusher-run broken stone, it is evident from the preceding paragraphs that the relative propor¬ tions of sand to stone may be properly graded in the field from day to day by employing in all cases as little sand as may be and avoid visible voids in the concrete. In practice, when proportioned in this way the mixture is found to be not only theoretically denser and stronger, but it works most smoothly in placing. Accordingly, in laying the lining in Jerome Park Reservoir, the inspectors were instructed to vary slightly from day to day the relative proportions of sand to stone, according to the way the material was running as it comes from 'die crusher, or the bank, while using at all times the same total volu me of total aggregate. Having selected a ratio of cement to total aggregate, this gives a better and easier method of proportioning in ci.ses where it is not considered economical to use more than two aggregates—that is, where artificial grading is pro¬ hibited—than any number of theoretical tests. This plan is also especially advantageous where the sand varies in fineness from day to day, for the finer the sand, the farther it will spread, and the less of it is required to fill the voids of the stone, and also the less of it should be used in proportion to the cement, since a mortar of fine sand is weaker than a mortar of coarse sand. Comparative Density and Strength of Similar Concrete with Different Percentages of Cement and 21-inch Stone Graded as an Ellipse and Straight Line. In Table 13 is tabulated the density and strength of beams in which different percentages of cement are used to the weight of the total dry material, ranging in percentages from 8 to 15- per cent, and 70 in proportions by weight from 1: 11.5 to 1: 5.7, thus covering all ordinary proportions in practice. As the percentage of cement is increased, the strength increases and in nearly similar ratios. Effect upon Density of Substituting Cement for Fine Sand. The average densities in Table 13 show an extreme variation of less than 1 per cent. This is in confirmation of the statement already made that a substitution of cement for fine sand having grains of the same diameter does not affect the density. This is also proved still more definitely in the volumetric tests of density. As the mechanical analysis curves in all the beam mixtures include both cement and aggregate, in the specimens with different percentages of cement the additional cement is actually substituted for sand hav¬ ing the same size grains. Complete Schedule of Transverse and Compressive Tests, 1905. In Tables 14a-e are given the complete results of the tests of concrete beams together with the reports of the compressive strength of the pieces of beam which were capped with neat cement. Nearly all the tests given in these tables have already been in¬ cluded in the short comparative tables, and most of the results of apparent value have been presented there. Because of the difference- in age of the transverse and compressive tests, no decisive ratio can. be calculated between the transverse moduli of rupture and the com¬ pressive strengths of the same specimens. The age is somewhat diffi¬ cult to allow for, as the growth of strength in tension and compres¬ sion are not necessarily in the same ratio. Most of the beams are in duplicate, as indicated by the bracket before the average modulus of ruture. As a rule, the columns are self-explanatory. Columns 7, 8 and 9 define the mechanical analysis curves where artificial mix¬ tures are employed. The analysis curves are all drawn in the dia¬ grams, Figs. 15 to 20. As has already been stated, a few of the curves have a parabola for their beginning, but this was discarded as giving less satisfactory results than the ellipse. The special ellipses used in beams No. 149 and 150 were discarded because there was not sufficient fine material to fill the voids. All of the other ellipses are drawn according to the regular plan adopted at the out¬ set of the experiments. TABLE 13. —Comparative Density and Strength of Similar Concrete with Different Percentages of Cement and 2i-lNCH Stone Graded as an Ellipse and Straight Line. 71 a a ° H p* Cfi . cy> 2 a a H ,2 H > M g ^ a b a^ HW OS os^ a PS ^ Qj a o *£1 ggo tfQ o CO M 2 t= g | Q s a g m a cs ◄ S I*. S W Eh £ g Z ^ pi a fH « S hPh a ShO a w § Qg° a a 5? a a ®.S a « gee -a cfi • © «a° Jh cS ® _£3 ; o ! OJ o 250 ID 00 (M oo | CO CD 1 CD i> o 05 00 00 & GO CO ID GO S!T g? 00 cc oc I GO 36 | CD GO ID ID 00 8 CO 00 § GO S § 00 V Qj Pi TJH CO CO rr PO Tf S «f5 °° ^’i ’i • i> CO O 05 ~TC5 C« OJ 03 (M uo .co c Aeoeo G °.cocoeococo J> - • ^ ..t- - co oo t- oo tj< co ■<* a-ga a 133 £ O 73 -a =a 0) ® a P a.2 Ptra 03 08 03 03 03 05 03 o’ o' o’ o’ d o d o’ o" o* o’ >> , „ . c3_ „ . - - - eq- - - D >4 M- ® ■ o cc Cl ?H £S =2 S ® p « « ss8 TH .g8 TH 48ii§Sj ^ ria r -oo „ +3 © W S3 ge a&H >>% ce £ M c ® CD £•-9 O cu oja 72 The weights in pounds, Columns 10 to 14, are calculated directly from the beam mixing data. The total weights per cubic foot of the different specimens follow the same general relation as the densities already referred to, the aggregate having coarsest stone giving the heaviest concrete, and the ideal mixes being usually heavier than the mixes of natural proportions. The Cowe Bay materials give a slightly lighter concrete than the Jerome Park mixtures because of their lighter specific gravity, even although they are denser. The weights of the concrete as mixed run from 146 to 162 pounds per cubic foot. Column 15, as stated in the foot-note, represents the minimum theoretical weight of the concrete, assuming that this is made up of the dry materials plus a quantity of water for chemical combination equal to 10% of the weight of the cement. This theoretical mini¬ mum is from 8 to 12 lb. below the actual weight as mixed. Column 16 gives the theoretical weight, assuming that the voids are all filled with water, and a comparison of this column with column 14 and also an inspection of column 21 indicates that the air voids in most of the mixtures are inappreciable, the difference being nearly within the limits of error which must be allowed for in the weights and measurements. The calculated volumes of material in one cubic foot of beam, columns 17 to 21, are really absolute volumes, the total dry volume, column 19, representing the sum of the ratios of the elementary volumes of the cement and of the aggregate to the unit volume, thus giving the density of the mixture. Inspection of the quantity of water in the beams, columns 13 and 20, reveals that the smaller stone, that is, the i-in., requires the most water in gaging, and the coarsest stone, the 2^-in. the least. The Cowe Bay gravel and sand require less water than the broken stone and screenings. The total volume of voids per cubic foot, column 24, gives merely the complements of the densities in column 19. The voids in the cement, column 22, are based upon the voids in the two neat beams given first in the table, it being assumed in the concrete speci¬ mens that the voids due to the cement are in the same ratio to the weight of the cement per cubic foot of concrete as the voids in an average neat beam are to the weight of cement in this neat beam. TABLP with Different Percentages ms. Spans, 68 and 32 Compressive Strength of Concrete Prisms, 6x6 X 18 Inches. Average Age, 140 Days. 1 1 25 | 26 | 27 » 1 —\ 29 30 1 1 2 1 3 i 4 \ 5 6 Modulus of Rupture. w Compressive Strength. £> S ai M Pounds per square inch. pd © rQ g Pounds per square inch. P s AS S3 8 ce r a 5 a | © tuo 1 ® ® © © % i & a a © § © = 4-4 4) P3 © be <1 o o' fc a 1 | '3 § e3 © ► <1 Ph © ® tf-. ® « <4-1 o o’ 0 a M ce § 3 a a g tuo c3 c © > 219 201 2(2 90 3 295 276 285 219 135 141 142 139 2 2 2 2 1 775 1 375 1 565 1 875 1 410 1 265 1 550 1 590 1 320 1 560 1 770 91 91 90 3 3 3 248 290 300 220 246 280 J-251 288 Surface rocky. 201 202 210 210 1 ooU 203 90 qa 3 O 298 OQ Q 234 OKK J-276 203 141 2 1 745 1 555 1 650 204 Ol K 91 /C 3 /Cvo 261 4DD 241 249 204 215 141 136 2 2 1 520 1 435 1 345 1 475 1 445 /ClD 211 90 q a 1 9 196 196 ono 1-205 211 139 2 1 175 1 085 1 130 212 207 «7U 91 O 3 Q 251 Surface rocky. 1 265 1 320 1:8*° 10.2 1 " 0.10 33.2 14.5 127.6 142.1 10.3 152.4 143.3 153.3 .075 .745 .820 .163 .983 .064 .116 .180 90 3 800 280 288 210 139 2 1 666 l 875 1 660 1 770 1:2":6 88 10.2 « 1 " 14.5 127.5 142 0 9.7 151.7 143.2 153.2 .075 .745 820 .156 .976 .064 .116 .180 90 3 298 234 i 208 141 1 745 1 r 1:2 93 :5 88 10.2 14.4 126.7 141.1 9.8 150.9 142.3 152.6 .075 .740 !815 .1.-.; .'.>72 .064 .121 .185 90 2 298 255 204 141 1:8 80 10.2 1 " "56' '37\7* Ellipse. 14.4 126.7 141.1 11.0 152.lj 142.3 152.6 .075 .740 .815 .176 .991 .064 .121 .185 91 3 261 241 249 215 136 2 1 545 1 345 1 445 1:8 80 10.2 h" .075 34.4 13.9 122.3 136.2 12.8 149.0 137.3 149.6 .072 .715 .78? .205 .992 .061 .152 .213 90 1 196 196 i 211 139 1:8 80 10.2 j" .075 34.4 13.7 120,4 134.1 14.9 149.0, 135.2 148.1 .071 .705 .776 .239 1.015 .060 .164 90 3 817 202 r 205 1 180 1:2 W3 :5 88 10.2 f 13.8 121.3 134.1 12.4 147.5 135.2 147.6 .073 .710 .7(-3 .199 .982 .062 .155 i217 91 3 838 210 (.910 207 140 2 1 885 1 080 1 285 1 310 1:2":5 88 10.2 14.0 123.0 137.0 12.8 149.31 138.1 150.0 .072 .719 .791 .197 .988 .061 .148 .209 91 3 235 192 208 140 2 1 205 1 115 1 175 1:6 87 12| 21" 0.20 33.8 Ellipse. 19.2 131.3 150.5 9.4 159.9 152.0 158.9 .099 .767 .SI Hi .151 1.017 .084 .050 .134 90 3 418 853 879 Slightly rocky. 218 185 2 2 230 2 150 1:6®* 15.3 H" 0.20 33.8 22.7 126.0 148.7 9.6 158.8j 150.5 157.9 |.117 .735 .154| 1.006 .099 .049 .148 91 3 445 891 409 216 180 2 2 440 1 840 2 140 Note — Volumes, cubic feet per 100 lbs. as mixed : Cement, 1.00; Jerome Park screenings (Crusher Rimi, 1.06; Jerome Park stone, 1.03: mixture Jerome Park screenings and stone varies with mixture; Cowe Bay sand (natural) in - „ ■Travel 0 '17 Specific gravity: Cement, 8.10; Jerome Park screenings varies with size; Jerome Park stone, 2.78; mixture Jerome Park screenings and stone, 2.77: Cowe Bay sand, 2.65; Cowe Bay gravel, 2.65. Weights per cubic foot iN nnmi r'l' •' l on- Jerome Park screenings (Crusher Run), 94.8; Jerome Park stone, 9.70; mixture Jerome Park screenings and stone varies with mixture; Cowe Bay sand (natural), 0.90; Cowe Bay gravel, 102.7. Tensile strength cement pounds ner so iar« imli ' Neat 7 days, 662; 28 days, 695; 8 months, 720. F H men. ’ (lol 15 = Col. 10 X 0.08 4- Col. 12. Col. 16 = Col. 12 + (Col. 24 X 62.4). Col. 15 represents theoretical minimum, assuming &X of water for chemical combination. Col. 16 repx-esents theoretical maximum, assuming voids filled with wntu,- * Including cement. r See separate table for equations. • TABLE 15. - Composition and Transverse Strength of Beams of Neat Cement and 1:9 (by and Jerome Park Screenings and 2J-Inch Stone. Beams 6 X 6 X 72 Inches. Weight) Concrete Made with Various Brands of Cement Spans, 68 and 30 Inches. Age, 90 Days. Compressive Strength of Concrete Prisms, 6 X 6 x 18 Inches Average Age, 140 Days. Reference numbers. ~ 2 3 4 6 6 7 8 9 10 n 42 1 18 | 14 1 15 16 17 18 19 20 1 21 22 | 23 | 24 25 26 27 | 28 29 1 1 2 1 3 1 4 1 6 1 8 Brand of Cement. Proportions by weight. Kind of stone. Kind of sand. Maximum size of stone. ] Uniformly graded above diameter, inches. Per cent, finer than diameter in Col. 7. Curve below diameter in Col. 7. Weight in Lbs. of Materials in 1 Ft. of Beam as Mixed. :cu. Calculated Volume in Cubic Feet of Materials in 1 Cu. Ft. of Beam as Mixed. Volume op Voids in 1 Cu. Ft. Modulus of Rupture. Reference numbers. Compressive Strength. Age, days. No. of breaks. Pound per square s inch. 6 < No. of breaks. Pounds per square inch. Cement. Aggregate. Total dry mixed. Water. Totals. •g a •«! Minimum. Maximum, j Cement. 1 Aggregate. | Total dry. Water. | Total. Cement. I Aggregate. Total. Maximum. Minimum, 1 £ ■5 Maximum. 1 I a 1 a Average. 210 Giant Neat Cement. 104.5 104.5 29.6 134.1 112.9 133.1 .541 .541 .475 1.016 .459 .... .459 89 3 1 177 929 l 01(1 210 100.5 100.5 30.4 130.9 108.5 180.4;. 530 •••• .520 .487 1.007 .480 "" .480 91 3 70S 744 199 103.4 103.4 29.7 133.1 111.7 132.4 .535 .535 .485 1.020 .465 .465 91 3 591* 260 888 Atlas 101.0 101.0 28.5 129.5 109.1 129. 1 .560 .550 .450 1.060 .450 .450 91 3 852 827 880 238. 280 Universal 100.8 100.8 30.9 131.7 108.9 129.4 .541 .541 .495 1.036 .459 .459 91 8 766 717 748 239 210 Empire 101.9 101.9 31.2 133.1 110.0 131.2 .530 .530 .482 1.012 .470 .470 91 3 1 087 955 1 028 240 211 Iron Clad 96.0 96.0 2s. 9 124.9 103.7 126.6' .510 .510 .463 .973 .490 .490 90 662 598 241 Hudson River 101.2 101.2 31.2 132.4 109.3 130.1 .537 .537 .600 1.037 .463 .463 90 3 608 613 632 242 Pennsylvania 97 3 97.3 27.8 125.1 105.1 127 9 .510 .510 .445 .955 .490 .190 90 3 790 703 737 243 \\ Vlltt'lot t e 91 9 91 9 31.0 122.9 99 3 133 '.1 .488 .488 495 .983 .512 .... .512 90 3 774 509 244 215 Atlas 1:9 J. Park. J. Park. 0.20 37.8 Ellipse. 14.7 132.0 146.7 10.7 157.4 147.9 1 156.71.078 .762 .840 .171 1.011 .004 |:096 .160 88 191 191 | 315 187 2 1 295 1 285 1 290 216 14.2 128.0 142.2 12.0 154.2 143.3, 153.7, .076 .740 .816 .193 1.009 .062 .122 .184 88 2 858 192 I 246 137 2 1 806 1 160 1 260 247 Universal 14.2 127.9 142.1 12.0 154.1 148.2 153.1 .076 .748 .824 .192 1.016 .064 .112 .170 88 2 190 156 | 247 136 2 1 140 960 1 060 248 14.3 188.7 143.0 11.8 154.8 144.1 154 .2 .077 .744 .821 .189 1.010 .065 i.114 .179 93 1 160 160 , 248 135 2 1 060 885 976 219 Empire 14.3 128.3 142.6 11.4 154.0 143.7 154.0 .075 .743 .818 .183 1.001 .067 .n.j .182 93 3 195 1 249 185 3 1 050 910 980 250 “ 14.3 128.8 143.1 11.9 155,0 144.2 1544 .075 .744 .819 .190 1.009 .067 ,m .181 93 3 232 f 230 250 185 2 1 3110 995 1 125 851 Ironclad 14.3 127.3 141.6 11.9 153.5 142.7 153.4 .076 .735 .811 .191 1.002 .073 .116 .189 91 3 228 187 1 251 133 2 1 145 890 1 020 252 14.2 142.2 11.8 154.0 143.3 1 153.7; .075 .740 .815 1.004 .072 .113 .185 91 350 186 353 133 2 1 075 890 980 268 Hudson River 14.3 128> 11.9 154.9 144.1 | 154.3 1 .075 .743 .818 >91 1.0091.065 .117 .182 92 2 213 194 1 917 ! 253 134 2 1 150 1 070 1 110 254 14.3 128.5 142.8 12.2 156.0 143.9 ! 154.2 .075 .742 .817 .195 1.012 .065 .118 .183 92 1 214 214 ,217 254 134 2 1 135 1 080 1 110 255 Pennsylvania 14.1 127.0 141.1 11.6 152.7 142.2 153.1 .074 .733 .807 .186 .993 . 071 .122 .193 91 3 224 224 , 255 133 2 1 390 1 340 1 365 14.4 129.7 144.1 11.8 155.0 145.3 155.11 .075 .748 .189 1.012 .072 .105 .177 91 3 215 215 | 256 133 2 1 500 1 400 1 450 951 Wyandotte 14.5 130.4 144.9 11.6 156.5 146.1 155.4'. 077 .753 !830 .186 1.016 .091 .079 .170 91 3 230 230 | 257 133 2 1 :*45 1 206 1 276 258 14.6 130.9 145.5 11.8 157.3 146.7 156.0 .078 .754 .832 .189 1.021 1 .082 j .086 .168 91 3 205 205 | ' 258 133 2 1 335 1 220 1 225 Notk. -Specific Gravity: Giant, 3.10: Atlas, 3.02; Universal, 2.98; Empire, 3.08; Ironclad, 3.00; Hudson River, 3.04; Pennsylvania, 3.06; Wyandotte, 3.02. Col. 15 — Col. (10 x 0.08) -4-Col. 12. Col. 16 = Col. 12 4- (Col. 24 x 62.4). Col. 16 represents theoretical minimum, assuming: 8^ of water for chemical combination. Col. 16 represents theoretical maximum, assuming voids filled with water. Aggregate in above table is similar to that described in previous table. * Not included in average because the fractured surfaces snow many air holes and lumps of cement. TABT/ IXED with Different 6 X 6 X 72 ( Compressive Strength Concrete Prisms, 6x6 X 18 Inches. Average Age, 140 Days. 27 28 | 29 ulus of Rupture. 178 179 180 181 182 191 183 184 185 186 190 193 197 189 217 220 221 222 ! Pounds per square inch. 268! 208 2931 135 / 235! 223 f 1 245 223-} 334 292 289 Surface rocky I j j 188|‘! 272 ) 262 | 276 361 281 I 346 [ 301 f 331! 300 / 305 270 f 374! 365 221i 188 186 158 223 139 1:2 1:2 250 244 243 265 238 f 176 f 211 j 228"/ 274! 228 329-j 302 -[ 369 204 171 184 230 -j 242-' Slightly rocky Surface rocky Slightly rocky. 178 179 180 181 182 191 183 184 185 186 190 193 197 217 220 221 222 234 ooo © 5b < No. of breaks. Pounds per square inch. Maximum. S' s Average. 1 | 1 2 | 1 410 1 250 1 330 j 1 1 710 1 710, 1 710 2 1 180| 1 090, 1 135 2 1 495 i| 1 410 1 455 2 1 440| 1 315; 1 375 2 1 o65 i 1 450, 1 505 2 1 920 1 1 745 1 830 2 1 715 1 605; 1 660 2 1 795 1 1 750 1 770 ! 2 1 370 1 630 1 500 1 2 1 425 1 395 1 410 2 1 100 940: 1 020 • • ■ • J 2 965 915 940 2 970 970 970 •••■ 2 1 880 1 555 1 715 2 1 140 1 260 1 350 .... 1 2 1 455 1 175 1 315 •••• 2 1 485 1 310 1 395 ■ J 2 1 800 1 800 1 800 11 1 A AZ. 1 ■* t At: 1 TABLE 14 c.— Composition and Transverse Strength of Concrete Beams Made with Different Aggregates of Various Analyses Mixed with Different Percentages of Cement Tested at Jerome Park Reservoir by the Aqueduct Commission during the Year 1905. Beams, 6 X 6 X 72 Inches. Spans, 68 and 32 Inches. Giant Portland Cement. Age, 90 Days. ( '(impressive Strength Concrete Prisms, 6x6 X 18 Inches. Average Age, 140 Days. 7 8 9 10 i' 12 13 14 15 16 17 18 19 20 21 1 Uniformly graded above diameter, inches. ! Per cent, finer than diameter in Col. 7* Curve below diameter in Col. 7.1 Weight in Lbs. of Material in 1 Cu. Ft. of Beam as Mixed. Calculated Volume in Cubic Feet of | Material in 1 Cu. Ft. of Beam as Mixed. J Cement. Total dry mix. Water. Totals. As mixed. Minimum. Maximum. Cement. Aggregate. Total dry. 1 Total. 0.20 85.8 F.llipse. 18.5 129.6i 148.1 9.1 157.2 i 140.6* 157.8 .096 .748! -11 .146 .990 0.20 37. 8 18.7 130.8 140.5 11.8 161.3 ; i5i.o 158.7 .097] .7551 .852| .189 1.0411 0.20 37. 8 18.1 127.5 145.6 11.3 156.9 147.0 156.1 .094 .7371 .831 .161 1.IH2 0.20 39.4 18.4 1 1 128.8 147.2 9.7 156.9 148.7 157.8 .095 .7481.838 .155 998 0.20 18.1 127.0 145.1 10.8 ■ 1 146.5 155.8 .094 .785 .829 .173 1.002 0.20 41.3 17.5 122.8 14D.:: ll 2 151.5 1 141.7 152.8 .091 .709 .800 .180 980 0.20 37.8 21.8 123 6 145.4 11.6 157.0 147.1 155.1 .130 .718 .8-15 186 | 1.031 0.20 37.8 125.0 147.0 10.9 157.9 148.8 155.6 .139 .722 .861 .175 1 .036 0.20 -I 8 123.4 145.2 11.3 156.5 146.9 156.1 .113 .713 .826 .181 1 1.007 0.29 1 39.4 , 21.7 122 ■ 1 144.6 11.4 156.0 146.3 155.0 .124 .710 .884 .188 j 1.017 0.20 41.8 21.8 126.5 148.3 11.5 159.8 150.0 156.0 .116 .7H() .876 .184 1.060 0.20 ! 37.8 11.6 133.5 145.1 10.6 155.7 146.0 155.6 .060 .: 7i .170 1.001 0.20 | 11.6 133 0 144.6 11 4 156.0 145.5 155.4 .060 1.767 .183 1 .01" 0.20 1 41.3 11.5 182 5 114.0 11.8 155.8 144.9 154.9 ; .060 .766 ]826 .189 1 .015 0.20 82.1 15.2 128.0 143.2 8.3 151.5 144.4 152.2 ! .079 .776 .855 .133 .’■IK* 0.20 ; 32.1 , 15.9 14.7 133.7 124.0 149.6 138.7 9.5 9.8 159.1 148.5 150.9 139.9 157.0 149.4 1.082 i.076 .810 .750 .892 826 .152 .157 1.044 .<188 14.7 124.0 138.7 148.6 139.9 149.4 .076 .750 .826 .159 .985 0.20 37.8 Ellipse. 15.3 129.0 144.3 9.2 153.5 145.5 158.0 ‘.079 .781 .860 .147 1.007 0.10 | 31.2 ; 14.9 126.0 140.9 9.8 150.7 142.1 150.9 .077 .764 .841 .157 .998 |_ J. Park. 28 24 25 26 Volume of Voids in 1 Ci. Ft. Modulus of Rupture. 075 .156, RS 8 i 066 .148 02 j 2 090 .169 92 j 8 I 081 .162 91 ! 8 092 .171, 92 8 1281.2001 91 2 .045 .155 91 8 ,021 .139 91 2 .078;.174 90 8 026! .124 118 .169] 122 .178, '4! 884 2821 | . 292 188 ('. i Slightly i 289 , 262 ; 270 846! 801 f 831 800 i 305 270 \ Compressive Strength. 245 Surface rocky.I 17; 1 410 1 250 1 380 1 7101 1 710 1 710 1 180 1 090! 1 185 I 495 1 410 1 455 1 440, 1 816 1 375 1 565 1 450, 1 505 1 920 1 745 1 880 1 715; 1 605 1 660 1 795! 1 750 1 770 1 370 1 680 | 500 1 425 1 895 1 410 1 100] 940 1 020 965 916 040 Note.— Volumes cubic feet per 100 lbs. as mixed: Cement, 1.00; Jerome Park screenings (Crusher Run), 1.06; Jerome Park stone, 1.08; mixture Jerome Park screenings and stone varies with mixture; Cowe Bay sand (natural), 1.11; Cowe Bay ursvel, 0.97. Specific gravity: Cement, 3.10; Jerome Park screenings varies with size; Jerome Park stone, 2.78; mixture Jerome Park screenings and stone, 2.77; Cowe Bay sand, 2.65; Cowe Bay gravel, 2.05. Weights per cubic foot as mixed: Cement. 1110; Jerome Park screenings (Crusher Run), 94.8; Jerome Park stone, 97.0; mixture Jerome Park screenings and stone varies with mixture; Cowe Bay sand (natural), 9.90; Cowe Bay gravel, 102.7. Tensile strength cement, pounds per square inch: Neal ; (lavs, 562 ; 28 days, 695; 8 months, 720. Col. 15 = Col. 10 X 0.08 -f Col. 12. Col. 16 = Col. 12 + (Col. 24 x 62.4). Col. 15 represents theoretical minimum assuming of water for chemical combination. Col. 16 represents theoretical maximum, assuming voids filled with water. * Including cement. TABLE 14 d. —Composition and Transverse Strength of Concrete Beams Made with Different Aggregates of Various Analyses Mixed with Different Percentages of Cement Tested at Jerome Park Reservoir by the Aqueduct Commission during the Year 1905. Beams, 6 x 6 x 72 Inches. Spans, 68 and 32 Inches. Giant Portland Cement. Age, 90 Days. Compressive Strength of Concrete Prisms, 6x6 X 18 Inches. Average Age, 140 I >ays. 23U 281 238 3 | 4 5 6 7 b •o Itf ii 1 •L it f! ° a 08 *§ 1 ii i" 1 If 1 | S 2 o 3 ii £ 05 10.6 j Cowe Bay. Cowe Bay. i " 0.10 io!o I in n nin* 10.6 r 0.075 10.6 r 10.6 181 2}" 6.20 15.9 2J" 0.20 8.5 2j- )*.2u 10.6 21" i) 20 10.6 21" 0i20 10.6 0.10 10.6 | 1 " 0.10 10.2 J. Park. “ 21" 0.20 10.2 .. 2j" 0.20 10.2 “ 21" 0.20 10.2 2 j" 0.20 10.2 2J" 10.2 2j- ======== = --. : - o6 Ellipse. Ellipse. 16 I 17 1 18 ] 19 | 20 19.8 23.0 12.4 I 15.8 15.7 ! 15.1 ! 15.2 14.6 | 15.8 148.4 142.11 143.4 Calculated Volume ii Cubic Feet of Material in 1 Cu. Ft. of Beam as Mixed. 9.2 152.6 144.6, 152.5 145.8 130.0 141 148.lt 139.7 149.4 155.0 147.2 154. 154.0 146.6 153. 155.2 146.6, 158.6 168.6' 150.7 1 157.0 149.7 155.6 161.4 143.31 151.5 153.5 144.o; 152.3 142.9 10.2 150.31 10.0 153.1 144.lj 153.7 160.3 151.5; 158.5 173 1.005 147! 1.002 158! .993 165 . 992 Volume of Voids in 1 Cu. Ft. i 1 II- 004 .979 . 068 .979 . 064 1.014 .084 1 .014 .1C.) 1.025 . 054 1.018 .067 .990 . 064 Modulus of Rupture. .196 90 .175 90 .185 90 .183 »0 .129 91 039!. 108 047 .116 087|.153 077].144 010 .174 084 .181 382 258 258 2981 258 / 232 214 f 44 " 282 262 272 225, 199 216 170 170 I | 182 182 f 178 ’ 326 272 II 352 816 ; ( 1' 176 150 168 l. 263 226 1) 052 5 KM.j' 270, 242 1 256 {Sort^very 880 898 iSDrface y Verylt h 11 270 I I I slightly ) 289 ( rocky. ™ I I | Surface 1 J I rocky. | 210 i 172 1 [ 1 , 2 j 8 4 1 6 1 8 t | Comprks8ivk Strength. ; 1 I | Pounds a 3 per square inch. 8 8, is a | • £ - 0 E & 1 - a 3 226 146 2 1 505 1545 1 555 227 1 147 2 2 005 1000 1 660 228 I 146 2 1 305 1 185 1 275 235 | 141 2 1 ego 1 1 580 259 ; IBS 2 1 650 1 575 1 610 286 j 141 2 1 205 1 026 1 145 m 141 2 980 895 885 221 146 2 2 110 1 07" 1 890 146 2 2 815 1 7 < W > 2 040 288 i 143 2 1 045 940 990 229 i 143 2 1 lift 1 27U 1 840 280 j 143 2 1 790 1 170 1 480 281 143 2 1 ; (0 1 185 1 610 282 142j 2 980 1 0<10 200 141 2 1 385 1 280 1 280 209 1891 2 1 685 1 605 1 620 213 1361 2 2 000, 1 800 1 900 214 136 1 2 1 925 1 S26 1 875 205 142 2 1 480 1 120 1 800 206 1401 2 1 435 1 265 1 350 ...oi Noth. Volumes, cubic feet per 100 lbs. as mixed: Cement. 1.00; Jerome Park screenings (Crusher Run). 1.06: Jerome Park stone. 1.03; mixture Jerome Park screenings and stone varies with mixture: Cowe Bav sand (natural). 1.11; Cowe Bav „ *™ vuy: ^ em « nt • 839: Jerome Park screenings varies with size: Jerome Park stone, 2.78; mixture Jerome Park screenings and stone. 2.77; Cowe Bay sand, 2.65: Cowe Bay gravel, 2.65. Weights per cubic toot as mixed : Cement. .irllHUu K SlrccUinffS ( Lrusllpr Klin > H* lAPiime Pori* ctnna 07 IV mivfnpo Tammo PopI- e/muMiinoc and afnna uapiac nritV mivtneA- D^.. ,.n n d /nntiunl ) O O i. Dr,,, »M»nal 1 0f) r 7 'Tnnnllc ■ ■ « ’ , 7 days, 562; 28 days, 696; 8 months. 720. Col. 15 = Col. 10 X 0.08 4- Col. 12. Ci * Including cement Col. 12. Col. 10 = Col. 12 -4- (Col. 24 x 62.4). Col. 15 represents theoretical minimum, assuming 8-V of water for chemical combination. Col. 16 represents theoretical maximum, assuming voids filled with water. TA^nalyses Mixed with Year 1905. dulus of Rupture. 210 199 144 146 148 149 155 157 150 156 158 145 147 151 152 159 160 26 27 28 29 Pounds per square inch. 177 282 278 284 258 282 2 26 280 240 222 , 2231 ....I 929 744 248 231 246 253 244 199 246 248 201 241 222 209 203 198 j-919 276 251 265 269 ] j-243 J 269 {■247 212 30 Surface rocky. Excess of fine sand. Surface rocky. Very rocky. Surface very rocky. I Surface [ rocky. Surface rocky. Surface rocky. Compressive Strength of Concrete Prisms, 6 X 6 X 18 Inches. Average Age, 140 Days. 1 2 3 4 5 1 6 Compressive Strength. u % 8 p S3 Pounds CQ per square inch. © © a Age. £ r© a 0 a p ® bo Sh © © Pn O o' a 1 3 a •a § © > < 210 199 144 146 148 149 154 1 1 310 1 310 1 310 155 142 3 1 580 1 250 1 425 157 142 3 1 490 1 305 1 425 150 149 1 1 405 1 405 1 405 156 142 2 1 325 1 235 1 280 158 145 141 2 1 225 1 185 1 205 147 151 147 2 1 155 1 015 1 085 152 147 2 1 060 980 1 020 159 141 2 920 795 855 160 141 2 895 785 840 Compressive Strength TABLE 14 a._ Composition and Transverse Strength of Concrete Beams Made with Different Aggregates of Various Analyses Mixed with of Concrete Prisms, Different Percentages of Cement, Tested at Jerome Park Reservoir by the Aqueduct Commission during the Year 1905. 6 X 6 X 18 Inches. Beams, 6 X 6 X 72 Inches. Spans, 68 and 32 Inches. Giant Portland Cement. Age, 90 Days. Average Age, 140 Days. 2 8 4 5 6 7 8 9 10 11 j 12 13 14 15 j 16 17 18 19 1 20 21 22 23 24 25 1 26 27 1 28 29 so 2 | 8 : 4 1 5 1 6 Proportions by weight. Cement to total dry 1 materials, X. o a 1 o 2 1 Maximum size of stone. ( Uniformly graded above diameter, inches. Per cent, finer than diameter in Col. 7.* Curve below diameter in Col. 7.t Weight in Lbs. of Materials in 1 Ft. of Beau as Mixed. Cu. Calculated Volume in Cubic Feet of Material in 1 Cu. Ft. of Beam as Mixed. Volume of Voids in 1 Cu. Ft. Modulus of Rupture. Remarks. | Reference numbers. | Compressive Strength. Pounds per square inch. Age. No. of breaks. | Pounds per square inch. Cement. j Aggregate. | Total dry mixed. H eC Totals. j As mixed. | Minimum. | Maximum. | 1 Cement. 1 Aggregate. Total dry. Water. i 5 Aggregate. Total. 1 Age, days 1 o' i | Maximum. Minimum. 1 Average, j Maximum. Minimum. Average. | Neat. 104.5 00.0 104.5 29.6 134.1 112.9 133.1 .541 .000 .541 .475 1.016 .459 .000 .459 5.89 8 1 177 929 1-919 210 . 1:9 10 J. Park. J. Park. 2 i" 0.10 29.9 Parabola. 15.0 00.0 100.5 135 0 150.0 10.2 160! 2 1U3.0 13U.4 151. 2\ 158.9 .520 . 000 .078 .7.80 .520 . 487 .858:.16S 1.007 1.021 .480 .066 .000 .076 .480 .142 91 90 3 303 248 276 Surface rocky. 199 1 144 :::::: 1:9 10 8i" 0.10 31.4 15.4 138.2 153.6 9.3 162.9 154.8 166.1 .080! .799 .879 .149 1.028 .067 .054 .121 90 3 266 231 251 1 Excess of 146 . 1:9 10 2j" 0.20 35.8 Ellipse. 15.0 135.0 150.0 10.2 160.2 151.2 158.9 .078 . 780 .858 .163 1.021 .066 1.076 .142 90 3 282 246 265 Surface rocky. 148 1 1 810 1 810 1 810 1:9 10 2} " 0.20$ 36.9t “ * 14.9 138.7 148.6 9.9 158.5 149.8 158.0 .077 . 772 .849 .159 1.008 .065 .086 .151 91 2 278 258 269 1 Very rocky. ( Surface 149 I.' 64 1:9 10 2}” 0.20 37.8 15.0 134.5: 149.5 10.8 160.3 150.6 158.5 .078 . 777 .855 .173 1,028 .006 .079 .145 90 3 284 244 J very rocky. 155 1 142 3 1 580 1 250 1 425 1:9 10 2i" 0.20 37.8 14.9 134.01 148.9 10.0 158.9 150.1 158.1 .077' .775 .852 .160 1.012 .065 .083 .148 90 8 258 199 ] Surface 157 142 3 1 490 1 305 1 426 J { roeky. 1:9 10 2j" 0.20$ 38.6t “ t 14.8 133.3 148.1 8.9 157.0 149.3 157.6 .076! .771 .847 .143 .990 .064 .089 .153 91 2 293 246 269 150 149 1 1 406 1 405 1 405 1:9 10 2 r 0.20 39.4 14.6 181.2 145.8 '•‘.6 155.4 147.0 156.3 .075 . 757 .832 .153 .985 .064 .104 .168 90 3 282 248 Surface rocky. 156 142 2 i :i2f, 1 285 1 280 1:9 10 2r 0.20 39.4 14.5 131.0 145.5 9.8 155.3 146.7 156.0 .075 . 757 .832 .157 .989 .064 .104 .168 90 3 251 201 158 141 2 1 225 1 185 1 205 1:9 10 1" 0.10 34.2 Parabola. 14.3 129.01 143.3 11.8 155.1 144.4 155.8 .074 . 725 .7991.189 .988 .063 !.138 .201 90 286 241 264 145 1:9 10 1 " 0.10 35.0 14.4 130.1 144.5 9.9 154.4 145.7 155.2 .075 .753 .828.159 .987 .064 .108 .172 90 2 220 222 224 147 1:9 10 0.10 35.9 Ellipse. 14.3 129.0 143.3 11.9 155.2 144.4 154.5 .075 .745 1 .820 .191 1,011 .064 .116 .180 90 3 230 209 151 147 2 1 155 1 015 1 1 086 10 0.10 35.9 14.3 129.0 143.3 11.9 155.2 144.4 1 54.5 .075 . 745 .820 .191 1.011 .0641.116 .180 90 3 240 203 152 147 2 1 060 980 1 020 1:9 10 0.10 37.7 13.9 124.8 138.7 12.8 151.5 139.8 151.6 .072 .720 .792 .205 .997 .061 .147 .208 90 2 222 198 Surface rocky. 159 141 2 920 795 856 1:0 10 1 " 0.10 37.7 M.O 125.5 139.5 13.8 153.3 140.6 152.1 .0731.725 .798 .221 1.019 .062 ! .140 .202 90 2 223 202 s 21 160 141 2 89S 785 840 1:9 10 i " 0.075 36.3 13.4 120.8 134.2 12.9 147.1 135.3 148.7 .069 .698 .767' .207 .974 .059 1 .174 .233 91 2 174 172 j 104 137 2 m i 770 800 1:9 10 l" 0.075 36.3 13.5 121.2 134.7 13.0 147.7 135.8 149.0 .070 . 700 .770 .208 .978 .057 .171 .230 91 3 2-:*J 198 Surface rocky. 165 187 2 96E i 955 I 900 1:9 10 i" 0.075 38.2 13.3 120.2 183.5 12.8 146.3 134.61 148.2 1 .069 .695 .764 .205 .969 .059 1 .177 .236 90 3 273 195 208 153 181 2 89E > 85C 1 870 Note.— Volumes, cubic feet per 100 lbs. as mixed: Cement, 1.00; Jerome Park screenings (Crusher Run). 1.0G; Jerome Park stone, 1.03; mixture Jerome Park screenings and stone varies with mixture; Cowe Bay sand (natural), 1.11; Cowe Bay gravel, 0.97. Specific gravity: Cement, 3.10; Jerome Park screenings varies with size; Jerome Park stone. 2.78; mixture Jerome Park screenings and stone, 2.77; Cowe Bay sand, 2.65; Cowe Bay gravel, 2.66. weights per cubic foot as mixed: Cement, 1,00; Jerome Park screenings (Crusher Run), 94.8; Jerome Park stone, 97.0; mixture Jerome Park screenings and stone varies with mixture; Cowe Bay sand (natural), 0.90; Cowe Bay gravel, 10.27. Tensile (strength cement, pounds per square inch: Neat, 7 days, 562: 28 days, 695; 8 mouths, 720. Col. 15 = Col. 10 X 0.08 + Col. 12. Col. 16 = Col. 12 + (Col. 24 x 62.4). Col. 15 represents theoretical minimum, assuming 8 X of water for chemical combination. Col. 16 represents theoretical maximum, assuming voids filled with water. * Including cement, t See separate table for equations. t.Not an ideal ellipse. (Per cent, on 0.20 ordinate is giveD for comparison with other curves, but tangent point is at about 0.29.) Compressive Strength TABLE 14 b. —Composition and Transverse Strength of Concrete Beams Made with Different Aggregates of Various Analyses Mixed with of Concrete Prisms, Different Percentages of Cement, Tested at Jerome Park Reservoir by the Aqueduct Commission during toe Year 1905. 6 X 6 X 18 Inches. Beams, 6 X 6 X 72 Inches. Spans, 68 and 32 Inches. Giant Portland Cement. Age, 90 Days. Average Age, 140 Days. 2 3 | 4 5 6 i 7 8 I 9 10 | 11 1 | >2 i 13 j 14 ! 15 16 17 18 | 19 20 j 21 i 22 23 24 25 | 26 27 | 28 29 80 1 2 I 8 1 4 1 5 1 6 Proportions by weight. 1 Cement to total dry materials ,. Kind of stone. 1 © p a I Maximum size of stone. 1 Uniformly graded i above diameter, inches. 1 Per cent, finer than diameter in Col 7.* Curve below diameter in Col. 7.+ Weight in Lbs. of Materials in 1 Ft. of Beam as Mixed. Cu. Calculated Volume in Cubic Feet of Material in 1 Cu. Ft. of Beam as Mixed. Volume of Voids in 1 Cu. Ft. Modul US OF Rupture. Remarks. j Reference numbers, j Compressive Strength. Age, days. ! No. of breaks. Pounds per square 1 inch. 1 | No. of breaks. Pounds per square Inch. l 6 bl 1 4 | Total dry mix. cij £ Totals. 1 3 | f eg s 1 O ! Aggregate. | Total dry. j | Water. 1 £ j Cement. 1 | Aggregate. | 3 & | Maximum, j | Minimum. j Average. J Maximum. | Minimum. | Average. 1:9 10 J. Park. J. Park. J- 0.075 38.2 Ellipse. 13.7 123.0 136.7 13.2 149.9 137.8 150.4 .071 .710 .781 .212 .993 .060 .159 .219 90 3 212 197 208 154 142 2 890 870 880 1:9 10 i” 0.075 39.8 | 13.3 119.5 132.8 14.6 147.4 133.9 147.8 1.069 .690 . .234 .993 .058 . 183 .241 90 8 25S 218 195 185 2 1 005 1 015 1:9 10 “ “ 1” 0.075 39.8 , 13.3 119.5 132.8 14.4 147.2 133.9 147.8 .069 .690 .759 .231 .990 .058 .183 .241 90 2 305 205 J227 190 135 2 1 015 910 905 1:2* :6 s 10 - 2J” ! 14.6 131.3 145.9 9.6 155.5 147.1 156.8 .076 .758 .884 .154 .988 .064 . 1 < >2 . 166 90 8 207 $76 | 168 135 2 786 800 1:2* :6* 10 “ 2)" | 14.4 130.0 144.4 10.7 155.1 145.6 155.3 .075 .751 .m .172 .998 .064 .110 .174 90 3 201 161* 135 2 990 901) 910 1:8:6 10 2j- :::::::::::: ! 14.2 127.7 141.9 8.9 150.8 143.0 153.7 .074 .737 .811 .143 .954 .063 .126 .189 91 3 311 225 166 137 2 1 210 i 000 1 185 1:8:6 10 2J- . 1 14.2 127.8 142.0 9.0 151.0 143.1 153.2 .074 .738 .812 .144 .950 .063 .117 .iso 90 1 224 224 ] ^ 107 136 2 1 065 965 1 015 1:9* :6* 10 .. 1 •• 14.2 127.8 140.0 14.6 156.6 141.1 151.5 .074 .736 .810 .232 1.042 .063 .121 .184 90 2 168 138 | 172 148 2 875 855 865 1:8»:6* 10 1 " 14.1 127.2 141.3 11.2 152.5 142.4 153.3 .073 .735 .808 1180 .988 .062 130 192 91 2 187 ill 1 15 173 145 i 2 905 711) 855 1:8:6 10 j 14.0 126.3 140.3 14.0 154.8 141.4 152.6 .072 .730 .802 .226 1.028 .061 .137 r> 90 3 1 Mi 170 188 2 920 800 890 1:3:6 10 1 " ! 13.8 125.4 139.2 12.5 151.7 140.3 151.9 .071 .725 .796 .200 .996 .060 .144 .204 90 3 199 160 ( 176 171 138 2 845 850 1:8* :5? 10 1 " 13.7 123.4 137.1 13.6 150.7 138.2 150.6 .071 .713 .784 .218 1.002 .060 .156 .216 91 3 192 166 j 174 145 2 890 795 845 l;8*:5* 10 13.8 124.0 137.8 13.2 151.0 138.9 151.2 .071 .715 .786 .211 .997 .060 .154 .214 91 3 227 171 175 145 2 980 875 930 1:2*:6* 10 4" i 13.5 121.9 135.4 14.0 149.4 136.5 149.6 .070 .703 .773 .224 .997 .059 1.168 M 90 3 211 191 204 Siightly rocky. 192 187 2 i 106 875 990 1:8:6 10 J" 1 18.5 121.6 1 135.1 12.4 147.5 136.2 149.3 .070 .702 .772 .199 .971 .059 .169 .228 88 2 13.5 100 176 143 2 090 070 680 1:8:6 10 r 13.5 121.0 | 134.5 13.5 148.0 135.6 148.9 .070 . 700 .770 .216 .986 .059 .171 .230 88 8 175 169 f 15 177 143 2 885 740 790 1:8*:5* 10 4" "ZZ 18.3 119.3 i 132.6 14.4 147.0 133.7 1 147.7 .069 .689 .758 .231 .989 .059 .183 .242 90 8 156 145 152 194 135 2 830 815 825 1:9 Uniform stone. 10 21" ,'20" 87.8 Ellipse. 14.9 134.0 i 148.9 9.6 158.5 150.1 158.1 .077 .775 .852 .154 1.006 .065 .083 .148 90 2 269 244 257 Slightly rocky. 187 141 2 i 086 1 105 1 350 i i;0 “ »» 10 1 " .10 87.7 | 13.9 195.fi 1 139.5 12.8 152.3 140.6 152.2 .072 .725 .797 .205 1.002 .061 .142 .203 90 2 230 229 188 141 '.*70 915 950 l 1 i;0 u »» 10 r .075 39.8 l “ 1 134 121.C l 134.4 14.3 148.7 135.5 148.8 1 .069 .700 .769 .229 .998 .069 .172 .231 91 2 181 179 180 198 145 2 900 875 890 Note —Volumes; cubic feet per 100 lbs. us mixed: Cement, 1.00; Jerome Park screenings (Crusher Run). 1.06; Jerome Park stone, 1.03; mixture Jerome Park screenings and stone varies with mixture; Cowe Bay sand (natural), 1.11; Cowe Bay gravel,0.97. Specific gravity: Cement, 3.10; Jerome Park screenings varies with size; Jerome Park stone, 2.78; mixture Jerome Park screenings and stone, 2.77; Cowe Bay sand, 2.65; Cowe Bay gravel, 2.65. Weights per cubic foot as mixed: l enient, 1.00: Jerome Park screenings t, Crusher Ruu), W.8; Jerome Park stone, 97.0; mixture Jerome Park screenings and stone varies with mixture; Cowe Bay sand (natural), 0.90; Cowe Bay gravel, 10.27. Tensile strength cement, pounds per square iueh: Neat, T days, 5(53; S3 days, 6115; 8 months, 720. * T * , Tt 10 x + Col. 12. Col. 16 = Col. 12 4- (Col. 24 x 62.4). Col. 15 represents theoretical minimum, assuming 8% of water for chemical combination. Col. 16 represents theoretical maximum, assuming voids filled with water. * Including cement. $ Omitted in average. Column 23 is the difference between columns 24 and 22. In determining the average strength of each mixture, column 29, all of the breaks of the two beams, which are in duplicate, are aver¬ aged, that is, if one of a pair of beams has only two breaks, while the other has three, the five breaking strengths are added together, and divided by five, instead of assuming that the two breaks give the I average of one beam and the three breaks the average of the other beam. The method followed is considered more accurate than the other, because a beam which can be broken only once or twice is ■ possibly imperfect, and therefore this beam should not have quite so 1 large a place in the average as the beam with three breaks. The compressive strength of the concrete prisms made by capping ft two pieces of each beam are given at the right-hand of each sheet of t Table 14. These prisms were broken by Prof. Frederick L. Pryor at I the Stevens Institute of Technology, Hoboken, N. J., and although j the heads of the machine were fixed, the ends of the specimen were sufficiently parallel to give good results, nearly all the pieces break- I ing in a manner normal to long prisms. The breaks, as is usual i with such specimens, were more longitudinal than is the case with [cubes, where two pyramids are generally formed with their bases ■ against the heads of the machine. ■ Comparative Density and Strength of Beams Made with Dif¬ ferent Brands of Cement. Table 15 gives the results of tests of neat cement beams, and of concrete, made with a number of brands of cement selected as repre¬ sentative of different sections of the United States. The samples of these cements were purchased a year before testing, and on this account the relative results may be questioned somewhat. How¬ ever, as the cements were packed in barrels and stored in the labora¬ tory at Jerome Park, it is probable that the results are equal to that which would be obtained from fresher samples. The results indicate that the Giant cement used in the 1905 experiments is equal to any of the other brands, and has normal density. A comparison of the density of the neat beams of Giant cement in Table 15 with the density of the neat beams of the same brand tested in 1904 shows a much greater density for the 1905 cement. The average density of 74 the latter in the neat beams is about 0.53, while the density is about 0.49 in the 1904 tests. The effect of increased density of any cement upon its real value has yet to be determined, but a comparison of the two series of tests shows in this case that the cement giving paste of the less density produced the poorer concrete. Permeability Tests. The results of the permeability tests have emphasized the fact of how little is known of the action of concrete in resisting the flow of water. Examination of the tests, which are given in full, Table 16, indicates in general that, using different proportions and different sizes of the same class of materials, the laws of water-tightness are somewhat similar to those of strength; If the percentage of cement be the same the specimens having the greatest density are usually most water-tight, and in the specimens having similar density but different percentages of cement, the water-tightness increases with the percentages of cement. The ratios, however, are very different from the ratios of either density or strength, a slight difference in the composition producing a tremendous effect upon the water¬ tightness. Different kinds of aggregate also, for reasons not yet ex¬ plained, produce very different results in water-tightness. The results of the tests are discussed in detail on pages 78 to 83. An important result of the permeability tests has been the evolution of an apparatus, described in succeeding paragraphs, by means of which almost any character, shape or thickness of specimen may be sub¬ mitted to water pressure. The advantages of the apparatus lie in the exposure of the entire top surface of the specimen to the water pressure, the coating of the sides of the specimen with neat cement so as to confine the flow, and the discharge of the water through the bottom of the specimen. Method of Making Permeability Tests. The apparatus used in making permeability tests and the prepara¬ tion of the specimen is indicated in Figs. 21 and 22. The method of preparing the specimen which was found satisfactory after a number of trials is illustrated in Fig. 21. A piece of a concrete beam, about 6 in. square and 17 in. long, obtained by breaking the 75 ‘beam in the regular manner in the testing machine, is scored over the surface of its four sides by a hammer and cold chisel, so as to give rough and uneven surfaces and afford a better adhesion to the cement coating. The specimen is then immersed in a can of water and soaked for 24 hours. It is then taken from the water, placed in a wooden mold so constructed that the specimen can set upright in it, and leave a space on the four sides, and a mound of sand is formed on the top and held in place by a wood strip, first thoroughly soaked with water. The surface of the sand is covered with a piece Fig. 21.— Method of Preparing Specimen for Permeability Test. of tissue paper to prevent the cement flowing into it. 1 000 grammes of fine sand which passes a No. 30 sieve and is caught on a No. 40 are required. On top of the sand is placed a 4-in. iron flange with a i-in. nipple 4| in. long screwed into it and made up in a tight joint so that there can be no leakage along the thread. Neat cement mixed to a paste about as stiff as can be conveni¬ ently handled and compacted, is poured into the mold around the specimen and over the sand, thus forming a dome above it. About 76 83 lb. of cement and 21 lb. of water are used for the casing*. When pouring the cement four pieces of |-in. galvanized ribbon wire or some other form of reinforcing metal are placed in pairs, two at right angles to the other two, over the top of the specimen and down along the side, as shown in Fig. 21, page 75, to reinforce the cement and prevent the blowing off of the dome, which occurred in one of the experimental tests. As the cement is compacted, water rises- to the surface and bubbles of air come up through the cement and water. The casing is not considered complete until this bubbling ceases. The surface of the specimen, after it is complete, is covered with miost sand, and this is kept wet by a very slow flow from a water cooler. At the end of 24 hours, the neat cement has thor¬ oughly set, and the specimen can be removed from the mold and buried in damp sand until ready to test. Apparatus for Testing Permeability .—The apparatus for testing, together with the specimen itself, is shown in Fig. 22. The speci¬ men is placed upon a tin funnel, set in a wooden frame which rests upon any suitable foundation. The pipe projecting through the top of the specimen is connected by a union coupling with the bottom of the air pressure tank, which is nearly filled with water before be¬ ginning the test. The pressure is raised by the hand pressure air pump shown in the photograph, and maintained at any desired* pres¬ sure for any required period. The time is read by a stop-watch read¬ ing for convenience to hundredths of minutes, instead of to sec¬ onds, which is started at the beginning of the tests, and allowed to run as long as the experiment is continued. The water is caught in a bottle placed below the funnel under the specimen, and weighed at intervals by substituting a second bottle. Recording the Data .—The method of recording the data of the tests is shown in a typical form. Table 17. Convenient periods of flow for most of the specimens were found to be five minute inter¬ vals, this giving sufficient time to weigh the bottle, and being fre¬ quent enough to obtain the average rates of flow. It was found that about five minutes were required for the flow at any given pressure to become constant, and therefore the flow during the first five min¬ utes at any given pressure is not included in figuring the average flow. Fig. 22.—Apparatus for Testing Permeability of Concrete. c* Rate op Flow of Water,, Grams per Min. •m -bs aad -sq[ 08 1 Y | SS‘8‘83 : * Including cement. this and the Jerome Park mateilalf 61 * 0611 ^^ 6 C6ment ^ ^ ncrease( ^ ,a ea( ^h case to balance the difference in specific gravity between 8 "UT bs aad -sqi 09 IV 50 05 CD {- DriOOO®MHiOT)'t.t'OOli30r(ONOON(X)SJOl»MO' CO(N •uitd ‘sjRaddR aaqi-BAv aunx »o ^ io : • • H« . . ID • w rJSS eocCli '*if55; ec ' l>e,:i * i5i >' :,, ®* loeo0| 30oot-»05005 • • <33C? 0 05 T^OJOJO • CO s^Rp ‘aSy ONIS05MlODf'r«ffiONlO®Mt-Ot'THHNN®NNI»i- iO Calculated Volume in Cubic Feet op Material in One Cubic Foot op Beam as Made. •joa iro irjoj, 1.028 1.011 .978 .988 .956 1.042 1.028 .997 .971 1.006 1.001 1.041 1.031 .988 .983 1.007 1.005 1.025 1.014 1.014 1.025 .975 1,010 .976 .968 1.017 1.006 spiOA irjoj, i2£SS£9S°3- lt '* GOGlOOSO ?! l ^tf3 T J , oa>C5»£>coJ>iooof-eooQ ^|®g350Q00505g4N^O^iC^t>^5D04OC0(Ma0i0Q0g«0Tji 5 •jajR^ CS •^ap irjox 1 ?5 S 2 S! S'* 0 NMCQNT-iWlOinoOOJt-iiONMiflOOMNJJ }S £? £* goI? XS S £~ J> »o o? m m ot co oo t> to co i> r-< m e* i> to o ao ao t» ao go oo go i-1-oo go ao oo oo ac ao ao oo oo oo ao oo oo ao t—oo ac 3 •ajRSaaSSy i^i2SSS9SSS2237?* n 3r l,R ‘ 1 ^«30i- | irt^*nxcoooj»oeoi-ir' i>!^Ojocococooo{>|^-»0'—it^iOxmococoosT^o-Tfiojoeo o •juauiao I Sj22S?2S!5?291' < 2 i>c> ®®® i >' ! t‘oo5oioxKio®N C5 '& "OK 'I°0 ui ‘uiRip MDpq aAariG Ellipse. . Ellipse. Ellipse, Ellipse. Ellipse. Ellipse. I. 00 *1 'ON TO ni •uiuip nnqj jauy ‘judo ja» • gC? lO z m 2 | H oi CO . S3 + Ss S o o£ a cu Ift- ® ^ S3 ® o 5z 3 S I .gift gft gft g ft - ft ft *-5 1-5 : [S >*>43 43 +3 +3 ft :goqftft ft ft ft q/ ft ft ft ft »-5 •sjaqxunu aouajajaH Based on a unit of 100 lb. cement per cu. ft. Fig. 23. Apparatus for Volumetric Tests of Density of Cement. I 83 further experiments. The specimen with Jerome Park screenings has substantially the same permeability as the specimens Nos. 8 and 9, which contain very fine sand, and yet the strength of whose mortars, as would naturally be expected and as indicated in tests 2 and 3, is extremely low. The concrete with Cowe Bay sand, specimen 7, is less permeable than specimen 6 with Jerome Park screenings, thus tending to confirm the results of the regular tests in Table 16. For some unexplained reason the most permeable specimen is that con¬ taining a combination in equal parts of Jerome Park screenings and one of the fine sands, while the least permeable specimen is a combination of Jerome Park screenings and another sand which is even slightly finer than the former. Volumetric Tests of Density of Neat Cement and Mortar. During the progress of the experiments at Jerome Park frequent tests were made of the density of the cement which was being used and also the densities of various mortars. For this purpose the cylinder apparatus described for the density experiment upon con¬ crete was used, and also tests were made upon a smaller scale with a 200 c. c. graduate for the mold. The apparatus used for these tests on the small scale and the Jackson specific gravity flask, also con¬ stantly employed, are showm in Fig. 23. Effect of Different Percentages of Water Upon Volume and Density of Portland Cement Mortars and Concretes. Table 21 and Fig. 24 illustrate the variation in volume of paste produced from the same weight of the same cement with differ¬ ent percentages of water. It also shows the difference in volume between the fresh paste as mixed and the same paste after standing for a period of about 3 hours and compacting by occasional shaking and tapping of the mold. The tests were made in the graduates as described above. By adopting a standard method in all the tests and maintaining a uniform consistency, good results were obtained. When sand is mixed with the cement, there is much less variation in the volumes of the mortars due to different percentages of water 84 than in neat pastes, but the variation is generally very appreciable, although less with coarse than fine sands. TABLE 21. _Effect of Different Percentages of Cement upon the Volume of Neat Portland Cement Paste. 1 1 2 f 3 5 6 ! 7 8 9 10 11 12 Experiment number. Nominal mix. Kind of cement. •2 VOLUME OF g I Paste. 1 • i Absolute Volume. Weight of cement. Water used in mixin Fresh. Final Compacted. Per cent, of i mixing Fresh. Final compacted. Water. Cement (density). Water. Cement. (density). Grams. Grams. c. c. c. c. w . c. w . c. 463 Neat. Giant. 300 60 20 165.8 163.8 .362 .584 .354 .590 464 300 69 23 ! 168.7 167.7 .391 .546 .404 .574 465 u 300 78 26 | 174.0 172.3 .438 .543 .441 .560 467 u 300 96 32 ! 183.2 179.3 .504 .508 .508 .534 447 tl 300 150 50 244.0 221.0 .613 .395 .571 . 435 449 300 300 100 390.6 214.3 .772 .248 .569 .436 The resulting volumes of different concretes of the same materials and proportions but with varying percentages of water are nearly constant provided sufficient water is added to permit thorough com¬ pacting. When an excess of water is used, it is nearly all expelled as the solid materials settle by gravity. There is, however, a slight variation even in concrete. A dry mixture, provided sufficient water is added to rise to the top on hard ramming, can be compacted some¬ what more than a concrete of medium consistency, while a very wet concrete is apt to occupy a very slightly larger volume and hence be less compact than a medium concrete. In most cases the differ¬ ence between the medium and very wet is only noticeable by ex¬ tremely accurate measurements. Tensile Tests of Neat Giant Portland Cement Used in Concrete Beams, 1905. The following series of tests, Table 22, was made upon the cement which was used in the beam specimens. 85 TABLE 22. —Tensile Strength of Neat Giant Portland Cement Used in Concrete Beams, 1905. First Series. Date tested. Age. ! Average tensile strength, lb. per sq. in. March 7. 1 day 252 13. 7 days 568 April 3. 4 weeks 733 May 1. 8 “ 760 29. 12 “ 733 June 26. 16 “ 855 July 24. 20 “ 774 August 21. 24 “ 746 Second Series. Date Tested. Age. Average tensile strength, lb. per sq. in. March 23.... 1 day 218 29 ... 7 days 556 April 19.... 4 weeks. 658 May 17.... 8 “ 686 June 14.... 12 k ‘ 708 July 12.... 16 “ 656 August 9.... 20 “ 738 September 6.. : 24 “ 754 Note.—E ach value is an average of 5 briquettes. Mechanical Analysis of Average Jerome Park and Cowe Bay Materials Used in Beam Tests. The figures in Tables 23 and 24 show the average mechanical analyses of the Jerome Park crushed stone and screenings and the Cowe Bay gravel and sand which were used in the mixtures made up by natural proportions. Each of these analyses is an average of a number of samples. TABLE 23.— Average Analysis of Jerome Park Material. Sieve numbers. I Diameter open- j ing in sieves. 2J4" stone, per cent, passing. 2.25 2.25 100.0 1.50 1.50 78.4 1.00 1.00 41.8 0.75 .75 29.9 0.60 .60 18.4 0.45 .48 11.4 0.35 .36 6.8 0.27 .29 3.6 0.20 .20 1.9 0.15 .16 1.3 0.10 .10 1.0 No. 10 .075 .67 15 .046 .61 20 .034 .58 30 .020 .55 40 .016 .48 50 .014 .47 74 .0071 .32 100 .0058 | .27 150 .0036 .19 200 .0027 1 .01 1" stone, per cent, passing. J4” stone, per cent, passing. Screenings, per cent, passing. 100.0 71.5 44.2 100.0 27.3 IOii.O 99.7 16.3 59.6 99.1 8.6 31.6 96.9 4.5 16.7 92.8 3.1 11.4 87.2 2.4 8.7 80.7 1.6 5.8 76.7 1.4 5.3 70.8 1.39 5.0 66.3 1.32 4.8 55.8 1.15 4.2 47.8 1.12 4.1 43.2 0.76 2.8 21 0 0.64 2.3 15.4 0.45 1.6 6.7 0.24 0.8 2.3 86 TABLE 24. —Avebage Analysis of Cowe Bay Sand and Stone. Sieve numbers. Diameter opening in sieves. 2J4” stone, per cent, passing. 2.25 2.25 100.0 1.50 1.50 94.2 1.00 1.00 81.8 0.75 .75 67.3 0.60 .60 44.8 0.45 .48 26.4 0.35 .36 14.0 0.27 .29 8.6 0.20 .20 5.05 0.15 .16 3.40 0.10 .10 1.76 No. 10 .075 1.57 15 .046 1.38 20 .034 1.28 30 .020 1.09 40 .016 0.98 50 .014 0.80 74 .0071 0.47 100 .0058 0.43 150 .0036 0.26 200 .0027 0.01 1” stone, stone. Sand, per cent. per cent. I per cent. passing. passing. | passing. 100.0 100.0 100.00 82.27 54.77 32.28 100.0 100.0 97.9 ioo.oo 96.2 17.12 53.03 95.0 10.51 32.57 93.7 6.17 19.13 W.8 4.16 12.88 88.9 2.15 6.67 1 85.1 1 .92 5.95 78.9 1.69 5.23 1 68.3 1.56 4.85 i 62.4 1.33 4.13 49.9 1.20 3.71 37.3 .98 3.03 1 31.8 .57 53 1.78 1.63 7.1 4.6 .32 .98 2.3 .01 .04 1.8 Elasticity Tests. Figs. 25 to 29 show the results of tests of elasticity of several of the capped prisms. These tests were made at the Stevens Institute, Hoboken, N. J. Headings of compressive deformations were taken at increments corresponding to unit increments of 100 lb., and m most of the tests the load was released at 500 and 1000 lb. and the permanent set recorded. In a few cases shown by zigzag lines the effect of repeated stress was also measured. The results are tabu¬ lated in Table 25. Notice in the diagram that nearly all the curves run up sharply from 100 lb., the first load applied, indicating that the modulus between 100 and 200 lb. is not greater than at the next higher load¬ ings. After deducting the set, the deformation (shown by curve labelled “Elastic”) in the majority of the specimens at first in¬ creases uniformly, the curve being in these cases a straight line m its lower portion. The point of tangency with the decided curve above is sometimes considered the elastic limit of the concrete. HJL9N31 UNO Nl N0ISS3ydW0D APPLIED LOADS POUNDS PER SQUARE INCH Fig. 27,-Total and Elastic Compression of 6-In. by 6-In. by 18-In. Concrete Prisms Made of Giant Portland Cement and Cowe Bay Sand and Gravel. Average Age, about 140 Days, ( 1 ) ©' © 8 © tM © tf 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 92 TABLE 25.— Modulus of Elasticity of Concrete Prisms. Tests for deformation. Jerome Park Reservoir. 1905. (2) (3) (4) (5) (6) (7) (8) (9) (10) (ID 1 (12) (13) Beam number. (See Table 14.) Proportions by weight. Cement to total dry material. Kind of stone. Kind of sand. Maximum size of stone. Uniformly graded above diam. Per cent, finer than diam. in Col. 8. Density. Ultimate strength. Modulus of elasticity. Modulus taken between limits. No. % In. In. % Lb. per Lb. p°r Lb. per sq. in. sq. in. sq. in. 155A 1:9 10 ! Jerome Park. Jerome Park. j-2.25 0.20 37.8 .855 1 425 2 143 000 100-1 000 151A 1:9 10 1.00 0.10 35.9 .820 1085 1 702 000 100- 500 165A 1:9 10 ** 0.50 0.075 36.3 .770 960 1 436 000 100- 500 192A 1:2 5 :6 5 10 0.50 Natural Mix .773 875 1 274 000 300— 500 167 1:3:6 10 2.25 .812 965 2 425 000 100- 500 170A 1:3:6 10 u 64 1.00 .802 860 1 798 000 100— 500 176A 1:3:6 10 0.50 “ .772 670 878 000 200- 500 193A 1:11 5 8 2.25 0.20 37.8 .831 940 1 750 000 100- 500 155A * 1:9 10 2.25 0.20 37.8 .855 1 425 2143 000 100—1 000 179 A 1:7 12 y 2 2.25 0.20 37.8 .852 1 710 2 250 000 200— 800 183A 1:5 6T 15 u Cowe Bay. 2.25 0.20 37.8 .845 1 920 4 650 000 100— 500 234D 1:8* 3 10.6 j Cowe Bay. j-2.25 0.20 37.8 .860 1 800 2 867 000 100-1 500 233A 1:10 76 8.5 2.25 0.20 32.1 .871 940 2 252 000 100— 500 217A 1:8 43 10.6 2.25 0.20 32.1 .855 1880 3 920 000 100— 500 224A 1:6 s1 13J4 ** 2.25 0.20 32.1 .865 2 110 3 675 000 100—1 ooo 225A 1:5 30 15.9 2.25 0.20 32.1 .867 2 315 4 273 000 500-1 000 221A 1:2 81 :5 62 10.6 2.25 Natural Mix .826 1 455 3 260 000 100- 600 227D 1:2 81 :5 62 10.6 1.00 “ .832 2 065 3 100 000 100-1 500 213A 1:8 50 10.2 -j Jerome Park. Cowe Bay. [-2.25 0.20 I 33.8 .873 2 000 3 480 000 100- :>00 218A 1:6 87 12M 2.25 0.20 1 33.8 .866 2150 3 830 000 100— 800 216A 1:5 54 15.3 2.25 0.20 1 33.8 .852 2 440 3 508 000 200—1 600 205A 1:2 92 :5 88 10.2 2.25 Natural Mix .831 1 480 3 037 000 300- 500 203A 1:2 92 :5 88 10.2 1.00 .820 1 555 2 550 000 100- 700 207A 1:2 92 :5 88 10.2 46 0.50 .783 1 145 2 206 000 100— * Beam 155 repeated here to facilitate comparison. The table shows several interesting comparisons: 1. —The modulus increases with the maximum size of stone. Compare respectively specimens 1, 2, 3, 5, 6, 7, 17, 18, 22, 23, 24. 2. —The modulus increases with the percentage of cement. Com¬ pare respectively 8, 9, 10, 11, 13, 14, 15, 16, 19, 20, 21. There are one or two exceptions to this rule, but the general trend is unques¬ tionable. 3. —In general the modulus of the Cowe Bay gravel and sand is higher than the modulus of similar specimens mixed with Jerome Park sand and screenings. 4. —The modulus of specimens with mixture of Jerome Park stone and Cowe Bay sand is in general slightly lower than similar specimens of straight Cowe Bay material. (T5507.)