The book of ice-cream

The determination of the composition of dairy products is a very simple yet important part of the ice-cream business. The fat and the solids not fat are the constituents usually determined. The most accurate method by which to buy all raw materials is on the basis of their composition. The finished product should be tested to determine its composition, regardless of whether or not it is sold on this basis. The testing of the finished product is necessary to check up the amount of materials used. By this means an accurate cost account can be kept and any variation in the composition of the product quickly discovered.

THE BABCOCK TEST⁠[43]

The amount of fat in milk, cream and skim-milk can be ascertained quickly and accurately by means of the Babcock test. The essential requirements are that the operator be careful not to break the glassware and that the measurements are accurately made.

142. Testing whole milk for fat.

—The sampling is the most important operation of the test. “The sample to be tested should be thoroughly mixed before it is measured out. Mixing is done by shaking the vessel in which the milk is contained, or better still, by pouring the milk from one vessel into another. The fat in milk is lighter than the other constituents and soon rises to the surface. Unless great care is exercised an unfair sample will be taken. If the sample is an old one, such as a composite sample, it should be heated to a temperature of not over 85° F. in order to soften the fat. The sample should not be heated above 85° F., since the fat is likely to separate in the form of an oil and when so separated it is impossible to remix it evenly throughout the sample.”

143. Composite samples of milk.

—The purpose of taking composite samples is to reduce the labor and expense of testing. The true composite sample consists of aliquot portions of milk of several deliveries from the same patron. Composite sample jars must have a tight seal in order to prevent evaporation of moisture. Pint jars sealed with glass stoppers, cork stoppers, metal caps or screw tops may be used for this purpose. Bottles with paper caps and jelly glasses with tin lids do not furnish tight seals; they should not be employed.

A separate jar is used for each patron and each jar must bear the respective patron’s number. The jars should be thoroughly clean and, in order to guard against errors, they should be arranged on convenient shelves near the weigh-can in numerical order, grouping the jars of patrons of the same route together. Correct composite samples may be obtained by the use of a milk thief or a graduated pipette. If the milk thief is used, it is inserted into the weigh-can of milk of the entire delivery of one patron. The milk in the tube rises to the level of that in the weigh-can. The milk thief is then emptied into the sample jar. In case the graduated pipette is employed, a certain quantity of milk is taken for every pound delivered by the patron (usually about .1 cubic centimeter for every pound of milk delivered). The milk thief is the handier instrument of the two, but when the amount of milk delivered by different patrons varies considerably, the samples of milk from the larger producers are often too ample to be practical.

Other so-called composite samples are taken by using the same measure for all milk receipts. In this case a small dipper holding about one ounce is commonly employed. With this dipper a sample of milk is taken daily from the weigh-can of each patron’s milk and transferred into the sample jar. This method of composite sampling is not mathematically correct and the results tend to be less reliable, although experimental data show that the results average practically the same as when aliquot portions are taken. Evaporation causes the percentage of fat and other solids to increase, yielding misleading tests. It also tends to dry the milk on the surface, causing the formation of a tough, leathery layer. In this condition it is difficult to secure a representative portion for the test. This can be prevented: by giving the sample jar a gentle rotary motion after each addition of milk; by replacing the stopper properly after each addition of the milk; and by protecting the sample from excessive heat. Fermentation may be prevented by the addition of a small amount of preservative, such as corrosive sublimate, potassium bichromate or formaldehyde. It is usually best to have the temperature of the milk from 60° to 70° F. when measured into the test bottle; however, variation within reasonable limits will not affect the test since the coefficient of expansion of the milk is not high enough appreciably to affect the amount measured by the pipette.

144. Measuring the sample.

—The instrument for measuring the milk for the test is called a pipette. (Fig. 62.) It has only one graduation, 17.6 cubic centimeters, equivalent to 18 grams. The sample is measured by drawing the milk above the graduation and then placing the index finger over the end of the pipette. By carefully releasing the finger, the column of milk can be lowered until the bottom of the meniscus is on a level with the 17.6 c. c. mark on the pipette. It is absolutely necessary that the mark on the pipette be held on a level with the eye, so as to show when the column of milk is on a level with the mark. The milk is then transferred from the pipette to the test-bottle. (Fig. 63.) The pipette and the test-bottle should be slanted so that the milk will run down the bottle neck and not be forced out by the air coming from the bottle. Whole-milk test-bottles are of two kinds, those reading as high as 10 per cent and graduated in fifths, and those reading as high as 8 per cent and graduated in tenths. In each case the graduations give readings directly in terms of percentage, since the graduated part of the neck is made to hold a column of fat which is a definite percentage of the weight of the milk taken.

145. Adding the acid.

—Sulfuric acid is added to the milk in the test-bottle, by means of a special measure which has only one graduation, 17.5 cubic centimeters. (Fig. 64.) The purpose of adding the acid is to destroy all the milk solids except the fat, which it does by moist combustion. In this process great heat is produced. This is advantageous, since the fat must be kept in a liquid condition in order to perform the test properly. The neck of the test-bottle gives percentage readings only when the fat is in a liquid condition. In adding the sulfuric acid, the bottle should be slanted, the same as in adding milk. As the acid is poured in, the bottle should be revolved so that the acid will wash down any milk that adheres to the neck. If this is not done, the milk dries on the neck and is lost in the test; it also causes a cloudy bottle-neck and obscures the fat column when the test is completed. The acid and milk should be mixed thoroughly as soon as the acid is added to the bottle, else portions of the sample might be charred and so lock up small particles of fat. It is well to mix the contents of the bottle for at least half a minute after all the milk has apparently been dissolved by the acid. The mixing is done by holding the bottle by the neck between the thumb and the index finger, and giving it a rotary motion from the wrist (Fig. 65); if an up-and-down motion is used, the contents of the bottle are likely to be spilled.

The strength of the acid is reckoned in terms of its density, which should be 1.82 to 1.83. A special instrument is used for testing the density, and, since this instrument is seldom available in a dairy or a creamery, one of the best ways of testing the acid is actually to perform a test with it and note the results. The acid should be of such strength that it will turn the contents of the bottle to a dark brown as soon as mixed, and the mixture should turn an intense black after standing for about one minute. The best acid is colorless, yet it may be fairly dark and yet be fit for use. The acid should never contain any undissolved material, since this is likely to rise with the fat and obscure the reading.

146. Whirling the sample.

—After the acid and milk are thoroughly mixed, the samples are ready for whirling. The centrifuges used are of three main types (Fig. 66), those driven by hand power, by steam, and by electricity. The steam machines usually are considered best, since with them it is easy to maintain the proper temperature during the process of whirling. The hand and electric machines should perform equally as good work, provided a high enough temperature is maintained to keep the fat in a liquid condition. The frame of the hand machine should always be filled with hot water before the bottles are whirled. In case of the four-bottle machines, which have no frames, the bottle cups, which are made large for that purpose, should be filled with hot water. Great care should be taken to have the machines balanced; by this is meant that for every bottle on one side of the machine there should be a bottle on the opposite side. The machines should also be well oiled, especially those driven by steam, which, because of the heat, soon dry out.

The sample is whirled for five minutes and then filled with hot water to the base of the neck, then whirled for two minutes and hot water again added so as to bring the fat within the graduated part of the neck. The sample is then whirled for one minute in order to bring all the fat into the graduated neck. Some operators of the Babcock test make two separate runs instead of three, filling the bottles to within the graduated neck after the first run. While this may give fairly good results, it is better to make three separate runs as indicated above and fill to the base of the neck the first time. This washes the fat free from any sediment and gives a clearer reading than would otherwise be obtained.

147. Reading the test.

—The sample should be read at once, before the fat column has had time to cool. In reading, the bottle should be held between the thumb and the index finger and the fat column should be on a level with the eye. The fat column in a whole-milk bottle is not large enough to be greatly affected by temperature unless it is extremely hot or cold. With a steam centrifuge, the temperature may be extremely high and thus the reading may be slightly increased. This danger may be avoided by allowing the bottles to stand for a minute at room temperature before reading. There is greater danger of reading the fat column at too low than at too high a temperature. It does not take long for the fat column to harden, and if the room is at all cold it is safer to set the test-bottles in water at about 140° F., having the water come above the fat column in the bottle. The extreme points of the column should be included in the reading (Fig. 67) since this method makes up very closely for minute particles of fat which are not brought to the surface during the process of testing.

There are two methods of reading the percentage of fat in the neck of the bottle. The first is to obtain the difference between the bottom and the top of the fat column; if, for example, the bottom of the fat column rests on the 1.7 per cent mark and the top on the 5.8 per cent mark, the percentage of fat is 4.1 (5.8 - 1.7 = 4.1). The second method of reading is to count the whole percentage and the tenth percentage marks covered by the fat column. Some operators make use of dividers in reading the fat column. The exact space covered is obtained, and one point of the dividers is placed on the zero mark and the other point against the graduated mark. The latter point will indicate the percentage of fat. There is no objection to this method, provided the fat column is never measured when it is above or below the graduations on the neck of the bottle. This is often done, yet there is no certainty that the space above or below the graduations is of the same size as it is within the graduations; in fact, it is usually larger. It is, therefore, easy to see the inaccurate results that may be obtained by taking a reading when the fat column is without the graduated part of the bottle neck.

148. Appearance of a completed test.

—In a completed test the fat should be straw-yellow in color; the ends of the fat column should be clearly and sharply defined; the fat should be free from specks and sediment; the water in the neck just below the fat should be clear; and the fat should be in the graduated part of the neck. Some of the defects and remedies are explained in the following paragraphs.

If the fat column is too dark in color, the acid may have been too strong, or too much may have been used, or the temperature of the milk and the acid may have been too high just before mixing, Mixing too slowly might also permit charring of part of the fat. The charred or darkened condition of the fat may be corrected to some extent by using less acid, by cooling both milk and acid below 60° F. just before mixing, and by rapid, vigorous mixing continued for about a minute after all casein has been dissolved.

If the fat column is too light in color, the acid was either too weak or too cold. This condition may be corrected to some extent in succeeding tests by using more acid and by having the milk and the acid at a slightly higher temperature when brought together.

If the acid is not of the correct strength (specific gravity 1.82 to 1.83), it will be difficult to obtain a correct test, but the trouble may be overcome partially by using more acid when it is weak and less when it is too strong.

149. Care of the test-bottles.

—As soon as all the bottles are read, they should be emptied. If allowed to stand until cold, they are more difficult to clean. The cleaning will be accomplished much more easily if the bottles are shaken violently up and down as the contents run out. A viscous sediment is formed by the action of the sulfuric acid on the milk, and the hot acid helps to loosen this if the bottles are well shaken. All Babcock glassware should be kept clean and bright. This can be done best by washing in hot water and washing-powder, and then rinsing in hot water. If many bottles are employed, a block and a top-board are very useful. The block has holes bored in it, of a size just large enough to hold the bottles, and it may be made to contain any desired number. The holes in the top-board are large enough to admit the passage of the necks through them and the board rests on the shoulders of the test-bottles. In using this block and board, a number of bottles can be emptied at once and the hot bottles will not burn the hands of the operator.

150. Testing cream.

—In testing cream there are three main factors to be considered: first, taking the sample; second, getting the correct quantity of cream into the test-bottle; and third, correct reading of the completed test.

151. Cream testing apparatus.

—There are several forms and sizes of cream test-bottles. (Fig. 68.) The six-inch nine-gram bottles are preferable, especially for use in hand testers. This form has a scale graduated to read from 0 to 50 per cent, the smallest scale divisions equaling .5 of 1 per cent.

The balance for weighing cream test samples should be sensitive to .1 of a gram. There are several different types on the market.

An ordinary four-quart pail would serve as a vat in which to bring the fat in the cream test-bottle to the proper temperature before adding the meniscus remover and reading the test. The vat should be of such depth that when it is nearly full of water and the cream test-bottles are placed upright in it, the upper surface of the water and of the fat columns will be on about the same level.

The thermometer should be of a form that registers each temperature degree between the freezing and boiling points of water. That would permit of its use for a variety of purposes.

152. Sampling cream.

—Cream differs from milk in containing a higher percentage of fat. Cream testing 30 per cent of fat would contain 70 per cent of skimmed-milk substance, or milk-serum. Before sampling, the fat should be distributed evenly by thorough mixing or pouring. If the cream is old or lumpy or some has dried on the container, it should be warmed to about 95° F. and the lumps passed through a strainer before mixing. Then about two ounces should be placed in the sample bottle.

153. Making the cream test.

—The test sample must be weighed instead of measured because:

1. The percentage of fat and the specific gravity of cream vary widely, and the weight of a definite volume would vary accordingly.

2. Cream may contain bubbles of air or of carbon dioxide.

3. Cream varies so widely in viscosity (sticky quality) that the amount delivered or the amount remaining in the pipette would be unknown.

In testing cream 9 grams are used. The bottle should be balanced on the scales, and a 9-gram weight placed on the opposite side. The sample is mixed thoroughly, and by means of a pipette the cream transferred to the test-bottle until the scales exactly balance. About 9 cubic centimeters of water are next added to the test-bottle. (This water may be measured with sufficient accuracy in the acid measure by filling it a little over halfway to the mark.) About 15 cubic centimeters of the acid should be placed in the test-bottle, and the contents mixed thoroughly. The cream and acid mixture should not turn black, but should remain coffee color. About 15 cubic centimeters of acid give the proper concentration to dissolve the solids not fat, since the fat forms such a large part of the mixture and does not go into solution. The bottles should be centrifuged and the water added exactly as in testing whole milk.

154. Tempering the fat and reading the percentage.

—When the last whirling is completed, the test-bottles should be transferred to the tempering vat containing water held at a temperature of 140° F. The water should be tempered in advance, and should be deep enough to surround the necks of the bottles to the top of the fat columns. After four minutes the bottles should be taken from the water, and the meniscus remover added at once by placing the tip of a dropping pipette containing some of the substance against the inside of the bottle’s neck, which is held in a slightly slanting position. The red liquid is allowed to run slowly down the inside of the neck and spread over the fat to a depth of about one-fourth of an inch. It should not mix with the fat.

The meniscus remover is made from a purified mineral oil that has been colored red with alkanet root. It is sometimes called glymol. When placed on the top of a fat column in a cream test-bottle, it flattens the curved surface, which is known as the meniscus. The test should be read immediately by subtracting the number on the scale at the bottom of the fat column from the number at the line of division between the fat and the meniscus remover. (Fig. 69.) Thus, if the bottom line of the fat column reads 12 and the line between the meniscus remover and the fat at the top, 39, the percentage of fat would be 27.

155. Testing skim-milk.

—A special bottle is used for testing skimmed-milk. (Fig. 70.) The graduated neck of the test-bottle has a very small bore in order to measure the fat accurately. A second neck with larger bore is attached to provide a convenient means of filling the bottle. The smallest divisions on the scale usually indicate .01 of 1 per cent, but on some bottles .05 of 1 per cent.

The same care is necessary in mixing and sampling skimmed-milk and buttermilk that is required for whole milk, and the same pipette is used in measuring out the sample. The skimmed-milk is added to the test-bottle through the larger neck. Since a little more acid is necessary thoroughly to free the fat in skimmed-milk, the measure should be filled to about a quarter of an inch above the mark. About one-half of the acid should first be added, and the mixture shaken thoroughly; then add the remainder, and again shake it vigorously for about a minute. One should avoid throwing undissolved casein into the small neck while mixing the milk with the acid. The bottles are then centrifuged and filled in the same manner as in testing whole milk, except that the first whirling should be continued for ten minutes instead of five, in order to bring up all the smaller fat globules. The percentage of fat is read immediately on completing the final whirling.

156. Modifications of the Babcock test for ice-cream.

—The Babcock test as already explained cannot be used to test ice-cream, because it contains a large percentage of sugar. This sugar would char or burn and so interfere with the reading of the test. The following⁠[44] are three modifications of the Babcock test.

157. The glacial acetic and hydrochloric acid test.

—“A representative sample of the ice-cream is taken and melted and thoroughly mixed; a 9 gram sample is weighed into an 18 gram Babcock cream test bottle. A mixture is prepared using equal parts of glacial acetic acid and concentrated hydrochloric acid. Twenty cubic centimeters of this acid mixture is added to the 9 gram sample of ice-cream in the test bottle and is then well shaken. The bottle is placed in a water bath of 120° F. to 130° F., and shaken at intervals until a brown color appears. It is then placed in the Babcock centrifuge and the test completed in the same way as for testing cream and the reading multiplied by two.”

158. The sulfuric acid test.

—“To make the test with sulfuric acid, a 9 gram sample is weighed into an 18 gram test bottle. About 9 cubic centimeters of lukewarm water is then added to dilute the sample, in order to have about 18 cubic centimeters of mixture in the bottle. The sulfuric acid is then added slowly, a little at a time, at minute intervals, shaking well after each addition until a chocolate brown color appears in the bottle. No definite amount of acid can be stated as the quantity will vary with different ice-creams. As soon as the chocolate brown color appears in the ice-cream a little cold water may be added to check the action of the acid. The bottle is then placed in the centrifuge and the test completed in the usual way. The reading is multiplied by two.”

159. Acetic and sulfuric acid test.

—“Weigh a 9 gram sample of ice-cream that has been thoroughly mixed. About 9 cubic centimeters of water are then added to dilute the sample. Add 5 cubic centimeters of acetic acid and then add carefully 6 to 8 cubic centimeters of sulfuric acid. Centrifuge, and then add water the same as in other tests. If using an 18 gram bottle, multiply the reading by two, to obtain the per cent fat in the ice cream. A 9 gram bottle which is graduated to give the percentage of fat directly needs no correction when reading.”

160. The lactometer.

—Because not only the fat but all the solids are utilized in the ice-cream, it is important to know the amount of total solids and the solids not fat in the milk. This is ascertained by determining the specific gravity of the milk and knowing the fat-content; the solids not fat can then be calculated. The specific gravity of liquids is measured by an instrument called a hydrometer. Its use is based on the fact that when a solid body floats in a liquid, it displaces a volume of liquid equal in weight to its own. Hydrometers are in many cases so made that the specific gravity can be read at the point where the scale is even with the upper surface of the liquid. A hydrometer especially adapted to milk is called a lactometer. There are two in common use, the Quevenne and the Board of Health.

The Quevenne lactometer is a long slender hollow piece of glass weighted at the bottom to make it float in the milk in an upright position. (Fig. 71.) The upper end is slender and contains the scale which is graduated from 15 at the top to 40 at the bottom. Each reading on the scale corresponds to the point marked specific gravity on a hydrometer, except that the figures are not complete. For example, 15 on the Quevenne scale means a specific gravity of 1.015; a reading of 30 means a specific gravity of 1.030, and so on. The Quevenne lactometer is graduated to give correct readings at 60° F. The milk should be at this temperature; if above or below this, a correction must be made to the reading. The temperature should not be more than 10 degrees above or below 60° F. The correction for each degree in variation can be made by adding or subtracting 0.1 from the lactometer reading, as the case may be. If the temperature is above 60° F., the correction is added to the lactometer reading; if below 60° F., it is subtracted. The reading should be taken when the lactometer is floating free in the milk. The scale is read exactly at the surface of the milk. The better lactometers have a thermometer with the scale just above or opposite the lactometer scale.

The Board of Health lactometer is very similar to the Quevenne except that the scale is graduated from 0 to 120. (Fig. 72.) The point on the scale that floats at the surface in water is represented by 0, and 100 represents the specific gravity of 1.029. On the Board of Health lactometer the 100 degrees or divisions from 0 to 100 equal 29 divisions on the Quevenne. Therefore, one division on the Board of Health equals 0.29 of a division on the Quevenne. To convert Board of Health reading to Quevenne, multiply by 0.29 and to convert Quevenne to Board of Health, divide by 0.29. The correction for temperature above or below 60 F. is made the same as with the Quevenne, except 0.3 is added or subtracted from the lactometer reading instead of 0.1 as with the Quevenne.

161. Calculating the solids not fat in the milk.

—When the lactometer reading and the fat-content of the milk are known, there are several formulas for calculating the solids not fat. In the following, L equals Quevenne lactometer reading at 60° F.; F. the percentage of fat in the milk, and S. N. F. the solids not fat in the milk.

162. Testing milk for acidity.

—Several tests on the market are used to determine the amount of acid in milk. Each is based on the principle of chemistry, that acids and alkalies tend to neutralize each other. The acidity of milk is of two kinds, apparent and real. The apparent acidity is due to the acid reaction of the acid phosphates and casein. It usually varies from .08 to .1 per cent. The real acidity is due to the action of the bacteria on the milk-sugar. It is usually assumed when determining the acidity of milk, that all the acidity is due to the presence of the lactic acid.

The process by which the acidity is determined is called titration. A known quantity of milk is placed in a cup or flask and an alkali of known strength measured into it by means of a burette. (Fig. 73.) The unit of measure is the cubic centimeter. The burette is usually graduated into tenths of a cubic centimeter. The point at which all the acid in the milk is neutralized by the alkali is told by means of an indicator. The one commonly used is phenolphthalein. This is colorless in the presence of acid and pink in the presence of alkali. If two or three drops of indicator are put in the milk, the color will not change because it is acid. When just enough alkali has been added to neutralize the acid, the color will change to pink. The alkali should be added slowly and gradually the acid will be neutralized by the alkali until at last a uniform pink color appears, which will slowly fade away. All the acid has been neutralized and the amount of alkali used should be read from the burette, when the first change to a uniform pink color is noted.

The different acid tests on the market are sold under various trade names, such as Nafis, Manns, Marschalls and Farringtons. Each is based on the same principle but uses different amounts of milk and alkali solutions of various strengths. However, in each test the amount of milk and strength of the alkali solution are such that the number of cubic centimeters of alkali used are read directly as the percentage of acid in the milk. This eliminates all calculations. If the strength and amount of the alkali solution required to neutralize the acid in the milk is known, and the amount of milk used, accurately measured, the percentage of acid can be calculated.

It is a chemical fact that one cubic centimeter of a normal solution of alkali will neutralize exactly .09 grams of lactic acid. In actual practice an alkali solution weaker than a normal solution is employed. This is because the latter is so strong that only a small amount would be used, hence a small variation in the amount would make a big variation in the final percentage. A ¹⁄₁₀ or ¹⁄₂₀ normal solution (expressed n/10 or n/20) is commonly used. One cubic centimeter of a n/10 alkali solution would neutralize .009 grams of lactic acid. An example will illustrate how to figure the results. Suppose it took 4 c. c. of n/10 alkali solution to neutralize the acid in 18 grams of milk. What is the percentage of acidity in the milk? One cubic centimeter of n/10 alkali will neutralize .009 grams of lactic acid. Four cubic centimeters will neutralize 4 × .009 = .036 grams of acid; .036 grams of acid divided by 18, the grams of milk used, multiply by 100, equals .20 per cent acidity in the milk. This may be expressed thus:

Then the above problem would be expressed thus:

163. Test for formaldehyde.

—Sometimes formaldehyde is added to the milk to preserve it. It can be detected easily when making the Babcock test. The required amount of milk is measured with the pipette into the test-bottle and a few drops of ferric chloride added. The required amount of sulfuric acid is next put in. If formaldehyde is present, a lavender-colored ring will appear between the layer of acid and the layer of milk. If the contents of the bottle are slowly mixed, the dissolving casein will take on a lavender color. The test will not work if the milk is too old or too much of the formaldehyde has been added. Because of the presence of ferric salts in the sulfuric acid as impurities, it is not always necessary to add the ferric chloride although it is best to do so.

164. Test for boiled milk.

—It is often desirable to know whether or not milk has been boiled. The following test will give this result: Two sets of reagents may be used: (1) hydrogen peroxide, potassium iodide and starch, (2) hydrogen peroxide and paraphenylenediamine hydrochloride. In milk there is an enzyme glactase which may be destroyed by heat. When the milk has not been heated, this enzyme sets free the oxygen from the oxidizing agent. In case of the first materials, the glactase splits up the hydrogen peroxide. The free oxygen splits up the potassium iodide and liberates free iodine. The starch in the presence of free iodine turns blue. In the second case, the free oxygen acts on the paraphenylenediamine hydrochloride and turns the solution blue. In either case if a blue color results, the milk has not been boiled. Hydrogen peroxide often contains sulfuric acid. When this is the case, the reagent is useless for the test with starch as the free acid would break up the potassium iodide. If this condition exists a blue color would result, whether or not the milk had been pasteurized.

TESTING BUTTER FOR FAT, MOISTURE, AND SALT

When large quantities of butter are used in the making of ice-cream, it is important that it be tested. Sweet or unsalted butter is best adapted for the making of ice-cream. If a sample is suspected or tastes salty, a test should be made to determine the exact percentage. In order to make cream of a desired percentage of fat, the composition of the butter must be known.

165. Preparing the sample.

—Experiments⁠[45] indicate that the salt and moisture in the butter are not uniformly distributed. This shows the need of careful sampling and preparation of the sample before testing. The sample should be placed in a wide-mouth ground-glass stoppered glass jar. The bottle should be kept stoppered to prevent evaporation. A hardwood stick is best for stirring. The bottle containing the sample to be tested should be warmed to a temperature of 110°-120° F. until the butter is the consistency of thick cream. This may be done by placing the bottle in warm water. While it is being warmed, it should be stirred to obtain a uniform mixture. It should not be heated too much or the water and fat will separate and it is almost impossible to mix them again. When the butter is about the consistency of thick cream, it should be cooled and stirred thoroughly while cooling. This insures a uniform composition in the butter. The cooling should continue until the butter is quite firm.

166. Testing butter for fat.

—After the sample has been prepared to test as outlined above, 3-5 grams of this butter should be weighed into a cream test-bottle. The addition of warm water, warm enough so that the fat will melt, will bring the weight of the butter and water to approximately 18 grams. Sufficient acid to give a light brown color should be added. It will take less acid than for cream because there are fewer solids not fat. The procedure is the same as in testing cream for fat. After the test has been read, the percentage of fat in the butter must be calculated.

167. Testing butter for moisture.

—Several moisture tests⁠[46] are on the market. The following is a very simple one:

The apparatus used is an alcohol lamp, iron stand, asbestos sheet, hot pan lifter, aluminum cup for holding the sample, and a very sensitive scale. To make the test, 10 or 20 grams of the prepared sample of butter as described in paragraph 165, should be weighed into the aluminum cup. The cup should be dry and about the same temperature as the room. The alcohol lamp is then placed under the iron stand and the asbestos sheet on the stand. The lamp is lighted and the cup put on the asbestos sheet. It is well to light the lamp at least two or three minutes before placing the cup on the asbestos in order to heat it and save time. The heat of the flame may be increased or diminished by raising or lowering the wick. The cup should always be handled with the hot pan lifter, as by so doing it will be kept clean and errors in weight due to dirt on the cup will be avoided.

While the sample is heating, it should be shaken from time to time as this breaks up the blanket of casein on the surface and hastens the escape of moisture. As soon as the casein has lost its snow-white color, the cup should be removed from the flame. When the moisture has all been driven from the sample, a slightly pungent odor may be noticed. This may also be used as a guide to tell when the sample has been heated enough. The foam begins to subside at this point. Often one or two small pieces of casein are slow to give up their moisture. This is indicated by the snow-white color of the pieces. Evaporation can be hastened by shaking the sample with a rotary motion and thoroughly mixing these pieces with the hot liquid. If this is not done, one might have to heat the sample so long that some of the fat, which had already given up its moisture, would volatilize.

After all the moisture is driven off, the sample is allowed to cool to room temperature. While cooling, the cup should be covered with something (a sheet of paper will do) to prevent the sample taking up moisture from the atmosphere. After cooling, the cup is placed on the scales. The sample is lighter than before heating, because it has lost its moisture. The loss in weight divided by the weight of butter taken gives the percentage of moisture in the sample of butter.

168. Testing butter for salt.

—The following test has been devised by H. C. Troy of Cornell. The materials used are: One ten cubic centimeter burette graduated to tenths of a cubic centimeter; Babcock milk pipette; one white cup; one pint bottle marked to show the line at the upper surface of the liquid when the bottle contains 300 cubic centimeters; standard tenth normal silver nitrate solution (dissolve 17.5 grams of so-called chemically pure silver nitrate in water and make the volume up to 1000 cubic centimeters); 10 per cent solution of potassium chromate for indicator.

To make the test, three or four ounces of the butter should be softened by warming to a pasty condition in a fruit jar or wide-necked bottle. It should be mixed thoroughly with a table knife or strip of wood in order evenly to distribute the moisture. Ten grams of the mixed butter should be weighed into a dish and washed with hot water into the pint bottle. (If a moisture test was made on ten grams of the butter, the substance remaining in the cup may be used for the salt test.) Enough hot water should be added to bring the surface up to the 300 cubic centimeters mark on the bottle, the stopper placed in the bottle and shaken vigorously for about half a minute. The bottle should rest for about five minutes, and a Babcock milk pipette of the watery portion drawn (17.6 cubic centimeters) and placed in a white cup. Three or four drops of the potassium chromate solution should be added, stirred, and run in the standard silver nitrate solution from the burette, with constant stirring until the color of the substance in the cup changes to a permanent brownish red. On the burette scale the amount of standard silver nitrate solution used may be read.

Each one-tenth of a cubic centimeter of standard silver nitrate solution employed equals one-tenth of 1 per cent of salt in the butter.

169. Test for viscosity.

—It is often desirable to test milk, cream or the ice-cream mix for viscosity. There are several viscometers for the purpose. The simplest way to determine the viscosity is to heat and draw out the end of a pipette so that it has a very small opening. The pipette can then be filled with the material to be tested. The length of time required to empty the pipette determines the viscosity. The more viscous the material, the longer it takes to run out of the pipe. In order to make comparisons, the materials should be at the same temperature each time. This is a very important factor.

170. Standardization.

—One of the main requirements for a successful ice-cream business is uniformity of quality. In order to obtain this, it is necessary to have a product each time containing the same percentage of fat. As it is impossible always to secure cream of a uniform fat-content, the cream and milk used in the ice-cream must be standardized.⁠[47]