( Originally Published 1939 )
In order to make good milk powder, it is necessary to start with milk of good quality. Usually milk powder is made from surplus milk of the bottled-milk supply. The equipment for the cheaper grade of powder (called roller powder) is not expensive, and a unit occupies only a relatively small amount of space. These conditions make it possible for even the small plant operator to convert his sur-plus milk into a good quality of powder where he could not afford to install the larger and more expensive equipment for manufacturing condensed or evaporated milks. Unfortunately, much of the roller powder is of low grade because it is made in small plants which are not equipped with the proper machinery or staffed with qualified operators to produce a uniform grade of high-quality powder.
Spray process. In this process, the milk is sprayed or atomized into a drying chamber where a counter-current of heated air evaporates the moisture, and where the dried particles of milk fall to the bottom of the chamber somewhat like snow. There are several different ways of spraying the milk into the drying chamber, chief of which are the forcing of milk under a pressure of about 1500-4000 pounds per square inch (according to the various processes) through a small orifice, and also the delivery of milk at the center of a rapidly rotating disc where it flows by centrifugal force to the circumference and there is thrown off as a fine spray. Different adaptations of these principles to the several types of drying chambers constitute the distinctive commercial processes. The best known of these are the Gray-Jensen and the Merrell-Soule systems.
In the Gray-Jensen system, the heated milk is pumped into a tank (called a collector) where it is heated by the hot discharged air from the drying chamber, and given a preliminary concentration. From these tanks the hot milk is pumped through an homogenizer at a pressure of about 2500 pounds per square inch into the upper part of a drying chamber shaped like an inverted cone. The descending shower of milk spray encounters a rising current of hot air. The moisture is carried off in the hot air current while the powder falls to the bottom of the cone and is discharged by a ramshorn pump into a hopper for barreling.
The Merrell-Soule process may or may not effect a preliminary concentration. The milk is sprayed under pressures of about 2500-3000 pounds per square inch through one or more nozzles set horizontally at one end or arranged around the four walls of a drying chamber. The heated air enters through ports in the floor or low on a wall. At the end of the day's operations, the operator enters the room and rakes the pile of powder into barrels.
Roller process. The roller process is more standardized. The milk is distributed to the outer surface and top of a horizontally rotating iron drum, heated inside with steam, so that the milk is evaporated to dryness during a part of the rotation of the drum. The sheets or flakes of dry milk are shaved off the roll by a knife which is set against the surface. Sometimes these rolls are arranged in pairs, so close together that the space between them is like a trough which supplies the milk to both the rolls. Here the milk is given a preliminary heating and a degree of concentration before the drying operation begins. A small amount of milk is dried on a steam-heated roll or drum under a vacuum, hence the designation of "vacuum drum" as against "atmospheric roller" powder.
Sometimes milk is given a preliminary concentration in a vacuum pan in order to increase the capacity of the drying plant. Such powder is always darker and less soluble than the powder that is dried without this treatment. After drying, the milk powder is reduced to uniform particle size. Spray powder requires no grinding. It is bolted and then packed into its final containers. Roller powder comes off the rolls in irregular wrinkled sheets. These must be ground to powder consistency, bolted, and then packed—the powder for human consumption usually in 100-pound cans or paper-lined barrels, and that for animal feed in burlap bags.
The powders can usually be classified as spray or roller by their appearance, independent of their laboratory examination, but some roller powders are now so finely ground that they cannot be distinguished from spray powders by merely feeling their texture or by their appearance. However, microscopic examination or a determination of their solubility index serves to distinguish them at once. A spray powder is lighter in color and packs heavier in a barrel than roller powder. Spray powder is more soluble in water and has a more natural flavor when reconstituted to the consistency of normal milk. The properties of vacuum roller powder are intermediate between spray and atmospheric roller powder.
Powder market. Most of the milk powder on the market is made from skimmed milk. The powdering of whole milk is mechanically possible, but the powder itself does not keep well on account of its high content of butterfat, averaging about 30 percent. There is a tendency for this powder to become rancid, especially if copper or iron were picked up during the manufacturing process. Great progress has been made in preparing a satisfactory marketable product by using stainless-steel equipment to preclude the metallic contamination, and then packing the powder in cans in an atmosphere of inert gas, such as nitrogen.
Several brands of standardized natural milk as well as synthetic mixtures prepared especially for infants and sick people are on the market. These are usually made with skimmed milk as a base to which is added a definite and constant percentage of butterfat and sometimes other constituents of special nutritive value.
Relatively very little milk powder is packed for retail consumption. No practical means has been devised to package the milk powder so that it can be marketed in fresh condition and cheap enough to create a popular demand. The channels of human consumption are the baking, the candy, and the sausage trade. The ice-cream industry uses some in times of milk shortage. The product itself may be perfectly sanitary, but it usually imparts a "powdery" flavor to ice cream unless it is fresh.
In order to facilitate commercial transactions and to set up standards of quality as a basis of price schedules, the dry-milk industry organized the Dry Milk Institute, Chicago, Ill. This organization has established and defined the several types and grades of powder, and devised laboratory tests for determining them. Copies of their standards, including laboratory methods of analysis, are available at the request of anyone interested. These grades are described on page 227, and the laboratory procedure for determining them is outlined on page 231.
RELATION TO THE PUBLIC HEALTH
Composition and standards. Dried-milk products have an average composition as tabulated in Table XII.
COMPOSITION OF DRY MILK PRODUCTS
Water Protein Fat Lactose Ash
Dry whole milk 4.0 27.2 26.0 36.8 6.0
Dry cream 0.7 13.4 65.2 17.9 2.9
Dry skim milk 4.0 37.4 1.0 49.2 8.4
The American Dry Milk Institute reports the normal variations in the composition of dry skim milk as shown in Table XIII.
VARIATIONS IN COMPOSITION OF DRY MILK PERCENT
Acidity (lactic on reconstituted basis) 0.09—0.20
Total protein (N X 6.38) 30—38
Lactose (by polariscope) 47—53
Total ash 7—8
The U. S. Department of Agriculture has adopted the following standard for powdered milk:
Dried Milk. The product resulting from the removal of water from milk. It contains not less than 26 percent of milk fat, and not more than 5 per cent of moisture.
Dried Skimmed Milk. The product resulting from the removal of water from skimmed milk. It contains not more than 5 percent of moisture.
The Commission on Milk Standards of the New York Milk Committee went farther and recommended that labeling and sanitary provisions be included in the regulatory control of dried milk. "Recombined milk" or "reconstituted milk" should be the label designation of the product made by dissolving milk powder in water. The Commission holds that all such products should be delivered to the purchaser as meeting all the requirements as to cleanliness, bacterial count, and chemical composition of fresh milk of the same grade or class, and should be labeled to indicate correctly their true character. It made further recommendation that the labels of packages of dried milk declare the quality of the original milk in terms of present grades and standards, the heat treatment, any possibly added product, and the date of manufacture.
The American Dry Milk Institute has adopted quality standards and grades 5 embracing the following provisions for the production and sale of dry skim milk (the production of dry whole milk is very small) :
1. All dry skim milk for human consumption shall conform to all federal and state regulations.
2. It shall be produced in clean, sanitary factories by work-men free of communicable disease.
3. It shall be made from freshly skimmed milk to which no foreign substance has been added (e.g., neutralizing agents, etc.) and which has been pasteurized at a temperature of 142° F. for 30 minutes or its equivalent in bacterial destruction.
4. It shall be uniform in composition and light in color.
5. It shall be sweet and clean in flavor and odor, substantially free from brown specks, and entirely free from rancid, tallowy, fishy, cheesy, soapy, or other objectionable flavors or odors.
6. The package shall protect the powder from contamination with dirt, foreign matter, and odors and moisture.
The Institute then designates the specific requirements for the several grades of powder for human consumption as in Table XIV.
The Institute committee points out that, whereas formerly an inferior quality of product unfit for human consumption had been used for animal feeding, it is now recognized that nutritive quality must be considered also in this industry. The committee has recommended the same sanitary requirements as for powder for human consumption and specific requirements as in Table XV. Milk powder for animal feed is usually packed in bags.
Epidemiology. The only epidemic which has been found by the author to be attributed to powdered milk is an outbreak of enteritis reported by Dick, Dick, and Williams.' A special type of milk powder known as protein milk' had been used in a Chicago institution for homeless infants. Eighty-eight babies became ill and 27 died. The outbreak was found on autopsy to be caused by the Morgan dysentery bacillus. When prophylactic measures were adopted for eliminating contamination from the fingers of the nurses, and using boiled fresh bottled milk, fresh breast milk, or protein milk freshly prepared, the epidemic quickly subsided. Examination of the powdered milk 8 showed that the total count was rather high, that it contained a variety of living bacteria such as spore-bearing sarcinae, staphylococci, Gram-negative bacilli, and lactic acid bacilli, and that the product was heavily contaminated with a green-producing streptococcus. A study of the data shows that no pathogenicity was found or even looked for. This neglect, together with the correction of obviously poor sanitary practices, raises great doubt as to whether the epidemic should be charged to the milk.
Microbiology. Several investigators have studied the effect of the drying process on the microflora of milk. Delephine artificially inoculated milk with tuberculous milk. This milk when injected into guinea pigs caused the development of tuberculosis both before and after roller-drying. Delephine considered that the heat treatment in any stage of the operations, was not sufficient to kill all the bacteria and that there was an appreciable contamination after manufacture.
Not satisfied with the conclusion of this investigation, Hunwicke and Jephcott inoculated two milks with human and bovine tubercle bacilli respectively. These milks were then dried on rolls according to factory practice in respect to steam pressure, temperature, and rate of flow of the milk. They inoculated two samples from each batch into two guinea pigs respectively. All were negative. They then inoculated milks with non-sporing organisms: Escherichia coli, Aerobacter cloacae, an unknown encapsulated organism with fat envelope producing a red pigment, Sarcina lutea, Staphylococcus albus, and long chain streptococci. After drying on the rolls as above, all samples were negative except the A. cloacae which had the excessive count of 2,000,000,000 organisms before drying. Other tests showed that Bacillus mycoides and Bacillus subtilis survived in large numbers. The authors concluded that the survival of some sporing types is not of epidemiological significance because no sporing strains which would be likely to be found in milk are pathogenic when administered orally, although they add that the presence or absence of sporing bacteria in dried milk serves as a valuable indication of the sanitary quality of the original milk, especially as Clostridium welchii is a particularly resistant spore bearer of intestinal origin.
Shrader and his colleagues examined 100 samples of milk powder collected from the original packages (cans, drums, and barrels) in scattered shops in the city of Baltimore, Md. They injected reconstituted samples into guinea pigs and then autopsied them to deter-mine whether any pigs had developed lesions of tuberculosis. Their detailed report shows that there was not a single positive case, indicating that the powder on the market was free of viable tubercle bacilli. Subsequently, Fuller and France found similar results on spray and roller powders.
The milk powdering operation is quite effective in reducing the number of living bacteria in the original milk. Supplee and Ashbaugh have shown that, although the bacterial content of the liquid milk may run from 1 1/2 to 345 million organisms per milliliter, the powder taken directly from the rolls ran from 200 to 1350 organisms per milliliter. They showed that, during the regular commercial operations of sifting, bolting, and packaging, there may be a material increase of bacterial content, but that improved sanitation in plant operations will enable such a plant to operate regularly with a count of only several hundred organisms per gram of powder. They conclude that counts over 1000 per gram (by the Just process) indicates recontamination. Surviving organisms die off rapidly during storage after 2 to 4 months.
The Huckers have studied the bacteriological quality of infant food preparations and conclude that it is reasonable to require that such a product should contain not more than 10,000 organisms per gram, and that the practice of sanitary methods in plant operations will greatly reduce the bacterial content. Giblin and von Pourtales examined nearly 200 samples of milk powder and found none to contain any pathogenic hemolytics or any tubercle bacilli by subcutaneous inoculation into the groins of guinea pigs. The average bacteria counts in the various brands ranged from 1000 to 13,000 organisms per gram.
At the time that the studies of the above investigators were made, the bacteriological quality of some of the milk that went into the manufacture of powder was not so good as it might be. Bacterial plate counts of the powder usually were low, but often the original milk contained excessive counts. The work of Shrader and associates showed that this heavy bacterial contamination is accompanied by a high content of decomposition products indicated by the determination of free ammonia.
Nutritional value. The work by Price, Clark, and Collins 16 has clearly shown that powdered milk has a high nutritive value for infant feeding. They fed modification of cows' milk in comparison with modifications of whole milk powder to more than 200 infants, and found that the powder is safe and in some cases seems to have a distinct therapeutic value.
Tobey has reviewed the studies of several investigators on the feeding of about 600 infants on various dried-milk preparations. In general, they found that the milk is low in bacterial content, is safe, is easily digestible, and produces robust babies. McCord reports 18 that his experience of 14 years in the use of an irradiated dry milk in infant feeding indicates its particular value in the feeding of prematures, in the convalescent treatment of colitis patients, and in the prevention of rickets.
Although the above work has shown that powdered milk is nutritionally adequate, comparative studies have indicated that the drying operation measurably affects the digestibility of the different kinds of powder according to the degree of the heat treatment.
Nevens and Shaw found that the protein of fresh whole milk was more completely digested than that of powdered whole milk. The fat and sugar of the two were equally digestible.
A more extensive study by Fairbanks and Mitchell showed that there was no appreciable reduction in digestibility between the control raw skim milk and the choice commercial roller or preheated spray-process milk powders, although the biological value of the protein (the percentage of the absorbed nitrogen that is retained for both maintenance and growth) was lowered from about 90 for the fresh milk to about 82. The preheat treatment caused a partial destruction of the cystine. When the temperature of drying in the roller process was increased until perceptible scorching occurred, the digestibility was decreased, and the biological value of the milk proteins was lowered to 70 or less, owing to the increased destruction of lysine. They point out that such products were of no value as supplements to the proteins of cereal grains. However, the net energy value did not seem to be appreciably affected. They showed that the solubility of the total solids and the nitrogen products of dry skim milk powders was greater for spray-process than for roller-process powders, and decreased in the latter with the intensity of the drying operations. The brightness of the spray powders was greater than that of the roller powders, and in general this decreased as the heat treatment increased. However, Fairbanks and Mitchell state that these color and solubility differences were not reliable criteria of the changes occurring in the nutritive value of the proteins. This work substantiates the economic soundness of the commercial practice of discounting scorched powder.
Sanitary quality. Generally speaking, the sanitary conditions under which milk powder is made approach those of bottled milk. This is due to the fact that powdered milk is made almost exclusively as a by-product of the surplus milk for bottling. It is not an established industry like the evaporated-milk industry which supports several well-known brands. Milk powder constitutes in large degree the final and cheaper outlet for surplus milk, at whatever price the greatly fluctuating market will bring. The powdering operations are not under official regulatory control, and approved practices are not always followed. However, the milk powder made by some reputable companies under advertised brands are the equal of any bottled milks in sanitary quality of raw supply and plant operations.
Types of spoilage. Powdered milks may spoil through the development of off-flavors associated with rancidity. Whole milk powder spoils so quickly after manufacture that only relatively small amounts are made. Dry skimmed milk develops a tallowy, sometimes chalky, flavor on long standing, especially if the moisture content becomes high. It may absorb moisture and become caked or lumpy. Weevil infestation is occasionally encountered. Black specks throughout a package may be due to soot, dust, and other such debris, or to careless handling on the drying rolls whereby burnt particles of powder may fall into the finished product. Sometimes maladjustment of the spray-drying operations causes the formation of large particles or lumps of powder which may not have been effectively removed in the screening or bolting operations.
Chemical examination. Preparation of the sample. The sample should be sifted through a 20-mesh sieve and stored in an airtight container
Moisture. This determination is made by weighing 1 to 1.5 grams into a previously weighed dish, heating in vacuum oven at the temperature of boiling water to constant weight, under a slow current of dry air.
Fat. A 1-gram sample is accurately weighed into a small beaker, diluted with 10 milliliters of water, and transferred to a Röhrig tube or similar apparatus. Ten milliliters of alcohol are thoroughly mixed in, and the determination of the fat then proceeds as for milk.
The Standards Committee of the American Dry Milk Institute has worked out a laboratory procedure for making the several determinations which fix the grade of a given sample of milk powder.
Sample. The sample of about 1 pound is shipped in single or double friction-top cans and is thoroughly mixed.
Moisture. The water is determined directly by distillation into a graduated receiving tube where its volume is read directly. An alter-native method provides for heating the sample in a vacuum oven at 100° C. and bubbling a slow current of washed, dry air across the sample.
Butterfat. This is the same as the official Roese-Gottlieb method.
Solubility index. A sample of 20 grams is added to 200 milliliters of filtered tap or distilled water, thoroughly mixed, and then centrifuged in a graduated tube so that the volume of insoluble material can be read directly from the graduations.
Titratable acidity. The powder is reconstituted by a solution of 20 grams of powder in 200 milliliters of distilled water, and titrated with 0.1 N sodium hydroxide, with phenolphthalein as indicator.
Foreign sediment and black specks. The sediment is collected onto a disc or allowed to settle to the bottom of a glass container. This is given a rating in comparison with a standard illustration.
Flavor and odor. After the sample has been reconstituted to normal milk consistency, it is examined organoleptically and classified as good, fair, or bad.
Bacteriological examination. The American Dairy Science Association has published a suggested procedure for the bacteriological examination of dry milk.
Sample. Handling of the sample must be done as quickly as possible because milk powder absorbs moisture from the air. Samples for analysis should always be taken at points beneath the exposed surface.
Colony counts. The Dry Milk Institute Method 22 provides for reconstituting the sample by dissolving 10 grams in 100 milliliters of sterile distilled water, and then determining colony counts by the official method. The American Dairy Science Association recommends 23 that 10 grams of powder be dissolved in 90 milliliters of sterile distilled water.
Direct microscopic count. The procedure for liquid milk may be followed for reconstituted dry milk. It cannot be relied on as an adequate procedure for determining the bacterial count of milk powders.
Special microbiological examination. Hemolytic streptococci may be determined on the reconstituted dry milk by means of blood agar plates. Thermophilic bacteria are determined on tryptophane-yeastdextrose agar plates, and incubated at 55° C. for 48 hours 25 Yeasts and molds are determined sometimes as an index of sanitary conditions, and the procedure is the same as for butter.
Control procedure. The same conditions obtain for the regulatory supervision of milk powder as for evaporated and condensed milks. Control must rest exclusively on the examination of samples as the product is offered for sale. Milk-powder plants do not come under the permit system, Und are not regularly inspected by regulatory officials. They may be located far from the metropolitan centers where public-health sanitary regulations are enforced, except those plants furnishing milk for bottling.
1. Associates of Rooms, Fundamentals of Dairy Science, 2nd ed., Reinhold Publishing Corp., New York, 1935.
2. American Dry Milk Institute, J. Dairy Sci., 13, 320 (1930).
3. Service and Regulatory Announcements, Food and Drug 2, fifth revision, November, 1936.
4. Commission on Milk Standards, U. S. Pub. Health Service Reprint 634, 1921.
5. E. C. THOMPSON, A. H. JOHNSON, and M. KL0SER, J. Dairy Sci., 17, 419 (1934).
6. G. F. DICK, G. H. Dick, and J. L. WILLIAMS, Am. J. Diseases Children, 35, 955 (1928).
7. J. A. ToBEY, .17th Ann. Rept. Internat. Assoc. Dairy and Milk Inspectors, 1928, p. 265.
8. G. F. DICK and G. H. DICK, Am. J. Diseases Children, 34, 1041 (1927).
9. S. DELEPHINE, Food Report 21, Local Govt. Board, 1914.
10. R. F. HUNWICKE and H. JEPHCOTT, J. Dairy Sci., 8, 206 (1925).
11. J. H. SHRADER, C. L. EWING, F. A. KoRFF, and L. W. CoNN, Am. J. Hygiene, 8, 386 (1928).
12. J. E. FULLER and R. L. FRANCE, Mass. Agr. Exp. Sta. Bul. 293, 1933, from Chem. Abs., 28, 6863 (1933).
13. G. C. SUPPLER and V. J. ASHBAUGH, J. Dairy Sci., 5, 216 (1922).
14. G. J. HucKER and A. M. HUCKER, N. Y. State Agr. Exp. Sta. Tech. Bul. 154, 1929.
15. J. GIRLIN and J. H. VON POURTALES, Am. J. Diseases Children, 41, 1100 (1931).
16. W. H. PRICE, Pub. Health Repts., 35, 809 (1920) ; T. CLARK and S. D. CoLLINS, ibid., 37, 2415 (1922).
17. J. A. TOBEY, Arch. Pediat., 50, 183 (1933).
18. M. M. MCCORD, ibid., p. 873.
19. W. B. NEVENS and D. D. SHAW, J. Nutrition, 6, 139 (1933).
20. B. W. FAIRBANKS and H. H. MITCHELL, J. Agr. Research, 51, 1107 (1935).
21. Methods of Analysis, 4th ed., Association of Official Agricultural Chemists, Washington, 1935.
22. American Dry Milk Institute, Inc., Chicago.
23. Committee on Bacteriological Methods of Analyzing Dairy Products, J. Dairy Sci., 15, 383 (1932).
24. Standard Methods of Milk Analysis, 6th ed., American Public Health Association, 1934.
25. C. M. SORENSEN, Food Research, 3, 421 (1938).
26. Committee report, J. Dairy Sci., 16, 289 (1933).