UNIVERSITY OF CALIFORNIA COLLEGE OF AGRICULTURE AGRICULTURAL EXPERIMENT STATION BERKELEY, CALIFORNIA FLAVORS OF MILK AND THEIR CONTROL C. L. ROADHOUSE AND J. L. HENDERSON BULLETIN 595 SEPTEMBER, 1935 UNIVERSITY OF CALIFORNIA BERKELEY, CALIFORNIA CONTENTS PAGE Introduction 5 Review of literature 5 Experimental procedure 7 Taste of milk of individual cows 8 Experiments with roughages 11 Influence of alfalfa hay and green alfalfa on the flavor and odor of milk . 11 Effect on flavor of milk when feeding oat hay with alfalfa hay .... 12 Effect of feeding musty cheat hay on the flavor and odor of milk ... 13 Effect of corn silage on milk flavor 13 Effect of green pasture feeds on milk flavor 14 Rapidity of absorption of feed flavor by milk 15 Influence of concentrate feeds on the flavor of milk 16 Wheat bran 16 Dried beet pulp 17 Rolled barley 17 Wet brewers' grains 18 Concentrate feed mixtures 19 Taste and flavors of milk due to abnormal conditions of the cow .... 20 Udder disturbance as influencing the taste of milk 20 Advanced lactation as influencing the taste of milk 20 Rancid flavors in milk 21 Flavors due to oxidation of milk fat 22 Flavors of milk resulting from its contact with certain metals .... 22 Flavors caused by exposure of milk to sunlight 23 Flavor report of milk samples exhibited at the California State Fair ... 25 Flavors absorbed by milk in handling and storage 25 Summary 26 Acknowledgments 28 Literature cited 29 Digitized by the Internet Archive in 2012 with funding from University of California, Davis Libraries http://www.archive.org/details/flavorsofmilkthe595road FLAVORS OF MILK AND THEIR CONTROL C. L. ROADHOUSE 2 and J. L. HENDERSON 3 INTRODUCTION The nutritive value and wholesomeness of market milk are controlled largely by state dairy laws and city ordinances which define the require- ments as to minimum fat, solids-not-fat, and maximum bacteria content. Palatability, however, does not lend itself to legal definition and regu- lation that can be enforced. In determining palatability and aroma, the consumer can employ the senses of taste and smell for the judging of milk quality. If the consumption of fluid milk is to be maintained, the product must be kept uniformly palatable from day to day. In recent years the value of milk in nutrition has been so well explained by scientific workers that officials of the public schools have attempted to stimulate the habit of milk drinking among children. Health authorities and many physicians have encouraged the regular use of milk by adults also, because of its value in nutrition — its calcium and vitamin contents, the quality of its protein, fat, and carbohydrate constituents. If milk has an unpleasant flavor, many persons will not use it, or they will take only limited quan- tities when a larger amount would be beneficial. Children are especially sensitive to off-flavors. Since the milk distributors in large cities do not usually produce the milk that they sell, they should check for taste and flavor all that they purchase and should withhold from distribution any that is not pleasing to them. REVIEW OF LITERATURE The feed consumed by the cow was doubtless the first source of flavors noted by consumers of milk. As early as 1829, William Harley (1)4 in his book, The Harleian Dairy System described a method for "preventing milk from tasting of turnips." Harley's recommendation was to "make a solution of nitre and put a few teaspoonfuls in the milking pail" or else u steam the vegetables before feeding them." He observed, "It is i Received for publication May 11, 1935. 2 Professor of Dairy Industry and Dairy Technologist in Experiment Station, s Associate in the Experiment Station. 4 Superscript figures in parentheses refer to "Literature Cited" at the end of this bulletin. [5] 6 University op California — Experiment Station chiefly common turnips and cabbages that give the strong taste to milk and bntter." In 1914, Ayres and Johnson (2) published a method for re- moving garlic flavor from milk and cream. Gamble and Kelly (3 - 4) apparently initiated the recent interest in the subject with their systematic study of the effect of silage upon flavor and odor of milk. In this respect they reported a wide variation among indi- vidual normal cows receiving the same feeds. Silage fed one hour before milking was reported to be absorbed so quickly that the flavor was no- ticeable in the milk. In order to avoid discernible feed flavors and odors, Gamble and Kelly recommended that silage be fed immediately after milking and that not more than 15 to 25 pounds of corn silage or 15 pounds of legume silage be fed twice daily after milking. Careful and prompt aeration of warm milk was found to remove silage flavors and odors permanently from slightly tainted milk and to reduce the degree of more pronounced silage flavors and odors. During the period 1923-1925, an intensive study of feed flavors in milk was conducted by C. J. Babcock of the United States Department of Agriculture. Investigations reported in a series of bulletins (5_9) con- firmed and extended the work initiated by Gamble and Kelly. Turnips, cabbages, and green alfalfa were found to produce objectionable flavors and odors if fed in the usual quantities to cows one hour before milking. Green alfalfa fed 5 hours before milking had practically no effect. Po- tatoes, green corn, green rye, and green cowpeas fed in the usual quan- tities one hour before milking produced only slight flavors and odors. Proper aeration was found to reduce strong off-flavors and odors caused by feeding green alfalfa and to eliminate slight off-flavors and odors. Garlic flavor and odor were detected in samples taken one minute after feeding garlic but were scarcely perceptible in other samples drawn 7 hours later. Roadhouse, Regan, and Mead (10) in 1926 confirmed the conclusion of Gamble and Kelly that there is a marked difference in the flavor of milk from individual cows. Alfalfa hay and green alfalfa were fed to a group of cows at periods of from 1 to 4 hours before milking. The most pro- nounced flavor was produced when 20 pounds of green alfalfa was placed before the cows 2 hours before milking; a less pronounced flavor when the alfalfa was placed before the cows 3 hours before milking; and a still less noticeable flavor when the feed was given 4 hours before milking. Factors other than feed were observed by various investigators to cause flavor and odor defects. Hammer and Cordes (11) in 1920 reported that exposure of milk to sunlight produced tallowy flavors — a statement later confirmed by Frazier (12) and by Tracy and Ruehe. (13) Henderson Bul. 595] Flavors of Milk and Their Control 7 and Roadhouse (14) found that the induction period of milk fat was short- ened by exposure to sunlight. Lea, (15) exposing milk fat to direct sun- light, observed an increased rate of peroxide formation, indicating oxi- dation. Hunziker, (16) Seligman, (17) Guthrie, (18) Guthrie, Roadhouse, and Richardson, (19) and others have studied, in recent years, the influence of metals on the flavor of milk and milk products. Roadhouse and Koestler (20) investigated the taste of milk freshly drawn from the cow. The effect of feed and other secondary contribu- tors to flavor were eliminated by withholding feed during the 5-hour period before milking. They found the chloride-lactose relation in milk to be one of the most important bases of milk taste. They define taste as the impression perceived when the milk is taken into the mouth; flavor as a combination of sensations of taste perceived in the mouth with those of smell, produced through the medium of the inner nasal passages. This distinction between taste and flavor has been observed in this bulletin. Flavor defects caused by rancidity have been reported in the litera- ture. (21,22) Bacterial activity, aside from the common souring of milk, has also been associated with flavor development in milk. (23,4) EXPERIMENTAL PROCEDURE Taste, flavor, and odor were determined organoleptically, being re- sponses of the sense organs of taste and smell ; they have not been meas- ured by physical or chemical methods. The judging of milk for flavor is paralleled in the evaluation of other beverages, such as tea, coffee, and wine, which are also graded for flavor by the organoleptic method. Ex- perienced judges, accustomed to the same standards for flavors, have agreed remarkably in the scoring of a given set of samples. Individuals vary in their ability to detect certain tastes; some, for example, cannot recognize bitterness, because of an inherited failure of their taste or- gans. (24) Physiologists (25) report that the organs of taste in the mouth can recognize only the sensations of sweetness, bitterness, saltiness, and sourness. Flavors in milk caused by feed, absorbed gases, or products from the deterioration of fat or protein may be detected by the sense of smell alone, especially if the milk is warm. The score card of the United States Department of Agriculture (26) allowing 15 points for a perfect score on flavor was used as the standard for the earlier work reported in this bulletin. The score card now in general use, however, is that of the American Dairy Science Associa- tion/ 2 " which allows 25 points for a perfect score on flavor. In order that the flavor scores throughout the bulletin might be comparable, the earlier scores were converted to the later standard in the following man- ner : A score of 13.5 for flavor on the 15-point score card was considered 8 University op California — Experiment Station to be equivalent to 23 on the 25-point card, since these scores are the minimum established by custom for milk that is above criticism on fla- vor. Class intervals of 0.15 point above and below 13.5 on the first score card mentioned were considered to be equivalent to 0.5-point class in- tervals on the 25-point score card, and conversions were made accord- ingly on this basis. Since the standards of the judges may vary slightly from day to day with their physical condition and, perhaps, other factors, the numer- ical value of the flavor scores should be considered as relative rather than absolute. Control samples, as a rule, were collected, in order to re- duce to the minimum the variation in standard. The control samples were obtained from the same animals before the experimental feeds had been given, and they were collected after all feed had been withheld for more than 5 hours before milking. In the experiments to determine the influence of feed on flavor, only cows producing good-flavored milk were selected. The experimental rations tested were fed in definite amounts, and the cows were milked at definite periods after the feeds were placed before them. The milk was drawn into new, well-tinned pails; and when the oxidation of milk fat was a factor, the milk was stored in amber- colored glass bottles immediately after milking, in order to avoid any influence of sunlight on flavor. When milk was being scored for tastes, flavors, and odors, the samples were heated to a temperature between 100° and 110° F. When the series consisted of more than two samples, the samples were scored as un- knowns to the judges. The quality of the flavor or odor present was desig- nated as (a) desirable, (b) acceptable, or (c) undesirable; while the in- tensity of the flavor or odor was designated as (a) slight, (&) distinct, or (c) strong. In each case, an attempt was made to describe the nature of the taste, flavor, or odor present. In some samples no flavor was evi- dent while the milk was being consumed, but a slight flavor, designated as "after-flavor," remained in the mouth after swallowing. TASTE OF MILK OF INDIVIDUAL COWS Earlier in this bulletin, reference has been made to the variations in the taste of milk produced by different cows. The results of the experiments conducted at this station are shown in table 1, reporting the lactose, chloride, and taste of milk from 24 cows. Peed was withheld from the animals during the 5-hour period before milking in order that the pri- mary taste of the milk (20) might not be masked by secondary flavors such as those caused by feed. The cows were selected from different herds and show variations in the chloride-lactose number and taste scores of the milk. Bul. 595] Flavors of Milk and Their Control The ratio of per cent chloride to per cent lactose multiplied by 100 is known as the chloride-lactose number. (28) A low chloride-lactose num- ber is usually associated with a desirable primary taste of the milk, whereas a high chloride-lactose number, caused by a high chloride and a low lactose content, is ordinarily accompanied by a salty taste. As table 1 shows, the samples of milk with the higher scores on taste had a TABLE 1 The Lactose and Chloride Percentages and Taste Scores of Milk of Individual Cows Cow No. Lactose, per cent Chloride, per cent Chloride- lactose number* Taste score Judges' comments 1 5.34 0.056 1 24 24.5 Excellent taste 2 4.87 0.077 1.58 23.0 Very sweet taste, smooth body 3 5 03 069 1.37 23.0 Pleasing taste 4 4.75 0.085 1.79 23 Good taste, no odor 5 4.84 0.078 1.61 23.0 Very good milk, pleasing taste 6 4.94 0.083 1.68 22.5 Pleasing taste 7 4.74 0.074 1.56 22.5 Very good taste 8 4.98 0.076 1 53 22.5 Slightly astringent 9 4.90 0.107 2.19 22 5 Very good taste 10 4.70 0.120 2.55 22 Very slightly salty 11 5.08 0.075 1.48 22.0 Thin body 12 4.50 0.133 2.95 21 Slight aftertaste 13 4.64 0.105 2.26 21.0 Slight aftertaste 14 4 24 0.138 3.26 21 Slightly salty and aftertaste, not pleasing 15 4.19 0.127 3.03 21.0 Slight aftertaste, slightly salty 16 4.09 131 3 20 21.0 Bitter, slightly salty 17 4.74 0.091 1.93 20.5 Slightly bitter aftertaste 18 4.28 0.148 3.46 20.5 Slightly salty 19 2.97 0.167 5.64 20.0 Salty 20 3.74 0.168 4.50 20 Strongly salty, aftertaste 21 4.68 0.101 2.16 20.0 Disagreeable taste and odor 22 4.82 0.092 1.91 19.0 Disagreeable taste 23 4.16 0.162 3.88 19.0 Salty (udder trouble) 24 4.48 0.120 2.68 15 Rancid * The chloride-lactose number is determined by dividing the chloride percentage by the lactose per- centage and multiplying by 100. (28> relatively low chloride-lactose number. The salty taste is usually accom- panied by a chloride content of more than one-tenth per cent. The rela- tion between the chloride, lactose, and taste scores of the different milk samples may not correspond in all cases, for occasionally a defect in taste or flavor of a secondary nature may be present in addition to a high chloride-lactose ratio. The rancid flavor of the milk of cow 24, table 1, is an example of this condition. After the first two or three months following freshening, the lactose content of the milk of an individual cow usually decreases slightly until the end of the lactation period. In some instances it may decrease to a point that influences adversely the primary taste of the milk. Tests were 10 University of California — Experiment Station TABLE 2 Lactose and Chloride Content of Milk During Different Stages of the Lactation Period Cow No. and period of lactation 130; first three months 130; second three months 130; third three months 473; first three months 473; second three months 473; third three months 473 ; to end of lactation period 421; first three months 421; second three months 421; third three months 421; to end of lactation period Average per cent lactose 4.88 4.76 4.58 5.05 4.96 4.77 4 24 5 05 4.97 4.98 3.77 Average per cent chloride 0.088 0.103 0.106 0.048 0.056 0.076 0.103 0.061 0.071 0.077 0.120 made on three representative cows as shown in table 2. A larger number of cows were tested and they showed similar trends. To determine further the variations in the tastes and flavors of milk from different cows, samples were collected from 536 animals represent- ing- 12 herds that supply the University Farm creamery at Davis with milk for manufacturing purposes. All feed was withheld from these cows during the 5-hour period before milking in order to eliminate its influence on flavor. The results of this survey are summarized in table 3. Of the samples of milk collected from the 536 different cows, 467 TABLE 3 Survey of Twelve Commercial Dairy Herds Giving Percentage of Cows Producing Abnormal Tastes or Flavors Herd Number of cows milked Salty taste Rancid flavor Other "off" tastes or flavors Total abnormal tastes or flavors No. Number Per cent Number Per cent Number Per cent Number Per cent 1 17 2 11.8 4 23.5 0.0 6 35 3 2 35 9 25.7 4 11.4 4 11 4 17 48.6 3 20 0.0 0.0 0.0 0.0 4 17 6 23.5 2 11.8 2 11.8 10 58.8 5 12 2 16.7 3 25 5 41.6 6 36 2 5.6 0.0 2 8.4 4 13 9 7 66 5 7.6 1 1.5 0.0 6 9.1 8 82 4 4.9 3 3.7 3 3.7 10 12 2 9 24 0.0 1 4.2 3 12 5 4 16.7 10 39 0.0 2 5.1 0.0 2 5.1 11 168 3 17.8 0.0 0.0 3 1.8 12 20 1 5 0.0 1 5.0 Total 536 34 17 17 68 Weighte d average 6.34 3 17 3 17 12.68 Buii. 595] Flavors of Milk and Their Control 11 (87.32 per cent) were normal, and 68 (12.68 per cent) produced milk with abnormal tastes or flavors. Of the latter animals, 34 (6.34 per cent) produced milk with salty taste, 17 (3.17 per cent) with rancid flavor, and 17 (3.17 per cent) with other abnormal taste or flavors. The salty tastes and rancid flavors varied from very slight to distinct. These sam- ples were collected during the winter months, and a larger percentage of the cows than usual were advanced in lactation. All samples exam- ined for rancidity were held at room temperature for 6 hours before be- ing scored. EXPERIMENTS WITH ROUGHAGES Judging from the literature published and from the early experiences of the California Agricultural Experiment Station staff, roughages were considered most important among the feeds used in dairy rations in their influence upon the flavor of milk. With this assumption, experi- ments were planned in which the feeds studied were those commonly used in California as roughages. These included oat hay, alfalfa hay, corn silage, and green feeds consisting of freshly cut barley, alfalfa, filaree, foxtail, and wild oats. Influence of Alfalfa Hay and Green Alfalfa on the Flavor and Odor of Milk. — Table 4 shows results of a representative feeding trial in which are reported the influence of alfalfa on the flavor and odor of milk. The group of cows used for these trials consisted of 8 Jerseys, 1 Guern- sey, 1 Ayrshire, and 1 Holstein. In every instance when all feed was withheld from the cows during the 5-hour period before milking, the flavor scores were high; the average score was 24. When 15 pounds of alfalfa hay was given 2 hours before milking, feed flavor was reported in every instance, and the average score was 21.5. The feeding of 34 to 40 pounds of green alfalfa 2 hours before milking gave a strong feed flavor in every instance, with an average score of 20.25. As these results show, alfalfa hay and green alfalfa when fed to cows in full rations 2 hours before milking will cause distinct and undesirable feed flavors. Clover hay and clover pasture, which were also studied, were found to cause a flavor similar to that resulting from the feeding of alfalfa. Table 5 shows how varying amounts of green alfalfa and alfalfa hay, fed to a group of cows two hours before milking, influence flavor and odor. As shown by the data presented, as little as 10 pounds of green alfalfa or 5 pounds of alfalfa hay fed to cows 2 hours before milking will cause a distinct feed flavor and odor, undesirable in market mill?:. When larger quantities of these feeds are given, the flavor is more pronounced. Re- cent experiments of Weaver, Kuhlman, and Fouts (29) with feeding al- 12 University op California — Experiment Station falfa hay before milking confirm former results published from this Station (10) and also the experiments reported in this bulletin. To determine the influence of the time of feeding alfalfa before milk- ing on the flavor of milk, 9 selected cows were given 25 pounds of green alfalfa at intervals of 1, 2, 3, 4, and 5 hours before milking. Control sam- ples were also taken from the same cows when they had received no feed TABLE 4 Influence of Alfalfa Hay and Green Alfalfa on the Flavor of Milk When Fed Two Hours Before Milking Number of cows fed Hours fed before milking Feed used Flavor score Judges' comments 12 5 24 12 5 24.0 12 5 (Controls) 24.0 Very good flavor 12 2 22 sirable 12 2 21 5 sirable 12 2 21.0 able 12 2 Green alfalfa, 34-40 pounds... 20.5 Strong feed flavor and odor, undesir- able 12 2 Green alfalfa, 34-40 pounds .... 20.0 Strong feed flavor and odor, undesir- able for 9% hours before milking. The flavor scores of the milk produced by these animals for each trial were averaged and were for the respective time intervals as follows: for 1 hour, 21.27; 2 hours, 21.31; 3 hours, 22.08; 4 hours, 21.97; 5 hours, 22.66; and for 9% hours, 23.25. In the feeding trials described above, the flavor scores were not strictly proportional to the intervals of feeding before milking. The reason may be that different cows vary in the time required to consume a given quan- tity of feed. Table 6 shows the time that the same cows spent in consuming 25 pounds of green alfalfa. The data presented in this table show the variation in time required by individual cows to consume a given quantity of feed. The first cow finished in about half the time needed by the last cow. Effect on Flavor of Milk when Feeding Oat Hay with Alfalfa Hay. — Some dairymen cooperating with the Station realized that clover, al- falfa, corn silage, and some other feeds affected the flavor of milk if con- sumed by the cows before milking. They inquired whether some other feeds, not contributing to the flavor of milk, could be given at the same Bul. 595] Flavors of Milk and Their Control 13 time as the flavor-producing feeds, thus modifying the flavor defect. Table 7 shows the results of feeding oat hay to cows with and without alfalfa hay. When 5 pounds of good-quality oat hay was fed to the cows 2 hours be- fore milking, there was no noticeable influence on the flavor of the milk. When 8 to 9 pounds of oat hay was given, there was a very slight after- flavor, not undesirable. When 5 pounds of oat hay was fed together with 7 pounds of alfalfa hay, the resulting flavor was not improved over that produced when 7 pounds of alfalfa hay was fed alone. These results indi- TABLE 5 Influence of the Quantity of Green Alfalfa and Alfalfa Hay on the Flavor of Milk When Fed Two Hours Before Milking Feed used Average flavor score Judges' comments 22.0* 21.25* 22. Of 20. 5t 23.5* 23. 5f * Average of 9 cows, t Average of 3 cows. cate that a given quantity of flavor-producing feed, such as alfalfa, con- tains a definite amount of flavor-imparting material and that the flavor is not modified by oat hay, if the same quantity of flavor-imparting feed is given in each case. Effect of Feeding Musty Cheat Hay on the Flavor and Odor of Milk. — Four experimental animals were fed cheat hay in which a musty odor was observed. The milk, when examined, had a musty flavor and odor. When this feed was discontinued, the musty flavor and odor were not observed. Effect of Corn Silage on Milk Flavor. — In feeding trials to determine the influence of moderate quantities of corn silage on the flavor of milk, 5 to 12 pounds of good-quality corn silage was given to cows 1 and 2 hours before milking. Five pounds fed 1 hour before milking did not impart a distinct feed flavor or odor but did cause an after-flavor, notice- able when the milk was compared with control samples from the same cows when the feed had been withheld for 5 hours before milking. Ten pounds of corn silage, fed 1 hour before milking, produced a distinct feed flavor, considered undesirable for market milk. Gamble and Kelly (3) have reported that 30 pounds of corn silage fed 1 hour before milking gave a decided feed flavor and odor. 14 University of California — Experiment Station Effect of Green Pasture Feeds on Milk Flavor. — To determine the in- fluence of other green feeds on flavor, two selected cows producing good- flavored milk were given as much green feed as they would consume dur- ing the 2-hour period before milking. The green feeds used consisted of barley, wild oats, foxtail (Hordeum murinum), and filaree. In the ex- periments previously reported, the maximum feed flavor was present in milk when such feeds were placed before the cows 1 to 2 hours before milking. The cows selected had just previously been used for a different TABLE 6 Variation in the Time Required for Different Cows to Consume 25 Pounds of Green Alfalfa (Average of 5 trials each) Cow No. Average daily milk produc- tion, pounds Minutes required to consume feed Cow No. Average daily milk produc- tion, pounds Minutes required to consume feed 179 6.7 19.9 25.8 27.4 30 7 35 36 37 37 39 403 25.9 29.7 28.5 16.7 42 500 427... 48 501 497. .. 62 494 493. . .. 68 498 experiment in which feed was withheld for 5 hours before milking, and their milk in that experiment was found to be free from abnormal flavor or odor. The quantity of green feed consumed, the amount of milk produced, the flavor scores, and the judges' comments of a representative feeding trial are shown in table 8. The green barley used in this experiment was just coming into head. As shown by table 8, green barley in the quantities used influenced the flavor when given to the cows 2 hours before milking. The feed flavor was distinctly noticeable, being strong enough in some samples to be con- sidered undesirable in market milk. In similar experiments with the feeding of pasture crops other than green barley to the same cows the following results were obtained from four trials with each feed. When an average of 24.4 pounds of green wild oats was consumed 2 hours before milking by cows averaging 14.6 pounds of milk per milk- ing, the flavor score was reduced from 23 to an average of 22.06. The fla- vor of the milk varied from a slight to a distinct feed flavor, in most in- stances considered undesirable in market milk. When an average of 19.9 pounds of green foxtail feed was consumed 2 hours before milking by cows averaging 15.3 pounds per milking, the flavor score of the milk was reduced from 23 to an average of 22.37. A Bttl. 595] Flavors of Milk and Tllfe ir Control 15 slight feed flavor was detected in each case, fout was less noticeable than when the cows consumed 35 pounds of green barley or 24.4 pounds of green wild oats 2 hours before milking. In most cases the milk was ac- ceptable from a commercial standpoint. Foxtail grows vigorously on Pa- cific Coast pasture lands during the spring months. When 30 pounds of green filaree was consumed 2 hours before milking by cows averaging 14.6 pounds of milk per milking, the flavor score was reduced from 23 to an average of 21. The feed flavor varied from dis- TABLE 7 Influence of Oat and Alfalfa Hays on the Flavor and Odor of Milk* Feed used Hours fed before milking Average flavor score of milk Judges' comments (Controls) Oat hay, 5 pounds Oat hay, 8-9 pounds Oat hay, 5 pounds; alfalfa hay, 7 pounds Alfalfa hay, 7 pounds 23.0 23.0 22.5 21.0 21.0 No feed flavor nor odor No feed flavor nor odor Slight after-flavor, slight odor, acceptable Distinct feed flavor and odor, undesirable Distinct feed flavor and odor, undesirable Average scores of milk from 3 cows;'triaIs~with"17 cows in another experiment gave similar results. tinct to strong and in every instance was considered to be objectionable either in market milk or in milk used for manufacturing dairy products. During the spring months, when filaree is present in pasture feeds, the cows should be removed from the fields at least 5 hours before milking if feed flavors are to be avoided. Rapidity of Absorption of Feed Flavor by Milk. — The rapidity with which feed flavors are absorbed by milk after the feed has been con- sumed by the cow is emphasized in an experiment in which alfalfa juice was given to four cows as a drench. (30) The juice extracted 5 from 25 pounds of green alfalfa was given to the cows by forced feeding just be- fore milking. Cows producing milk with good flavor were selected, and all feed was withheld for 5 hours before milking, in order to eliminate other sources of feed flavor. A sample with normal flavor was secured by partially milking the animals before giving the alfalfa juice. Other por- tions of the milk in the udder were drawn every 5 minutes for 30 minutes after drenching. In some instances samples were drawn at the end of 45, 60, and 120 minutes after drenching. The milk samples collected were scored for flavor within 2 hours after the last sample had been drawn. Judging from table 9, sufficient feed-flavor-producing constituents s The freshly cut alfalfa was frozen by placing it in a room at a temperature of -5° C to break the cell walls of the plant tissues. It was then passed through a food chopper. The juice was extracted from the fiber by pressure and then filtered. 16 University of California — Experiment Station are present in the juice of green alfalfa to impart a feed flavor to the milk as soon as 20 minutes after the material is consumed by the animal. When alfalfa juice was given as a drench, the most distinct flavor was present in the milk drawn within 45 to 60 minutes; but when natural green feed was consumed, as shown in previous experiments, the flavor was most prominent in milk drawn approximately 1 to 2 hours after- ward. The absorption of feed flavor into the blood and milk from the natural green feed is apparently less rapid. TABLE 8 Flavor Scores of Milk from Cows Fed Green Barley Two Hours Before Milking* Cow No. Pounds of feed Pounds of milk per milking Flavor scores Judges' comments 421 30 14 5 22 Slight feed flavor and odor, slight after- flavor 607 30 15.4 21.0 Distinct flavor and undesirable after- flavor 421 35 12.6 21.5 Distinct feed flavor, slight odor, unde- sirable 607 35 13.9 21.0 Distinct feed flavor, unde ira! le 421 35 13 21.5 Distinct feed flavor, slight odor, unde- sirable 607 35 12.8 22.0 Distinct feed flavor, slight odor 421 35 15 3 21.5 Distinct flavor, slight odor, undesirable 607 35 15.9 22.0 Distinct feed flavor, slight odor, slight after-flavor * Control samples scored 23 or higher in every case. INFLUENCE OF CONCENTRATE FEEDS ON THE FLAVOR OF MILK Experiments were undertaken to determine the influence of certain con- centrate feeds more commonly used in the rations of dairy cows in Cali- fornia, on the flavor and odor of milk. The feeds studied were wheat bran, rolled barley, dried beet pulp, co- conut meal, soybean meal, cottonseed meal, raisins, wet brewers' grains, and a mixed concentrate ration. Various quantities of each feed were given in order to determine the minimum amount of feed, if any, that would cause a flavor in the milk. The feeds were given to the cows both 1 hour and 2 hours before milking. Wheat Bran. — When three cows, in five different trials, received wheat bran in quantities varying from 4% to 7 pounds 1 hour before milking, the flavor of the milk was pleasing. There was more flavor when the bran was consumed than there was in the control samples produced the pre- vious day when feed had been withheld during the 5-hour period before milking. The flavor, however, was pleasing rather than objectionable. Bul. 595] Flavors of Milk and Their Control 17 Dried Beet Pulp. — The flavor of the milk produced by cows fed 3% pounds of dried beet pulp 1 hour before milking (table 10) was as good as that of the control samples; but when 5% pounds of dried beet pulp was consumed 1 hour before milking, a slight flavor and after-flavor were noticeable, though not objectionable. When 5 pounds of dried beet pulp TABLE 9 Flavor Scores and Comments on Milk Drawn Before and After Drenching Cows with Alfalfa Juice* Cow No. Interval (minutes) after drenching before milk samples were drawn Flavor score Judges' comments Control sample, taken before drenching \ 5 23 23.0 22.5 22.5 22.0 22.0 22.0 23.0 23.0 23.0 23.0 22.0 21.5 21.0 22.0 22.5 No feed flavor 10 No feed flavor, slightly astringent 153 < 15 20 Slight feed flavor 25 30 Feed flavor and odor Control sample, taken before drenching! 5 No feed flavor 10 No feed flavor 15 No feed flavor 428 < 20 Slight feed flavor and odor 30 45 Strong feed flavor and odor 60 120 Very slight feed flavor and odor * Data from two representative cows, only, are given; similar results, however, were obtained when two other cows were given alfalfa juice as a drench. t Five quarts of liquid expressed from 25 pounds of alfalfa was used as a drench. Time required to drench was 5 minutes. Cow was maintained on oat hay and concentrate ration. No feed given within 5 hours before drenching. X Cow was maintained on alfalfa hay and concentrate ration. Other conditions were the same as with cow 153. was fed 2 hours before milking, the flavor was less pleasing than when the feed was withheld for a 5-hour period, but was probably not suffi- cient to be noticed by the average consumer. It is assumed that 5 pounds of beet pulp would be the maximum amount likely to be used in average dairy rations. When 7 pounds of dried beet pulp was fed 1 hour before milking, a slight feed flavor and after-flavor were readily detected. The milk lacked the pleasing flavor observed in the control samples. Similar results were secured when 7 pounds of dried beet pulp was fed 2 hours before milking, but such milk would ordinarily be entirely acceptable to the consumer. Rolled Barley. — The flavor of the milk produced by cows fed 3% and 4% pounds of rolled barley 1 hour and 2 hours before milking (table 11) equalled that of the control samples. Seven pounds gave a noticeable flavor when the samples were examined at 100 °F, but it was not objec- tionable from a commercial standpoint. The flavor was more pronounced 18 University of California — Experiment Station when the feed was given to the cows 1 hour rather than 2 hours before milking. Wet Brewers' Grains. — Since wet brewers' grains have recently be- come available as a dairy feed in certain localities, inquiries have been received concerning their effect on the flavor of milk. To secure inf orma- TABLE 10 Scores of Milk from Cows Fed Dried Beet Pulp One and Two Hours Before Milking Interval Pounds Cow No. (hours) be- tween feeding and milking of dried beet pulp fed (dry basis) Score Judges' comments 165 5 (Control) 24.0 Good flavor and odor 165 1 3H 23.5 Good flavor and odor, slight after-flavor 165 1 5K 22.5 Slight flavor and after-flavor, not ob- jectionable 165 1 7 22.5 Slight but distinct feed flavor 165 2 7 22.5 No distinct feed flavor but not quit equal to control sample 165 2 5 23.0 No feed flavor, not objectionable but not equal to control sample 167 5 (Control) 23.5 Good flavor and odor, no after-flavor 167 1 3H 23.5 Pleasing flavor and odor, very slight after-flavor 167 1 5^ 22.5 Slight after-flavor, slightly astringent 167 1 7 22.5 Slight but distinct feed flavor 167 2 7 22.5 Slight feed flavor, no odor 167 2 5 23.0 No feed flavor, not so pleasing as con- trol sample 475 5 (Control) 24.0 Excellent flavor, no odor 475 1 3H 23.0 Slight feed flavor, not equal to control sample 475 1 5V 2 22 5 Distinct feed flavor and slight after- flavor 475 1 7 22.5 Distinct feed flavor and odor 475 2 7 22.5 Slight flavor, no odor, slight after-flavor 475 2 5 23*5 No feed flavor, but not so pleasing as control tion on this subject, 3 selected cows producing good-flavored milk were given fresh wet brewers' grains in quantities of 12, 15, and 20 pounds, 1 and 2 hours before milking. In every instance the flavor scores of the samples were equal to the controls taken from the same cows when they had no feed 5 hours before milking. The feed used contained approx- imately 23 per cent dry matter. When fed in quantities up to 20 pounds, the wet brewers' grains pro- duced no objectionable flavor or odor in the milk. Because of the high moisture content of fresh brewers' grains, fermen- tations may start quickly and may change the quality of the feed if it is not consumed within a few days after it is taken from the malt vats. This feed also molds readily, thus losing its palatability. The influence of Bul. 595] Flavors of Milk and Their Control 19 these fermentative changes on the flavor of milk was not determined. Concentrate Feed Mixtures. — Some other feeds investigated were not sufficiently palatable when fed alone to be consumed in quantities that would influence the flavor of the milk. They were therefore fed in com- bination with other feeds and are briefly reported in the following para- graphs. TABLE 11 Scores of Milk from Cows Fed Rolled Barley One and Two Hours Before Milking Interval Amount of Cow No. (hours) be- tween feeding and milking rolled barley fed, pounds Score Judges' comments 165 5 (Control) 24 Good flavor and odor, pleasing after- flavor 165 1 3H 23.5 Very slight flavor, no after-flavor 165 1 4^ 23.0 No objectionable feed flavor, slight after-flavor 165 1 7 22.5 Slight feed flavor, not objectionable 165 2 7 23 Very slight feed flavor , not obj ectionable 167 5 (Control) 23.5 Good flavor and odor 167 1 m 23.0 Slight flavor and after-flavor 167 1 VA 22.5 Slight feed flavor 167 1 7 22.5 Slight feed after-flavor, not objection- able, slight odor 167 2 7 23.0 No objectionable flavor, no odor, slight after-flavor 475 5 (Control) 24.0 Pleasing flavor, no odor 475 1 3M 23.5 More flavor than control sample, slight odor 475 1 4^ 24.0 Pleasing flavor, very slight odor 475 1 7 22.5 Distinct feed flavor, not objectionable, slight odor 475 2 7 23.0 Slight feed flavor, not objectionable Coconut meal, 2% pounds with wheat bran 2% pounds, fed 1 hour be- fore milking, caused a very slight feed flavor, probably not objectionable to the average consumer. Soybean meal, 3 pounds with wheat bran 1 pound, fed 1 hour before milking, resulted in a product of very satisfactory flavor. A grain mixture consisting of rolled barley, 100 parts, wheat bran 50 parts, coconut meal 20 parts, salt 2 parts, and bone black 3 parts, fed to each of three cows, in 5-pound quantities, 2 hours before milking, caused a very slight after-flavor in the milk, rendering it less pleasing than the control samples; but the extent of the flavor was considered to be of little importance from the consumer's standpoint. When iy 2 pounds of this concentrate mixture was given 2 hours before milking, the after-flavor was more pronounced. When 4 pounds of the concentrate mixture described in the preceding 20 University of California — Experiment Station paragraph was fed with 1 pound of cottonseed meal, all the samples gave a satisfactory flavor but left an unpleasant after-flavor. One pound of walnut meal added to 4 pounds of the concentrate mix- ture mentioned above did not influence the flavor. When 2 pounds of the mixed concentrate ration was fed with 3 pounds of soybean meal, the flavor was slightly different from that of the con- trols, but was not at all objectionable. When 5 pounds of equal parts of Muscat and Thompson Seedless raisins were fed 2 hours before milking, the flavor of the milk was not affected. TASTE AND FLAVORS OF MILK DUE TO ABNORMAL CONDITIONS OF THE COW Udder Disturbance as Influencing Taste of Milk. — When cows are affected with udder disturbances resulting from inflammation of the udder tissues, the milk may become abnormal in composition, and the taste may be influenced somewhat. After chronic inflammations of the udder, the milk from the affected quarters may taste salty or otherwise abnormal for the remainder of the lactation period. Saltiness is the most common taste defect resulting from this cause. The chloride percentages of the milk in the quarters of the udders of three cows that have not been affected with mastitis was compared with that in five other cows which had had one quarter affected with mastitis. The results are shown in table 12, together with the lactose percentages and the chloride-lactose numbers of the milk produced in each quarter of the udder. Cows 240, 427, and 428 had never been reported by the herdsman as having had mastitis, whereas the other five cows had suffered from the infection during the lactation period in which the milk was analyzed. The infections were chronic rather than acute, and at the time the sam- ples of milk were tested, the milk appeared normal. The milk samples secured from the affected quarters of each cow, however, were salty in taste, had high chloride and low lactose percentages, and consequently high chloride-lactose numbers, when compared with the milk produced in the other quarters of the udders, or with that produced by the cows that had never been affected. Advanced Lactation as Influencing the Taste of Milk. — The milk of animals advanced in lactation, particularly those which have been milk- ing more than a year, often shows changes in composition that give it a less pleasing taste. The salty taste is the most common defect resulting Bul. 595] Flavors of Milk and Their Control 21 from this cause. When it accompanies advanced lactation, the milk from all quarters is affected. The recommended procedure is to let the cow go dry. When the animal again freshens, the milk will usually be normal in composition and flavor. Rancid Flavors in Milk. — Occasionally, rancid flavor and odor are ob- served in milk produced by individual cows. In most instances the milk has a normal flavor when drawn, but becomes rancid 6 to 12 hours later. Milk of this character has been shown to contain an enzyme, (21) lipase, TABLE 12 Chloride and Lactose Percentages in the Individual Quarters of the Udder Cow No. Days from calving Quarter Lactose, per cent Chloride, per cent Chloride- lactose number ( LF 5.12 0.062 1.21 240 (control) 101 1 RF 5.09 0.061 1.20 ] LR 5.11 0.060 1.18 ( RR 5.13 0.060 1.18 f LF 4.93 0.064 1.29 427 (control) 175 1 RF 4.98 0.060 1.21 ] LR 5.00 0.059 1.18 [ RR 5.02 0.061 1.21 ( LF 4.88 0.076 1.56 428 (control) 152 i RF 4.90 0.071 1.45 ] LR 4.87 0.069 1.41 ( RR 4.87 0.070 1.40 f LF* 4.10 0.121 2.95 421 180 J RF I LR 4.90 4.98 0.071 0.075 1.45 1.50 { RR 5.01 0.083 1.66 ( LF 4.71 0.085 1.80 473 154 1 RF* | LR 3.79 4.70 0.135 0.085 3.57 1.81 ( RR 4.70 0.086 1.82 ( LF 4.85 0.093 1.89 226 23 1 RF I LR 4.96 4.94 0.090 0.089 1.82 1.80 [ RR* 4.35 0.125 2.85 f LF 4.38 0.096 2.19 371 94 1 RF ] LR 4.43 4.26 0.097 0.114 2.19 2.68 [ RR* 3.68 0.153 4.10 f LF* 4.43 0.121 2.73 441 146 1 RF 4.76 0.073 1.53 | LR 4.74 0.075 1.58 [ RR 4.70 0.084 1.78 Quarters of udder affected with mastitis. 22 University of California — Experiment Station which changes the fat of the milk by hydrolysis; some of the products of hydrolysis give a rancid flavor and odor. The early pasteurization of such milk destroys the enzyme and prevents rancidity. This condition has been observed more commonly in cows advanced in lactation which have been milking for a year or more without freshening. Cows milked less than a year but soon to freshen may also produce milk that becomes rancid. Such milk has been found in different herds located in several counties in northern California; also in Switzerland, where one of the present authors observed its characteristics. (22) The lipolytic enzyme that causes milk from certain cows to become rancid will cause rancidity in normal milk with which it is mixed. In one instance 2 per cent of ran- cid milk containing lipase caused rancidity within 15 hours in normal milk with which it was mixed. Cream separated from such milk and shipped to the city without pasteurization may become rancid before it reaches its destination. The remedy for rancid cream is the delivery of the milk immediately after each milking to the country receiving station, where it should be separated immediately and the cream pasteurized before shipment to the city. Cows producing milk that later becomes rancid should be eliminated from the milking herds. A sample may be collected from each string of 25 or 30 cows by mixing a small amount of milk from each cow in a quart bottle. The milk collected should be left standing at room temperature for 12 to 15 hours, during* which it will usually become rancid if one or more cows producing milk containing lipase are present in the string. After the string responsible for rancid milk is located, a sample of milk from each cow should be allowed to stand 6 to 12 hours at room tempera- ture and then examined for rancid flavor and odor. The identification and elimination of such cows should overcome rancidity in fresh milk and cream. Market milk would have a more dependable quality if pro- ducers would examine their cows at intervals to identify abnormal tastes and flavors. FLAVORS DUE TO OXIDATION OF MILK FAT Flavors of Milk Resulting from its Contact with Certain Metals. — An oxidized flavor that frequently occurs in milk and cream has sometimes been called metallic because it can usually be traced to the contact of milk with certain metals. The flavor varies in intensity according to the stage of its development. Other terms used to describe it are "cappy," "papery," "oily," and "tallowy." Certain metals are acted upon by the milk so that a small amount of the metal is dissolved. The metallic salts thus formed may give a taste Bul. 595] Flavors of Milk and Their Control 23 properly designated as metallic; some of them, however, act as catalysts, hastening the oxidation of the milk fat. The products of oxidation are the source of the substances that cause the flavor here designated as "oxidized." Oxidized flavor in milk is probably more often observed since pasteuri- zation has become generally adopted, because the metals are more sol- uble in hot milk. Another factor, since refrigeration is more commonly used by the consumer, is that milk is now kept longer, so that the oxi- dized flavor has sufficient time to develop. An extensive study of the influence of certain metals on the flavor of milk has been carried out at this Station. Milk was exposed to the va- rious metals used in dairy equipment. (19) Chrome-nickel-iron consisting of approximately 18 per cent chromium and 8 per cent nickel alloyed with iron, commonly known as stainless steel, when properly fabricated and finished, was not acted on by milk. No flavor change has been ob- served in milk processed in contact with this alloy. Pure aluminum, tin, and glass-enameled metals were shown to have no weight losses when exposed to hot or cold milk; and oxidized flavors did not develop in sweet milk exposed to such surfaces. A chromium-nickel alloy consisting of 12 per cent chromium, 5 per cent iron, and 83 per cent nickel, properly fabricated, has also proved resistant to corrosion by milk and does not influence the flavor. Copper and copper alloys such as ambrac, brass, bronze, monel metal, nickel silver, and Waukesha metal showed weight losses when exposed to fresh milk and caused oxidized flavors in milk that was brought in contact with them. Pure nickel, Ascoloy, and Enduro A were found to cause slight oxidized flavors in the experiments with some of the milk samples. Oxidized flavor occurs more readily when warm or hot milk is brought into contact with such metals. In occasional instances, cows have produced milk that became oxidized after one or two days' storage although there was no contact with metals or with sunlight. The cause of this autoxidation has not been determined. G-uthrie and Brueckner (31) have recently investigated the cow as a source of "oxidized" flavors in milk. As a result of the experiments reported by this Station and by other institutions, several manufacturers have used chrome-nickel-iron and chromium-nickel alloys in milk vats, sanitary pipes, milk pumps, and other dairy equipment. Noncorrodable material is recommended for these purposes whenever it can be practically applied. Flavors Caused by Exposure of Milk to Sunlight. — Several experi- ments have shown that exposure to direct or indirect sunlight soon de- velops in the milk a definite oxidized or tallow-like flavor, recognized by those who know the flavor of normal milk and easily discernible to the 24 University of California — Experiment Station average consumer, though he may not know the cause. When milk in clear glass bottles is exposed to direct sunlight during delivery or, when delivered to the householder, is left standing in sunlight even for a short time, the flavor will change sufficiently to affect palatability. This flavor has been noted in milk exposed to direct sunlight for as short a time as 10 minutes or even to indirect light for 45 minutes during the summer months. Sunshine flavor in milk is largely controlled if the householder takes the milk indoors as soon as it is delivered. Such flavor may result also in the milk-processing plant if the sun is allowed to shine against milk being pasteurized or cooled. The distinction between oxidized flavor resulting from exposure to sunlight and flavor resulting from contact with certain metals is that the former develops during the period of exposure, whereas the latter de- velops slowly, not being noticeable until after 1 to 3 days in storage. Further work should be carried out to determine whether changes other than oxidation of fat account for the rapid development of flavor in milk after exposure to sunlight. TABLE 13 Taste and Flavor Criticisms of 349 Milk Samples Entered at the California State Fair, 1930-1934 Samples submi tted Per cent of samples criticized for and flavor defects aste Per cent of total Year Total entries Per cent raw Per cent pas- teurized Feed flavor Salty taste Rancid flavor Oxidized flavor Heated flavor entries abnormal 1930 64 57.8 42.2 9.4 6 3 28.2 0.0 43.9 1931 74 67.6 32.4 12.1 13 2.7 17.6 2.7 36.4 1932 57 68.4 31.6 10 5 0.0 7 3 29.6 0.0 47 4 1933 63 71.4 28.6 9 5 16 3 2 23.8 1.6 39.7 1934 91 71 4 28.6 16.5 0.0 4.4 22 11 44 Average 67.6 32 4 11.6 0.6 4.8 24 2 1.1 42 3 Distribution of taste and flavor defects among raw and pasteurized entries Feed flavor Salty taste Rancid flavor Oxidized flavor Heated flavor Year Per cent of raw Per cent of pas- teurized Per cent of raw Per cent of pas- teurized Per cent of raw Per cent of pas- teurized Per cent of raw Per cent of pas- teurized Per cent of raw Per cent of pas- teurized 1930 13.5 3.7 0.0 10.8 13.5 48.2 0.0 1931 18.0 0.0 2.0 4 12.0 29.2 8.3 1932 12.8 5.5 0.0 10.2 25.6 38.9 0.0 1933 13 3 0.0 2.2 4.4 20.0 33.3 5 5 1934 23.1 0.0 0.0 6.2 20.0 26.9 3.8 Average 16.1 1.8 0.9 7.1 18.2 35 3 3.5 Bttl. 595] Flavors of Milk and Their Control 25 FLAVOR REPORT OF MILK SAMPLES EXHIBITED AT THE CALIFORNIA STATE FAIR During the five-year period 1930-1934, 349 samples of milk were entered in the California State Fair contests. The judges' criticisms of taste and flavor of these samples, based on the score card of the California State Department of Agriculture, are summarized in table 13. The samples were scored by three judges, one of whom served in the five contests, one in four, and another in three. Approximately two-thirds of the samples were raw milk. The samples were 24 to 54 hours old when scored, their age depending on the location from which they were shipped. They may be regarded as representative of the milk the housewife would use 1 or 2 days after delivery. It will be noted that an average of 42.3 per cent of the entries were criticized for a taste or flavor defect. When these defects are further segregated as to occurrence in raw and pasteurized milks, practically all the feed, salty, and rancid flavors were found to occur in the raw milk, whereas the principal criticisms of the pasteurized samples were those of oxidized and heated. Among the pasteurized milk entries 35.3 per cent of the samples were reported as oxidized; among the raw samples, only 18.2 per cent. As these results show, abnormal flavors in milk are of general occur- rence. The recommendations listed in the summary of this bulletin as to feeding, detection of abnormal milk, and the avoidance of exposure of milk to certain metals should be noted if the general milk supply is to be improved in flavor. FLAVORS ABSORBED BY MILK IN HANDLING AND STORAGE As has long been recognized, milk may absorb the odors of some fruits, vegetables, chemicals, and other materials. In addition, it readily ab- sorbs odors from paint, coal-tar disinfectants, gasoline, kerosene, some fly sprays, and tobacco smoke. The amount of odor absorbed from the presence of these substances, where milk is being handled, depends on the efficiency of ventilation. Milk stored in a refrigerator box may also take up the odors from fish and other foods, cooked or uncooked. 26 University of California — Experiment Station SUMMARY According to experiments at this Station and elsewhere, certain feeds, when consumed by cows during the 5-hour period before milking, im- part to the milk a flavor that varies in intensity according to the kind of feed, the quantity consumed, and the length of the time between feeding and milking. Full rations of alfalfa hay, green alfalfa, clover hay, or corn silage fed 1 to 2 hours before milking produced strong, undesirable feed flavors and odors. As the interval between feeding and milking increased, the intensity of the feed flavor decreased. When these feeds were withheld during the 5-hour interval before milking, objectionable feed flavors and odors were eliminated. Smaller quantities of the flavor-producing roughages — for example, 5 pounds of alfalfa hay, 10 pounds of green alfalfa, or 10 pounds of corn silage — fed 1 or 2 hours before milking imparted to the milk a distinct and undesirable feed flavor, evident to one accustomed to judging milk. When green barley, wild oats, foxtail, and filaree were fed to cows 2 hours before milking, in quantities required for satisfactory nutrition and as sole sources of roughage, undesirable feed flavors varying from slight to strong were imparted to the milk in every instance. The inten- sity of the flavor varied according to the kind of feed used. Cows being grazed on these feeds should be removed from the pasture at least 5 hours before milking, if such flavors are to be avoided. When the juice expressed from 25 pounds of green alfalfa was ad- ministered as a drench to each of four cows, feed flavor was detected in the milk 20 minutes later. The flavor was most prominent in the milk drawn 45 minutes after the juice was administered, less noticeable in the milk drawn after 2 hours. Tame oat hay was found to give only a slight after-flavor in milk when 8 to 9 pounds was fed to cows 2 hours before milking. When fed in a mixture with 7 pounds of alfalfa hay, it did not modify the intensity of the alfalfa flavor. Improperly cured hay having a musty odor was found to transmit a musty flavor to the milk. The usual concentrate feeds — rolled barley, coconut meal, soybean meal, cottonseed meal, wheat bran, and dried beet pulp — when fed 1 or 2 hours before milking, in quantities used by the average commercial dairyman, did not give milk sufficient flavor to make it undesirable to the average consumer. Rolled barley and beet pulp, however, fed alone in 5-pound quantities or more, 1 and 2 hours before milking, gave either a detectable flavor or after-flavor; but the judges believed that these Bitl. 595] Flavors of Milk and Their Control 27 would not be noticed in cold milk by the average consumer. Wheat bran seemed to improve the flavor of the milk when fed in 5% to 7-pound quantities 1 hour before milking. It gave more flavor to the milk than was present in the control samples, and the flavor was reported as pleasing. Some feeds caused a distinct feed flavor that could be observed while the milk was being drunk. In other samples the feed flavor was not ob- served during drinking; but after the milk was swallowed, an after- flavor was detectable. Probably the difference in the impressions be- tween flavors and after-flavors is caused by the quantity of flavor-pro- ducing feeds consumed by the cow. Salty taste was observed in milk from certain cows that were ad- vanced in lactation and also from one or more quarters of udders pre- viously affected with mastitis. The first type of salty milk may be de- tected by examining for taste the milk of all cows advanced in lactation. After cows producing such milk are identified, they should be elimi- nated from the milking herd. Saltiness of milk resulting from mastitis is more easily detected from the physical condition of the udder and usually also from the appearance of the milk itself. Usually one cannot recognize salty taste in mixed milk produced by several cows, because of the dilution of abnormal with normal milk. Rancid milk was produced by certain cows that had been milking for longer than the usual lactation period. Occasionally such milk appears in the first months of lactation. The cause of rancidity that quickly de- velops in certain milks is the enzyme lipase present in them at the time they are drawn from the cow. Lipase is normally present in blood, but the conditions controlling its presence in milk are not known. The de- tection of cows producing such milk and their elimination from the milk- ing herd is desirable, since a small quantity of rancid milk contains enough lipase to cause rancidity when it is mixed with a larger volume of normal milk. Pasteurization of milk within a few hours after milking will destroy the activity of the lipase and prevent the development of rancidity. Oxidized flavor develops in milk that has been in contact with certain corrodable metals. Copper and its alloys have been found to be the most common cause of oxidized flavors. Exposure to sunlight also causes an oxidized or tallowy flavor. The use of noncorrodable metals and the pro- tection of milk from strong sunlight during processing and delivery will eliminate the causes of these defects. 28 University of California — Experiment Station ACKNOWLEDGMENTS Grateful acknowledgment is given to the Division of Animal Husbandry for cooperation in supplying the animals for the feeding experiments; and to Professor W. M. Regan for helpful suggestions in some of the experimental procedures. Bul. 595] Flavors of Milk and Their Control 29 LITERATURE CITED i Harley, William. 1829. The Harleian dairy system. 288 p. James Ridgway, Piccadilly, London. 2 Ayers, S. Henry, and W. T. Johnson, Jr. 1914. Removal of garlic flavor from milk and cream. U. S. Dept. Agr. Farmers' Bui. 608:1-4. s Gamble, J. A., and Ernest Kelly. 1922. The effect of silage on the flavor and odor of milk. U. S. Dept. Agr. Dept. Bui. 1097:1-24. 4 Associates of Lore A. Rogers. 1935. Fundamentals of dairy science. 2d. ed. 616 p. (See specifically p. 360-363, 591-593.) Reinhold Pub. Corp., New York City, s Babcock, C. J. 1923. Effect of feeding turnips on the flavor and odor of milk. U. S. Dept. Agr. Dept. Bui. 1208:1-7. 6 Babcock, C. J. 1923. Effect of feeding green alfalfa and green corn on flavor and odor of milk. U. S. Dept. Agr. Dept. Bui. 1190:1-12. 7 Babcock, C. J. 1924. Effect of feeding cabbage and potatoes on flavor and odor of milk. U. S. Dept. Agr. Dept. Bui. 1297:1-12. s Babcock, C. J. 1925. Effect of garlic on the flavor and odor of milk. U. S. Dept. Agr. Dept. Bui. 1326:1-10. 9 Babcock, C. J. 1925. Effect of feeding green rye and green cow peas on the flavor and odor of milk. U. S. Dept. Agr. Dept. Bui. 1342:1-8. io California Agricultural Experiment Station. 1926. California Agr. Exp. Sta. Rept. 1925-26:62. ii Hammer, B. W., and W. A. Cordes. 1920. A study of brown glass milk bottles. Iowa Agr. Exp. Sta. Research Bui. 64:100-111. 12 Frazier, W. C. 1928. Defect in milk due to light. Jour. Dairy Sci. 11:375-379. is Tracy, P. H., and H. A. Ruehe. 1931. Flavor development in market milk. Jour. Dairy Sci. 14:250-267. 14 Henderson, J. L., and C. L. Roadhouse. 1934. Factors influencing the initial induction period in the oxidation of milk fat. Jour. Dairy Sci. 17:321-330. is Lea, Colin H. 1934. Estimation of aldehydes in rancid fats. Indus, and Engin. Chem. (Analyt. Ed.) 6:241-246. is Hunziker, O. F. 1923. Selection of metals in the construction of dairy equipment. World's Dairy Cong. Proc. 2:1189-1202. 17 Seligman, Richard. 1923. Milk and metals. World's Dairy Cong. Proc. 2:1202-1212. 30 University op California — Experiment Station is Guthrie, E. S. 1916. Metallic flavor in dairy products. New York (Cornell) Agr. Exp. Sta. Bui. 373:609-643. 19 Guthrie, E. S., C. L. Koadhouse, and G. A. Richardson. 1931. Corrosion of metals by milk and its relation to the oxidized flavors of milk. Hilgardia 5(14) :425-453. (Contains bibliography of 35 references in the subject.) 20 Roadhouse, C. L., and G. A. Koestler. 1929. Contribution to the knowledge of the taste of milk. Jour. Dairy Sci. 12: 421-437. 2i Palmer, L. S. 1922. Bitter milk of advanced lactation. Jour. Dairy Sci. 5:201-211. 22 Koestler, G. A., C. L. Roadhouse, and W. Loertscher. 1928. Zur Kenntnis der Sekretion lipolytish aktiver, sogenannter "ranziger" Milch. Landw. Jahrb. Schweiz 1928:927-966. 23 Sadler, W. 1929. Flavour defects in high grade milk. Sci. Agr. 10:111-119. 24 Blakeslee, A. F. 1930. Genetics of sensory thresholds : taste for phenylthiocarbanide. Natl. Acad. Sci. Proc. 18:120-130. 25 Parker, G. H. 1922. Smell, taste and allied senses in the vertebrates. 192 p. (See specifically p. 132-158.) J. B. Lippincott Company, Philadelphia. 26 Kelly, E., and R. J. Posson. 1926. How to conduct milk and cream contests. U. S. Dept. Agr. Dept. Cir. 384:1-22. 27 Babcock, C. J., and C. S. Leete. 1929. How to conduct milk and cream contests. U. S. Dept. Agr. Dept. Cir. 384:1-22. 28 Koestler, G. A. 1920. The chloride-lactose number of milk. Mitt. Lebensmtl. Untersuch. u. Hyg., Eidg. Gsndht samt. [Switz.] 11:154. 29 Weaver, E., A. H. Kuhlman, and E. L. Fouts. 1935. The effect of alfalfa hay on milk flavor. Jour. Dairy Sci. 18:55-61. 30 Roadhouse, C. L., and J. L. Henderson. 1932. The relation of the soluble portion of alfalfa to the rapid absorption of feed flavor in milk. Jour. Dairy Sci. 15:299-302. 3i Guthrie, E. S., and H. J. Brueckner. 1934. The cow as a source of "oxidized" flavors in milk. New York (Cornell) Exp. Sta. Bui. 606:1-10. 15m-9,'85