LIBRARY. 4A 8: M COLLEGE. CAMPUS. R60-113'6-6m EXAS AGRICULTURAL EXPERIMENT STATION _A. B. CONNER, DIRECTOR COLLEGE STATION, BRAZOS COUNTY, TEXAS ULLETIN NO. 534‘ NOVEMBER, 1936 DIVISION OF CHEMISTRY FACTORS WHICH MAY AFFECT THE HARDNESS 0F COTTONSEED CAKE Ll B RARY Agricultural & Mechanical College of Texas If tatlun, Texas. _ y ' AGRICULTURAL AND MECHANICAL COLLEGE OF TEXAS T. O. WALTON, President Specimens of the same slab of cottonseed cake varied from 2600 to 3500 pounds in crushing strength. No constant regions of hardness or softness were found. No definite relations were found between the chemical com- position and the hardness of the cake. A rapid change in moisture content of the cake resulted in a lowering of the crushing strength, contrary to expectations. Meats with high moisture content produced a soft cake. Storage for approximately two years had no appreciable eifect upon the hardness of the cake. The direction of the testing load and the degree of smooth- ness of the test specimen had a considerable effect upon the apparent hardness of the cake. Modified Brinell, schleroscope, abrasion, and impact tests were found unsuitable for testing the hardness of the cake. The crushing test, as already stated in a previous publication, was found to be satisfactory. The rate of application of the crushing load had no apparent effect upon the results of the crushing tests. CONTENTS Introduction 5 Some factors afiecting hardness 5 Relation of position in slab of cake 5 Relation of chemical composition 7 Relation to the change in moisture content of the cake ____________ __ 9 Relation to moisture content before pressing ______________________________ __ 11 Relation to press box temperature 11 Effect of storage of the cake 12 Relation to specific gravity of cake 12 Some factors affecting the test 14 The direction of the testing load 14 Smoothness or roughness of top of specimen __________________________ _-.__ 15 The effect of the rate of application of load 16 The relation of the unit deformation of 1" cores under a load of 1000 lbs. to their crushing strength 18 Some methods tried in addition to the crushing and ball tests-__-_-_____- 19 The modified Brinell test 20 The modified schleroscope test 21 The impact test 23 The abrasion test 23 References 27 BULLETIN NO. 534 NOVEMBER, 1936 FACTORS WHICH MAY AFFECT THE HARDNESS OF COTTONSEED CAKE By G. S. Fraps, Chief, and C. D. Marrs, Assistant Chemist, Division of Chenzistry. The hardness of cottonseed cake as related to its suitability for feeding has been discussed in a previous publication (Bulletin N0. 523). As there pointed out, some lots of cracked cottonseed cake are so hard that animals do not eat them readily, and it is desirable to have a method of testing cake to ascertain the hardness in this respect. The force required to crush cottonseed cake between flat surfaces was adopted as the best method for securing the desired information regarding hardness. Tenta- tively, cracked cake with a crushing strength of less than 400 pounds was classed as soft for cows; cake with a crushing strength of 400 to 1500 pounds was classed as medium hard; cake with a crushing strength of 1501 to 2500 pounds was classed as hard, and that with a crushing strength of over 2500 pounds was classed as very hard. This classifica- tion refers to the cracked or cut cake as sold for feeding purposes. In connection with the work referred to above, a. study was made of some of the various factors which might affect the hardness of cottonseed cake. Other methods besides the crushing test referred to were also studied. It is the object of the present bulletin to present the results secured in this study. SOME FACTORS AFFECTING HARDNESS Some of the factors relating to the location of the sample taken from the slab, storage, etc., were studied as described below. Relation of Position in Slab of Cake to Hardness It is generally believed by oil mill men that the outside ends of a slab of cottonseed cake are softer than the inside portions. Since it seemed possible that the hardness varies in different portions of the same slab of cake, tests were madeto ascertain the variation in the crushing strength of different portions of the same slab of cake. The slabs of cottonseed cake were mapped, so that the position of the specimen tested could be recorded, and samples from regularly spaced positions in the cake were tested. Two of the slabs were tested by crushing specimens one inch square, and the third was tested by crushing and ball tests on alternate one-inch cores. The methods used have already been described. There was a wide variation in the hardness of the samples taken from different parts of the same slab of cake. The crushing strength of the one inch squares varied from 2600 lbs. to 3800 lbs. in one slab and from 2700 lbs. to 3700 lbs. in the other. The crushing strength of the cores 6 BULLETIN NO. 534, TEXAS AGRICULTURAL EXPERIMENT STATION varied from 1140 lbs. to 2630 lbs., and the ball test required from 245 lbs. to 455 lbs. to break the core. Fig. 1. Lines of equal crushing strength 0f one-inch squares cut from slab of cottonseed cake N0. 1. The results of the tests were plotted upon a map of the slab, and lines, similar to contour lines, were drawn through points of equal average crushing strength. These lines are shown for two slabs in Figures 1 and 2. It is clear from these that there is no definite relation between Fig‘. 2. Lines of equal crushing strength of one-inch squares cut from slab of cottonseed cake No. 2. the regions of average strength‘ in the two slabs. It is also evident that there are no special regions of hardness or softness. FACTORS WHICH MAY AFFECT HARDNESS OF COTTONSEED CAKE 7 The Relation of Chemical Composition to the Hardness Many oil mill operators think that the oil content of cottonseed cake Two theories have been advanced is related to the hardness of the cake. as to the cause of this effect: inhibitor of excessive cementing of the particles of meal in the cake; and second, the oil content of the cake is a measure of the pressure attained in the presses used for extracting the oil, so that the cake is harder when the pressure is greater. first, the oil acts as a lubricant and as an TABLE 1. Comparison of crushing test with chemical composition Average Nitrogen Number Range'of Labora- Crushing Protein Fats Crude Free Water Ash of Crushing tory Test 0 0 Fiber Extract 0 o Cores Test Number (1bs.) 0 % Averaged (lbs.) 41038 710.0 41.56 6.14 11.55 28.64 6.60 5.51 7 180 41042 851.5 45.62 6.34 10.61 25.64 6.50 5.29 .. 41049 928.8 45 .28 6.45 9.67 25.73 7.28 5.59 8 90 41026 972.2 42.40 6.33 12.46 26.11 7.56 5. 14 13 200 41349 1020.0 43.13 5.77 11.32 28.08 6.58 5.12 8 200 41353 1114.5 42.44 5.64 12.20 27.94 6.70 5.08 11 200 40998 1126.0 44.36 7.30 10.65 24.96 7.44 5.29 15 280 41043 1128.3 45.38 5.94 10.36 26.55 6.47 5.30 .. 41053 1162.3 45.71 6.86 9.91 24.78 7.09 5.65 11 200 41050 1166.5 45.50 6.02 9.68 25.56 7.81 5.43 10 200 41035 1210.7 43.58 6.40 9.62 27.34 7.15 5.91 15 200 41039 1225.0 41.14 6.00 11.76 27.87 7.66 5.57 10 200 41022 1241.4 41.52 5.87 12.65 27.11 7.16 5.69 11 200 41020 1292.5 41.33 5.73 12.37 27.76 7.03 5.78 12 195 40996 1294.0 44.13 6.04 10.03 27.15 7.24 5.41 .. 41054 1323.3 46.22 6.93 9.82 24.44 6.83 5.76 9 100 41033 1325.5 43.79 6.80 9.52 27.18 6.63 6.08 11 200 41352 1335.0 42.45 5.63 12.00 28.39 6.55 4.98 10 200 40994 1343.8 43.69 6.03 11.72 25.86 7.24 5.46 12 125 41346 1356.7 42.45 5.96 11.57 28.42 6.47 5.13 9 190 41000 1404.0 43.92 6.85 10.01 26.73 7.31 5.18 21 370 41009 1423.0 44.98 7.49 10.02 25.23 6.83 5.45 10 200 41024 1424.6 44.00 6.23 11.82 25.00 7.73 5.22 14 140 41029 1472.8 44.35 6.51 9.86 26.50 6.73 6.05 9 200 41017 1511.1 45.15 6.52 10.27 24.89 7.81 5.36 9 160 41005 1543.3 45 .03 6.52 10.65 25 .34 7.07 5.39 9 230 41002 1557.5 44.86 6.78 11.15 25.23 6.82 5.16 10 300 41057 1562.7 40.72 5.73 13.07 27.74 7.08 5.66 11 200 40992 1618.3 44.87 6.24 9.97 26.92 6.59 5.41 9 180 41010 1638.8 44.63 6.71 9.74 26.44 7.15 5.33 12 175 41030 1667.8 44.43 6.05 9.64 26.57 7.22 6.09 9 150 41013 1751.1 44.28 6.01 10.89 25.81 7.84 5.17 9 210 41018 1795.0 45.16 6.58 9.85 25.29 7.78 5.34 9 145 41006 1798.7 44.41 6.45 10.50 26.22 6.98 5.44 8 200 41058 1808.0 40.65 5.50 13.19 28.03 6.89 5.74 ~ 1O 200 40990 1838.6 44.73 6.17 8.95 27.77 6.68 5.70 11 535 41014 1966.6 44.88 6.02 11.04 25.13 7.76 5.17 9 200 41046 2746.6 45.18 5.21 11.98 25.77 6.87 4.99 9 580 Other oil mill operators have ness of the cake. Twenty-five cores Were cut in Experiments held that the protein content, the crude fiber content, and the moisture content of the cake are related to hard- were made to ascertain these relations. a line from each slab of cottonseed cake to be tested. These cores were crushed in the SouthWark-Emery testing 8 BULLETIN NO. 534, TEXAS AGRICULTURAL EXPERIMENT STATION machine and the crushing strength of each core was recorded. The cores Were arranged in the order 0f the magnitude of their crushing strength, Crushing strength and chemical composition averaged in TABLE 2. groups of 5 samples Average Nitrogen Crushing Protein Ether Crude Free \Vater Ash Test % Extract Fiber Extract 0 % (1bS-) % o o 856.6 43.71 6.31 11.07 26.53 6.99 5.38 1105.5 44.01 6.20 10.69 27.06 6.77 5.29 1201.2 43.45 6.23 10.72 26.53 7.37 5.65 1314.1 43.58 6.23 10.73 26.98 6.85 5.60 1392.8 43.81 6.51 11.03 26.25 7.11 5.29 1529.5 44.02 6.41 11.00 25.94 7.10 5.52 1694.2 44.67 6.32 10.02 26.21 7.31 5.47 2079.9 43.78 5.79 10.68 26.29 7.12 5.33 TABLE 3. Comparison of ball test with chemical composition Average Nitrogen Number Range of Labora- Ball Protein Fat Crude Free Water Ash of Ball tory Test % % Fiber Extract % % Cores Test Number (1bs.) % % Averaged (lbs.) 41040 220.0 45.58 6.55 10.59 25.32 6.65 5.31 9 50 41051 225.6 45.43 6.66 9.84 25.49 7.07 5.51 9 45 41047 228.5 45 30 6.88 10.05 24.94 7.51 5.32 10 35 41034 233.6 44.13 6.52 10.00 26.35 7.09 5.91 11 60 41007 245.0 44.49 6.73 9.52 27.17 6.76 5.33 9 55 41048 263.5 45.10 6.93 9.53 25.08 7.26 5.38 10 30 41052 276.5 45.36 6.91 9.95 25.27 6.98 5.53 10 40 40997 288.4 43.75 5.05 10.92 21.67 13.17 5.44 18 60 40999 291.4 44.38 6.59 8.77 27.54 7.43 5.29 16 65 41041 306.5 45.48 5.96 10.57 26.24 6.48 5.27 11 60 41015 306.6 45.49 6.70 10.17 24.57 7.76 5.31 9 60 41008 315.4 44.37 7.06 9.90 26.14 6.87 5.66 15 6O 41025 322.0 43.20 6.53 11.71 25.50 7.81 5.25 12 45 41016 346.6 44.45 6.56 9.99 25.76 7.87 5.37 9 20 41031 347.7 43.75 6.17 9.95 27.51 6.59 6.03 9 6O 41003 349.6 44.66 6.86 11.46 24.62 7.29 5.11 12 60 41348 356.5 43.04 5.74 11.80 27.98 6.38 5.06 10 60 41036 358.8 41.48 6.70 11.03 28.42 6.82 5.55 9 60 41351 360.0 42.57 5.87 11.99 27.91 6.68 4.98 11 60 ‘41011 361.1 44.52 6.19 11 14 25.94 6.83 5.38 9 60 40995 363.0 44.49 6.41 10.69 25.62 7.40 5.39 14 85 41027 370.0 44.62 6.12 10.21 25.96 7.01 6.08 11 60 41019 382.7 41.29 5.81 12.26 28.20 6.84 5.60 13 45 41023 388.9 43 70 6.68 11.51 25.19 7.75 5.17 14 90 41350 406.0 42.65 6.00 11.76 27.96 6.62 5.01 10 40 41021 414.5 41.34 5.68 12.36 27.82 7.17 5.63 13 40 41012 415.8 45.51 6.07 10.96 25.48 6.80 5.18 12 40 41004 418.0 45.19 6.59 10.62 25.02 7.12 5.46 10 60 41032 418. 7 44.29 6.25 10.20 26.62 6.55 6.09 11 60 41001 421.2 45.38 6.32 11.22 24 89 7.04 5.15 13 65 41055 428.9 41.27 5.65 13.06 27.70 6.71 5.61 14 65 41347 430.5 42.21 5.66 11.69 28.64 6.56 5.24 10 90 40993 432.3 44.08 5.77 11.68 24.21 8.72 5.54 13 60 41037 439.2 41.45 5.78 11.81 28.60 6.81 5.55 12 50 40989 449.0 45.26 5.98 9.02 27 30 6.63 5.81 14 110 40991 463.8 45.45 6.75 9.69 26.09 6.60 5.42 12 30 41028 467.5 44.53 5.79 9.61 27.07 6.88 6.12 8 80 41056 482.5 41.24 5.54 12.57 28.46 6.53 5.66 8 60 41044 489.0 45.03 4.83 11.83 26.27 7.07 4.97 11 60 41045 l 582.0 44.59 4.84 11.75 26.73 7.09 5 .00 10 | 100 FACTORS WHICH MAY AFFECT HARDNESS OF COTTONSEED CAKE 9 a and cores of each test which did not vary too widely from the aver-age were selected for chemical analysis, Table 1 contains the results of the crushing tests and the chemical analysis. Table 2 gives the averages by groups of five samples. There appears t0 be no consistent relation between the crushing tests of the icores and their chemical composition. The samples with the highest ‘* crushing strengths contain the lowest fat, but the samples with the next to the highest crushing strength have the same fat content as those with the lowest. TABLE 4. Ball test and chemical composition averaged in groups 0f 5 samples Average Crude Nitrogen of Ball Protein Fat Fiber Free Water Ash Test % % % Extract % % (lbs) % '5 7‘ 230.5 44.99 6 67 10.00 24 85 7.01 5.48 r 286.1 44.81 6 39 9.95 25 16 8.26 5.38 ; 327.7 44.25 6 60 10.34 25 89 7:38 5.52 357.2 43.25 6 27 11.48 26 97 6.80 5.21 382.1 43.35 6 21 11.28 26 59 7.12 5.45 417.6 44.35 6 18 11.07 25 97 6.94 5.50 436.0 42.85 5 77 11.45 27 29 7.09 5.55 490.9 45.17 5 55 11.09 26 92 6.83 5.43 A similar comparison was made between the ball test and the chemical fcomposition. Table 3 contains the results, and Table 4 the average in {groups There seems to be a distinct relation between the fat content ‘grand the ball strength of the core, but this relation is not consistent. In fmost cases, the harder samples seem to have the lower fat content. p It is seen in Table 4 that the fat content of the cores is lowest when {the ball test is highest. The Moisture Content of the Cake The theory of some of the mill operators that cottonseed cake becomes fharder after storage for a considerable time suggested that experiments ibe run to determine if this change in hardness is caused bIy a change in Ythe moisture content of the cake. Series of 24 one-inch cores were prepared in the usual manner. The léeven-numbered cores of three series cut from the same sample of cake vwere placed in a vacuum oven for varying lengths of time in order to remove varying amounts of moisture. The odd-numbered cores of these “series were tested for hardness in the usual manner. The even-numbered icores of three series of cores cut from another slab of cake were placed in a closed container over water for varying lengths of time in order to “dd varying amounts of moisture to the cores. C) .Moisture determinations were made on the corings of each 12 cores ut, and the value found was taken to be the average original moisture Content of the cores used. The final moisture content was calculated f om the original moisture content and the difference in weight before 10 BULLETIN NO. 534, TEXAS AGRICULTURAL EXPERIMENT STATION and after treatment. After weighing, the even-numbered cores were crushed in the usual manner. Similar determinations were made with the ball test. In another experiment, three series of 15 cores each were cut from each of three cakes of cottonseed. The first of these series was dried in a vacuum oven at 100° C for 5 hours before testing, the second series was tested in the same condition as the original cake, and the third series was placed over water in a closed vessel for 5 hours before testing. After each core was crushed, it was placed in a numbered tin box, Gill style, with a tight cover so that the moisture content of the core at the time of crushing could be determined. This moisture was calculated from the loss in weight of the crushed core which was dried for 5 hours in a vacuum oven. TABLE 5. Effect of change in moisture upon crushing strength Untreated Cores Treated Cores . Percentage Number Percentage change in of cores Crushing Crushing change in crushing tested Moisture strength Moisture strength moisture strength per cent pounds per cent pounds 12 8.66 2492 2.84 . 2049 -—-5.99 -——17.78 12 8.68 2055 4.68 1587 —-4.27 —-22.77 12 9.11 1843 8.67 1830 —-0.54 —-0.71 12 8.53 2393 9.33 2166 0.89 —-9.49 12 8.57 2465 10.95 1528 2.54 —-38.01 12 8.58 2149 11.25 1274 2.98 -—40.71 The results of the crushing tests are shown in Table 5, and those of the ball test in Table 6. TABLE 6. Efiect of change in moisture upon ball test Untreated Cores Treated Cores Number Percentage of cores Percentage change in tested Moisture Ball Moisture Ball change in ball test each in corings test in cores test moisture determination per cent (lbs.) per cent (lbs.) 12 9. 7O 266 2.20 243 —7.67 ——8.65 12 9.34 265 4.14 255 —-—5.43 —3.77 12 9.19 272 7.09 228 —2.57 —-16.18 12 10.12 263 12.54 253 2.78 —-3.80 12 9.64 263 13.15 242 4.02 --7.97 12 9.50 289 16. 38 161 7.67 -—44.35 In each of the above cases, it will be noted that the cake was hardest under the original conditions. A rapid change of the moisture content seems to soften the cake. The crushing strength and the ball test were lower when the cake was dried, or when it took up water. Taking up moisture had a greater effect than drying. Table 7 shows the result of the second series of tests. The results are the same as those previously found except in the case of one of the three FACTORS WHICH MAY AFFECT HARDNESS OF COTTONSEED CAKE 11 samples, N0. 40633-B. With this sample the average hardness of the dried cores was greater than the hardness of the original cores. TABLE 7. Effect of moisture content upon crushing strength Cores dried Cores not treated Cores moistened Laboratory Number Crushing Crushing Crushing Moisture strength Moisture strength Moisture strength per cent per cent per cent per cent per cent per cent 40640A 4.01 2338.0 6.31 2521.0 A 8.24 2134.6 40633B 2.09 2496.6 7.84 2116.6 11.34 1308.6 40634B 2 .09 2368.0 7.13 2428. 7 _ 8. 79 2102. 7 The Effect of Moisture Content Before Pressing Upon the Hardness of Cottonseed Cake Some mill operators think one of the causes of the production of hard cottonseed cake is the introduction of too much moisture during the cooking process. Tests were made of the effect of the quantity of moisture in the meal before the meal was pressed into cakes. Samples of the cooked meal were taken from the cake-former in the Bryan Cotton Oil Mill, and samples of the corresponding cake were tested. The moisture in the cooked meats was increased as much as possible, until the meats had a tendency to crawl in the press-cloths. The results are shown in Table 8. Within the limits of this experiment the cake becomes softer as the moisture in the unpressed meats increases. The specimens in which the moisture in the unpressed meats exceeded 6.00% were so esoft and waxy that no point of failure could be noted on the testing machine. The specimens seemed to flow under pressure and to exhibit no typical failure cracks. Additional work is desirable. TABLE 8. Effect of moisture in unpressed meats upon hardness of resulting cottonseed cake Moisture per Crushing Average Laboratory Number of cent in strength ball test Number specimens unpressed meats (lbs.) (1bs.) 38943 8 4 2720 511 38942 8 4 83 2886 480 38941 8 5 8O 2154 440 38944 8 5 93 1561 306 38946 8 6 74 10w 298 38945 8 8 low 318 Effect of Press Box Temperature upon Hardness L Some mill operators think that hard cottonseed cake is produced if {the temperature of the press boxes in the hydraulic presses has been Epermitted to go too high. Experiments were made in the Bryan Cotton Oil Mill to determine the ijeffect of press box temperatures upon the hardness of the cake. Ther- 12 BULLETIN NO. 534, ’I‘EXAS AGRICULTURAL EXPERIMENT STATION TABLE 9. Effect of press box temperature upon hardness of resulting cottonseed cake . I . Average Average Number of Press box crushing ball Laboratory specimens temperature test test Number tested °F (lbs.) (lbs.) , . 38941 8 96° 2154 - 440 38942 8 106 2886 480 38943 8 114 2720 511 38944 8 | 124 1561 306 38945 8 136 low 318 ss946 s I 132 10“; 29s mometers were placed in the press boxes and were read when the cake was removed from the press. The cake from the press was saved for testing.- The results are shown in Table 9. As the press box temperature increases above 114° F, the hardness of the cake decreases. The cake was soft and waxy at 136° and 132°. It was also harder at 106° F than at 114° F. This factor needs further study. Eflect of Storage on the Hardness It has been suggested that cottonseed cake becomes harder after long storage than it was when first pressed. An experiment was designed to ascertain the effect of storage. Samples of slab cottonseed cake which had been stored for twenty- three months in a covered wooden box in a dry basement were tested. In Table 1O the crushing tests are compared with the crushing tests which were made on these samples when received. It is evident from the results that, in general, storage did not increase the hardness of the cake. Although some of the samples had a higher average crushing strength after storage, most of the samples had lower crushing strengths. In practically every case, however, the differences between the tests were within the differences which could have been caused by the location of the test specimen in the slab. Relation of Specific Gravity of Cake to Hardness It was noticed that some specimens of the cottonseed cake tested seemed to be more dense than others. The specimens seemed to vary from a fine grained, very tightly compacted mass to a coarse grained, very loosely compacted mass. This suggested that the density or specific gravity of the specimens might be related to the hardness of the cake. Two experiments were made to ascertain the relationof the specific gravity to the crushing strength. Cores were cut from a sample of cottonseed cake, and the end surfaces of the cores were sanded smooth and parallel. FACTORS WHICH MAY AFFECT HARDNESS OF COTTONSEED CAKE 13 TABLE 10. Effect of storage 0n crushing strength Laboratory Before storage After storage Number pounds pounds 39679-3 83S 1099 39680-A 948 1123 39680-3 963 1323 40703 1.071 1296 40633-3 1143 1475 40634-151 1169 835 40640-3 1181 1192 40643-C 1199 1171 39680-C 1234 1092 40704 1234 1123 40643-3 1322 1419 40633-3 1405 1321 40633-41 1415 1404 40640-A 1416 1624 40633-3 1533 1450 40633-17 1557 1629 40718 1604 1429 40633-H 1655 1619 40643-A 1670 1057 40624-A 1679 1262 40633-O 1722 1702 40633-G 1796 1776 Average . . . . . . . . . . . . . 1352 1337 In the first experiment the specific gravity was obtained for each core by first weighing the core in air, dipping it in hot paraffin, and then obtaining the loss of weight in water. Crushing tests were then run on each core. In the second experiment, one-inch cores were cut from a cottonseed cake which had been found by previous tests to have below the average hardness. The fiat ends of these cores were sanded smooth and parallel by a sanding wheel. The cores were brushed well with a brass wire brush to remove all loose particles and were then weighed and the weight of each was recorded. The specific gravity was determined by the following method: A 50 cc lipless, tall-form beaker was placed in a 4" evaporating dish. The beaker was filled to overflowing with mercury. Down on top of this beaker was forced a pyroxylin plate which had 3 tacks pushed through it to form a triangle of such size that all 3 tacks would rest on one of the cores. In this way the excess mercury was removed from the beaker. One of the cores was then placed on top of the mercury and was forced down by the three tacks in the pyroxylin plate until the plate again rested on top of the beaker. The overflowing mercury was caught in the evaporating dish, transferred to a tared 25 cc beaker, and weighed. The weight of this mercury divided by 13.595 was recorded as the volume of the core. The results of determinations of the specific gravity of three sets of six cores by the above method showed that the error in computing the specific gravity of the cores of cottonseed cake by this method is less than 0.5 per cent. L ‘I B R 1A F: v Agricultural 8r ti“ are“? cairn... a; ma... i . as ‘$153M u“ ' . , pfigffffirw {fzfig-F-ann. ‘P’ 14 BULLETIN NO. S34, TEXAS AGRICULTURAL EXPERIMENT STATION FortIy-five cores were prepared from cake No. 40653 and their specific gravity was determined by the above method. The same cores were crushed, and the results 0f the tests were averaged in groups of 9. The same procedure w-as repeated for cakes No. 40633A and 40633G, except that twenty-five cores were used and the tables are. therefore, averaged into groups of five determinations each. TABLE 11. ‘Comparison of specific gravity and crushing strength of cottonseed cake 40633A 40653 40633G Group Average Average Average Number » crushing Average crushing Average crushing Average strength specific strength specific strength specific (lbs.) gravity (1bs.) gravity (lbs) gravity 1 2020 1.2880 1976 1.3222 2198 1.2896 2 2128 1.2867 2243 1.3230 2438 1.2940 3 2240 1.2946 2412 1.3175 2593 1.3071 4 2340 1.2991 2534 1.3216 2693 1.3077 5 2562 1.3057 2709 1.3157 2838 1.3118 Average. . . 2255 1.2946 2375 1.3200 2551 1 3022 Table 11 shows the comparison of the averages of the separate groups from each of the samples tested. An inspection of this table shows that for two of the samples, Numbers 40633.4 and 40633G, the average crushing strength for each group increases regularly as the average specific gravity of the group increases, but that the reverse is true for Sample 40653, and also that these results are irregular. There seemed to be a relation with two of the slabs, but not with the other one. SOME FACTORS WHICH AFFECT THE METHOD OF TESTING Some of the factors which afiect the method of testing were studied as described below. The Direction of the Testing Load The process of manufacturing cottonseed cake is such that the cake tends to have well defined bedding planes which are analogous to the bedding planes of sedimentary rocks. It was decided, therefore, to ascer- tain the differences in crushing and ball strength when the testing forces acted in a line parallel to the bedding‘ plane and when they acted in a line perpendicular to the bedding plane. A jig for cutting 1" squares from cottonseed cake was arranged so that all specimens tested would be of the same size and shape. This jig was made similar to the box used by a carpenter for fitting moulding. Fifty specimens were cut from each sample in four adjacent rows. The tests were arranged so that the two adjacent specimens were given dif- ferent tests. FACTORS WHICH MAY AFFECT HARDNESS OF COTTONSEED CAKE 15 TABLE 12. Relation of Ball and Crushing Tests of 1" squares tested flat and edgeways Crushing strength Laboratory Percentage strength on Number edge is of strength flat Flat On edge 40654 3371.2 2652.0 78.7 40629 3070. 4 2019.6 65 . 8 40718 3373.8 2166.2 64.2 Ball strength 40956A 580.8 310.2 53.4 40650 493.4 269.0 54.5 Table 12 shows the results of the work, The crushing strength was much lower when the sample was tested on edge than when it was tested flat—that is, along the bedding plane. The crushing strength of the sample on edge was about two-thirds that of the sample placed flat in the machine. The force required of an animal to break the cake would seem, therefore, to depend upon how the crushed cake is presented to the teeth, but it is likely the piece would be rolled around in chewing so that all sides would be presented. Effect of Smoothness or Roughness of Ends of Specimens In some of the experimental work, the samples of cottonseed cake were sanded smooth. The following experiments were made to ascertain the effect of this procedure. A series of forty-five cores was cut from three cakes of widely different crushing strength as determined by previous tests. The cores were cut in three adjacent lines of fifteen cores each and were numbered in the order of their removal from the sample. The odd-numbered cores were tested with their plane surfaces rough; the even-numbered cores were sanded on a sand wheel until their plane surfaces were smooth and parallel. This arrangement was made so that average results of the crushing tests on the cores could be compared. The results of these tests are shown in Table 13. The experiment was repeated on two rows of 10 cores each from three cakes. One row of cores from each cake was tested in the condition in which they were taken from the cake, and the other row of cores from each cake was tested after the ends of the cores had been sanded smooth and parallel. The results of these tests are also included in Table 13. The crushing strength of the smooth cores is much greater than that of the rough cores. The crushing strength of the rough cores varied from 40.2 per cent to 70.5 per cent of the crushing strength of the sanded cores. It is seen that the standard deviations for the rough cores in four cases are extremely high, and that the standard deviations for the sanded 16 BULLETIN NO. 534, TEXAS AGRICULTURAL EXPERIMENT STATION TABLE 13. Comparison of crushing strength of cores with rough ends and wvith smooth ends Rough ends Smooth ends Percentage crushing strength for rough ends Number Average Per- Number Average Per- is of of crushing Standard centage of crushing Standard centage crushing specimens strength deviation deviation specimens strength deviation deviation strength tested (lbs) is of tested (1bs.) is of- for smooth average average ends 23 1027 166 16.18 22 2360 258 10.92 43.50 1O 1126 139 12.38 1O 2085 185 8.89 53.95 10 1129 208 18.39 10 2110 187 8.85 53.50 23 1145 147 12.85 22 2187 281 12.81 52.35 23 1349 247 18.34 22 3352 391 11.65 40.20 23 1415 338 23.85 22 2255 202 8 .96 62.75 10 1436 260 18.23 10 2908 183 6.31 49 . 40 23 1573 286 18.11 22 2602 334 12.82 60. 40 23 1796 223 12.40 22 2552 225 8.04 70. 38 cores vary from 6.31% to 12.82% of the mean. The data also suggest that, although the testing of the rough cores gives a more accurate indication of the hardness of the cake as fed to the cows, smooth cores may be better for the purpose of experiments on the factors affecting the hardness of the cake. This suggestion is also given because the theory of testing materials indicates that, for comparative results, the testing load must be applied uniformly ‘over the surface of the test specimen. This is obviously im- possible with the rough cores of cottonseed cake. The theory of loads concentrated at a point on a surface (1) is that maximum stresses are set up along the approximate surface of a cone with the apex of the cone at the point of application of the load and with the angle at the apex of the cone of approximately 110°. The rough projections of the unsanded cores tend to concentrate the testing loads at these projections and the forces within the core tend to split it rather than to cause it to fail by straight compression. The Effect of the Rate of Application of Load of Cottonseed Cake Upon Its Crushing Strength It has long been known that the rate of application of a load in testing material by compression materially affects the Iyield point and ultimate strength of the material (2). It was decided, therefore, to make tests to determine the effect of this factor upon the ultimate strength of specimens of cottonseed cake. Since there is a probability of a great difference in the ultimate com- pression strength of two cores of cottonseed cake taken from adjacent positions in the slab, and since compression tests with two different speeds cannot be run on the same specimen, it was necessary to devise some means of approximating the ultimate compression strength of a specimen in terms of other neighboring specimens. FACTORS WHICH MAY AFFECT HARDNESS OF COTTONSEED CAKE 17 Series of 24 cores were cut in a line lengthwise of the cake and as close together as the thickness of the \metal in the core drill would permit. A complete series of cores was tested at a constant rate or" de- formation of .4 inch per minute, and another series was tested at a rate of .2 inch per minute. TABLE 14. Effect 0f rate of application of load upon crushing strength and ball test Speed of Crushing Table . (In. per Min.) .4 .2 .2 .1 .1 .05 Number of Specimens Tested: Crushing Test pounds. . 12 12 12 12 12 12 Ball Test pounds... . . . . 21 24 . . . . . . . . .. 12 . . . . . . . . .. 12 Average load at Failure: Crushing Test pounds. . 1540 1432 1368 1340 1331 1420 Ball Test pounds . . . . . . . 404 392 . . . . . . . . . . 369 . . . . . . . . . . 339 Standard Deviation of Load at Failure: Crushing Test pounds. . 157 210 212 266 186 195 Ball Test pounds . . . . . . . 68 48 . . . . . . . . . . 33 . . . . . . . . . . 40 Percentage standard devia- tion is of average: Crushing Test pounds. . 10.02 14.63 15.50 19.82 13.95 13.71 Ball Test pounds . . . . . .. 16.83 12.24 . . . . . . . . .. 8.95 . . . . . . . . .. 11.80 Average total deformation (inches): Crushing Test . . . . . . . .. .1521 .1412 .1089 .1225 .1270 .1510 Ball Test . . . . . . . . . . . . . . .074 .080 . . . . . . . . . . .073 . . . . . . . . . . .082 Average load applied per second (lbs): Crushing Test . . . . . . . .. 67.5 33.8 41.7 19.8 17.5 7.84 Ball Test . . . . . . . . . . . . .. 36.7 16.3 . . . . . . . . .. 8.42 . . . . . . . . .. 3.46 Average time required to apply failing load (see): Crushing Test . . . . . . . .. 22.8 42 .4 32.7 67.5 76.0 181.1 Ball Test . . . . . . . . . . . .. 11.0 24.1 . . . . . . . . .. 43.8 . . . . . . . . .. 98.0 The crushing strengths of each two consecutive odd-numbered cores were averaged and the average strength was compared with the strength of the intermediate even-numbered core. A comparison of the average strength of two consecutively even-numbered cores with the intermediate odd-numbered core was made. The results are shown in Table 14. As the tabulation shows in each case that the error caused by the assump- tion that the ultimate compressive strength of any one specimen in the series is equivalent to the average of its proximate neighbors in the series is considerably less than the standard deviation, it is assumed that this method of computing the ultimate compression strength of a speci- men of cottonseed cake in terms of specimens from adjacent positions is sufficiently correct for the purpose of this study. Series of 24 cores were prepared from a slab oi cottonseed cake in the manner already described. The odd-numbered specimens were tested 18 BULLETIN NO. 534, TEXAS AGRICULTURAL EXPERIMENT STATION at one rate of deformation and the even-numbered specimens of the same series were tested at another rate. An inspection of Table 14 shows that in general the ultimate crushing strength of the cores of cottonseed cake increases as the rate of appli- cation of the load increases. This fact suggests that there may be a certain amount of plastic flow of the material under the slow application of the load. TABLE 15. Efiect of rate of application of load upon crushing strength and ball tests Crushing strength Ball test Sprafed o crushing Number Average Standard Per cent Number Average Standard Per cent table of crushing deviation deviation of ball deviation deviation (in. per min.) samples strength of crush- is of samples test of ball is of tested (lbs.) ing test average tested (lbs.) test average .4 23 1413 319 22.57 24 404 68 16.83 .4 12 1540 157 10.02 12 419 53 12.65 . 2 24 1386 285 20.55 24 392 48 12 .24 .2 12 1432 210 14.63 24 348 76 21.81 .2 12 1368 212 15 .50 12 346 44 12.71 .2 12 430 56 1310 .2 12 392 34 8.67 .1 24 1338 196 14.65 24 350 46 13.12 . 1 12 1340 266 19 .82 12 342 36 10.51 .1 12 1331 186 13.95 12 333 34 10.20 .1 12 402 37 920 .1 12 369 33 8.95 .05 24 1343 260 19.31 24 310 42 13.54 .05 12 1420 19S 13. 71 12 376 40 10.63 .05 .. 12 339 40 11.80 Table 15, however, shows that on specimens tested from the same slab of cottonseed cake the ultimate crushing strength of the cores under different rates of application of load does not vary more than the variation shown for the ultimate strengths of the cores from different positions in the slab tested at the same rate of application of load. It may be assumed, therefore, that for all practical purposes the rate of application of load to the cores of cottonseed cake has no effect, within the limits of these experiments, upon the ultimate crushing strength of the cake. The Relation of the Unit Deformation of 1" Cores under a Load of 1000 lbs. to Their Crushing Strength Elastic materials under compressive or tensile loads are deformed in proportion to the load until the elastic limit of the material is reached (1). The deformation per unit of load then increases rapidly until the speci- men under test fails. Because of the varying dimensions of test specimens of cottonseed cake, the deformation per unit of length was used as a measure of comparison of two specimens. Cores on which the original surfaces were left intact were crushed in the testing machine with the crushing head of the machine moving at a FACTORS WHICH MAY AFFECT HARDNESS OF COTTONSEED CAKE 19 known constant speed. The time required for the applied load to change from nothing to 1000 pounds was checked by a stop watch, and the deformation of the specimen was computed by multiplying this elapsed TABLE 16. Comparison of unit deformation under a load of 1000 lbs. to crushing strength Unit de- Unit de- Unit de- Unit de- formation Crushing formation Crushing formation Crushing formation Crushing Number at strength at strength at strength at strength 1000 lbs. (lbs.) 1000 lbs. (lbs.) 1000 lbs. (lbs.) 1000 lbs. (lbs.) (in./in.) (in./in.) (in./in.) (in./in.) 1 .129 1000 .089 1380 .064 1570 . 115 1630 2 .105 1120 .098 1420 .096 1570 .080 1640 3 .121 1140 .091 1440 .062 1580 .062 1670 4 .108 1220 .113 1440 .076 1580 .114 1700 5 . 093 1240 . 056 1490 .066 1590 . 070 1740 6 . 125 1250 .077 1490 .095 1590 .076 1760 7 .101 1260 .115 1510 .115 1590 .092 1760 8 .114 1320 .043 1530 .069 1600 .068 " 1790 9 .092 1340 .063 1540 .072 1610 .075 1790 10 . 090 1360 . 081 1540 . 079 1620 . 069 1810 11 .113 1370 .061 1570 .065 1630 .047 1820 Mean . . . . . . 108 1238 081 1480 078 1594 079 1737 1 087 1820 .059 2010 .049 2110 053 2410 2 051 1850 . 063 2010 . 065 2170 052 2470 3 080 1880 .054 2030 .071 2180 O56 2490 4 049 1890 . 088 2030 . 069 2210 059 2490 5 O74 19 10 . 058 2040 . 059 2220 064 2490 6 061 1930 .054 2050 .058 2250 047 2510 7 094 1940 .074 2060 .065 2260 079 2550 8 065 1950 . 075 2060 . 071 2290 078 2610 9 091 1970 . 045 2070 088 2290 . 1O . 050 2010 . 065 2070 . 059 2300 11 .053 2010 .056 2100 .067 2310 Mean . . . . . . O69 1924 063 2048 066 2235 .061 2503 time by the velocity of the crushing table. The unit deformation was obtained by dividing the total deformation in inches by the original thickness of the specimen in inches. Table 16 shows the results of the comparison of the unit deformation under a load of 1000 pounds with the ultimate crushing strength of 1" cores of cottonseed cake. The determinations were arranged in the table in the order of the magnitude of the crushing strength of the cores. The determinations were then averaged in groups of eleven determina- tions and these averages were examined for possible relationship between the factors. An inspection of the table seems to indicate that the hard specimens are deformed less than the soft specimens under the load of 1000 pounds. OTHER POSSIBLE METHODS FOR TESTING HARDNESS As explained in Bulletin 523, the crushing test was considered to have the best relation to the hardness of cottonseed cake. Other methods for testing the hardness of cottonseed cake were tried. They include the so-called “bootheel” test, the abrasion test, the Brinell 20 BULLETIN NO. 534, TEXAS AGRICULTURAL EXPERIMENT STATION test, the schleroscope test, the impact test, the “tooth” test, and the ball test. The bootheel test, the tooth test, and the ball test were described in Bulletin 523. The Modified Brinell Test The definition of hardness includes the physical properties of resistance to penetration and to permanent deformations. One of the accepted tests for penetration which is used in the metal industries is the Brinell test (3). This test is made by forcing a 10 mm steel ball into the surface of a material to be tested by a standard load. The depth of penetration is computed from the diameter of the indentation and the diameter of the ball. The ratio of the load to the depth of indentation is the measure of the hardness of the material. An attempt was made to determine the hardness of cottonseed cake bIy means of a modified Brinell test. A steel ball 15/32 inch in diameter was mounted in a section of steel shafting in such a manner that it could be forced into the specimens of the cake by means of the Olsen machine. The specimens for use in this experiment were pieces of cake cut in squares with dimensions 2" x 2". The surfaces of the squares were sanded to a smooth finish before the test was made. The preliminary tests of the Brinell hardness of cottonseed cake were unsatisfactory because of the lack of accurate means of measuring the diameter of the indentation caused by the steel ball. A Beggs Deformeter Microscope was later available, and was used for further tests of the method. The Beggs Deformeter microscope is a 10-power microscope fitted with a micrometer eyepiece. This eyepiece contains a set of crosshairs Which may be moved across the field by means of a micrometer screw. The head of this screw is divided into 100 parts. The crosshairs are fitted with an index which moves along a divided scale in the field of the eyepiece. This arrangement permits the measurement of two diameters of a circle at one setting of the specimen provided the specimen is placed so that the circle is tangent to both crosshairs when the movable scale is at the zero index. The crosshairs are moved so that the horizontal and vertical crosshairs traverse the indentation to be measured. The scale and micrometer head are read when each crosshair has traversed the circle. The mean of these readings is recorded as the diameter of the circle. The perimeter of the indentation caused by the ball is not a true circle, because the hard particles in the cake are forced down into the soft parts of the cake. The difliculty in reading the true diameter of the indentation was overcome in part by marking with chalk the surface to be tested. This procedure gave a clearly defined indentation because the oil in part of the specimen under the ball was pressed out and obliterated the chalk in that area. The indentation had a perimeter closely approximating a true circle. The reading of the two diameters FACTORS WHICH MAY AFFECT HARDNESS OF COTTONSEED CAKE 21 and the estimation by eye of the position pf the true circumference of the indentation reduced the error in reading the diameter. The indentations were made in the smoothed surfaces of 12 one-inch cores cut from each sample. They were made by means of a steel ball 15/32 inch in diameter and under a load of 50 kg. After the Brinell tests were made, six specimens from each sample were tested by crushing and the remaining six specimens from each sample were tested by means of the ball test. A comparison of results is shown in Table 17. TABLE 17. Comparison of modified Brinell tests with ball and crushing tests Average Average Average Laboratory Number diameter ball test crushing test Number of cores of indentation on cores on cores tested (mm) (lbs) (lbs) 39679-A 12 2.863 527 3078 38570 12 3.168 493 3427 38564-C 12 3.218 343 2131 39680-C 12 3.295 415 2020 38592 12 3.401 360 1733 39680-13 I " 12 3.592 373 1505 The results of the Brinell test seem to correlate with those of the crushing test. One set of the five sets of results, that of Sample No. 38570, is not in accord with this conclusion. The Brinell test, however, does not seem to correlate as closely with the ball test as it does with the crushing test, for two of the sets in this comparison are not in accord. The Brinell test did not seem to offer a practical method of testing cottonseed cake. The Modified Schleroscope Test The schleroscope (4) test consists of the dropping of a steel ball upon the test specimen and the measuring of the height 0t the rebound of the ball. This has been used by the ball-bearing industry for the testing of the hardness of ball-bearings and for the automatic classification of the balls into three classes: balls that are too soft, balls that have the right hardness, and balls that are too hard. The classification is obtained by arranging for the balls to roll down a chute, strike a hard steel plate, and bound to three slots arranged at different distances from the point of rebound of the balls from the steel plate. The hard balls bounce to the most distant slot, the balls which have the right degree of hardness bounce to the middle slot, and the balls which are too soft bounce t0 the slot which is nearest to the point of rebound of the ball from the steel plate. The schleroscope test and many modifications of the test are used in the metal industries to test the hardness of finished products that may not be destroyed through testing. Steels and other alloys are tested by this method. 22 BULLETIN NO. 534, TEXAS AGRICULTURAL EXPERIMENT STATION This test was modified for use with cottonseed cake so that the steel balls rolled down a tube of a certain length, struck the specimen, and bounced to a piece of paper covered by a sheet of carbon paper. In this way, the point at which the ball struck was recorded and the distance the ball bounced could be measured. The length of tube used in this experiment was 40 cm. and the angle of incidence of the ball on the specimen was 60° from the vertical. TABLE 1S. Comparison of average modified schleroscope test with ball and crushing test 40 cm tube and 60° angle of incidence Average ' Average Average Laboratory Schleroscope test. Ball test Crushing test Number (cm) (lbs.) (lbs.) 39680-B 31. 6 373 1505 38570 31.6 493 3427 38592 35.3 360 1733 38564-C 36.5 343 2131 39679-A 36 9 527 3078 39680-C 3 415 2020 o The results are shown in Table 18. It is seen that there is no rela- tion indicated between the schleroscope test and the ball or crushing test. It was decided, however, to make further tests of this method before discarding it as impractical. A series of 10 cores was cut from each of three cakes and plane sur- faces of these cores were sanded smooth and parallel to each other. The TABLE 19. Comparison of schleroscope tests to crushing strength of cotton- seed cake 20 cm tube and 45° angle 0f incidence Laboratory Number 40634~B 40633-B 40643-A Specimen Trial Number Number Average Average Average Schlero- Crushing Schlero- Crushing Schlero- Crushing scope strength scope strength scope strength test (lbs.) test (lbs.) test (lbs.) (cm) (cm) (cm) 1 18.9 2400 18 5 2300 18.9 2680 2 19 7 2400 19 . 0 2490 18 . 7 2880 3 18.2 2610 17.8 2540 19.1 2960 4 19.0 2660 17 . 6 2600 18.1 3040 5 18.3 2680 19.4 2660 18.4 3060 6 19.3 2690 17.7 2690 18.8 3180 7 18.5 2900 18.5 2700 19.1 3340 8 19.3 2930 18.4 2740 18.5 3360 9 19.1 2990 18 .8 2820 19 .0 3390 10 18.2 3020 18.1 2990 18.6 ' 3440 Mean . . . . . . . . . . . . . . 18.85 2728 18.38 2653 18 72 3133 schleroscope tests were similanto those previously described, except that the tube was 20 cm long and inclined at an angle of 45°. Four tests were made on each core and the core was then crushed in the usual FACTORS WHICH MAY AFFECT HARDNESS OF COTTONSEED CAKE 23 manner. The results of the tests are shown in Table 19. These results confirm the conclusion that there is no relation between the schleroscope hardness and the crushing strength of the cake. The Impact Test The impact test (5) is made by dropping a standard weight upon the specimen either from a fixed height or from varying heights. The number of strokes of the weight required to shatter the specimen is recorded as the hardness of the specimen. The impact test is primarily intended for the testing of highway materials, such as crushed stone and gravel, which are required to bear the continued impact of passing vehicles. A series of tests was made on cores from the same sample of cake in order to compare the hardness of 1" cores of cottonseed cake as measured by a Page Impact Machine with the hardness as measured by the ball and crushing tests. The Page Impact Machine is designed to measure the toughness of rock and gravel used for paving roads. The machine is so constructed that a 2 kg. weight is raised and dropped vertically at regular intervals upon a 1 kg. striker which rests upon the specimen to be tested. The specimen rests on a fixed anvil which has a plane surface. The Page Impact Machine is so arranged that the hammer may be raised to heights which increase by one cm. for each blow, or that it may be raised the same height for each blow. Both methods were used in the tests of cottonseed cake. In these tests, the energy required to fracture the specimen is reported as the impact test of the specimen. This energy is computed by multiplying by the weight of the hammer the total distance through which the hammer falls in order to fracture the specimen. Cores were cut from various samples of cottonseed cake and were tested in the rough condition in which they were when cut. An equal number of cores was tested by ball, crushing, and impact tests. Impact tests were run by permitting the hammer to drop from varying heights for each blow, ‘from a constant height of 6 cm. for each blow, and for some of the tests from a constant height of 4 cm. for each blow. Table 20 shows the comparison of the ball tests with the impact tests, and Table 21 shows the comparison of the crushing test with the impact tests. There appears to be no relation between either the ball or crushing tests and the two kinds of impact tests. i The Abrasion Test The abrasion test (5) is made by determining the ‘weight lost by specimens of the cake when ground in a ball mill under special standard conditions. This loss in weight indicates to a certain degree the amount 24 BULLETIN NO. 534, TEXAS AGRICULTURAL EXPERIMENT STATION TABLE 20. Comparison of impact tests with ball test on 1" cores 01f cottonseed cake Average Impact tests (Kg. crn) Number of Average ball cores tested test Varying Constant drop Constant drop (lbs.) drop of of hammer of hammer hammer , (6 cm) (4 cm) 12 186 92.3 94.8 12 221 119.7 113.0 12 245 45.3 54.0 .... 25 267 117.6 51.8 236.5 25 291 94.6 42.7 151.4 12 323 102.6 122.3 25 364 154.6 142 . 7 290.2 12 370 145.8 147.0 25 383 113.1 136.3 284.8 12 390 187.5 246.0 .... 12 425 120.8 42.0 25 445 128.0 56.2 293.6 12 460 92.5 86.0 12 488 35.7 57.0 25 524 72.5 42.7 113.3 12 531 87.2 105.0 of cohesion between the particles of the tested specimen. This test is used for the determination of the ability of certain highwary materials, such as stone and gravel, to stand up under the continued abrasion of the wheels of passing vehicles. TABLE 21. Comparison of impact tests with crushing tests on 1" cores of cottonseed cake Averag¢ Impact tests (Kg. cm) Average Number of crushing ' cores tested test Varying » Constant drop Constant drop (1bs.) drop of of hammer of hammer hammer (6 cm) (4 cm) 12 710 92.3 94.8 . 12 835 92 .5 86.0 . . 12 937 119.7 113.0 . 12 948 120.8 42 .0 12 963 45.3 54.0 25 1094 113.1 136.3 284.8 12 1216 102.6 122.3 25 1234 117.6 51.8 236.5 12 1234 145.8 147.0 25 1405 94.6 42.7 151.4 25 1415 154.6 142.7 290.2 25 1604 128.0 56.2 293.6 12 1876 187.5 246.0 12 2068 87.2 105.0 25 2326 72.5 42.7 113.3 12 2768 35.7 57.0 Six series of one-inch cores were prepared from one slab of cottonseed cake. These cores were cut and sanded smooth and true to size. Each of the twelve cores in a series was tested by means of the ball test, and the average ultimate load of the twelve was taken as the ultimate load of the series. The ball test was carried just to failure so that the cores still appeared intact. FACTORS WHICH MAY AFFECT HARDNESS OF COTTONSEED CAKE 25 Each series of cores was then weighed as a whole and was placed in a ball mill with 10 one and one-half inch iron balls. The mill was run at TABLE 22. Comparison of ball test with loss 0t weight of cores under abrasion test Number of Average ball test Loss in weight of cores specimens (lbs.) per cent 12 350 5 .80 12 37S 8 .96 12 386 5.88 12 390 5 . 36 12 393 6.95 a speed of 25 revolutions per minute for one hour, after which the cores were taken from the mill and reweighed. Table 22 shows the results of the test. ' It is evident that there is no relation between the ball test and the abrasion test described. The abrasive hardness of cottonseed cake was also compared with the crushing strength of the cake for 6 one-inch cores placed in a ball mill with five 1%” iron balls. The mill was run at a speed of 25 R.P.M. for one hour and the loss of weight of the cores was determined. The cores TABLE 23. Comparison of crushing test with loss of weight of cores under abrasion test Number of cores Average crushing test Loss in weight of cores tested (lbs) per cent 6 1956 1 92 6 2017 2 60 2333 1 were then crushed in the usual manner and the average crushing strength of the cores was recorded. The results of the test are shown in Table 23. It is evident that there is no relation between this abrasion test and the crushing test. ACKNOWLEDGMENT The authors wish to acknowledge the cooperation of Mr. F. D. Fuller of the Division of Feed Control Service, Profs. J. J. Richey, J. T. L. McNew, and C. E. Sandstedt of the Civil Engineering Department, and Mr. J. Webb Howell and Mr. Jack Howell of the Bryan Cotton Oil Mill, as well as of a number of cottonseed oil millers who furnished samples and information. ‘ 26 BULLETIN NO. 534, TEXAS AGRICULTURAL EXPERIMENT STATION SUMMARY Various factors in the composition and the manufacturing of cotton- seed cake were investigated to ascertain their relation to the hardness of the cake. No definite regions of hardness or softness were found in slabs of the cake. Hard spots and soft spots were found both in the central portions and in the end portions of the cake. There seemed to be a slight relation between the fat content of the cake and the hardness, but not sufficient to be definite. There seemed to be no relation between the protein, crude fiber, or ash and the hardness of the cake. A rapid and radical change in the moisture content of the cake re- sulted in a lowering of the crushing strength. In experiments made to determine the effect of the moisture content of the unpressed cottonseed meal upon the hardness of the cake, it was found that, within the limits of this experiment, the meats with a com- paratively high moisture content gave a soft and waxy cake. Hard or very hard cake was not obtained in this experiment. The temperature of the press box during the pressing of the cake affected the hardness of the cake to some degree. Soft and waxy cake was made with the press boxes at high temperatures while hard cake was made with press boxes at lower temperatures. This factor requires further study. Storage for approximately two years in a dry place apparently has no effect upon the hardness of the cake. There seems to be only a slight relation between the specific gravity of the cake and the hardness. The cake gave a lower crushing test if the crushing force was applied parallel to the bedding plane of the cake than if the crushing force was applied across the bedding plane. Those specimens which were sanded smooth before testing had a higher crushing strength than those specimens which were not smoothed. The rate of application of load within the limits of the experiment did not affect the apparent hardness of the cake. Harder specimens were deformed less at the crushing load than were the softer secimens. The modified Brinell method and the schleroscope method of testing the hardness of the cake were found unsuitable for the testing of the cake. There was no relation between the crushing strength and the force required for the failure of the test specimens under impact and under steadily applied loads. Therewas also no relation between the loss of weight of the test specimens under abrasive tests and the crushing strength of the specimens. FACTORS WHICH MAY AFFECT HARDNESS OF COTTONSEED CAKE 27 REFERENCES Batscén, 119g}? and Hyde, J. H. Mechanical Testing. E. P. Dutton and 0. . Abrams, D. ‘A. Effect of rate of application of load upon compressive sltzrggfthlgi); concrete. American Society for Testing Nlaterials. ‘Brinell Hardness Testing of Metallic Materials. American Society for Testing Materials Standards. 1933. . Shore, A. F‘. The Schleroscope. American Society for Testing Materials Proceedings. 10:490. 1910. i-Test for the Toughness of Rock. American Society for Testing Materials ‘ Standards. 1933.