~U.S.;-DEPARTMENT. OF AGRICULTURE, BUREAU OF ANIMAL INDUSTRY—Bulletin No. 58. S$ D. E, SALMON, D, V. M.,, Chief of Bureau. | 253 | .W 38 > = PA ; ee Seah ee ee ¥. THE FAT TESTING OF CREAM me BABCOCK METHOD. ere = is ED. H. WERSTER, /nuspector and Dairy Expert, DAIRY DIVISION, BUREAU OF ANIMAL INDUSTRY, =? eat ; AND ky ks A ; ; PoE | . > > EOO meals Betas ° . 2 99a ©33 3 a55 aa 2_— | ,2 BD\e : Wf ime Nie Bes ea iee d . { ‘ ? ) ) ) \ > ; < aS NY pe . ere a ? a ) Pa) nA = bo RR rat . 2>8 ’ LETTER OF TRANSMITTAL. U. S. DEPARTMENT OF AGRICULTURE, BUREAU OF ANIMAL INDUSTRY, Washington, D. C., February 4, 1904. str: I have the honor to transmit herewith the manuscript of an article on ‘‘The fat testing of cream by the Babcock method,” pre- pared by Hd. H. Webster, inspector and dairy expert, and C. E. Gray, special agent, under the direction of Henry E. Alvord, Chief of the Dairy Division of this Bureau. Mr. Gray, chemist, of Topeka, Kans., was temporarily appointed as special agent to assist in this work. It is evident from investigations conducted in the field in connection with the gathering of data for this volume that much misapprehen- sion exists in regard to the proper way of making the Babcock test with cream, and that loose and inaccurate methods of doing the work are universally in vogue. This matter becomes of much importance when large operations are involved, consequently the information and directions submitted in the following pages are of value to creamery men and others interested in the subject. I therefore recommend the publication of the work as Bulletin No. 58 of the series of this Bureau. ' Respectfully, D. E. SALMON, | Chief of Bureau. Hon. JAMES WILSON, Secretary. 3 Dy—bdl1 (ae aaa Coe o See ee ee oe es ene 35.2 7.8 | 28.8 29.6 | Boye gate ee bridle fe eT Me ata aris oe 710.4 28.2 8 29.2 Dole Bec bh Eek Se ape RCE i a 140.8 27.8 28.0 27.0 ROosetcqyert.. 0st Sees PPE Nee, Ble eee 30.2 28.2 31.0 | 34.0 BiG SS Se ae eS Se ets | een 70.4 27.8 35.4 | 41.6 1D See dre paar Bh Ree I CIES ANY ree ors 140.8 27.4 28.4 | 33.3 Noveover cts oe eto ee ae 35.2 | 28.2 (@)x cen (a) Diss eos epee ee ene Oe ae "0.4 27.8 43.6 80.8 1 By RN geet tk MAR A iC on CANS a RENN Bet Site 140.8 28.0 30.8 45.3 Series C. | Tight Cover. sie ae en 85.2 3.8 3.8 | 3.8 WO ek eA AE ee eee ae 70.4 4.0 3.9 | 3.8 D0 ers SS i ee Ne Ea ata are 140.8 3.8 3.8 | 3.8 Loose cover 220 2 eo ue orgie 35.2 3.8 4.4 | 4.8 WIRE ten, oor DISS Pi tet tren gos 70.4 3.8 4.0 | 42 DG. aol Pay Seis ey ee 140.8 3.8 | 4.2 | 4.2 INO [COVER 2. a Ae as oes Nd yee Se ee ee 35.2] 4.0 (2) / (2) Do i Ges 8 ees Sar! 5 Oe eee 70.4 3.8 | 4.4 | 5.4 Dp. shee ee eee eee penne ee 140.8 4.0 4.0 4.5 a Sample dried. FAT TESTING OF CREAM BY BABCOCK METHOD. i The jars used in the above test were the ordinary half-pint sample jars used for composite samples. The covers were the ‘‘ Lightning” pattern. The jars classed as ‘‘ tight cover” werein good repair and the covers as tight as such jars usually are. The ‘‘loose-cover” jars had covers that did not fit well and were left somewhat loose. The ‘‘no-cover” jars were entirely open. Each jar contained the number of cubie centimeters of cream recorded under the heading “Quantity of cream or milk.””» On November 13 the cream was tested and put into the bottles. This test is recorded under the date named. On November 27 and December 4 the samples were again tested, and the records of these tests appear under these dates. These data are worthy of considerable study. The covers that were Supposed to be tight evidently were loose enough to let out some moisture. Thesmaller samples show the greatest loss. Those marked ‘‘Sampie dried” were so dried out that. they could not be tested at all. The sample jars can not be too carefully looked after, and yet it is no uncommon thing to see jars containing samples without cov- ers or with very imperfect ones. The result of such practices is inevitable. The creamery will lose money and will, in all probability, never know how it occurred. Preserving the sample.—The average sample in which a preserva- tive tablet has been placed shows the weakness of the methods of pre- serving that are commonly in vogue. A bright-red spot appears where the tablet lay in the bottle, the rest of the cream being in no way affected by the tablet. Such a method is a waste of tablets, and the cream sample is practically in the same shape it would have been had no attempt been made to preserve it. The cream is usually thick, or quickly thickens in the jar, and the tablet, although dissolved in it, is not mixed throughout the cream. As quickly as the operator has time after putting the first cream in a jar, if it is too thick, he should warm the bottle to about 100° F., thus bringing the cream to a fluid condition and enabling an easy mixture of the preservative and cream. Usually the next addition of cream can be quite readily mixed with the first. The mixing should be done immediately after adding the sample tothe jar. All the jars should be looked over at the end of the day, and any that have not mixed properly should be warmed and shaken up until the mixing is complete. In the summer time composite samples will sometimes show a sepa- ration into layers, the lower portion being whey and the upper part fat and casein. Such samples do not look well, but so far as obser- vations go this does not to any appreciable extent influence the test, provided they are thoroughly mixed before the sample is taken. In the winter season it has not been shown that a preservative other than the cold weather is needed. In shaking composite samples to mix preservative and cream always give a rotary motion to the hand, as this causes the cream to flow around the bottle. A violent agitation will often churn the contents, thus destroying to a great extent the usefulness of the sample. 14 BUREAU OF ANIMAL INDUSTRY. MEASURING OR WEIGHING THE SAMPLES. The same precaution used in procuring the sample in the first place must be used in getting it ready for measuring into the test bottle. Bring the cream to a uniform condition by warming the samples te 100° F. or a little more, then pour from one cup to another a sufficient number of times to insure a uniform texture of the cream. Be sure that no cream sticks to the side of the sampie jar. There will be no dried cream if the jar has been tightly covered. If dried cream is present the test will not be correct, because some moisture has evap- orated. If the cream is lumpy, pour it through a fine sieve and press the lumps through with the finger, as described in Bulletin No. 100 of the Vermont Experiment Station. After pouring the cream until it is uniform, measure quickly the desired quantity before any movement of fat begins toward the top. If the sample has in any way become churned, heat it warm enough to melt the butter fat and very quickly mix and measure out the desired amount. The test of such churned cream can only be taken as an approximation of the true result. There is only one correct way of securing a true sample, and that is to weigh the cream into the test bottle with a delicate pair of scales or balances. Anyone desiring correct results should not use the pipette measure for cream. ‘There are too many conditions entering into the problem to make the pipette measure at all reliable. The cream, in pouring to get a uniform sample, will incorporate more or less air. Very often it is partially sour when received and in this condition contains many gas bubbles. The specific gravity varies with its richness. Thick and viscous cream may contain large air bubbles in its body which do not move readily to the surface. Tables for correcting the error due to specific gravity have been compiled and are correct for that factor, but they do not and can not take into account the other factors, which are often of greater impor- tance than the specific gravity. The quantity of cream weighed into the test bottle does not matter, so long as the correct reading of the result can be obtained. Usually 9 or 18 grams are taken, but some- times a less amount is used. The details of making the test after weighing out the samples are of so common knowledge as to make it needless to discuss them here. It is sufficient to say that the desired end is a perfectly clear fat col- umn, so that the readings can easily and readily be made. READING THE TEST ACCURATELY. This seems the simplest step of all, and yet it has been proven the hardest. Mention has been made of the fact that creamery men were reading the cream test lower than standard instructions warranted. They doubtless reasoned that the test must be too high or they would get an overrun equal to that obtained under the whole-milk system. FAT TESTING OF CREAM BY BABCOCK METHOD. 15 Only one creamery in all the territory under investigation had made any attempt to discover why.the test seemed to be too high. ‘This creamery, through its chemist, whose work along this line is made use of in this bulletin, had worked out the table below showing that errors of a glaring character were being made in their methods of reading the test. There are a number of different types of cream bottles on the market, and the tests made were between different types of bottles in use at the creamery in question. Those used were the 30 per cent 9-inch, 50 per cent 9-inch, and 50 per cent 6-inch bottles. The first is not in general-use, but the last two are used quite extensively in Western creamery practice. A COMPARISON OF VARIOUS BABCOCK TESTS WITH THE TEST BY ETHER EXTRACTION.4@ The tests by the Babcock method were made as fol- lows: All bottles were tested for accuracy of gradua- tion before using. Three types were used: First, a bottle 9 inches in length, graduated to read two-tenths of 1 per cent and up to 30 per cent; this type of bottle will be referred to as the 30 per cent 9-inch bottle. Second, a bottle 9 inches in length, graduated to read five-tenths of 1 per cent and up to 50 per cent; this type of bottle will be referred to as the 50 per cent 9-inch bottle. Third, a bottle 6 inches in length, grad- uated to read five-tenths of 1 per cent and up to 50 per cent; this type of bottle will be referred to as the 50 per cent 6-inch bottle. All samples were weighed on ‘a Torsion balance, sensitive to 10 milligrams. The weights used vere tested for accuracy. All readings were taken au 130° F., this temperature being obtained - : : ; Fig. 1.—The ‘“* Me- by placing the bottles in water and heating to this — jiscus. section point. Three readings were taken from each test, the of a test bottle method being as follows: Reading A, from a to d; EAS Oe reading B, from b to d; reading C, from c to d. umnfrom differ- The crescent-shaped surface of the fat in the neck of ne soho the bottle, as shown at a0, is called the meniscus. The determination by the extraction method was made as follows: Samples of about 54 grams were weighed out into fat-free extraction thimbles, charged with ignited asbestos, then dried to constant weight at 100° C. The samples were then placed in a continuous ether extractor and the extraction carried on with ether that had recently been distilled over sodium. After six to eight hours’ extraction the ‘flasks containing the extracts were removed and replaced by other flasks and the extraction again continued for three hours. This extraction gave only a trace of fat. The extracted fat was dried to «Tests made by C. E. Gray. 16 BUREAU OF ANIMAL INDUSTRY. constant weight at a temperature of 100° C. Duplicate analyses were made of all samples. Only the averages are given in the table below: TaBLE IIl.—Readings of Babcock tests compared with similar figures obtained by ' ether extraction. Result by Sample. ether ex- traction. |Reading. Per cent. = pec he ae INO20 Si oe Nei ie ee ee eee eee 52.32 . ewe 2! (Ce eee Awe Se INGE 2) fe 25. See ie ce ee ie ed le en ee 43.41 ° ake 2a. (Canes ses FAs oue See NO.3 23.3. 3S Ee ee ee eee 39. 28 P HEN Ae Cea se NOe4 esis fe Rs 3 Ae een encarta 26.23 p ENS Ci reaots EAT Aes INO Seer ee Se UP ee eee red 23.42 Es ire Cen eas (Aer aea IN0.16) 222 55 eee. 5 ee eee eee 8:27 : heads Ce Seek jee Noha oe cr ee ee 8.30 f ener | 6.05 TaBLE III.—Summary of average differences between extraction and Babcock methods. [(+) denotes readings above and (—) readings below those obtained by extraction method. ] Babcock readings. 30 per cent Per cent. 52. 90 52. 50 52. 00 44. 20 43. 80 43. 40 39. 80 39. 40 39. 00 27.00 26. 60 26. 20 23. 90 23.50 23.10 8. 80 8. 40 8.00 9. 00 8.60 8. 20 30 per cent 9-inch. Per cent. : sample. 50 per cent di Results by Babcock method. 50 per cent | 50 per cent 9-inch bottle, 18-gram 6-inch bottle, © 18-gram sainple. Per cent. | Per cent. 53.20 52. 80 52.35 44.10 | 43.70 | 43. 30 39. 75 39.25 38.75 26. 80 26. 30 26.00 | 23.90 23.45 22.95 8.80 | 8.45 8.05 9.00 8.65 8.35 | inch. Per cent. +0. 62 | + .19 eek 50 53. 53.10 52. 40 44.45 43.80 43.25 40.00 39. 40 38.75 27.15 26. 60 26. 05 24. 20 23.80 23. 40 9.58 8.90 8.25 9.05 8.65 8.25 6-inch. 8 per cent Per cent. a =e 00 43 12 The differences in readings of 9-gram samples were doubled in order that they should appear in such form as to compare with the other. columns of the table, which represent 18-gram samples. A COMPARISON OF MILK AND CREAM READINGS. The lowness of the milk test as compared with cream readings is exemplified by the following table, which gives the summary of dif- ferences as compared with the test made in the 10 per cent milk bottle. . PLATE l. Buctcetin No. 58, B. A. I. TYPES OF CREAM AND MILK TEST BoTTLES USED IN EXPERIMENTS. ‘O14 ‘SSSNOSINAI ‘SNANTOD LV4 ONIMOHS ‘SATLLOG LSAL 1) GNV WvayO JO SadAL BULLETIN No. 58. B. A. I. PLATE Il. FAT TESTING OF CREAM BY BABCOCK METHOD. LZ The readings were made and the differences found in the same man- ner as in the comparison with the ether extraction figures, shown in Table ITI. TaBLE IV.—Ewcess of average cream readings A, B, and C over milk reading A. Cream reading. 30 per cent 50 per cent 50 per cent 9-inch. 9-inch. | 6-inch. eed MES aw RGSS: 17 RAR ae cies ec gy Sa ee a Re ee oat 40.65 | +1.50 fe ee emer See) <9. Seek ee A a ee ac ON ee: + .27 + .32 | +75 OO glen Sg ee Be SN Rr re aps, ES ein + .07 + .05 | + .20 Credit is due to Mr. Gray for suggesting this line of investigation. A eareful study of the foregoing tables will indicate that something wrong is pretty sure to occur when the Babcock test is applied to cream by a careless or inexperienced. operator. The readings were made at a, b, and ¢ in the above work because these points were quite . distinct and could be read without much chance of error. The results show that the probab:e cause of the differences in the readings of the different types of bottles lies in the meniscus. (See fig. 1, p.15.) Fol- lowing out this suggestion, aseries of experiments were undertaken at the Kansas Experiment Station. EXPERIMENTS AT THE KANSAS EXPERIMENT STATION. Through the courtesy of Prof. J. T. Willard, director and chemist of the Kansas station, Manhattan, Kans., the author secured the use of the station chemical laboratory and the cooperation of the assistant -hemist, Mr. R. H. Shaw, in carrying on further the work already « lined. Prof. J. O. Hamilton, assistant professor of physics in the v.ansas State Agricultural College, gave material assistance in the mathematical and physical problems involved. Experiments were repeated with the same results as obtained in the work already quoted and shown in Tables II, III, and IV. ERRORS CAUSED BY INCORRECT READING OF THE MENISCUS. The most interesting feature of the work was a study of the effect of the meniscus in bottles of different diameters on the reading of the test. A series of thirteen bottles were selected ranging in diameter of neck from 0.2338 inch to 0.5862 inch. The bottles were filled with water to some point in the lower part of the neck, and then 1 gram of pure butter fat weighed into the neck. The specific gravity of the fat was 0.9, and each gram should represent just 5.55 plus per cent butter fat. .The bottles were whirled in a hand tester and read at 120° F. Readings from extreme top to bottom of fat column were earefully made and the depth of meniscus measured and recorded in per cent of the scale on which the reading was made. The following 2218—No 58—04——2 D F BUREAU OF ANIMAL INDUSTRY. table gives the style of bottle, diameter of neck, the per cent of fat, the depth of meniscus in same terms as fat, and the difference from the true amount: TaBLE V.—Effect of meniscus in bottles of different diameter. | x lab: F s es Difference e 0. O lameter er cen er cen over true Style of bottle. bottle. | of neck. of fat. meniscus.| amount (5.55-+). ware eo 0), 2338 5.70 | 0:20:42), 05 Sida Cr) Sete he a airy cies | 12 . 2682 5.80 .30 25 3 . 2858 5. 90 || . 40 . 30 ; 42 . 3056 Pye Sc .39 a0 per cent 9neh creams oe aoe ee 5 ; 4 on eae ee : a ih . 3396 6.10 . 60 aoa F if 34 . 3916 6. 25 . 80 .70 50 per + Orin ch’ Greate 215 oa ees ' De Coes a Ares Caine 4100 6.25 - 90 70. & é f 45 . 9016 6.75 | 1.50 1.20 BO Ne ee cae anna | 44 | 5232 =.00. 1.80 1.45 25 . 5498 7.00 2.00 1.45 50 per cent 6-meh:cream — 22222 20 . 5664 7.25 | 2.23 1.70 19 | . 0862 7.40 | 2.2) 1.8 It will be seen in the above table that the 10 per cent milk bottles give a reading of 5.70 and 5.80, respectively, and that the 50 per cent 6-inch cream bottles give a reading of 7.00, 7.25, and 7.40, and that the meniscus of the 10 per cent bottles is 0.20 and 0.30, while for the 50 per cent 6-inch bottles it is 2.00, 2.25, and 2.25, and, further, that the 10 per cent bottles read 0.15 and 0.25 above the true amount and the 50 per cent 6-inch bottles read 1.45, 1.70, and 1.85 above the true amount. This is especially interesting from the fact that all of these bottles have been supposed to read alike. It will also be seen that the 10 per cent bottles read more than the true amount, whereas in general practice the reading of these bottles is supposed to indicate the amount exactly. In ordinary milk testing it is estimated that about two-tenths of 1 per cent adheres to the inside of the bottle below the neck, and in reading the per cent of fat the total length of the fat column will just give the desired reading. In the above table the fat was weighed into the neck of the bottle and was therefore known to be all in the neck, and the effect of the meniscus is thus as apparent in the 10 per cent bottles as with the others. : If we calculate that about the same amount of fat is held back in the bulb of all bottles, regardless of size of the neck, it will be seen that the differences—as shown in the last column of the above table— regularly increase as the necks of the bottles increase in diameter. Taking No. 13, the first bottle in the table, as a standard, the differ- ences from the true amount, here taken as 5.70, are seen to range from 0.10 in bottle No. 12 to 1.70 in bottle No. 19. This shows the actual errors that would constantly be made in regularly testing with these bottles. " i - FAT TESTING OF CREAM BY BABCOCK METHOD. 19 Tabie V shows that the depth of meniscus increases with the diam- eter of the neck, and that the true reading comes somewhere between the reading of the top and the reading at the bottom of the meniscus. A careful inspection of Table V will show that if from the depth of the meniscus there is deducted four-fifths of itself, the result will be very near the true difference in reading between that which the bottle actually gives and the true amount. Therefore the shortest rule that ean be given to correct the error due to the meniscus is as follows: Read the test from the extreme top to bottom of fat column; deduct: from this reading four-fifths of the depth of the meniscus, and add 0.2 per cent to the result. The 0.2 per cent is the amount supposed to adhere to the test bottle and is not shown in the fat column in the neck. In none of the cases in the table will this rule give exactly the right figure, but the differences are all, with one exception, less than one-tenth of 1 per cent. TABLE VI.—Showing corrected readings after deducting portion of meniscus, as per rule. Reading Number} poag pep th BAUS f oe | oe ead- (6) SO | fitths o Style of bottle. re ing. menis- | menis- pops uF menisc us saa cus. cus. ere. +0.% per ’ | cent. | Per cent.| Per cent.| Per cent. | Per cent. i | Per cent. : : | 13 Df 0.20 0.16 5. 54 5. 74 ee 12 5.8 30 24 5.56 5.76 3 5.9 40 aoe 5.58 5.73 4 | 42 5.9 . 50 40 5. 50 5.70 30 per cent 9-inch cream -_-....--..---- 5 6.0 50 40 5.60 5.80 7 6.1 . 60 48 5. 62 5. 82 ; 34 6. 25 80 . 64 5. 61 5. 81 ne Cer Oat (aorta | 32 6.25 20 2 5.58 5.73 | 45 6.75 1.50 1.20 aD 5a > it 6-1 h CYE¥2 30 te ee eee eC nn oe a em : ee ue ceca 3 44 7.00 AEGON ide 5.56 5.76 | 25 7.00 2.00 1.60 5. 40 5. 60 50 per cent 6-inch cream ___....______-- 20 7.25 2.25 1.80 5. 45 5. 65 19 7.40 2.25 1.80 5. 60 5. 80 Pee aati 1S GO ae LO? 86 Bas | 5. 14 ; | If the 10 per cent bottles are taken as correct and as standards in this case, the final average is on!y 0.01 per cent less than the average reading of the 10 per cent bottles, the latter being 5.75 per cent. | If a 9-gram sample is used and the result must be doubled, the 0.2 — per cent must not be added until the four- a ths meniscus is dedweted and this result multiplied by 2 The differences shown in Table VI occur because it is a practical impossibility to read the exact amount in any test. In the 10 per cent bottles and the 30 per cent 9-inch bottles tenths have to be estimated; in the 50 per cent 9-inch bottles anything less than 0.5 20 BUREAU OF ANIMAL INDUSTRY. per cent has to be estimated; in the 40 per cent 6-inch bottles used in this test anything less than 1 per cent has to be estimated; and in the 50 per cent 6-inch bottles used anything less than 0.5 per cent has to be estimated. All of this estimation makes not only possible but altogether probable errors in reading which range from less than 0.1 per cent to as much as 0.5 per cent. The space occupied by 1 per cent in the large-diameter bottles is so shallow that a slight tip- ping of the bottle away from the perpendicular may easily make a difference of from 0.25 per cent to 1 per cent in the reading. Applying the rule for reading the meniscus to the readings in Table II, the following is the result: TABLE VI!I.—Results of applying rule for reading meniscus to readings found in Table ITI. or Read- | Menis. | fifths of lees four Double | +0.2 per reading ede ae Style of bottle. ing.a cus. ee pole e fl reading.| cent. Ce : cus. tion). Per cent.| Per cent.| Per cent.| Per cent.| Per cent.| Per cent.| Per cent. 30 per cent 9-inch -.. 26.45 0.45 0. 36 26. 09 52. 18 52. 38 i ” per cent 9-inch -_- 53. 20 .85 68 52550 ioe see eee 52. 70 52.32 50 per cent 6-inch -.- 53. 85 1.45 1.16 yal 8S Bil ese 52. 89 30 per cent 9-inch _-- 22 10 50 40 21.70 43.40 43. 60 2 be per cent 9-inch -_- 44.10 . 80 . 64 AS AGH o eases 43. 66 43.4] 50 per cent 6-inch _-_- 44.45 1.20 96 43 AQ | see ee a 43.69 30 per cent 9-inch __- 19. 90 40 . 32 19.58 39.16 39. 36 3 ” per cent 9-inch -__- 39.75 1.00 . 80 88505 i= see 39. 15 39. 28 50 per cent 6-inch --_- 40. 00 1.25 1.00 a0 pe wa ie 39.20 | 30 per cent 9-inch __- 138.50 .40 . 32 13.18 26. 36 26.56 fs per cent 9-inch _-_- 26.80 . 80 . 64 265167) sae 26. 36 26.23 50 per cent 6-inch __- 27.15 1.10 ~ 88 265215 (See Ses 26.47 30 per cent 9-inch __- 11.95 .40 .32 11.63 23.26 23.46 | 5 ” per cent 9-inch __- 23.90 . 80 . 64 QaaCOn |e ee 23.46 23. 42. 50 per cent 6-inch __- 24.20 95 76 DE AAR 2 aes aie ans 23. 64 30 per cent 9 inch __- 4.40 -40 32 4.(8 8.16 8.36 | 6 be per cent 9-inch ___ 8.80 15 .60 Br OO ie cee 8. 40 | as 50 per cent 6-inch _-_- 9.55 1.30 1.04 Bre: | ee 8.71 30 per cent 9-inch __- 4.50 .40 . 82 4.18 8.36 8.56 7 }\50 per cent 9-inch __- 9. 00 65 52 SLA SOs 2. Soe ee 8.68 ; (50 per cent 6-inch._.| 9. 05 80 SBE OOS AT Phat Sc ae aThis column gives the highest reading (A) in each case. In the above table the column headed ‘‘ Reading” is the extreme reading, and the column headed ‘‘ Meniscus” is the difference between ‘‘a” and ‘“‘c,” as shown in fig. 1 (on p. 15). For the 30 per cent 9-inch bottles only one-half the amounts given in Table II are used, as one-half is equal to the 9-gram reading that was taken in this case. This is corrected to the proper amount in the column headed ‘‘ Double reading.” The column ‘‘ +0.2 per cent” gives the corrected reading, and the column ‘‘ True reading” gives the amount determined by the extraction method. FAT TESTING OF CREAM BY BABCOCK METHOD. 21 A rearrangement of the final results in the above table makes the following showing. The sum of the seven samples is taken and the average differences determined: TaBLeE VIII.—Showing average excess of Babcock readings over extraction readings. aol age | pean. 0.2 per Cor- |Average| Average Style of bottle. eeiOUe | cent ad-| rected pene differ- menis- | Gition. |reading.|amount.| ence. cus. ‘Per cent. |Per cent. |Per cent. |Per cent. Per cent. BBpeIeCMh O-tnehe. Fae see ras 2 eS LS 28.69 0. 20 28.89 28. 74 +0.15 MepeTeCy OANGN, on 18a we a oe eo 28.71 . 20 28.91 28. 74 hn 19 £1) TUG CS ra Snes Se eee ee ee ee 28.83 . 20 29.038 28.74 + .29 The column ‘‘Average reading less four-fifths of meniscus” is the average of double the 30 per cent 9-inch readings, and the 50 per cent 9-inch and 50 per cent 6-inch readings. To these is added the 0.2 per cent for the corrected reading. From the corrected reading is taken the average true amount, which gives the difference found in column ‘‘Average differences.” This is the average above the true amount for each reading ‘‘A” in the three classes of bottles rep- resented in Table II. It is doubtful in making these readings in Table II whether the meniscuses were read deep enough, as they were not read with a thought of this comparison in view. Another possibility enters into the proposition that will have to be left an open question for the time being; that is, whether 0.2 per cent is not too much to allow for the fat which does not rise into the neck when testing cream in the wide-necked bottles. In Table VI it must be remembered that the fat was weighed into the neck of the bottle, and that the amount was positively known to be in the neck. The table was arranged from Table V, with the supposition that 0.2 per cent would remain below the fat column in all the bottles, and that being the case, the assumption that in the 10 per cent bottles the read- ing would actually be the same as if the test had been made in the reg- ular way and read 5.7 per cent for No. 13, ete. While this is true for the 10 per cent bottles, it may not be true for the cream bottles, as Table VIII would seem to indicate. In view of the fact that nearly every error in sampling and making tests tends toward a larger read- ing than the correct one, the addition of the 0.2 per cent might safely _ be omitted and no injustice done to either party of the transaction. It is a common practice to use a 9-gram charge and double the read- -ing. This multiplies the error made in reading by 2. An examina- tion of the diameter of the bottles given in Table V will show the wide variations to be found in bottles of the same class. The 13 bottles used in this experiment were selected from 50 bottles taken- at random from a large stock, and they show almost every grade of variation. Z2 BUREAU OF ANIMAL INDUSTRY. It would not be a safe rule to deduct a certain amount from bottles of the same per cent reading, the differences between bottles of the same class being as great as between classes. ; THE LOWER MENISCUS. Among other things observed in making this test was the lower meniscus or line between the fat and the liquid below. When the bottles were whirled for five minutes in a tester in which the tem- perature remained constant, the line between the fat and water was a straight line. Immediately on cooling this line began to haye a curved shape, the center being lower than the sides. If the bottles were placed in a hot bath and the temperature raised above that at which they had been whirled, the line curved upward, the center becoming higher than the sides. These changes in the shape of the bottom line, or plane of division, were nearly equal in effect to the upper meniscus. If the line curved either way, the effect was to increase the reading. The curve downward, due to cooling, did not affect the reading as much as the curve upward, due to heating. As the fat is drawn down and the bottom reading made lower, the fat column is shortened at the top, and this in a measure equalizes the error. _When the center is moved upward as the result of expansion, the top is pushed up and the bottom line remains the same; the line that touches the glass is then the one from which the reading will be taken. This effect has been observed where a hand tester had a steam connection for heating the bottles. The operator would turn in the steam for the last whirling of one minute and then read the test. This sudden heating caused an expansion, and the movement of the fat and liquid on the surface of the glass being slower than that in the center, the bottom line of the fat column was curved upward, and consequently the reading was too high. ERRORS DUE TO EXPANSION BY HEAT. In testing cream the matter of expansion of the fat is of consider- able importance. An experiment made by Mr. Gray is here recorded which shows the effect of expansion in heating the fats to tempera- tures ranging from 110° to 140° F. Plan of test.—A sample of pure butter fat was separated from pure, fresh — butter. A determination of the specific gravity of this fat showed it to be normal (0.9004 at 100° C.) compared with water at the same temperature. Samples of the fat were weighed into the various types of test bottles and skim milk enongh added to each to make 18 grams. The fat and skim milk were mixed as thor-. oughly as possible and the test made as with cream. From this mixture a test was obtained that in appearance was very similar to the regular cream test. Three readings were taken from each test [as described in fig. 1 (p. 15) ], at tem- peratures of 140°, 130°, 120°, and 110° F. Six tests were made with each type of bottle, and the results averaged as in the following table. The cor:ect amount is very near 39.8 per cent. (That is, this would be the accurate reading or test for the fat actually present in this mixture or unratural cream.) FAT TESTING OF CREAM BY BABCOCK METHOD. 23 Tasie [X.—Average readings of six bottles at different temperatures, showing expansion, | Temperatures. Style of bottle. ‘Reading. | = | 140° F. | 130° F. | 120° F. | 10°F. Per cent. Per cent. Per cent. Per cent. Ree so ivye 20.1 20.0 19.9 110585 oo per. centoOanch, O-cram=-. 9.22 222-2 : Thee JIS AC gees bY Sc as ames (3 Bist 19.75 C--.-.-| 19.7 | 19.65 | 19.65 | 19.65 A eorre Pen tras .* 40° 0 39.8.7 ° 39.7 30 per cent 9-inch, reading doubled_____-.-.-____- : ares er 39.8 | 39.6 | 39.5 | 39.5 Cres 39.4) 393 | 39.8 39.3 (eee eg, UO SO I. 30.5 39.5 50 per cont 9-inch, 18-gram___.-.---.-.------------ . see | 394]; 393] 99.2 39.1 Resse) se 3RPyh | 38 9. Picea tec: ogg iy b AB sea Se 40.5) 40585°| +. 40.1 39.9 S0per cent.oimeh, 18-eram: —.__-_ 226. 4.322 seoc5- [PE ce ae | 39.9 | . 39.8 39.6 | 39.5 ; rete. | 89.0 | S800 es aeeT ob Se G TABLE X.—Average changes in readings at stated temperatures due to heat expansion. Temperatures. Style of bottle. | 140° to 130° F. | 130° to 120° F. | 120° to 110° F. | 140° to 110° F. ; PCr CCMie tle GECTAGents Per cent. Per cent. 30 per cent 9-inch, 9-gram ___________| 0.1 Ori 0.05 | 0.25 Sareea OU le Gi see ee | .2 S Baa a | 5 _ 50 per cent 9-inch, 18-gram -__-_-.. ----| 2 oes 0 4 50 per cent 6-inch, 18-gram __--.__--- | .2 Bare a2 6 Average full reading __..._____.._--- | 52 oa al ing 5 Theoretical difference.__.__.__------ | . 142 . 142 | .142 | 427 ena ee | Rival O71 O71 Ata ae The average increase in reading when the temperature is raised from 110° to 140° is 0.5 per cent, as taken from the readings in Table IX. The theoretical increase is 0.427 per cent. The difference between the practical reading and the theoretical calculation is but 0.073 per cent, which is less than can be read in either of the styles of bottles in the test. The coefficient of expansion of fat is given as 0.00064 for melted butter fat within the range of temperatures as found in common testers in use. For each change of 10° F. 1 ¢. ¢. of fat, or 5 per cent, will change 0.0178 per cent. In the above illustration there was 39.8 per cent, or nearly 8 c. c. of fat, and the change would be eight times 0.0178 per cent, or 0.142 per cent, as shown in Table X. Fifty per cent cream would be ten times 0.0178 per cent, or 0.178 per cent. Thirty per cent cream would be six times this factor, ete. 3 The correct temperature for reading is given as 120° F 3 and: in Table IX the readings, taken at 120° and corrected by the rule for / 24 BUREAU OF ANIMAL INDUSTRY. ealeulating the correct reading, give results nearest to the true amount, 39.8 per cent. It will be seen that the differences of reading, when 9 grams of cream are used, are just one-half of those from a full charge, or 18 grams, so far as temperature is concerned. It will also be seen that when the readings of the 9-gram charge are multiplied by 2 to obtain the correct reading, the difference is also multiplied by 2, and the final result is the same as though 18 grams had been used. There has been an objection raised against the long-necked cream bottle to the effect that so long a column of fat necessarily shows more expansion and greater care must be exercised to have the temperatures just right. A eareful consideration of the expansion table above will not bear out this assertion. Butter fat expands volumetrically. Cream-test bot- tles are graduated volumetrically and will take care of any expan- sion, no matter what the shape of the neck. With the long seale and the narrow neck the expansion is measured in just the same pro- portion as in a short scale and wide neck. The hand tester will, as a rule, give nearer the right results than the steam-turbine tester, because the fat from the hand tester is usually read nearer 120°. The steam machine may heat the bottles to 160° or 180° F., and if read at this temperature a little calculation will show the nature of the results. If the steam machine is run with the lid raised an inch or two, better results will be obtained. This will allow an air current to be drawn through the machine, and the temperature will thus be kept down. If bottles are placed in a water bath to bring them to the right temperature for reading, they should remain in the bath for twelve to fifteen minutes before being read to insure an even temverature of fat and liquid bclow the fat and to insure a straight line between the two. FINANCIAL RESULTS OF INCORRECT READING OF TESTS. It has been stated before that a difference in test of one-half of 1 per cent would amount to $1 for every 1,000 pounds of butter made if valued at 20 cents per pound. In the preceding discussion it is shown that in the matter of reading the test alone a difference of 1.7 per cent can be made. Figured on the same basis as the above, this would equal $3.40 per 1,000 pounds, or $68 per day on an output of 20,000 pounds. The methods of operating and the-basis of paying fcr butter fat in the centralizing plants give the farmer all there is in it. In other words, the farmer in most cases gets a price for a pound of fat or the cream containing it which, together with the transportation charges on this cream, will equal the selling price of a pound of butter. The creamery must make its profits and pay its running expenses out of the overrun it is able to obtain. If this overrun is cut down through FAT TESTING OF CREAM BY BABCOCK METHOD. 25 mismanagement of any kind, profits are gone. Instances are now on record where concerns have become bankrupt through bad manage- ment along these very lines. The test question is but one item in the way of successful operation, and the things pointed out in this bulletin are not all the corners that must be watched in getting a proper overrun; but so far as the testing goes, this bulletin points out the dangers which are of vital importance to the creamery inanager. The following are examples of shortages occurring on several systems in Kansas and Nebraska largely due to improper methods of testing: TABLE XI.—Showing losses occurring at cream-receiving stations in Kansas and Nebraska, 1902-03. Month. Stations. Pee Butter fat foe at factory. paid for, Pounds. Per cent. 1902. Number.| Pounds. Pounds. FIBERS se ee i A a pes hee er 174 | 572,986 | 581,424] 8,488 1.45 AS TOE CRS Ee aE Fe Sa Pes ea, Se gO 179 481, 216 483, 612 2,396 .49 Sembee Is eit. tien Ses) oe oa Yer el 178 399, 276 396, 575 a2,701 a. 67 CGO STE See eae ae eee tu meen sale 175 239, 847 244, 816 4,969 2.03 INIOme nese se ce edieeiy a Pe cl neue SUN Ee 175 191, 038 197, 672 6, 634 3.35 WWECOnMMe eee eevee Keno eS OLINy BAe ys ik ise 164 173, 692 181,598 7,906 4.35 1903. Wile os Bes See ak ee aoe en ae eee 109 377, 576 392, 891 15, 315 3. 89 BI fUntT Omer seein ake sae see Se eu 177 812, 871 817, 903 5, 032 .61 TOU Sasa Ain f= ese gee a a ne 181 655, 691 657, 806 2,115 32 SPUR Iess asses, oe Ne ee TN 192 439, 073 439, 886 813 18 TD) cea ai es NE Oe oe ee 21 64, 326 69, 550 5, 224 49 ISU ny pemteageta as UR LI MeN od ahs Foe Sl MER 259 794, 670 812, 303 17, 633 Peau LNTOUSAT IGS Bs Ba NI reap nee gic Os Ceed 257 761, 055 777, 486 16, 431 PIA aGain. One-half of 1 per cent is allowed for mechanical loss by some of the larger creameries. Out of the 13 cases represented above, there are four within this limit—one showing a gain of 0.67 per cent and one very close to the limit. The butter fat paid for and the butter fat received are found by testing at the station and at the factory. When this was checked against a low churn yield, this and the difference between the test called for an investigation. Not all of it can be laid to any one thing, but with careless sampling, loss of cream by spilling, ete. , reading the test at too high a temperature, and reading the full meniscus, the losses may befound. ‘The differences between the station and the fac- tory tests are in the main due to poor work in sampling and testing at the station. It is a noteworthy fact that the tests at the factory are made more carefully than those atthe stations. Thestation tests 2s a rule are made by pipette rather than scales. In most cases a cor- 26 BUREAU OF ANIMAL INDUSTRY. rection table is used to correct the errors made by pipette measure. Although this method gives lower results than with the seales, the faulty methods of sampling, combined with too high reading, still give the station a higher test than shown at the factory, where the samples are usually weighed and more carefully read. During the past year the creamery companies that have kept any records of their work have been cutting down this difference between tests and helping to correct the low churn results. It is hoped that the ideas given in this bulletin will be of some benefit to them; also to others who have not realized the possible loss along this line and the great probability that they have sustained it, although ignorant of the fact. SUMMARY. CONDENSED DIRECTIONS FOR MAKING FAT TESTS OF CREAM. SAMPLING: (1) Uniform composition and texture of cream is necessary. (2) This is obtained by pouring from one pail or can to another. (3) Frozen cream must be thawed before it can be sampled. (4) Churned cream can not be successfully sampled. (5) The tube sampler gives surest results. (6) The dipper sampler does well if the cream is thoroughly mixed. (7) Cream adhering to outside of tube should not get into sam- ple jar. (8) The tube should be blown out with steam or rinsed with hot water before using each time. (9) Keep the top of the tube open while it goes down, so it may fill as fast as lowered. KEEPING THE SAMPLES: : (1) Sample jars must have tight-fitting covers and be kept tight. (2) If cream is dried in bottles it is evidence that covers are not tight enough to prevent escape of moisture. (3) Preservatives must be thoroughly mixed with cream; if too thick, heat the jars. (4) Do not shake the bottle to mix the cream; give it a rotary - motion. (5) It is best to have samples protected from extreme heat or cold. (6) Churned cream gives only approximate results; dried cream gives too high results. ; (7) Extreme hot weather and lack of attention may cause sepa- ration of whey. (8) Do not take too large samples; it is a waste of cream. (9) Look after samples every day and see that they are in proper shape. ~I FAT TESTING OF CREAM BY BABCOCK METHOD. 2 PREPARING SAMPLE FOR MEASURING INTO TEST BOTTLE: (1) Sample must be absolutely uniform throughout. (2) Heat sample to about 100° F., or until it is quite fluid. (3) If sample is weighed, a much higher temperature may be used. (4) Pour from one cup to another until uniform. (5) The hotter the sample the more fluid it will be and the easier to make uniform. (6) Take care that no cream remains in sample jar adhering to the sides. | (7) If sample is lumpy, press lumps through a fine wire sieve (such as is used for a teapot strainer). (8) Melt any churned samples, mix, and sample quickly. (9) Make things convenient for this work and see that it is thoroughly done. MEASURING INTO TEST BOTTLE: (1) Weighing the sample is the only method that will give cor- rect results. (2) Use delicate balances and keep them in perfect order. (3) Test weights and scales for accuracy before using. (4) Torsion balances are very accurate; weigh one test at atime. (3) 5) Less than 9 grams may be used, but 9 or 18 grams are more : convenient. ; (6) Air and gas bubbles in cream cause pipette tests to be inac- curate. (7) Specific gravity of cream causes pipette tests of cream to be too low. (8) Tables for correcting specific gravity are in use, but they do not correct for error caused by air and gas. (9) Weighing corrects all difficulties due to specific gravity and air or gas in cream. (10) Use great care to get the weights exactly right. MAKING THE TEST: (1) If 18 grams of cream are used, add an equal weight of acid of 1.82 to 1.83 specific gravity. (2) If 9 grams of cream are used, add an equal amount of water, then add acid as for 18 grams. (3) Use enough acid to make a clear fat column; determine by trial. (4) Use condensed steam or rain water for filling bottles. (5) After adding acid, fill bottles at once to bottom of neck with water at about 120° F., and then whirl five minutes. (6) Then add water of same temperature to bring fat within seale, and whirl two minutes. (7) Keep the temperature down to 120° F. while ee ae BUREAU OF ANIMAL INDUSTRY. MAKING THE TEST—Continued. (8) Have a hole drilled in top of. tester insert thermometer. (9) Run the tester at as high speed as bottles will stand. (10) For hand tester put in boiling water when beginning the test till it nearly reaches the bottles. (11) For steam tester raise the lid slightly while making the test. (12) When through whirling keep tester closed, so as to main- tain heat even as possible. READING THE TEST: (1) See that line between fat and water is straight, and read from bottom to extreme top of fat column. (2) Read the depth of meniscus and deduct four-fifths of it from previous reading.