53 V Issued May 10, 1917. DEPARTMENT OF AGRICULTUEE, BUREAU OF ANIMAL INDUSTRY. A. D. MELVIN, Chief. CHEMICAL TESTING OF MILK AND CREAM. BY ROSCOE H. SHAW, Chemist, Dairy Division. WASHINGTON; GOVERNMENT PRINTING OFFICE. 1917. Digitized by tine Internet Archive in 2010 witli funding from Tine Library of Congress littp://www.arGliive.org/details/Gliemicaltestingo01unit A._i2. Issued May 10, 1917. u. s. (department of agriculture, BimEAU OF ANIMAL INDUSTRY., A. D. MELVIN, Chief. CHEMICAL TESTING OF MILK AND CREAM. BY ROSCOE H. SHAW, Chemist, Dairy Division. WASHINGTON; GOVERNMENT PRINTING OFFICE. 1917. '^ 5.^t!. n D. of d; JUN ^2 1917 1^^ CHEMICAL TESTING OF MILK AND CREAM.i CONTENTS. Page. Chemical nature of milk 3 Testing f or fat ... : 5 Testing milk for fat 1 5 The Babcock test 7 Testing cream for fat ] 8 Testing skim milk for fat 23 Testing buttermilk and whey for fat 25 Preserving samples 25 Cleaning the test bottles 25 Determination of total solids in milk 26 Determination of specific gravity of milk 30 Calculating total solids by formula 32 Determination of acidity of milk and cream 36 Manns's acidity test 37 Detection of preservatives 38 Chemicals and apparatus used in the chemical analysis of milk and cream 40 Comparison of metric and customary weights and measures 41 Comparison of Fahrenheit and Centigrade ther- mometer scales 41 CHEMICAL NATURE OF MILK. In order to follow intelligently the methods for testing milk and cream some knowledge of the chemistry of milk ia essential. From a chemical standpoint milk is a very complex substance. The component parts may, however, 1 This is a reprint, \Yith slight revision, of a publication issued Feb- ruary 17, 1916, under the same title. In its preparation free use has been made of the various publications on the subject, particularly "Testing Milk and Its Products," by E. H. Farrington and F. W. Woll (Madison, Wis., 1911), and "Modern Methods of Testing Milk and Its Products," by L. L. Van Slyke (New York, 1907). (3) be classified into a few well-marked groups, as follows: (1) Water, (2) fat, (3) nitrogenous constituents, (4) sugar, and (5 ) ash . The components other than water are collectively known as total solids or milk solids, and the solids other than fat as solids not fat. Milk serum, or more properly milk plasma, is the term used to denote the milk minus the fat; hence the terms serum solids and plasma solids are synonymous with solids not fat. lYater. — ^The water in milk varies from 82 to 90 per cent. The usual variation in mixed-herd milk is much less and is probably covered by 84 to 88 per cent. Fat.— The fat in milk— milkfat or butterfat— ip not in solution but exists as an emulsion of microscopic glob- ules so small that a single drop of average milk contains more than one hundred millions of them. These glob- ules, even in milk from one cow, are not all of the same size. Some may be two or three times the size of others, the average size depending upon several factors, the principal one of which is the breed of the animal. Chem- ically the fat is not a single compound but a mixture of several compounds known as glycerids. Some of thes'e glycerids are common to all fats, while others are pecuUar to butter. This fact is made use of in detecting oleo- margarin. Cow's milk usually contains from 3 to 6 per cent of fat, depending very largely upon the breed of the animal. Nitrogenous constituents.— These are principally casein and' albumin, with traces of less important nitrogenous compounds. The coagulum, or curd, produced when rennet, dilute acids, or certain other chemicals, are added to milk, is chiefly casein. Albumin is the flaky precipi- tate produced by heating whey or skimmed milk from which the casein has been removed. In constitution and behavior it closely resembles white of egg. Casein is not really in solution in the milk, but exists in an extremely fine colloidal condition in combination with some of the ash constituents. With an appropriate filter of clay it is possible to separate it from the water. Albumin is in true solution in the water of the milk. Frequently, but im- properly, the term casein is applied to all the nitrogenous constituents in milk. Sometimes the term total proteins is used in referring to the nitrogenous constituents takeii as a whole. The amount of casein in average cow's milk varies from 2 to 4 per cent and the albumin from 0.5 to 0.8 per cent. Sugar. — Milk sugar, or lactose, belongs to a group known as carbohydrates and is a white substance less sweet in taste than cane sugar. Milk sugar is broken up into lactic acid by the action of bacteria, this bringing about the sour- ing of milk. Milk sugar is in solution in the water of the milk and is present to the extent of from 3.5 to 6 per cent. Ash. — The ash, or the mineral part of milk, exists to the amount of about 0.75 per cent and consists largely of the chlorids and phosphates of sodium, potassium, magnesium, and calcium. AVERAGE CHEMICAL COMPOSITION. The table below gives the average of more than 5,000 analyses of milk at the New York State Agricultiu-al Exper- iment Station, Geneva: Per cent. Water 87. 1 Total solids 12. 9 Fat 3.9 Casein 2. 5 Albumin 7 Sugar 5.1 Ash 7 TESTING FOR FAT. In the following remarks on the testing of milk and cream the aim will be to present the subject in such man- ner that it may be followed by those who have had neither chemical training nor a course of any sort in milk testing. To those who have had such training the following pages will doubtless appear very elementary and overburdened with detail. TESTING MILK FOR FAT. Preparing the sample for testing. — As before mentioned, fat is not in solution in milk, but is in an emulsion of very fine globules. These, being lighter than the surrounding serum, tend to rise, carrying with them some of the other solids, resulting in the familiar creaming of milk. Before the test can be made a homogeneous mixture must be ob- tained. This can best be obtained by pouring the milk several times from one vessel into another. When the sample is small, beakers are convenient for this purpose, and if the sample has not remained in the container more than a few hours, pouring back and forth four or five times is sufficient. When, however, the sample haiB stood for some time in the container, the cream layer is liable to be hard and to adhere to the walls . This is particularly true of preserved samples. In such event it is well to place the container in warm water until the cream has become soft- ened and can then be easily removed. Care must be taken that none of the cream is left on the cover of the container; if, however, any is left, a brush such as ia used in cleaning beakers is useful in dislodging it. The sample must always be well mixed immediately before measming out a charge for testing. If several charges are to be measured out, the sample must be mixed each time. Thorough mixing is absolutely necessary for accurate work. Partially churned milh. — Mlk from some cows, notably of the Jersey breed, churns very easily and sometimes a too vigorous agitation in the mixing of such milk results in some of the fat collecting in small granules which refuse to emulsify again. This also frequently happens when the milk is sent a long distance in partially filled containers. These granules are easily recognized, and when they are present special treatment is required to prepare the sample for testing. A little ether equal in volume to 5 per cent of the milk may be added, and the container stoppered and vigorously shaken. The ether will dissolve the granules and the solution will mix with the milk. A fairly accurate charge may now be quickly removed, but the percentage obtained must be corrected for the volume occupied by the ether. Another and perhaps a better way to treat churned milk is to place the container in hot water until the milk has attained a temperature of about 110° F. In a few minutes at this temperature the granules will have melted. The container is then vigorously agitated and a charge for test immediately measured out. The partial churning of the samples is not a frequent occurrence and with proper care can always be avoided. When samples are to be sent a considerable distance, the containers should be completely filled so that no space is lej t at the top. A good way is to fill a bottle to overflowing with the mixed sample and then to insert a rubber or cork stopper haAang a hole about one-eighth of an inch in diame- ter. As the stopper goes to its place the milk will spurt out through the hole; the hole is then filled with a piece of glass rod or a wooden plug. When treated in such manner milk will not churn. Sour milk. — ^While the souring of milk does not affect the fat, it is impossible to obtain a representative charge from curdled milk without special treatment. In order to obtain a good mixture, it is necessary to dissolve the curd. This may be accomplished by adding 5 or 10 per cent by volume of a strong solution of either caustic soda or potash; strong ammonia water may also be used. The alkali must be thoroughly mixed with the milk imtil it is completely liquid. The charge for test must be immediately meas- ured out and a correction made in the final percentage for the volume occupied by the alkali solution. If desired, the powdered alkali may be added directly to the milk in small portions at a time, being sure that one portion is dis- solved before another is added, and agitating until the milk has become liquid. No correction is necessary for the volume occupied by the powdered lye. When making a fat test on milk containing alkali, special precautions must be observed in adding the sulphuric acid, as an ex- cessive amount of heat is generated and the contents of the test bottle may be thrown out. When alkali is used, slightly more acid is required. THE BABOOCK TEST. The Babcock test for fat in dairy products, named for its inventor, Dr. S. M. Babcock, chief chemist of the Wiscon- sin agricultural experiment station, is based upon the fact that strong sulphuric acid will dissolve the serum solids in milk and set the fat free from its emulsion. In conducting the test the charge is placed in a specially constructed test bottle and mixed with the proper quan- 83460''-17 2 8 tity of sulphuric acid. The acid performs other functions than the simple solution of the serum solids. Much heat is developed by its action, and this causes the fat globules to lose their individuality and run together, a condition which greatly facilitates the separation from the serum, and this separation is still further accelerated by the ^ _I0 =-9 E-8 =-7 5-5 5-4 m =2j =4J aj d dj Fig. 1.— Old type of Babcock milk-test bottle. C&-J ^--^ Fig. 2.— Type of Babcoek milk-test bottle conforming to the requirements of the United States Bureau of Standards, and showing graduations. increase in specific gravity of the serum caused by the presence of the heavy sulphuric acid. When the solution of the serum solids is effected, the complete separation of the fat and serum is accomplished by whirling in a centri- fuge. The fat is gradually driven into the graduated neck of the bottle and the percentage read directly. / Test bottles. — The Babcock-test bottle for milk, as shown in figure 1, consists of a body holding about 50 eubic centi- meters and the neck graduated so that the percentage of fat may be read directly. Seventeen and one-half cubic centimeters are used in the test, and this volume of average milk weighs almost exactly 18 grams. At the temperature at which the bottles are standardized the specific gravity of butterfat is about 0.9. Two cubic centimeters weigh twice 0.9, or 1.8 grams, which is just one- tenth of the weight of the charge used in the test bottle. The volume between and 10 per cent in the neck should, there- fore, be 2 cubic centimeters, if the bottle has been correctly standardized. Each unit per cent is represented by a volume of 0.2 cubic centimeters in the neck. The old types of bottles were 10 per cent bot- tles, the smallest subdivision being 0.2 per. cent. In the more recent typesi, notably those made to conform to the specifica- tions of the United States Bureau of Stand- ards, the necks are somewhat smaller in diameter and read only to 8 per cent, and the smallest subdivision is 0.1 per cent. (Fig. 2.) The 8 per cent bottle is consid- ered the more accurate of the two, and has come into more general use. Milk pipette. — The charge for the Bab- cock test for milk is measured rather than weighed, the measuring instrument being a pipette graduated to deliver 17.5 cubic centimeters of milk. These pipettes, filled to their graduation mark, hold 17.6 cubic centimeters. The extra 0.1 cubic centimeter is allowed for the milk which clings to the walls. Pipettes may be obtained which conform to the requirements of the United States Bureau of Standards. (Fig. 3.) U Fig. 3.— Pipette holding 17.6 cubic centi- meters, used in measuring milk in the Babcocktest. 10 Acid measure. — This may be either a simple glass cylin- der graduated to deliver 17.5 cubic centimeters^ or one of the more complicated devices shown in figures 4, 5, and 6. A convenient little device is the small glass dipper (fig. 7) by which the proper quantity of acid may be dipped out of a larger container and poured into the test bottle. The centrifugal machine. — This is commonly called the Babcock tester, and various types are on the market, rang- 17.5 c.c 17' Fig. 4.— Simple acid graduate. Fig. 5.— Burette for measuring the acid. Fig. 6.— a combined bottle pnd acid measure. ing from the small, two-bottle hand tester to the large steam turbine or electric tester, accommodating 24 or more bottles. (See figs. 8, 9, 10, and 11.) They all con- sist mainly of a horizontal revolving disk or wheel pro- vided with swinging sockets to hold the bottles. At rest these sockets allow the bottles to stand upright, but when in motion, the centrifugal force causes the sockets to swing outward, bringing the bottles to a horizontal posi- tiouj with the necks toward the center. Where steam 11 pressure is available, a steam turbine tester is strongly recommended for the reason that it maintains a uniform motion under a definite pressure and at the same time the steam keeps the bottles warm and supplies the hot water required. Whatever kind of tester is used, it must be firmly secured to a rigid support. There must be no shaking or trembling of the tester when in motion. ^ac?.— The acid used in the Babcock test is the commer- cial sulphuric acid, sometimes called oil of vitriol, and should have a specific gravity of between 1.82 and 1.83. It should be kept in glass bottles or jugs, preferably with glass stoppers. Rubber stoppers will last for a time, but the use of cork stoppers is not permissible, as cork is rapidly attacked by the acid. Owing to the property of sulphuric Fig. 7.— a dipper made entirely of glass and holding 17.5 cubic centi- meters for measuring acid in the Babcock test. acid of absorbing water from the air and thus diluting itself, it can not be kept in open containers. Sulphuric acid is an extremely corrosive liquid, which attacks the skin, the clothing, wood, and most of the com- mon metals. Should the acid be spilled on the clothing, it should be immediately washed off with plenty of water, and ammonia water applied; this in turn must also be washed off. Unless the acid is washed off immediately after contact with the skin, severe burns will result. Acid spilled on the table or floor may be neutralized with wash- ing soda or other alkali. Lead is the only common metal not attacked by this acid. If much testing is to be done, it is a good plan to cover the testing table with sheet lead. Testing strength of acid.— As already mentioned, the specific gravity of the sulphuric acid used should be between 1.82 and 1.83. It is much better to purchase it guaranteed of the proper strength than to bother with diluting the stronger acid. Creamery supply houses han- 12 Fig. 8.— a 2-bottle hand tester. die acid guaranteed to be of the proper strength, and if kept in well-stoppered containers it will not change. For the benefit of those who prefer to test the acid themselves, the following directions are given: Use of the add hy- drometer. — This is a hydrometer designed only for liquids hav- ing a specific gravity about that of concen- trated sulphuric acid. (See fig. 12.) It is standardized at 60° F. , and for correct results must be used with acid at that temperature only. The acid at this temperature is poured into the hydrometer cylinder and the hydrometer allowed to float in it. When it has come to rest, the point on the scale intercepting the surface of the acid indicates the specific gravity. If it is much under 1.82 it can not be used for testing milk, and should be discarded and a fresh lot of acid obtained. If it is above 1.83 it may be diluted with water until it is of the proper strength. There are two ways of doing this. The acid may be exposed to the air un- til it absorbs sufficient water to lower its specific gravity; this is the safest and best way if the specific gravity of the acid is not much above the standard. The second way is to mix the acid with a small quantity of water. A small quantity of water is placed in a bottle or jar and the acid poured into it. Never pour water into acid, as a serious accident may result. After the mixture has cooled to 60° F. it is again tested with the hydrometer and the process repeated if necessary. Fig. 9.— a hand tester for 12 bottles. 13 DIRECTIONS FOR MAKINO THE BABCOCK TEST WITH MILK. Measuring the charge. — Directions have already been given for preparing the sample for the test. The milk is poured from one container to another two or three times. The tip of the pipette is immediately in- serted and the milk sucked up with the mouth until it reaches a point well above the graduation mark on the stem; the dry fore- finger is then quickly placed over the mouth of the pipette. By slightly relaxing the pressure of the finger the milk is al- lowed to flow down until it just reaches the mark. The tip of the pipette is Fig. 10.— a type of steam tester with an arrangement for heating the water used in the test. now placed in the neck of the test bottle and the milk allowed to flow slowly down the side. The right way is to hold the pipette obliquely to the mouth of the test bottle as shown in figure 13. The wrong way is shown in figure 14. If the bottle and pipette are held in the latter position the neck of the bottle may clog up and some of the milk run over the top. Care must be taken that none of the milk is lost during the operation. "When -A type of electric tester. ^^^^^^ ^j^ ^^^ ^^^ ^^^ run out of the pipette, the last drop is forced out with a puff of the breath. Adding the acid.— The temperature of the milk when the acid is added should be between 60° and 70° F., and the acid should be at about the same temperature. Seven- FlG. 11.- 14 teen and one-half cubic centiineters of the acid are meas- ured out, and, with the bottle held at an angle, carefully poured down the side, the bottle being turned slowly at the same time so that any milk adhering to the neck will be washed down . For two reasons the acid must not be poured into the middle of the test bottle; first, because it Fig. 12.— Hy- drometer and cylinder used in testing sul- phuric acid. Fig. 13.— The right way of add- ing milk to the test bottle. (Farrington and Well, Testing MUk and Its Products.) may form a plug in the neck, which may be driven out by the expansion of the air below ; and second, because the acid may partially mix with the milk and produce black particles which do not dissolve and later interfere with the reading of the test. The acid and milk should now be in two distinct layers without much of a dark layer between them. 15 Mixing the acid and the mill. — The acid is now mixed •with the milk by giving a combined rotary motion and gently shaking with the hand grasping the neck of the bottle, with the mouth of the bottle held away from the operator. When once commenced the mixing must not be interrupted until the solution is complete. The first effect of the acid on the milk is a curdling, wliich is sub- sequently dissolved. As the solution progresses the color changes first to a light yellow, then to dark yellow, then through various shades of vio- let to brown and finally to dark brown, if the acid is of the proper strength and the milk and acid are at the right tem- perature when united. Too strong or too warm acid pro- duces a dense black. If the milk has been presented with formaldehyde, a longer time is required to complete the solu- tion, owing to the toughening of the casein by that preserva- tive. Common errors of begin- ners are failure to mix the acid thoroughly with the milk and to continue the shaking imtil the solution is complete. A good plan is to shake the bottle for a minute or so after the so- lution is apparently complete. Although not necessary, it is preferable to centrifuge the bottles immediately, though they may be kept 24 hours if desired, in which case they must be placed in water from 170° to 180° F. for 15 to 20 minutes before whii'ling. Centrifuging the bottles. — The bottles are now placed in the sockets of the centrifuge, taking care that they are 83460—17 3 ^ / Fig. 14.— The wrong way of adding the milk to the mUk bottle. (Farrington and Woll, Testing Milk and Its Products.) 16 equally distribiited about the wheel or disk so that the equilibrium of the latter is not disturbed. An even number of bottles should always be whirled. Should an odd number of tests be made a test bottle filled with water may be used to balance the machine. When the bottles are in place, the tester is covered in order to keep the bottles from getting cold and to protect the operator from flying glass and acid should any of the bottles break. The tester is now set in motion and the bottles whirled 4 to 5 minutes at proper speed. This will be sufiicient to bring prac- tically all the fat to the surface. In cold weather, if a hand tester is used, it may be necessary to pour hot water into the jacket of the tester to keep the bottles warm. Speed of centrifuge. — Farrington and Woll have calcu. lated the proper speed of testers with wheels of different diameters to be as follows: Revolutions of wheel Diameter of wheel in inches: per minute. 10 1,074 12 980 14 - - 909 16 848 18 800 20 , 759 22 724 24 693 Adding the water. — ^With the pipette or with the device for the purpose attached to some steam testers, or in any other convenient manner, hot water is added to the bottles until the contents come nearly to the lower part of the neck. The cover is now replaced on the tester and the whirling repeated for two minutes. Hot water is again added until the fat reaches a point below the highest gi-aduation mark on the neck. It must never reach the top mark, or some of the fat may be lost. This time the water should be dropped directly into the fat in order to clear the fat of the light, fiocculent material which may be entangled in it and which would later interfere with the reading of the test. The whirling is repeated for another minute. The temperature at which the readings are taken is between 130° and 140° F., and this should be borne in o 17 mind -when the water is added, the object being to add the water at such a temperature that the temperature of the fat' at the close of the last whirling will be between these two figures. The water used should preferably be soft water or con- densed steam. The use of hard water is liable to cause trouble on account of its carbonates; these are decomposed by the acid, liberating carbon dioxid, which may cause foam on the top of the fat column and obscure the menis- cus. If soft water or condensed steam is not available, hard water may be used if, before heating it, a few drops of sulphuric acid are added. Reading the -percentage.— li the test has been successfully conducted, the fat will be in a clear, yellowish liquid column sharply separated from the clear and nearly color- less acid solution immediately below it and with no foam on top. The bottles should be kept warm either in the tester or in warm water until read, and the readings should always be made at between 130° and 140° F. The fat at this temperature will, if other conditions have been cor- rect, have a well-defined meniscus at both the top and the bottom. The readings are made from the extreme bottom of the lower meniscus to the extreme top of the upper men- iscus. Figure 15 shows this graphically. An ordinary pair of dividers is useful in making this reading. The points are placed at the upper and lower limits, then lowered until one point is at the mark ; the other point will indicate on the scale at the correct percentage for the sample tested. In some steam testers where the exhaust steam escapes into the jacket and no ventilation is provided, the tem- perature of the bottles will be too high. In such case, the bottles must be allowed to cool to 130° to -140° F. by plac- ing them in water at that temperature for several minutes before making the reading. Imperfect tests. — If the foregoing directions have been strictly followed, a perfect test should result. It is not to be expected, however, that the beginner will always meet with success. The next two paragraphs may be helpful in locating the trouble. 18 6.. c- -c/ a.. An imperfect test is caused by one of three things: (1) Foam on the fat column obscuring the upper menis- cus; (2) a dark-colored fat column containing dark parti- cles and with dark particles obscuring the lower meniscus; (3) a light-colored fat column containing white, curdy- material obscuring the lower meniscus. The first is caused by using hard water. Any one or a combination of the following may cause the second trou- ble: (a) The acid was too strong; {b) too much acid was used; (c) the acid was too warm when added to the milk; {d) the milk was too warm when the acid was added; (e) the acid was dropped directly into-the milk ; (/) the mixing of the acid and the milk was interrupted before the solution was complete; or (g') the acid and milk were allowed to stand too long in the test bottle before being mixed. The tliird trouble is caused by one or more of the following: (a) The acid was too weak; (b) too little acid was used; (c) the acid was too cold when added to the milk; {d) the milk was too cold when the acid was added; or (e) the mixing .was not continued long enough to dissolve all the serum solids. Tested Bahcock glassware. — Babcock-test bottles and pipettes should always be tested and found correct before being used. It is now possible to purchase test bottles and pipettes which have been tested and approved by the United States Bureau Many States also have officials empowered to test and approve Babcock glassware. The best way is to purchase it already tested by the Bureau of Standards, or to have it made to conform to the requirements of that bureau and then tested by a State official. TESTING CREAM FOR FAT. While in a general way cream is tested by the Babcock test in much the same manner as milk, there are some Fig. 15.— Show- ing metliod of reading fat c olumn in milk testing. Read from a to 6, not o to c, nor o to d. of Standards. 19 modifications that must be observ^ed. The range of fat in cream, and consequently the specific gravity, is much greater than in milk, so that 17.5 cubic centimeters do not necessarily represent 18 grams, as in the case of milk. Cream also varies in consistency, some being thin and some thick; therefore in some cases much more would adhere to the walls of the pipette than in others. For these rea- sons cream can not be accurately measured. The charge for the test must be weighed into the test bottle. Cream-test bottles. — The cream-test bottles used in the Babcock test are of various designs. (See figure 16.) Those conforming to the requirements of the United States Bureau of Standards differ from milk bottles only in the gi'aduations and in the length and diameter of the neck. Test bottles are made for both an 18-gram and a 9-gram charge. Cream-test balances. — Several types of balances designed for weighing cream charges are on the market (figs. 17,18, and 19). The small torsion balances prove to be very sat- isfactory if care is taken that the important metal parts are not allowed to rust. Balances should be tested for sensitiveness from time to time and should always be kept in perfect condition. Preparing cream for testing. — The point never to be lost sight of in testing cream or milk is that the small quantity taken for the test must be truly representative. No matter how carefully the test is carried out, if the charge taken does not accurately represent the cream or milk to be tested, the results will be worthless. The prepai'ation of cream for testing does not differ materially from that of milk. The fat must be evenly distributed, and if there are no lumps this can be accomplished by poiuing from one receptacle to another, warming the cream slightly if cold. If lumps are present, it has been advised to pass the cream through a fine sieve, rubbing the lumps through with the fingers and then mixing as usual. If the cream has stood for some time in the sample jar, the top may have become hard, leathery, and difficult to remove. In this case, the jars should be set in warm water until the contents have reached 100° to 110° F., when the cream will be soft and can be easily removed. 20 Fig. 16.— Types of 9-gram and 18-gram cream bottles conforming to the reqitirements of the United States Bureau of Standards. 21 Weighing the charge. — After the sample has become homo- geneous throughout, the charge is quickly weighed into the test bottle. The weight of the charge depends upon the Fig. 17.— Type of knife-edge cream balance. style of bottle used. For this purpose the 9-gram bottle is recommended. A pipette is useful in conveying the cream to the test bottle, as the flow can be easily controlled Fig. 18.— Type of torsion balance for single bottle. and checked on the drop when the pointer of the balance indicates that the correct quantity has been run in. This weight must be exact, and some experience is necessary before the charges can be quickly and accurately weighed. 22 Completing the test. — -Instead of adding a measured quantity of sulphuric acid to the cream in the test bottle, as is done with milk, the best way is to add the acid until the mixture assumes the color of coffee to which cream has been added. ^ The quantity of acid required to produce this color varies with the percentage of fat in the cream. If the cream and acid when mixed are about 70° F., about one-quarter or one-half the regular quantity (4 to 8 cubic centimeters) of acid (specific gravity 1.82 to 1.83), depend- ing upon the percentage of fat, will be required for a 9- gram charge. After adding the acid to the cream, the Fig. 19.— Type of balance for several bottles. procedure up to the reading of the percentage is exactly the same as in the milk test. After the final whirling, the test bottles are submerged to a point above the fat column in water at 135° to 140° F. in a suitable tank. After re- maining in this tank for about 15 minutes they are re- moved and the readings quickly made. The important difference between reading the cream test and the milk test is that in the cream test the fat column included is from the bottom of the lower meniscus to the bottom, not the top, of the upper meniscus. (See fig. 20.) 1 O. F. Hunziker and H. C. Mills, Testing Creamfor Butter Fat, Indi- ana Agricultural Experiment Station, Bui. 145. June, 1910. 23 K^ Some operators prefer to destroy the upper meniscus by dropping into the bottle at this point a few drops of a liquid in which the fat is not soluble. Glymol (petrolatum liquidum, U. S. P.), known commercially as white min- eral oil, gives satisfactory results and may be purchased at almost any drug store. If desired it may be colored with alkanet root.^ If glymol is used, the fat column included in the reading is from the bottom of the lower meniscus to the line between the fat and the glymol. If the fat column is read with the upper meniscus intact, care must be taken that the eye is on a level with the points on the scale at which the readings are made; otherwise an error will be introduced. 16 /5 /