ABSTRACT POLYBROMINATED BIPHENYLS IN RAW MILK AND PROCESSED DAIRY PRODUCTS by Teiko Murata The present study was aimed at the investigation of the distribution of PBBs in dairy products and the effect of processing on the residue levels. Milk containing less than 0.3 ppm of physiologically incorporated PBBs from each of four dairy herds was separated by physical methods into skim milk and cream as well as being further processed into pasteurized whole milk, pasteurized skim milk, pasteurized cream, butter, cheddar cheese, freeze-dried whole milk, freeze—dried Skim milk, and freeze-dried cream. Spray-dried skim milk and whole milk and evaporated whole milk were manu- factured individually from milk of two of these herds. PBBs were extracted using AOAC procedures for chlorin— ated hydrocarbon pesticides and quantitated with electron capture GLC. The PBB contents of the by-products, butter— milk and cheese whey, and the raw whole milk, skim milk, and cream were also measured. Moisture and lipid contents of all Teiko Murata the manufactured products were also determined. PBBs were found to be concentrated in the high-fat content products and skim milk, buttermilk, and cheese whey had very low levels of these compounds on a total weight basis. However, when the concentration of the PBB residues in these products was expressed on a fat basis, they had slightly higher levels than that which occurred in the pas- teurized whole milk suggesting PBBs may be associated with lipoprotein or may be soluble in the serum of milk. The variations in levels among the raw and processed dairy products were not significantly different when the PBB levels were expressed on a fat basis. Condensation was not effective in the removal of PBBs from whole milk while spray-drying appeared to significantly promote losses of PBBs from whole milk (30 - HO%) and skim milk (60 - 69%). The higher losses of residue observed in spray-dried skim milk suggest PBBs are more easily removed from low-fat content products and when the concentration in the original product is low. POLYBROMINATED BIPHENYLS IN RAW MILK AND PROCESSED DAIRY PRODUCTS by Teiko Murata A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Food Science and Human Nutrition 1976 ACKNOWLEDGMENTS The author expresses her gratitude to Dr. Mary Ellen Zabik and Dr. Matthew J. Zabik for their patience, help- fulness, and guidance throughout the entire graduate pro- gram, and for their valuable assistance in the preparation of this thesis. Appreciation is extended to Dr. Lawrence E. Dawson for the advice and encouragement given as a member of the guidance committee and to Dr. Walter M. Urbain for the guidance given in the initial work. Special gratitude is expressed to Dr. James R. Kirk for his assistance and help during processing of the dairy products. Appreciation is also extended to Mrs. Melissa Shafer for the elaboration of the computer program for analysis of the data. Special thanks are extended to CODOT for the financial support throughout the entire graduate program. Sincere appreciation is extended to my friends and fellow graduate students in the Department of Food Science and Human Nutri- tion for their encouragement, interest in the project, and suggestions in the preparation of this dissertation. Finally, I would like to thank my parents for their con— stant encouragement and support away from home. 11 TABLE OF CONTENTS Page Acknowledgments . . . . . . . . . . . . . . . . . . . ii List of Tables . . . . . . . . . . . . . . . . . . . v Introduction . . . . . . . . . . . . . . . . . . . . 1 Review of Literature . . . . . . . . . . . . . . . . A Polybrominated Biphenyls A Physical and Chemical Properties - Uses A Residues in the Environment . . . . . 5 Residues in Biological Systems 7 Toxicity of PBBs . . . 8 Polychlorinated Biphenyls — Chlorinated Hydro- carbons . . . . . . . . . . . lO Occurrence and Reactivity . . . . . . . . . 10 Milk Contamination . . . . . . . . . . . . . 12 Effect of Processing on Residues . . . . . . . . 1“ Experimental Procedure . . . . . . . . . . . . . . . 18 Separation by Physical Methods . . . . . . . . . 19 Pasteurization . . . . . . . . . . . . . . . . . l9 Butter . . . . . . . . . . . . . . . . . . . . . 20 Cheese . . . . . . . . . . . . . . . . . . . . . 2O Freeze— —Drying . . . . . . . . . . . . 22 Spray-Drying and Condensation . . . . . . . . . . 23 Chemical and Residue Analysis . . . . . . . . . . 2“ Moisture Analysis . . . . . . . . . . . . . . 2A Lipid Analysis . . . . . . . . . . . . . . . . 2A Residue Analysis . . . . . . . . . . . . . . . 26 GLC Analysis . . . . . . . . . . . . . . . . . 27 Analysis of the Data . . . . . . . . . . . . . . 28 iii Results and Discussion Distribution of Fraction . . Moisture and Lipid Analysis Residue Analysis . . . . Distribution of PBB . . . Effect of Processing in the PBB. Levels Summary and Conclusions Suggestions for Further Research References iv Page 30 3O 35 38 38 an 60 6A 65 Table 10 ll 12 LIST OF TABLES Specific industrial us 5 and approximate allo- cations of Fire Master BP-6 produced (Kerst, 197“) o o o o o o o o o o o o o o o o o o o o Recoveries of PBB (Fire MasteéE)BP-6) from milk at different added levels by the Food and Drug Administration method . . . . . . . . . . . . Yields, percentage distribution and recovery of milk fractions . . . . . . . Yields, percentage distribution and recovery of cream fractions in butter making procedure Yields and percentage distribution of milk frac- tions in cheese making procedure . . . . Average moisture contents, expressed as percent of total weight . . . . . . . . Average fat contents, expressed as percent of total weight . . . . . . . . . . . . Total PBB in milk and its distribution and recovery in the physically separated milk frac- tions . . . . . . Total PBB in cream and its distribution and recovery in butter making . . . . . Total PBB in milk and its distribution and recovery in cheese making . . . . . PBB residues in milk and related dairy products (ppm based on total weight) . PBB residues in milk and related dairy products (ppm based on solids) . . . . Page 29 31 33 3A 36 37 39 Al A2 A5 A6 Table 13 1A 15 16 l7 l8 19 20 Summary of the significant differences of the PBB residue contents between products . PBB residues in milk and related dairy products (ppm based on total lipid content) . . . . Effect of condensation on the PBB residue levels of pasteurized whole milk (ppm based on solids) Effect of condensation on the PBB residue levels of pasteurized whole milk (ppm based on total lipid content) 0 I O O O O O O O O O O 0 Effect of spray-drying on PBB residue contents of pasteurized whole milk and skim milk of Herd 1 (ppm based on solids) . . . . . . . . . Effect of spray-drying on PBB residue contents of pasteurized whole milk and skim milk of Herd 1 (ppm based on total lipid) . . . . . . . . . . Effect of spray-drying on the PBB residue con— tents of pasteurized whole milk and skim milk of Herd 2 (ppm based on solids) . . . . . . Effect of spray-drying on the PBB residue con- tents of pasteurized whole milk and skim milk of Herd 2 (ppm based on total lipid) . . vi Page A7 A8 52 52 5A 55 56 57 INTRODUCTION The growing concern over environmental contamination has undoubtably stemmed from the rapid influx of synthetic compounds and their accompanying waste materials. Of these compounds, much attention has been focused on the halogenated hydrocarbons. Organochlorine pesticides and polychlorobi- phenyls, with their lipophylic properties, have been found in the fat of virtually all animals. They resist biological degradation and consequently are expected to be present in the ecosystem for an extended period of time. Occurrence of chlorinated hydrocarbon residues are well documented (Abbott et al., 1966; Anderson et al., 1969; Holden et al., 1967; Holmes et al., 1967). Polychlorinated biphenyls (PCB) have been used indus— trially in the manufacture of resins, rubbers, plastics, building materials, paper, electrical equipment, hydraulic fluids, etc. (U.S. Dept. of HEW, 1969) and have been mentioned as capable of extending insecticide kill-life (Sullivan et al., 1953; Lichtenstein et al., 1969). They were first identified by Jensen (1966) and since then PCB residues have been detected almost the world over, in fish, birds, and wildlife (Holmes et al., 1967; Reynolds, 1969; Jensen et al., 1969; Risebrough et al., 1968; Peakall et al., 1970; Anderson et al., 1 1969; Holden et al., 1967), and in humans (Musial et al., 197A; Savage et al., 1973; Doguchi et al., 1975; Biros et al., 1970). Human exposure to these contaminants has generally been through their residues in foods. Surveys of foods indicate varying amounts of organochlorine pesticide residues in many commercial foods (U.S. Dept. of HEW, 1969) and food contamina- tion by PCBs has been reported (Kuratsune, 1969; Kolbye, 1972; Fujiwara, 1975)- Polybrominated biphenyls (PBB), currently manufactured by Michigan Chemical Corporation, are members of the halogen- ated hydrocarbon class of compounds with properties and uses closely related to PCBs. They are not present in the eco- system in as large amounts as PCBs and would not have been a subject of great concern if the accidental contamination of Michigan cattle and other farm animals with this product had not occurred. PBBs were mistakenly used in place of magnesium oxide in dairy feeds (Jackson et al., 197A) and this incident has led to their occurrence in milk. The FDA has set the action- able PBBs residue level in milk at 0.3 ppm (fat basis). Although there is no evidence indicating low levels of con- taminant chemicals are detrimental to man, it seems prudent to minimize concentration in foods, until more conclusive data firmly establish safe levels. Since milk is one of the major sources of human foods, it is important to investigate the possibility of eliminating or reducing any contaminant level in this product. Some studies have shown the effectiveness of removal of halogenated com- pound residues from milk. Chlorinated hydrocarbon pesticides have been removed from milk by molecular distillation (Bills et al., 1967), by roller-drying of whole milk (Ruzicka et al. 1967), and by spray-drying of whole milk (Li et al., 1970). The distribution of PCBs in dairy products and the effective decrease of their residue concentration by heating skim milk was reported by Platonow et a1. (1971), however, information concerning the distribution of polybrominated biphenyls in dairy products and their removalchudng processing of milk is non-existent. Thus, this project is focused on the possi- bility of eliminating or reducing the PBB residue levels in the processing of PBB contaminated milk as well as studying their distribution in related dairy products. REVIEW OF LITERATURE Polybrominated biphenyls (PBBs) are fairly new chemi— cals introduced essentially as a fire retardant for the manufacture of safe end-products. They belong to the halo- genated hydrocarbon class of chemical compounds and share many chemical properties with polychlorinated biphenyls (PCBs) and organochlorine pesticides. Uses and characteristics of PBBs are presented in this review. Similarities of the properties of PBBs to PCBs are summarized. The parallelism between reactivity of PCBs and many of the chlorinated hydrocarbon pesticides has been extensively documented. Since there have been no studies on the effect of processing on PBB levels in foods, studies of the effect of processing of dairy products on the levels of PCBs and chlorinated hydrocarbon pesticides are also reviewed. Polybrominated Biphenyls Physical and Chemical Characteristics - Uses PBBs reduce the flammability of thermoplastics. This property, in addition to their low cost, has led to their production on a commercial scale. Michigan Chemical Corpora- tion currently manufactures a product under the name of Fire MasterG>BP-6 which is a mixture of brominated biphenyls with approximately 63% of hexabromobiphenyls, lA% of heptabromo- biphenyl, 10.5% of pentabromobiphenyl, 2% of tetrabromobi- phenyl, and others, unidentified (Kerst, 197A). This product is employed in industries as a plasticizer and fire retardant. Its specific uses are summarized in Table l. 8) Fire Master BP-6 is solid at room temperature with a softening point of 72°C. It shows very low solubility in water, 11 ppb at 25°C, but it is soluble in most organic sol- vents (Kerst, 197A). Studies concerning stability of brominated biphenyls indicate Fire Masterfi) BP-6 undergoes reductive dehalogena- tion when exposed to ultraviolet light under certain condi— tions (Ruzo et al., 1975). This reaction has also been reported to occur with mono- and di-brominated biphenyls (Bunce et al., 1975) and octabromobiphenyls (Norris et al., 197A). The photoreactivity of this class of compounds has been shown to be greater than that of the chlorinated biphen- yls. This increased reactivity is probably attributable to the lower C-Br bond energy (Ruzo et al., 1975; Bunce et al., 1975) which is 71 kcal/mole compared to the C-Cl bond energy of 85.6 kcal/mole (Kerst, 197A). Residues in the Environment Reports on the detection of PBB residues in the envir— onment are so far non-existent. However, since these com- pounds are closely related to PCBs, both in uses and chemical properties, it is expected they may escape into the Table 1. Specific industrial uses and approximate alloca- tions of Fire Master‘B BP-6 produced (Kerst, 197A). Industrial use App. allocation of tot 1 Fire Examples Master BP-6 produced, % Business machines and industrial equipment Electrical Fabricated products Transportation Miscellaneous Typewriter, calculator and A8 microfilm reader housings; business machine housings. Radio and TV parts, thermo- 35 stats, shaver and hand test housings. 12 Project housings, movie equipment cases. Miscellaneous small automo- l tive parts, i.e. electrical wire connectors, switch connectors, speaker grills Small parts for electrical A applications, motor hous- ings, components for indus- trial equipment. environment in similar routes as that of the P083. Residues in Biological Systems The accidental use of Fire MasteNBABP—6 in place of magnesium oxide in feed led to the inadvertent contamination of numerous Michigan dairy herds, swine, and poultry during the fall of 1973 and the winter of 197A (Jackson et al., 197A; Gutenmann et al., 1975; Detering et al., 1975; Prewitt et al., 1975). Jackson and Halbert (197A) were the first to document a toxic syndrome in a herd from one Michigan dairy farm caused by intake of polybrominated biphenyls contaminated feed. These compounds were identified by a U.S.D.A. laboratory and were further traced to the same manufacturer who supplied magnesium oxide to the feed company. By July, 197A, 92 food producing premises involving cattle, poultry, and swine were identified as contaminated by P883 and quarantined. A fur- ther euthanasia program was planned to eliminate the contam— inated animals. Several works concerning PBB residues have shown that chemical is excreted into the milk of cows (Gutenmann et al., 1975; Fries et al., 1975; Willet et al., 1975; Detering et al., 1975). Appearance of PBB in milk, plasma and feces were observed by Willet and Irving (1975) after intraruminal doses of the product were given to cows. In the experimental cow and sheep feeding with PBB at levels of 50 ppm, Gutenmann et a1. (1975) found the half—life for PBB in milk was about 10.5 days which would give, after 105 days, a total daily excretion of less than 0.01 mg/day of this compound. They also reported the distribution of PBB residues in the tissues with highest concentrations of PBBs being in fatty tissues. The content of PBBs in blood, milk, and body fat of cows contaminated 7 to 9 months previously showed that body fat and milk fat contained 600 and 300 times, respectively, of that which the blood contained (Detering et al., 1975). Fries and Marrow (1975) studied the excretion of Fire Master® BP-6, PCB, and DDE into the milk of cows. They found the rate of disappearance of PBB in milk was 58 days, with mode of excretion (decay rate and steady state excretion levels) being similar to those for PCBs and chlorinated hydrocarbons. A study conducted by Fries et al. (1973) concerning polybrominated biphenyls and PCBs in hens and cows, had pre- viously shown that residues of PBBs in animal systems behave similarly to the residues of PCBs with a comparable degree of halogenation. Toxicity of PBBs The varying degree of toxicity of polybrominated biphenyls in animals is directly correlated to the concen— tration of these chemicals at which the animals are exposed. The oral lethal dose for Japanese quail exceeds 1.9 g/kg (Strik, 1973), for rats the value found is in the range of 21.5 g/kg (Michigan Chemical Corp.), and the acute dermal LDSO for rabbits is between 2.15 - 10.0 g/kg (Michigan Chemical Corp.). Liver lesions were the most common symptoms of acute and chronic toxic effects seen in the Michigan cattle (Jackson et al., 197A). These were seen when the level of PBBs in the body fat were approximately 200 ppm. Hematomas and abscesses in the peritoneal and thoracic cavities of these cows were also observed. From field observations of PBB contaminated cattle, Prewitt et a1. (1975) reported that animals which had over 20 ppm in the milk fat, at parturition, had unrelaxed pelvic ligaments and the calves were stillborn or died shortly after birth. Metrites and retained placentas as well as liver and kidney adhesions were also common. Examinations of dead calves showed PBBs were transferred through the placenta (Detering et al., 1975). PBB concentration in the body fat of these calves was in the range of 50 to A00 ppm. Strik (1973) reported hexabromobiphenyl fed to Japanese quail and chicken induced porphyria, especially in the liver. Hepatic microsomal enzymes of Japanese quail were induced by dietary PBB at 10, 20, and 100 ppm (Babish et al., 1975). Hatchability was affected by the diet which contained 100 ppm PBBs and A0% of these embryos died in the first day or two of development. However, when the feed contained 10 and 20 ppm of PBBs no effect was observed. Farber and Baker (197A) found that hexabromobiphenyl 10 was a more potent microsomal enzyme inducer than PCB (Arochlor 125A) at 5 ppm level in rats. PBBs were also found in the blood of farmers exposed to the contaminated feed (Chem. Eng. News, 1975), but these farmers appeared to be healthy. Polychlorinated Biphenyls - Chlorinated Hydrocarbons Occurrence and Reactivity The world wide occurrence of PCBs in the marine and aquatic environments is fully documented. Studies indicate that sources of these chemicals are diffuse and generally associated with waste disposal materials (Carnes et al., 1972; Schmidt et al., 1971; Fujiwara, 1975; Martell, 1975). Toxicological studies with PCBs have shown they have similar mode of action to that of chlorinated hydrocarbons. They promote abnormal changes in the liver and induce many microsomal enzymes (Turner et al., 197A; Holub et al., 1975; Villeneuve et al., 1971; Grant et al., 1971). PCBs with higher chlorine content appear to have greater effect in enzyme activity (Rhee et al., 1973; Iverson et al., 1975) and are more resistant to biological degradation than the low chlorinated biphenyls (Greb et al., 1975; Yoshimura et al., 1975; Jan et al., 1975). A number of studies have shown that humans contain low 11 levels of halogenated compounds. Biros et a1. (1970) found polychlorinated biphenyls and DDE in human adipose tissue. Comparable results were obtained in Japan, in the work con- ducted by Fujiwara (1975). The adipose tissues contained PCBs, DDT, DDE, and several BHC isomers. He pointed out that the PCB levels were similar to the levels found in Germany but higher than those in the United States. Analyses of human blood showed varying concentration of PCBs, BHC, DDT, and DDE residues (Doguchi et al., 1975; Fujiwara, 1975). The levels of PCB were assumed to be associated to the degree at which the sampled people were exposed since the residue concentrations were proportionally higher in more severely contaminated areas. Savage et a1. (1973) detected PCBs in human milk in rural Colorado, although the amounts were not appreciable. How- ever, in Japan the PCB concentrations found in human milk were relatively high and it was estimated that babies would ingest between A.8 - 5.3 ug/kg body weight of PCBs per day (Fujiwara, 1975). Musial et a1. (197A) reported that milk from people of provinces of Canada contained PCBs as well as DDT, and DDE. These evidences, in addition to several documented works of the occurrence of organochlorine pesticides in human tis— sues (Brown, 1967; Abbott etzfln, 1968; Dyment et al., 1971) indicate the ultimate deposition of these chemicals in man. The sources of human contamination has generally been associated with ingestion of food containing halogenated 12 compounds (Kolbye, 1972; Kuratsune, 1969). Reported results of surveys of food items conducted in the United States (Duggan et al., 1967; Duggan and Lipscomb, 1969; Lipscomb, 1968) and in Japan showed that low levels of PCBs and several organochlorine pesticides are present in most of the commer- cial food. Marth and Ellickson (1959) have reported that 25 to 62% of market milk supplies contained varying amounts of chlorinated hydrocarbon insecticides. Cummings (1966) has shown that the organochlorine pesticides dieldrin, heptachlor epoxide, lindane, and DDT were present in dairy products from market basket studies, although the levels were very low. Milk Contamination Several studies have documented the transfer of chlor— inated hydrocarbon pesticides from the feed into the milk of dairy cows. It has also been shown that spray treatment of cows with these pesticides led to their appearance in the milk by entering into cows through the skin. Marth et a1. (1959) had suggested that insecticide residues in milk are primarily a result of spraying dairy barns and cattle and of ingestion of treated forages by dairy cattle. The major concern about chlorinated hydrocarbon pesti- cides is that once the cows are exposed to these chemicals the residues will be present in the milk for a long period of time as a result of storage in the fat body tissues and l3 gradual translocation into milk from these tissues. Similar behavior of polychlorinated biphenyl residues has been pointed out by Fries et a1. (1972). They reported that the rates of decline in milk concentrations of PCB and DDE residues after environmental exposure were identical. In the long-term studies of PCB residues retention and excre- tion by cows, Fries et al. (1973) found that the rate of decline in milk fat of PCB concentration was 50% within 15 days after removal of the chemical from the diet. By compari- son with DDE from previous data (Fries et al., 1969) they showed the similarity of behavior of DDE and PCB in the accu- mulation and excretion in cows. They also observed that the decrease, in concentration in body fat,paralleled the decrease in concentration in milk fat. Although retention and excretion of polychlorinated biphenyls into milk has been well documented, the precise sources of contamination are not known. Some studies have indicated contamination of cows might occur through ingestion of feed containing these compounds. Savage et a1. (1973) reported the presence of PCB residues in silage stored in pits and upright silos although they were at very low levels. Similar results were obtained by Skrentny et a1. (1971) from silage stored in upright silos. They stated that the PCBs had leached into the silage from the sealants used to coat the silo walls. Furr et a1. (197A) reported that waste paper has been incorporated into farm animal feeds to serve as a substitute 1A form of cellulose in the ration. In their experimental feeding the cows with waste paper, they observed PCBs were transferred from the paper to the milk. Moreover, there was an evident increase in the concentration with the increase in the length of time of feeding. I Stanovick et al. (1973) pointed out that foodstuffs might get PCB contamination from paperboard which contains recycled waste paper with appreciable amounts of carbonless "carbon" paper. They suggested PCBs would migrate from the paperboard into the food. Effect of Processing on Residues Studies concerning the possibility of removing halo- genated compounds from foodstuffs upon processing have stemmed from the possible chronic effects of continued assimilation and accumulation of these compounds in the body fat. Several studies concerning the effect of residues on dairy processing have been conducted. Kim (1969) reported that no effect was observed on the growth of lactic starter cultures 111 milk containing different organochlorine pes- ticides. On the other hand, Bradley and Li (1968) observed that dieldrin slightly reduced the production of acid by a lactic culture organism in making cheddar cheese. Pasteur- ization had very little effect on the DDT residues in dairy products (Mann et al., 1950). Langlois et a1 (196A, 1965) 15 manufactured several dairy products and reported that the concentrations of the pesticides studied remained fairly constant. However, they observed significant losses of DDT and lindane during the spray—drying of whole milk and loss of lindane only during the roller-drying. They also observed lower insecticide content in butter and cheese than the raw milk on a fat basis. Li and Bradley (1969) observed that ultraviolet light treatment of milk and butteroil reduced the levels of chlor— inated hydrocarbon pesticide residues, but the treatment rendered the milk unacceptable for consumption. Reduction of organochlorine pesticide residues upon freeze drying of milk at a system pressure of l x 10'3 torr was suggested by Kiermeier et a1. (1967). However, freeze drying at a pressure of 0.1 mm Hg and mild deodorization were not effective on the reduction of the dieldrin and heptachlor epoxide levels in butteroil (Kroger, 1968), but heating at very high temperature (30000) and steam deodorization at 180 - 195°C and 0.01 - 0.5 mm Hg significantly reduced the amount of the residues. Ledford et a1. (1968) found that commercial steam distillation-vacuum processing was also fairly effective on reduction of lindane, but little or no effect was observed for dieldrin, DDT and heptachlor in milk. Li et a1. (1970) reported that manufacture of 30% cream, butter, cheese, spray—dried whole milk, condensed whole milk, and pasteurization and sterilization in general did not affect the amounts of pesticide residues in these products. 16 On the other hand, spray drying reduced dieldrin, lindane, and chlordane, and sterilization of condensed milk signi- ficantly removed dicofol while storage of cheddar cheese for six months slightly reduced dieldrin content. In addition, they observed that skim milk, buttermilk and whey contained higher concentrations of these pesticides than the whole milk on a fat basis. Bills and Sloan (1967) found that molecular distillation at high temperature (200°C) and under reduced pressure (5 x 10.“ torr) was very efficient in removal of lindane, hepta- chlor epoxide, heptachlor, aldrin, DDT, DDD, and DDE from butteroil. A possible degradation of DDT and DDE by cheese micro- organisms was studied by Ledford and Chen (1969). Their results show that aerobic growth of geotrichum species caused almost complete disappearance of the DDT and DDE residues. Platonow and Funnell (1971) studied the effect of pro- cessing on the levels of metabolically incorporated poly- chlorinated biphenyls. They pointed out that the manufactur- ing processes did not significantly alter the residue levels. However, heating skim milk at 7000 and holding for 10 minutes reduced the amount of PCB by approximately 67%. They also observed that most PCB was present in the cream and whole milk although the actual amounts were not always proportional to the fat content. This fact was indicated by the higher concentration found in whey per fat unit than any other sample. 17 Losses of 15 to 25% of several organochlorine pesti- cides from milk during roller drying were reported by Ruzicka et a1. (1967). Residues in skimmed milk were negli— gible. Stemp and Liska (1966) found that during the processing of concentrated, dried, and sterilized dairy products A0-50% of the telodrin residues were destroyed in evaporated milk and 10—20% in the dried whole milk. However, methoxychlor was more stable under the milk-processing treatments. They also observed a higher concentration of telodrin in the cream on a fat basis when compared to raw whole milk while methoxychlor remained fairly constant. Comparable effects were obtained by Liska (1968) during heat treatment of milk containing chlorinated insecticide residues. Spray drying destroyed more than 80% of the lin- dane residues and heptachlor was destroyed more easily than heptachlor epoxide. He concluded that the amount of residue destruction varied with the treatment used and nature of the insecticide residue. The effect of storage of Monterrey and cheddar cheese on the chlorinated hydrocarbon residue level was found to be not effective by Montoure and Muldoon (1967). However, they observed a change in the residue distribution during manu— facturing of the cheeses. DDT was not detected in whey sam- ples at dipping while high levels of DDE and TDE were evident. In all other phases the distribution of the residues was quite similar. EXPERIMENTAL PROCEDURE In order to investigate the relationship between the polybrominated biphenyls (PBB) residue levels in milk and the processing of the milk, environmentally contaminated milk with less than the actionable level (below 0.3 ppm) of PBB was processed into a variety of dairy products. Many Michigan dairy farms had their herds accidentally contaminated through feedings with feed containing PBB. The Michigan Department of Agriculture has been monitoring these farms and the milk from them is being analyzed periodically. From these monitoring data, herds with low levels of PBB were identified. Milk was purchased from four of these herds to be processed into dairy products. It was collected at the farms in stainless steel milk cans and brought back to the laboratory (iced) and kept refrigerated until processing. In June, 1975, the Michigan Department of Agriculture data showed PBB levels for the milk collected as follows (Van Patten, 1975): PBB Level, ppm Herd Collection Date (fat basis) 1 June 26 0.223 2 July 6 0.138 3 July 21 Trace A July 23 0.132 18 l9 Separation of Milk Fractions by Physical Methods Milk was separated into skim milk and cream with a Westfalia disc-type separator, model LWA 205, at approxi- mately 6000 rpm. The separator was first warmed up with hot water (6000) and fed with milk at temperatures of 60- 62.8OC. The resulting cream was measured volumetrically and the yield of skim milk was estimated by the difference between the whole milk and cream. Duplicate quantitative portions of each fraction and whole milk were taken after thorough mixing and stored in glass bottles at -23OC for further analysis. Pasteurization Immediately following separation, the whole milk and skim milk were pasteurized at 62.80C (1A50F) for 30 minutes in a Kusel culture cabinet and the cream was pasteurized at 71.100 (1600F) for 30 minutes, according to the "Milk Ordin- ance and Code" of the U.S. Dept. of HEW (1953). A measured portion of the pasteurized cream was stored atZLSOC in a glass beaker, covered with wrapping film, and kept for at least 12 hours for further butter processing. Approximately one-quarter of the cream, whole milk, and skim milk was stored in freeze-drier trays at —23OC and the remaining portions were kept refrigerated for further 20 processing, after two aliquots were taken for lipid and residue analysis. Butter The separated cream fraction was allowed to reach SOC and then churned in a Kitchen-Aid, model K5-A, at approxi— mately 2A0 rpm for 30 minutes or until separation of the buttermilk from butterfat occurred. Following collection and yield measurement of the buttermilk, the butterfat was washed once with cold tap water, wrapped in a double layer of cheese cloth and gentle pressure was applied manually to remove the excess of water. The butter was wrapped in a plastic wrapping film and kept frozen at -23°C. Cheese A stirred curd cheddar type cheese was made from pas- teurized whole milk following the procedure outlined by Kosikowski (1966). Lactic culture and rennet extract were obtained from the Dairy Plant of Michigan State University. A stainless steel container with dimensions of 2A cm x 30 cm x 6 cm and equipped with a spicket was used as a cheese vat and heated with a water bath. Twenty grams of lactic starter culture (1% of Strepto- coccus lactis and Streptococcus cremoris secondary culture) 21 were added to approximately 2000 g of warm (31.10C), whole pasteurized milk. This was stirred well and left for 15-30 minutes at 31.100 (88°F), with continuous agitation. After the ripening period, 0.A ml of single-strength rennet extract diluted to 10 ml with distilled water were added, stirred and left for A5 minutes at 31.1OC (88°F), or until the appropriate strength of the curd for cutting was obtained. The curd was cut into small cubes with approximately 6.5 mm (l/A in) per side, healed, and then with continuous agitation of the curd the temperature was slowly raised to 38.8 - 39.A°C (102 - 103°F). After the temperature of 38.8 - 39.A°C was reached, the curd was cooked for approximately A5 minutes more or until an acidity of 0.13 - 0.1A% of lactic acid was attained. The whey was then drained to the level of the curd. The acidity was determined by titration of 10 m1 of whey with 0.1N NaOH using phenolphthalein as an indicator. The whey was completely drained after approximately 30 minutes and cooking was allowed to continue with agitation of the curd every 5 - 10 minutes until the percent lactic acid was 0.23 - 0.25%. During this period, an aluminum foil cover was used to keep a homogeneous temperature through the entire curd. After the recommended acid value was reached, 5 g of coarse cheese salt (2.5% of the finished product) were spread over the curd and stirred for 10 minutes. The stirred curd was placed in a circular metal mold with a proper sized cheese cloth fixed at the bottom and covered with the cheese cloth followed by an adapted wood rim. Pressures of 22 2A.8 torr (0.A8lb/sq in) and 38.3 torr (0.7A 1b/sq in) were applied consecutively for a one hour period each and then the cheese was dipped into hot water (60°C) for approxi- mately 3 minutes, placed back in the hoop, and soaked liber— ally with the hot water. By this procedure the curd was softened sufficiently so that it could be efficiently molded. A final pressure of A3.9 torr (0.85 lb/sq in) was applied over the cheese and left overnight. A fraction of the cheese whey at the time it was com- pletely drained was taken for analysis. The cheese was weighed and the yield of cheese whey was estimated by the difference of the whole milk and cheese. It was cut into two halves and wrapped in plastic wrapping film. One half was allowed to age at 2.5°C for approximately 2 months and the other half was stored at —23°C. Freeze-Drying Portions of the pasteurized cream (900 - 1100 g), skim milk (1000 - 1300 g), and whole milk (900 - 1300 g) were placed in freeze-drier trays, giving a depth of approximately 6.5 mm (l/A in). These were covered securely with plastic wrapping film, and frozen at -23°C overnight. The frozen samples were freeze-dried in a Virtis REPP freeze-drier, model FFDA2 WS, for 2A-30 hours with a system pressure of 5 u (5 x 10'3 torr). The temperature of the platin was 65.6 — 71.1°C (150 — 160°F) giving a final product temperature 23 of A8.8 - 51.6°C (120 - 125°F). The freeze-dried samples were removed from the trays, and stored in tightly sealed plastic bags at —23°C. Spray—Drying and Condensation Spray-drying and condensation processes were done for milk from Herds l and 2. The pasteurized whole milk and skim milk were forewarmed at A8.8°C (120°F) and dried in a Swenson spray-drier, model 175505. Air outlet temperatures ranged from 76.6 to 98.8°C (170 - 210°F) with the following speci- fic conditions being used: Herd 1 Skim milk Lot 1 87.8 - 93.3°C Lot 2 76.6°C Whole milk 87.8°C Herd 2 Skim milk Lot 1 87.8 - 93.3°C Lot 2 87.8 - 98.8°C Whole milk Lot 1 87.8 - 93.3°C Lot 2 85.0 - 90.5°C For the condensation process, the pasteurized whole milk was heated to 5A.5°C (130°F) and then condensed in a research model Rogers vacuum pan under vacuum of 635 torr (25 in of Hg) to approximately 35% total solids for the milk of Herd 2 and 60% total solids for Herd l. The evaporated milk of Herd l was further diluted to give approximately 30% total solids. Duplicate aliquots of the evaporated milk were taken and 2A stored in tightly sealed glass bottles at -23°C. The spray- dried samples were placed in plastic bags and kept frozen at -23°C until needed. Chemical and Residue Analysis All chemical and residue analyses were carried out in duplicate. The chemicals used were ACS reagent grade and the solvents were redistilled from glass. All glassware was acetone-rinsed followed by petroleum ether rinsing before“ being used. Standard solutions were prepared in petroleum ether using the commercial Fire MasterfixBP-6, hexabromobi- phenyl (Lot No. 51A3, Michigan Chemical Corporation, Chicago, Illinois). Moisture Analysis Moisture analysis was carried out in a Hotpack vacuum oven, model 633, at 100°C and a vacuum of 660 - 711 torr (26 - 28 in of Hg) to a constant weight. The percentage of total solids was expressed as the dry weight of the sample. Lipid Analysis Total lipid content of the samples was determined by two methods. One of these methods, the soxhlet method (official AOAC method) using petroleum ether was applied for 25 low moisture content samples after they were previously dried in a Precision Thelco air oven, model 18, at 70°C for 30 - 60 minutes. This method was also used for the dehydrated samples except for the spray-dried samples. The samples were extracted for 5 - 6 hours, the extract was first evaporated in a steam bath and then dried to a constant weight in a Lab line vacuum oven, model 3615, at 70°C and a vacuum of 660 - 711 torr (26 - 28 in Hg). The second method involved the extraction of the crude fat for the high moisture content samples and spray-dried samples (dissolved in 50 m1 of hot (60°C) distilled water) and utilized a petroleum ether-ethyl ether solvent mixture following the extraction procedure outlined in the Pesticide Analytical Manual (PAM) of the Food and Drug Administration, Section 211.13h (1971). The lipid fraction was extracted three times with petroleum ether-ethyl ether (1:1). Ethyl alcohol was proportionally adjusted to the volume of the samples, and potassium oxalate, approximately 1 g was used. The extracts were combined and washed with 500 m1 of dis- tilled water containing 30 m1 of saturated sodium chloride solution, the water layer was discarded, and then washed once with 200 ml of 1% sodium sulfate solution. No centrifugation nor further drying of the extract by passing through a column of anhydrous sodium sulfate was involved. From this extract a 10 m1 aliquot was taken and dried in a Labline vacuum oven at 70°C and vacuum of 660 - 711 torr (26 - 28 in Hg) for fat determination. 26 Residue Analysis Extraction of the lipid fraction for residue analysis was done as previously described. Specific sample weights for the PAM method were as follows: evaporated milk, 13.0 - 27.0 g, diluted 1:2 with distilled water; spray-dried sam— ples, A.0 — 7.0 g; whole milk, 33.0 - 58 g; skim milk, 56.0 - 78.0 g; buttermilk, 65.0 - 83.0 g; and cheese whey, 65.0 - 78.0 g. For the Soxhlet method the sample weights were: cream, 2.0 - 5.0 g; butter, 2.0 - 2.5 g; cheese, 1.5 - 5.5 g; freeze-dried whole milk, 3.5 — A.5 g; freeze-dried skim milk, A.5 - 6.0 g; and freeze-dried cream, 2.0 - 2.5 g. The clean up procedures used are described in the Pes- ticide Analytical Manual, Section 211.1Aa, petroleum ether- acetonitrile partitioning, and Section 211.1Ad, florisil column. Modifications in these procedures were as follows: acetonitrile extract was washed with A00 ml of 1% sodium sulfate solution containing 100 m1 of petroleum ether, shaking vigorously for one minute. The aqueous acetonitrile layer was discarded and the petroleum ether was collected in an erlenmeyer flask containing anhydrous sodium sulfate. The extract was concentrated to approximately 10 m1 using a Kuderna Danish concentrator and then placed on a 23 cm (8 in) long Florisil column and eluted with 300 ml of 6% ethyl- ether-petroleum ether. This eluate was concentrated to approximately 10 m1 as previously described and evaporated 27 to dryness using a N-Evap evaporator, model 111, under a stream of dried air. Before GLC analysis, exactly 1 ml of petroleum ether was added to each sample. GLC Analysis Gas chromatograph analyses were carried out using a Tracor GLC, model 560, equipped with a 63 Ni electron cap- ture detector. The GLC was interfaced to a Digital PDP-8— Pamila GC data system. The column for GLC was a pyrex col— umn, 1.83 m (6 feet) long x A.0 mm i.d., packed with 3% OV—l on chromosorb W 80/100 mesh H.P. The carrier gas was nitrogen with a flow rate of A0 ml/min. Temperatures at the injection port, column, and EC detector were 270°C, 2A0°C, and 300°C, respectively. Standards were injected at the beginning of each run, after every 5 - 6 samples and at the end of the run. Quantitations were based on the peak area of the standards (hexabromobiphenyl peak used) and concentrations were expressed on the wet weight, dry weight, and weight of the fat components. In order to determine the percentage of recovery of the PBB residues by the utilized method, different levels of PBB (Fire Master® BP-6) were added to raw whole milk at concen- trations of 0.06, 0.02, and 0.006 ppm based on the total weight. The samples were extracted and cleaned up in dupli- cate at each level according to the Food and Drug Administra— tion method with the same modifications as previously des- cribed for the samples. The recovered fractions and percentage of recovery are 28 shown in Table 2. The values obtained are slightly low in comparison to the percent recoveries obtained by Gutenmann et a1. (1975). The presence of the PBB residues was confirmed by ultraviolet and mass spectrometric analysis. Ultraviolet degradation test was applied to raw whole milk and freeze— dried whole milk of Herd l, raw cream of Herd 3, pasteurized skim milk of Herd 2, and spray-dried whole milk of Herd A. These samples and a PBB (Fire Master€>BP-6) standard solu- tion were exposed to ultraviolet light for 0, 15, and 30 minutes in a Rayonet photochemical reactor, model 1162, equipped with short wave (25A NM) UV lamps. Gas chromato- graphic analyses were carried out immediately after expo- sure with the same conditions as that for the samples. Mass spectrometric analysis was run on a pool of all of the extracted samples. A GC-MS—CPU system was used which con— sisted of a Beckman GC-65 gas chromatograph interfaced to a DuPont 2l-A90 Mass Spectrometer which was in turn interfaced to a Digital PDP-l2 computer. The mass spectra was obtained at an ionizing voltage of 70 eV with a source temperature of 210°C. Analysis of the Data The PBB concentrations based on solids, total weight and on fat were analyzed for variance. Duncan's multiple range test (1957) was used to sort out significant differences 29 Table 2. Recoveries of PBB (Fire Master8)BP-6) from milka at different added levels by the Food and Drug Administration method. PBB concentration in milk, PEmi wet weight basis Percent recovery of PBB, Added Recovered ppm, wet weight basis 0.006 o.oo265° AA.17 0.02 0.0131“ 65.70 0.06 0.0A295 71.58 a Raw whole milk, average fat content A.l%. b Values are average of two determinations, except one for the 0.02 ppm added level. revealed by the analysis of variance. The PBB residue contents of spray-dried products and condensed whole milk were analyzed by the t statistic (Snedecor and Cochran, 1967). RESULTS AND DISCUSSION In order to investigate the effect of processing on the polybrominated biphenyl levels and their distribution in dairy products, milk from four dairy herds containing low levels of these compounds (<0.3 ppm) was processed into cream, skim milk, butter, and cheddar cheese. The effects of pasteurization, freeze-drying, condensation, and spray— drying processes were also studied. Distribution of Fractions The raw whole milk of each herd was first separated into the cream and skim milk fractions by using a conventional separator. The yields and percentage distribution of these fractions are reported in Table 3. The relative distribu- tion of the milk components showed that the skim milk frac- tion accounted for most of the total weight with yields ranging from 85 to 93%. The percentage distribution of the cream fractions was fairly consistent for the herds except for Herd 1 in which case was lower. Variations in the feed— ing rate of the milk into the separator during the fraction- ation process can slightly affect the composition of the cream fraction such that better separation can be achieved 30 31 oa.ooa om.ooa H~.mm sm.mm mpo>00mm mm.:a No.0:mm mm.:H mH.Homm mm.HH mm.ommz sm.n mm.azwm smote mm.mm mm.mmmoa Ho.mm nm.m:oma ma.mm mw.mszzm om.mm mm.:mmmm xHHE Efixm mm.:s:ma Hm.m::mH mm.~HHmm am.moamm xHHE oopMCOHuompmca m w w m u a cameo: a mocmaoz n m osmaoz u H ocmaoz ococoQEoo whom xHHz .mCOHpommm xHHS go zpo>oomp 6cm soapsnappmfio oprcoopom .moaofiw .m manme 32 with slower feeding rates resulting in cream fractions with higher lipid content but lower percentage distribution. In this study, yields of 7 to 15% of the cream fraction were obtained. The yields and percentage distribution of the pasteurized cream fractions used in butter processing are summarized in Table A. It was observed that the yields of butter were pro- portional to the butterfat content of the cream. Therefore, a percentage distribution of approximately 73% of the butter from the cream of Herd 1 (which had the highest fat content) was obtained. The yields of butter from Herds 2 and 3 were quite similar being approximately 23%. However, an unexpected higher relative distribution of butter, from Herd A (approxi— mately 36%) was observed. After the churning period, the butter fractions were washed with cold tap water and the excess water was removed by wrapping the product in cheese cloth and applying pressure manually. Further moisture deter— mination showed the butter from Herd A to have a higher moisture content which contributed to an apparent higher yield. This variation probably resulted from the inadequate removal of excess water. As can be seen in Table 5, the yield of cheese manufac- tured from the pasteurized milk was fairly similar for three of the herds, ranging from approximately 10 to 11%, but that the yield from Herd 3 was slightly lower. Numerous factors may influence yields of cheese including casein and fat contents of the milk, quality of starter cultures and 33 OH.OOH ss.mm om.ooa ss.mm aso>ooom mm.so m.mao mo.os m.:ms om.ss H.mme ms.mm s.msm gangsoooom mo.mm o.o:m sfi.mm s.amm mo.mm c.mam so.ms 0.3sm sooosm m.mmm H.mmm m.m:m w.mmm smmso Umvmcofi pomefiwCD m m w w a cameo; u ocwaoz a ocwaoz a cameo: : m H coapoMLm whom Emomo .opsomoopa wcfixme popusn CH mcofipommm Emomo mo mpm>ooom 6cm soapSQHmefio oprcmoan .moamfiw .3 dance u 3 os.mm no.3ssa mm.Hm mo.mmma mm.mw mw.smsa ms.mm no.3ssa hos: choose om.HH om.mmm ms.m :m.:sfi HH.OH ma.mom AN.HH mm.mmm choose oo.ooom oo.ooom oo.ooom oo.ooom xaas oaosz w w w w a zonwaoz u moswaoz u wcwaoz u Horses: ocosoosoo sass whom .mmsoooopa wcfixme omooao CH mcofiuommm xafie mo :ofipsnfippmfio mmmpcoopoo ocm moamfiw .m oHnt 35 rennet extracts, and conditions of processing. Care was taken to maintain processing conditions as similarly as possible for the four herds. Therefore, variations seen in the yields might be primarily attributable to the varia- tions in the fat content which in turn affects the casein content (direct correlation). Moisture and Lipid Analysis Determinations of moisture and lipid contents were done on the raw milk and processed dairy products of each herd. Their individual and product averages are shown in Tables 6 (moisture content) and 7 (lipid content). In general, the average moisture content of the raw and pasteurized products were quite similar. Variability seen between the raw and pasteurized cream of Herd l was due to the techniques involved in the two milk fractionation batches. A second separation was done to obtain samples of raw cream and skim milk. Although the conditions of this fractionation process were maintained as closely as possible to those used in the first batch (which fractions were subsequently pasteurized), the feeding rates of these two batches may have differed. Consequently, variations in fat contents were also found between the raw and pasteurized skim milk of this herd. Differences in the lipid content of individual herds between raw and pasteurized whole milk and cream were primarily due to the difficulty found in homogenizing the thawed samples. 36 Table 6. Average moisture contentsa, expressed as percent of total weight. Herd Product Product 1 2 3 A average Whole milk (raw) 86.20 86.50 86.70 86.80 86.55 Whole milk (pasteurized) 85.70 87.80 88.60 87.10 87.30 Skim milk (raw) 90.20 91.20 91.00 90.80 90.80 Skim milk (pasteurized) 90.20 91.00 91.00 90.20 90.60 Cream (raw) 55.00 60.30 73.80 60.70 62.A5 Cream (pasteurized) A0.00 66.70 73.A0 65.70 61.A5 Butter 1A.00 23.00 30.50 27.A0 23.73 Buttermilk 88.20 90.20 89.60 90.00 89.50 Cheese (fresh) 36.60 39.00 38.50 36.70 37.70 Cheese (aged) 3A.60 37.60 36.50 35.80 36.13 Cheese whey 92.70 91.70 92.00 91.90 92.08 Freeze dried skim milk 2.30 2.30 2.90 2.30 2.A5 Freeze dried whole milk 1.90 1.50 A.30 2.50 2.55 Freeze dried cream 0.50 0.80 0.90 0.90 0.76 Spray gried skim milk 6.60 2.60 - - A.60 Spray dried skim milk0 3.00 2.10 — - 2.55 Spray gried whole milk 1.80 0.30 - — 1.05 Spray dried whole milke - 1.10 - - 1.10 Evaporated milk 72.20 6A.70 - - 68.A5 a Values are average of two determinations. D Air outlet temperature 76.6-93.300 (170-2000F). 0 Air outlet temperature 87.8-98.8°C (190-2100F). Air outlet temperature 85.0-90.5°C (185-1950F). e Air outlet temperature 87.8-93.3°C (190-2000F). 37 Table 7. Average fat contentsa, expressed as percent of total weight. Product Herd Product 1 2 3 A average Whole milk (raw) A.AO A.lO A.90 A.50 A.A8 Whole milk (pasteurized) A.56 3.6A 2.95 A.55 3.93 Skim milk (raw) 0.77 0.15 0.2A 0.15 0.33 Skim milk (pasteurized) 0.15 0.15 0.22 0.16 0.17 Cream (raw) 37.A3 31.60 17.20 32.70 29.73 Cream (pasteurized) 55.58 27.20 l9.A0 27.00 32.29 Butter 83.50 75.77 66.25 69.60 73.78 Buttermilk 0.25 0.63 l.AO 0.61 0.72 Cheese (fresh) 30.80 33.77 22.00 28.00 28.6A Cheese (aged) 29.10 32.A0 22.10 27.80 27.85 Cheese whey 0.A6 1.09 0.A3 0.52 0.63 Freeze dried skim milk 0.A3 1.1A 1.00 0.55 0.78 Freeze dried whole milk 30.90 16.80 2A.20 32.50 26.10 Freeze dried cream 91.30 78.50 71.15 78.90 79.96 Spray gried skim milk 0.97 1.98 - - l.A8 Spray dried skim milkc 1.2A 2.00 - - 1.6A Spray dried whole milkd 28.15 27.A0 — - 27.78 Spray dried whole milke - 26.00 — - 26.00 Evaporated milk 7.80 10.10 — - 8.95 a Values are average of two determinations. Air outlet temperature 76.6-93.3°C (170-2000F). ° Air outlet temperature 87.8-98.8°C (190-210°F). Air outlet temperature 85.0-90.5°C (185-1950F). 9 Air outlet temperature 87.8-93.3°C (190-2000F). 38 Care was taken to homogenize the samples before an adequate aliquot was weighed for further analyses, but variability in the data indicated that this procedure was not always satisfactorily achieved even though the samples seemed visibly homogeneous. Residue Analysis Determination of PBB levels in the raw whole milk following the general procedure for chlorinated hydrocarbon pesticides outlined in the Pesticide Analytical Manual showed levels of 0.307, 0.186, 0.093, and 0.21A ppm on a fat basis for Herds l, 2, 3, and A, respectively. These values proved to be roughly 1.5 times higher than those obtained by the Michigan Department of Agriculture, although the ratio was not estimated for Herd 3 since levels for this herd were cited as trace amounts. Distribution of PBB The total amount of the PBB residues in the raw whole milk and their distribution and recovery in the physically separated milk fractions are summarized in Table 8. The concentration of PBBs was generally very low in the skim milk being approximately 3 to A.5% of the total amount, while the cream fraction contained most of the residues (39-73%) reflecting the association of PBBs with the lipid 39 .m manme Ca axozm mm muzwfioz cofipommm pcm xHHE HmpOp on» so comma mum mosam> m os.sg oo.m= oo.ss sm.ms mno>ooom ms.as com.o ss.mm osc.o mfl.ms mmm.o os.ms mom.o smote om.m moo.o mm.= woo.o H:.: Hmo.o oa.s mmo.o xaae Eaxm swm.o oma.o mw:.o :ms.o AHHE UmeCOfipompmca we we we we cameo; u a ocwaoz u mocmaoz n mocwfioz a m ososoosoo osom AHA: .mCOHpompm xHHE popwsmamm mfiamofimmca onp 2H >Lo>00op Ucm soapsnfippmfio mpfi ocm xHHE CH mmm Hmpoe .m oaome A0 phase. The recovered PBB fractions ranged from A3 to 78%. The lower recoveries might be attributable to the loss of the residues during the fractionation process since apprec— iable quantities of milk fat remained in the discs and on the walls of the separator. The data in Table 9 show the distribution of PBBs in butter and buttermilk during manufacture of butter from pasteurized cream. The percentage recoveries of residues in each fraction as well as total recoveries are also shown. In this study, cream was churned in an institutional mixer to allow separation of the butterfat from the buttermilk. The butter was further washed with cold tap water to reduce residual amounts of buttermilk. The results, again, show that PBB distribution was associated with the lipid content of the fractions. Butter contained approximately 53 to 115% of the total PBB whereas the accumulation of residue in the buttermilk was much lower with a range of 0 to 8%. Recover- ies were approximately 115, 55, 62, and 61% for Herds l, 2, 3 and A, respectively. Losses of PBB could be due to the losses of butterfat and buttermilk in the churning process and the washing of the butter fraction since appreciable amounts of butter remained on the paddle of the mixer and in the cheese cloth used to remove the excess water from the washed butter. The distribution and recovery of the PBB contents of the pasteurized whole milk in the cheese and cheese whey fractions are shown in Table 10. Once again, the higher Al .: dance :H exonm mm munwfioz cofipompm ocw emopo HepOp map no woman who mosfim> m ms.om ss.ao ms.sm mo.mHH mpo>ooom wo.m Hoo.o om.w moo.o mm.a Hoo.o 00.0 000.0 xafiepoppsm mn.wm umo.o mm.mm mHo.o mw.mm wmo.o oo.mHH HmH.o poupsm moo.o wmo.o mmo.o mma.o Scope ompmcoapompmcs we we we we w o w o w m w o u an H 3 u p: H 3 a p: H 3 a m»: H 3 ucocoano m m H emomo whom .wcfixme poupsn cfi 3po>ooop ocm soapsnfipumfio mpfi cam emmpo Ga mmm Hapoe .m canoe A2 .oom: xHHe mo pcwHoz Hmp0p on» so comma mum mosHm> m mH.mm mm.mm mm.Hm mm.es pHo>ooom :m.m moo.o em.Hm Hoo.o os.sm OHo.o 00.: moo.o soc: choose ms.m: :Ho.o mm.ms moo.o mo.mm MHo.o oo.ms . smo.o choose mmo.o soo.o smo.o. smo.o xHHs copHQOHpompmcD we we we we a cameo: R cameo; a osmaos a seemed: a m H unoCOQEoo whom xHHz .wcmee omooso :H mpo>oomp cam COszoprmHo mpH new xHHe CH mmm Hapoe .OH oHomB A3 amounts of the residues found in the cheese fraction (A9 to 7A%) reflected the preferential distribution of PBBs in the lipid portion of the product. Although the average fat content of the cheese whey fractions was less than 1%, the relative amounts of PBB present in these fractions (A.5 to 38%) were somewhat higher than those in the butter- milk milk when compared to the respective lipid contents. Montoure and Muldoon (1967) had observed DDE and TDE to be present in higher amounts in cheese whey from Monterey and cheddar cheese manufacturing while DDT was not detected. The recoveries of the residues in this processing ranged from approximately 58 to 95%. The less than 100% recovery of total PBB from the cheese and whey fractions was probably due to losses of the residues along with the whey and fat in the salting and pressing process of the curd. It was found in the present study that PBBs are associated with the lipid phase of the products substanti- ating the similarities of these compounds to polychlorinated biphenyls and chlorinated hydrocarbon pesticides. Through physical separation of milk, Li et a1. (1970) and Langlois et al. (196A) showed that organochlorine pesticides were significantly associated with the fat fractions cf the cream and butter. Platonow et a1. (1971) observed in their study that physiologically incorporated PCBs had a strong preiilcc- tion for the lipid phase of the processed dairy products and accumulation of DDE and PCB residues in the milk fat have been pointed out by Fries et a1. (1973). Additionally, Ang AA (1970) indicated that butteroil contained more dieldrin than butter on a weight basis. Effect of Processing in the PBB Levels The PBB residue contents of the raw whole milk and the manufactured dairy products under pasteurization, butter and cheese making, and freeze-drying processes are illustrated in Table 11, on the total weight basis and in Table 12, on the total solids basis. Analysis of variance revealed signi- ficant differences in the PBB contents of the dairy products on both a total weight and total solids basis. Duncan's mul- tiple range test was used to pinpoint these significant dif- ferences among the products and these results are presented in Table 13. Butter contained significantly more PBB than buttermilk at the 5% level of probability on both the solids and total weight basis. Fresh cheese also contained more PBB than cheese whey. In addition, butter and freeze-dried cream had significantly (p<0.05) higher PBB levels than pasteurized cream when the residues were expressed on the total weight basis. The results presented in Table 1A show the PBB contents of the raw milk and dairy products on a fat basis. Analysis of variance showed the PBB levels expressed on a fat basis were run; significantly different in these dairy products. Mann et al. (1950) found pasteurization had very little effect on the DDT residues in dairy products. Although the differ- ences found in the present study were not significant, “5 .egg H00.0 ems» mmmH mo mosz> oquHUcH mosHm> 000.0 mmCOHumcHELopoo o3» mo mwmpo>w ohm mmsHm> 0 050.0 H 00H.0 02H.0 m50.0 NMH.0 H:N.0 Ewmho UmHhU mnmmhm 000.0 H H00.0 000.0 H00.0 H00.0 000.0 xHHe ewa 00pr mumopm 0m0.0 H zm0.0 mm0.0 Hm0.0 0N0.0 00H.0 xHHe 0H023 00pr onmonm H00.0 H H00.0 H00.0 H00.0 m00.0 H00.0 30:3 omoono ~H0.0 H :20.0 000.0 :m0.0 ~m0.0 :m0.0 mmoono vow< mm0.0 H Hm0.0 030.0 0N0.0 5:0.0 000.0 mwmmno Smmhm 000.0 H H00.0 H00.0 H00.0 H00.0 H00.0 xHHELoppsm :w0.0 H 00H.0 030.0 030.0 000.0 NHm.0 poppsm mm0.0 H 000.0 300.0 0H0.0 wm0.0 mmH.0 emopo UoNHpsmpmmm Hm0.0_H :m0.0 mm0.0 HH0.0 mm0.0 mmo.0 emcee 303 000.0 H 000.0 000.0 000.0 000.0 000.0 xHHE Efixm UmNHhsmpmmm 000.0 H 000.0 000.0 000.0 m00.0 000.0 xHHe eme 30m 300.0 H 000.0 H00.0 H00.0 000.0 NH0.0 xHHE mHonz UmNHhsmpwwm 300.0 + 000.0 0H0.0 m00.0 000.0 MH0.0 xHHe oHonz 30m COHpmH>mU Uhmccmpm : m m H + Upom posvopm some 00:00pm .Apanoz prop so comma egav mposvopa 3LH00 UmpmHoL 0:0 xHHe CH wosvaop mmm .HH mHDmB A6 mmoH no mosHm> opmoHccH moSHw> 000.0 .eaa mmcoHpmcHemopoU 030 no mwmpo>m H00.0 ems» ohm mosHm> 0 050.0 H 53H.0 H3H.0 m50.0 MMH.0 m3m.0 emcee cloU mummmm 000.0 H H00.0 000.0 H00.0 H00.0 000.0 xHHe eme ooHpU onoopm 0m0.0 H 000.0 mmo.0 mm0.0 0m0.0 00H.0 xHHe mHonz Uonv oncomm 000.0 H 3H0.0 500.0 000.0 5m0.0 MH0.0 mos: omomno mmo.o H 000.0 350.0 mmo.o mmo.o mmo.o onooco come 0M0.0 H m00.0 050.0 030.0 mwo.0 mmH.0 omooso amoem m00.0 H 000.0 000.0 MH0.0 500.0 000.0 xHHepoppsm 050.0 H mmH.0 00H.0 H50.0 3HH.0 mmm.0 moppsm 300.0 H 5MH.0 0MH.0 m50.0 3HH.0 3mm.0 emopo ooNHnsopmmm H0040 H 0mH.0 NMH.0 000.0 03H.0 30m.0 emopo zmm H00.0 H m00.0 m00.0 m00.0 m00.0 m00.0 xHHe eme ooNHmdopmmm H00.0 H m00.0 000.0 m00.0 m00.0 300.0 xHHe ewa 303 0m0.0 H 530.0 0m0.0 mH0.0 030.0 mwo.0 xHHe oHonz UmNHpsopmmm 5m0.0 + 000.0 m50.0 3m0.0 000.0 500.0 xHHe 0H033 303 COHpmH>oU onwccmpm 3 m m H + whom poscopm some nonvopm .AmuHHom co comma eaav mposoopa 5pHmo oopmHop 0cm xHHe HS” moSUmep mmm .mH oHnt A7. a Table 13. Summary of the significant differences of the PBB residue contents between products. PBB Product mean Statistical content (ppm) + standard significance based on Product deviation p<0.05 Butter (A) 0.136 1 0.079 AI>B Buttermilk (B) 0.009 1 0.003 Solid Fresh cheese (C) 0.082 i 0.038 C2>D Cheese whey (D) 0.01A i 0.009 Butter (A) 0.108 i 0.07A Ai>B Buttermilk (B) 0.001 1 0.000 Butter (A) 0.108 t 0.07A A3>E Pasteurized Total cream (E) 0.060 : 0.052 weight Fresh cheese (C) 0.051 : 0.025 CI>D Cheese whey (D) 0.001 : 0.001 Freeze dried cream (F) 0.1A6 : 0.070 F‘>E Pasteurized Cream (E) 0.060 1 0.052 a Duncan's Test (1957). A8 .mCOprcHE3opo0 033 30 0w030>0 03m mosH0> .m 0 300.0 H 333.0 333.0 303.0 003.0 300.0 sco3o 00330 cuoo33 330.0 H 000.0 000.0 300.0 000.0 300.0 3335 E330 06330 030633 300.0 H 003.0 003.0 003.0 033.0 33m.0 333s 03oz; 0o330 onoo33 330.0 H 033.0 003.0 003.0 000.0 330.0 00:3 000030 000.0 H 033.0 330.0 003.0 033.0 m3m.0 onocno come 030.0 H 033.0 003.0 003.0 0m3.0 030.0 onoono 30033 003.0 H 033.0 033.0 000.0 003.0 000.0 3333300000 300.0 H 333.0 033.0 030.0 033.0 000.0 300000 300.0 H 303.0 m33.0 000.0 333.0 m3m.0 smo3o 06333000000 000.0 H 033.0 003.0 003.0 333.0 030.0 50636 300 m50.0 H HOH.0 mHH.0 000.0 mHm.0 3mm.0 xHHe eme 00NH3303000 000.0 H 333.0 003.0 300.0 033.0 000.0 3335 E330 :00 300.0 H 003.0 033.0 030.0 303.0 000.0 3335 63oz; oou3hspomnm 000.0 + 000.0 330.0 000.0 003.0 300.0 3335 03oz: smm so3em3>oe 03000000 3 m m 3 + 0303 0000030 cmme 3050033 .33303300 0HQHH H0003 00 00000 egav 03000030 33H00 0030H03 030 xHHe 0H 0:0Hmo3 mmm .3H oHnt A9 pasteurization seemed to slightly decrease the concentration of PBBs in the whole milk and cream from 0.200 to 0.150 ppm, and from 0.170 to 0.16A ppm, respectively. A slight increase of PBB residues (0.111 to 0.161 ppm) was observed in the skim milk under pasteurization but this variation resulted from the difficulty found in the quantitation of the very low PBBs found in the skim milk variables. It appeared that butter— milk contained more PBBs on a fat basis than butter and pas- teurized whole milk. This fact was also observed for the pasteurized skim milk and whey when the levels of PBB were compared to the pasteurized whole milk. Li et a1. (1970) pointed out that buttermilk, skim milk, and whey contained higher concentrations, on a fat basis, of the studied chlor- inated hydrocarbon pesticides than did the whole milk. These authors suggested that the pesticides were associated with the lipoprotein portion of the products. Attempts to deter- mine the possible causes of the apparent higher quantities of PBBs in these fractions were not made. Ang (1970) found that refined skim milk contained more dieldrin than whole milk on a fat basis. She suggested that dieldrin was slightly soluble in the serum of milk, but she also showed that the levels in the skim milk and whole milk were comparable and that lower levels were found in buttermilk and butter serum. On the studies of the freeze-drying process, an appar- ent higher concentration of the PBB residues, on a fat basis, in the freeze-dried whole milk and cream was observed when compared to the levels in the pasteurized whole milk 50 and cream, respectively. In this aspect, the possibility of occurrence of contamination from the glasswares used in the extraction and clean up procedures of these samples was con— sidered, though it seems to be not very probable since glass- ware was carefully rinsed with acetone and petroleum ether before each use. On the other hand, gas chromatographic analyses of the freeze-dried samples showed the presence of several unidentified peaks with shorter retention times than the hexabromobiphenyl peak (quantitated peak) at relative concentrations as high as or even higher than the hexabrom- inated biphenyl peak. Data from the confirmation test by ultraviolet irradiation of freeze-dried whole milk showed the complete disappearance of these extraneous peaks after 30 minutes of ultraviolet light exposure suggesting that they may have been halogenated compounds. Although attempts to identify these compounds by mass spectrometry were not made, they could have been brominated biphenyls with different degrees of bromination formed under the freeze-drying condi— tions. If this is true, or if other compounds with comparable retention time to the hexabrominated biphenyl compound were present in the milk, the apparent higher levels of PBBS in the freeze—dried whole milk and cream may be justified. It should also be mentioned that quantitation of PBB in the freeze-dried skim milk was found to be extremely difficult due to the presence of interfering peaks with very close retention times to the hexabromobiphenyl peak. The effect of aging for a period of approximately two 51 months on the manufactured cheddar cheese showed very little effect on the PBB levels. In fact, the levels found in fresh cheese and aged cheese were comparable. The result of the present study closely agrees with the data of Montoure and Muldoon (1967) on the DDE, TDE, and DDT residue levels in aged Monterey and Cheddar cheese. In general, the relative concentrations of PBB, on a fat basis, showed that these applied processes were not effective in the removal of the PBB contents. This observa- tion supports the earlier findings of Platonow and Funnell (1971) for PCBs during processing of cream, cottage cheese, and skim milk. They reported these processings did not con- tribute to significant changes in the polychlorinated biphenyl levels in these products. Condensation has been shown to be effective in the reduction of telodrin residues in milk (Stemp and Liska, 1966). Pasteurized whole milk of Herds l and 2 was evapor- ated to study the effect of condensation of the PBB levels. Data illustrated in Tables 15 and 16 show the PBB contents on a solids and fat basis, respectively, of the pasteurized whole milk and evaporated whole milk from the individual herds. Results of the statistical analysis are also pre— sented in these tables. The concentration of the residues of the evaporated milk and pasteurized whole milk of Herd l were revealed to be significantly different at the 5% level of probability by the t statistic. However, it was observed that the evaporated milk had a higher PBB content than the 52 Table 15. Effect of condensation on the PBB residue levelsa of pasteurized whole milk (ppm based on solids) Liquid mean Evaporated mean t er standard deviation standard deviation statistic 1 0.0829 : 0.0022 0.108A 1. 0.0016 -13.257* 2 0.0AA7 + 0.002A 0.0A26 + 0.0086 0.333 a Values are average of two determinations. * Significant at the 5% level of probability. Table 16. Effect of condensation on the PBB residue levelsa of pasteurized whole milk (ppm based on total lipid content). Liquid mean Evaporated mean t Herd + + standard deviation standard deviation statistic 1 0.2598 1 0.0110 0.3870 1 0.0123 -10.902* 2 0.1509 1 0.0033 0.1A9l 1 0.0273 0.092 a Values are average of two determinations. * Significant at the 5% level of probability. 53 pasteurized whole milk had, showing an increase in the PBB levels of about 30—A0%. Data for Herd 2 showed the residue levels in the evaporated milk to be less than those for pasteurized whole milk but the difference was not significant. Earlier reports have emphasized that spray—drying was effective in the removal of chlorinated hydrocarbon pesti- cides from milk. Therefore, in the present study, spray- drying of whole milk and skim milk was done to study the effect of this processing on the PBB residues. Data given in Tables 17 and 18 shows the PBB residue contents based on solids and on total lipid content, respectively, of whole milk and skim milk of Herd 1, before and after spray—drying. Statistical analysis established that the levels in the spray- dried whole milk were significantly lower than the levels in pasteurized whole milk, on a solids basis, showing a reduc— tion of 26.5% of PBBS. Nevertheless, when the residues were expressed on a fat basis, the difference was not significant even though the levels in the spray-dried whole milk were lower. In considering the spray-dried skim milk at air out- let of 87.8 - 93.3°C, the levels in this product were less than the levels in pasteurized skim milk but the difference was not significant. Tables 19 and 20 show PBB levels on a solids and fat basis, respectively, in the pasteurized and spray-dried whole milk and skim milk of Herd 2. Significant reductions in the PBB residue contents by spray-drying were established at the 5% level of probability by the t statistic. The PBB levels 5A Table 17. Effect of spray-drying on PBB residue contentsa of pasteurized whole milk and skim milk of Herd 1 (ppm based on solids). Liquid mean Spray-dried mean t Prod ct + + u standard deviation standard deviation StatiStiC Whole b milk 0.0829 : 0.0022 0.0622 1 0.0011 11.902* Skim c milk 0.003A 1 0.0007 0.0035 : 0.0003 -1.857 Skim d milk 0.003A : 0.0007 0.0008 : 0.0003 A.828 a Values are average of two determinations. b Air outlet temperature 87.8°C (190°F). c Air outlet temperature 76.6°C (170°F). d Air outlet temperature 87.8-93.3°C (190—200°F). * Significant at the 5% level of probability. 55 Table 18. Effect of spray-drying on PBB residue contentsa of pasteurized whole milk and skim milk of Herd 1 (ppm based on total lipid). Liquid mean Spray-dried mean t PrOduCt + + stati tic standard deviation standard deviation 3 Whole b milk 0.2598 1 0.0110 0.217A : 0.0086 A.29A Skim 0 milk 0.2336 : 0.0611 0.3371 : 0.0202 -2.275 Skim d milk 0.2336 i 0.0611 0.0600 1 0.0208 3.80A a Values are average of two determinations. b Air outlet temperature 87.8°C (190°F). c Air outlet temperature 76.6°C (1700F). d Air outlet temperature 87.8-93.3°C (190-200°F). 56 Table 19. Effect of spray-drying on the PBB residue contentsa of pasteurized whole milk and skim milk of Herd 2 (ppm based on solids). Liquid meanb Spray-dried mean t Product + + standard deviation standard deviation statistic Whole 0 milk 0.0AA7 i 0.002A 0.0268 i 0.002A 7.A58* Whole d milk 0.0AA7 : 0.002A 0.0282 i 0.0006 11.175* Skim d milk 0.003A i 0.0001 0.0017 : 0.0006 3.952 Skim milk 0.003A : 0.0001 0.001A : 0.0002e 12.6A9* a Values are average of the determinations. PBB residue content of pasteurized whole milk and pasteurized skim milk. Air outlet temperature 85-90.5°c (185-1950F). Air outlet temperature 87.8-93.3°C (190-2000F). Air outlet temperature 87.8-98.8°C (190-2100F). * Significant at the 5% level of probability. 57 Table 20. Effect of spray drying on the PBB residue contents of pasteurized whole milk and skim milk of Herd 2 (ppm based on total lipid). Liquid meanb Spray-dried mean t + + Product standard deviation standard deviation statistic Whole 0 milk 0.1509 i 0.0033 0.0956 1 0.0101 7.360* Whole d milk 0.1509 : 0.0033 0.1071 1 0.0018 l6.A79* Skim d milk 0.2115 1 0.0013 0.0829 1 0.0252 7.207* Skim e milk 0.2115 : 0.0013 0.0662 : 0.0018 92.5A6** a Values are average of two determinations. PBB residue content of original product. Air outlet temperature 85-90.5°C (185—1950F). Air outlet temperature 87.8-93.3°C (190-2000F). e Air outlet temperature 87.8 —98.8°C (190-2100F). * Significant at the 5% level of probability. ** Significant at the 1% level of probability. 58 in the spray-dried whole milk were reduced by approximately 36% (fat basis) and A0% (solids basis) at air outlet temper- atures of 85 — 90.5°C, and 30% (fat basis) and 38% (solids basis) at air outlet temperatures of 87.8 - 93.3°C. Addi- tionally, it was observed that the spray-dried skim milk at air outlet temperatures of 87.8 - 93.3°C showed levels of PBB significantly different than the levels in the pasteur- ized skim milk, on a fat basis, with reduction of approximately 61% in the residues of the spray-dried product. Nevertheless, when expressed on a solids basis, the difference was not significant. Comparable losses of PBB from the skim milk on a solids and fat basis were obtained when the skim milk was spray-dried at air outlet temperatures of 87.8 - 98.8°C. It was estimated that significant losses of approximately 69% (fat basis) and 67% (solids basis) of PBB occurred during the spray-drying process at this specified condition. From these studies it was observed that losses of PBB from skim milk were greater than from whole milk, roughly 30 to A0% from whole milk and 60 to 69% from skim milk, suggesting these compounds were more easily removed from low fat content products and/or when the levels were lower. It is possible that some PBB is distributed in the serum phase of milk and that this quantity could be more easily removed than the PBB associated with the lipid phase. Addi— tionally, the greater surface area formed during the spraying of the liquid product and the relatively high temperatures at which the formed particles are exposed could facilitate a more easily volatilization of the PBBs. For the same 59 reasons, the findings of significant reductions in the PBB levels only in the spray-dried products may be justified and further indicate the data obtained in these studies with PBBs confirms earlier reports with halogenated compounds. Langlois et al. (196A) indicated that significant losses of DDT and lindane occurred during spray—drying treatment of whole milk. Reduction of dieldrin, lindane, and chlordane in the spray-drying of whole milk was observed by Li et a1. (1970). Telodrin residues were reduced by approximately 10 - 20% during drying of milk (Stemp and Liska, 1966). Comparable reduction effects were pointed out by Liska (1968) during the spray—drying of milk containing lindane, hepta- chlor, and heptachlor epoxide. SUMMARY AND CONCLUSIONS The effect of processing on the polybrominated biphenyl levels and their relative distribution in dairy products were studied. Milk containing less than 0.3 ppm (fat basis) was separated into skim milk and cream fractions followed by pasteurization of the raw milk and the fractions. These products were further processed into butter, cheddar cheese, and freeze-dried whole milk, skim milk, and cream. Spray- drying of whole milk and skim milk and condensation of whole milk were done for two of the herds. The relative distribution of milk components showed skim milk accounted for 85 - 93% while cream fraction ranged from 7 to 15%. The yields of butter were proportional to the fat content of the cream used for its manufacture and were approx- imately 73, 23, 23, and 36% for Herds l, 2, 3, and A, respec- tively. The yields of cheese were also proportional to the lipid content with percentage distribution ranging from approximately 9 to 11%. Moisture, lipid, and residue analyses were carried out for the raw whole milk and the processed dairy products. Moisture and lipid data were characteristic of the type of dairy product being produced. Recoveries of the total PBB residues during the physical separation of whole milk into skim milk and cream, as well as during butter and cheese manufacture ranged from A3 to 115%. Therefore, some reduction 60 61 of PBBs occurred as a result of loss of lipid during these processes. Skim milk contained negligible amounts of the PBB residues (3 - A%) while cream (39 - 73%), butter (55 - 115%), and cheese (58 - 95%) contained most of the residues. These data indicated the lipid solubility of PBBs as well as the expected accumulation of these compounds in the high-fat content products. From the stand point of minimizing contam- inant levels in foods, the physical separation showed promis- ing results, but precise information regarding economical feasibility of using only skimmed milk from PBBs contaminated milk for food consumption should be examined. Analysis of variance showed no significant difference in the PBB levels (fat basis) in the raw and pasteurized whole milk, cream, and skim milk, butter, buttermilk, cheese, cheese whey, and freeze-dried whole milk, skim milk, and cream. The pasteurized products had slightly lower amounts of the residues indicating pasteurization had little effect on the reduction of PBB levels. Buttermilk, skim milk, and cheese whey contained slightly higher concentrations of the residues than the pasteurized whole milk, on a fat basis, confirming the earlier findings of Li et a1. (1970). Significant differences in the PBB contents were estab- lished for the products when expressed on a total weight and solids basis. Duncan's multiple range test was used to sort out the differences among the products. The data showed PBB levels were proportional to the lipid content of the products. Comparable levels of PBB were found in aged and fresh 62 cheese indicating aging had very little or no effect on removal of these residues. Gas chromatographic analysis showed that only the freeze—dried products contained several unidentified peaks at abnormally high concentrations. Exposure of freeze—dried whole milk to ultraviolet light promoted the complete dis— appearance of these peaks suggesting they may have been brominated biphenyls with different degrees of halogenation. Whole milk from two of the herds was condensed to pro— duce evaporated milk with 30% solids. Use of the t statistic revealed condensation was ineffective in reducing PBB resi- dues. A corollary study investigated the possibility of removal of the PBB residues by spray-drying. Pasteurized whole milk and skim milk of Herds l and 2 were spray-dried using pilot plant equipment. Significant reductions in the PBB levels in the spray-dried whole milk (ppm based on solids) for both herds were established with losses of approximately 30 to A0% of PBBS. Spray-dried skim milk of Herd 2 contained significantly less PBBS than the pasteurized skim milk, based on the lipid contents with losses ranging from 60 to 69%. When expressed on a solids basis, the significant reduction showed to be approximately 67%. Significant losses of PBBS only during the spray-drying process indicated these compounds were more easily removed when the surface area of the milk particles was increased and exposed to relatively high temperatures, thus facilitating 63 greater volatilization of these residues. The findings on the spray—drying studies indicate this method of processing may have promise for the purpose of reducing PBB levels in contaminated milk. SUGGESTIONS FOR FURTHER RESEARCH The questions and observations raised in the-present study suggest the following additional research: 1 - The relative distribution of the metabolites of DDT (DDE and TDE) have been reported in various phases during the manufacture of cheese. Similarly, the association observed between higher levels of PBBs with cheese whey should be explored further with the purpose of clarifying the possible factors involved in this distribution. 2 - Sufficient quantities of milk should be freeze-dried to facilitate identification of the unknown peaks found in this study. If these compounds are partially debrominated PBBs, the implications of freeze-drying conditions for debromination of polybrominated biphenyls would warrant fur- ther investigation. The study should be extended to include a toxicological evaluation of lower brominated biphenyl compounds. 3 - The correlated higher reduction of PBBs in lower fat content products found during spray-drying should be elabor— ated further on a larger number of samples with varying £885 and lipid concentrations. A - The feasibility of PBB removal from milk by spray- drying should be investigated on a larger scale using commer— cial conditions. 6A REFERENCES REFERENCES Abbott, D. C., R. B. Harrison, J. O' G. Tatton, and J. Thomson. 1966. Organochlorine pesticides in the atmosphere. Nature 211:259. Anderson, D. W., J. J. Hickey, R. W. Risebrough, D. F. Hughes, and R. E. Christensen. 1969. Signifi- cance of chlorinated hydrocarbon residues to breeding pelicans and cormorats. Can. Field Nature 83:89. Ang, C. Y. N. 1970. Studies on the specific distribu- tion of dieldrin in milk fraction. Ph.D. Thesis, Michigan State University. Anonymous. 1975. Feed contaminant in farmer's blood. Chem. Eng. News 53(1): Feb. 2A. Babish, J. G., W. H. Gutenmann, and G. S. Stoewsand. 1975. Polybrominated biphenyls: Tissue distri- bution and effect on hepatic microsomal enzymes in Japanese quail. J. Agric. Food Chem. 23(5):879. Bills, D. D., and J. L. Sloan. 1967. Removal of chlorin- ated insecticides residues from milk fat by molecular distillation. J. Agric. Food Chem. 15:676. Biros, F. J., A. C. Walker, and A. Medbery. 1970. Poly- chlorinated biphenyls in human adipose tissue. Bull. Environ. Contam. Toxicol. 5:317. Bradley, R. L., Jr., and C. L. Li. 1968. The effect of dieldrin on acid development during manufacture of cheddar cheese. J. Milk Food Technol. 31:202. Brown, J. R. 1967. Organochlorine pesticide residues in human depot fat. Can. Med. Assoc. J. 97:367. Bunce, N. J., S. Safe, and L. O. Ruzo. 1975. Photochem- istry of bromobiphenyls: Steric effects and electron transfer. J. Chem. Soc. Perkin Trans. I:l607. 65 66 Carnes, R. A., J. U. Doerger, and H. L. Sparks. 1973. Polychlorinated biphenyls in solid waste and solid- waste-related materials. Arch. of Environ. Contam. Toxicol. 1:27. Cummings, J. G. 1966. Pesticides in the total diet. "Residue Reviews." V16 Springer-Verlag/Berlin— Heildelberg - New York. Detering, C. N., L. R Prewitt, R. M. Cook, and G. F. Fries. 1975. On the rate of excretion of polybrominated biphenyls by lactating holstein. J. Ani. Sci. A1 (ll):265. Doguchi, M., and S. Fukano. 1975. Residue levels of poly- chlorinated terphenyls, polychlorinated biphenyls and DDT in human blood. Bull. Environ. Contam. Toxicol. 13:57. Duggan, R. E., and K. Dawson. 1967. A report on pesticides in food. FDA Papers le. Duggan, R. E., and G. Q. Lipscomb. 1969. Dietary intake of pesticide chemicals in the United States (II). Pest. Monit. J. 2:153. Duncan, D. B. 1957. Multiple range test for correlated and heteroscidastic means. Biometrics 13:16A. Dyment, P. G., and L. M. Hebertson. 1971. Human milk and serum from Texas and human milk from New Guinea. Bull. Environ. Contam. Toxicol. 6:532. Fries, G. F., G. S. Marrow, and C. H. Gordon. 1969. Com- parative excretion and retention of DDT analogs by dairy cows. J. Dairy Sci. 52:1800. Fries, G. F., and G. S. Marrow.' 1972. Relationship of milk fat and body fat concentrations of organochlorine compounds in cows. J. Dairy Sci. 55:706. Fries, G. F., L. W. Smith, H. C. Cecil, J. Bitman, and J. R. Lillie. 1973. Retention and excretion of polybromin— ated biphenyls by hens and cows. Presented at the 165th meeting, American Chemical Society, Dallas, Texas. Fries, G. F., G. S. Marrow, and C. H. Gordon. 1973. Long— term studies of residue retention and excretiép by cows fed a polychlorinated biphenyl (Aroclor 125A). J. Agric. Food Chem. 21:117. Fries, G. F., and G. S. Marrow. 1975. Excretion of polybrom- inated biphenyls into the milk of cows. J. Dairy Sci. 58(6):9A7. 67 Fries, G. F., G. S. Marrow, C. N. Detering, L. R. Prewitt, and R. M. Cook. 1975. Distribution of polybrominated biphenyl residues in tissues of dairy cattle. J. Dairy Sci. 58(5):76A. Fujiwara, K. 1975. Environmental and food contamination with PCBs (polychlorinated biphenyls)in Japan. Sci. Total Environ. A(3):219. Furr, A. K., D. R. Mertens, W. H. Gutenmann, C. A. Bache, and D. J. Lisk. 197A. Fate of polychlorinated biphen— yls, metals, and other elements in papers fed to lacta- ting cows. J. Agric. Food Chem. 22:95A. Grant, D. L., W. E. J. Phillips, and D. C. Villeneuve. 1971. Metabolism of a polychlorinated biphenyl (AroclorR 125A) mixture in the rat. Bull. Environ. Contam. Toxicol. 6:102. Greb, W., W. Klein, F. Coulston, L. Goldberg, and F. Korte. 1975. Beitrage zur okologischen chemie LXXXIV_.__114 Metabolism of lower polychlorinated biphenyls C in the Rhesus monkey. Bull. Environ. Contam. Toxicol. 13(A):A71. Gutenmann, W. H., and D. J. Lisk. 1975. Tissue storage and excretion in milk of polybrominated biphenyls in rum- inants. J. Agric. Food Chem. 23(5):1005. Holden, A. V., and K. Marsden. 1967. Organochlorine resi- dues in seals and porpoises. Nature 216:127A. Holmes, D. C., J. H. Simmons, and J. O' G. Tatton. 1967. Chlorinated hydrocarbons in British wildlife. Nature 216:227. Holub, B. J.,.I Piekarski, and K. Nilsson. 1975. The effect of a PCB (2,A,2',A'-tetrachlorobiphenyl) on lipid-synthesizing enzymes in rat liver microsomes. Bull. Environ. Contam. Toxicol. 1A:A15. Iverson, F., D. C. Villeneuve, D. L. Grant, and G. V. Hatina. 1975. Effect of Aroclor 1016 and l2A2 on selected enzyme systems in the rat. Bull. Environ. Contam. Toxicol. 13(A):A56. Jackson, T. F., and F. L. Halbert. 197A. A toxic syndrome associated with the feeding of polybrominated biphenyl- contaminated protein concentrate to dairy cattle. J. Am. Vet. Med. Ass. 165:A37. 68 Jan, J., M. Komar, and M. Milohnoja. 1975. Excretion of some pure PCB isomers in milk of cows. Bull. Environ. Contam. Toxicol. 13:313. Jensen, S. 1966. Report on a new chemical hazard. New Scientist 32:612. Jensen, 8., A. G. Johnels, S. Olsson, and G. Otterlind. 1969. DDT and PCB in marine animals from Swedish waters. Nature 22A:2A7. Kerst, A. J. 197A. PBB. Presented to the Michigan Envir— onmental Review Board, September 23. Kiermeier, F., M. Rudolphi, and G. Wildbrett. 1967. Behavior of chlorinated hydrocarbon insecticide residues during freeze-drying of milk. L. Lebensm.- Unters, Forsch. 133:78. In Chem. Abstr. 67:10AA3U. Kim, S. C. 1969. Relationship between some insecticides and lactic culture organisms in milk. Ph.D. Thesis, Michigan State University. Kolbye, A. C. 1972. Food exposures to polychlorinated biphenyls. Environ. Health Perspect. 1:85. Kosikowski, F. 1966. "Cheese and Fermented Milk Foods." pp. 212-21A. Published by the author, Ithaca, New York. Kroger, M. 1968. Effect of various physical treatments on certain organochlorine hydrocarbon insecticides found in milk fat. J. Dairy Sci. 50:32A. Kuratsune, M. 1969. An epidemiological study on Yusho. Fukuoka Acta Med. 62:513. Langlois, B. E., B. J. Liska, and D. L. Hill. ‘196A. The effects of processing and storage of dairy products on chlorinated insecticide residues. 1. DDT and lindane. J. Milk Food Technol. 27:26A. Langlois, B. E., B. J. Liska, and D. L. Hill. 1965. The effects of processing and storage of dairy products on chlorinated insecticide residues. II. Endrin, dieldrin, and heptachlor. J. Milk Food Technol. 28:9. Ledford, R. A., J. H. Chen, and W. F. shipe. 1968. Effect of direct steam heating and vacuum treatment on levels of pesticide residues in milk. J. Dairy Sci. 51:219. Ledford, R. A., and J. H. Chen. 1969. Degradation of DDT and DDE by cheese microorganisms. J. Food Sci. 3A:386. Li, C. F., and R. L. Bradley, Jr. 1969. Degradation of chlorinated hydrocarbon pesticides in milk and butter- oil by ultraviolet energy. J. Dairy Sci. 52:27. 69 Li, C. F., R. L. Bradley, Jr., and L. H. Schultz. 1970. Fate of organochlorine pesticides during processing of milk into dairy products. J. Assoc. Off. Anal. Chem. 53:127. Lichtenstein,E. P., K. R. Schultz, T. W. Fuhremann, and T. T. Li Ang. 1968. Biological interaction between plasticizers and insecticides. J. Econ. Entomol. 62:761. Lipscomb, 0. Q. 1968. Residues in food and feed: Pesti- cide residues in prepared baby foods in the United States. Pest. Monit. J. 2:10A. Liska, B. J. 1968. Effects of processing on pesticide residues in milk. J. Anim. Sci. 27:827. Mann, H. D., R. H. Carter, and R. E. Ely. 1950. The DDT con- tent of milk products. J. Milk Food Technol. 13:3A0. Martell, J. M., D. A. Rickert, and F. R. Siegel. 1975. PCBs in suburban watershed, Reston, Va. Environ. Sci. Technol. 9(9):772. Marth, E. H., and B. E. Ellickson. 1959. Insecticide resi— dues in milk and milk products. III. Insecticide residues in dairy products and associated problems. J. Milk Food Technol. 29:179. Musial, C. J., O. Hutzinger, V. Zitko, and J. Crocker. 197A. Presence of PCB, DDE and DDT in human milk in the provinces of New Brunswick and Nova Scotia, Canada. Bull. Environ. Contam. Toxicol. 12:258. Norris, J. M., J. W. Ehrmantrant, G. L. Gibbons, R. J. Kociba, B. A. Schwetz, J. Q. Rose, 0. G. Humiston, G. L. Jewett, W. B. Crummett, P. J. Gehring, J. R. Tirsell, and J. S. Brosier. 197A. Toxicological and environmental factors involved in the selection of decabromobiphenyl oxide as a fire retardant chemical. J. Fire Flammability Combustion Toxicol. 1:52. Peakall, D. B., and J. L. Lincer. 1970. Polychlorinated biphenyls. Another long-life widespread chemical in the environment. Bio. Sci. 20:958. Pesticide Analytical Manual, V.I., U.S. Department of Health, Education and Welfare, Food and Drug Administration, Washington, D. C., revised 1971, Sections 211.13h, 211.1Aa, and 211.1Ad. 70 Platonow, N. S., H. S. Funnell, D. H. Bullock, D. R. Arnott, P. W. Saschenbrecker, and D. G. Grieve. 1971. Fate of polychlorinated biphenyls in dairy products pro- cessed from the milk of exposed cows. J. Dairy Sci. 5Azl305. Prewitt, L. R., R. M. Cook, and G. F. Fries. 1975. Field observations of Michigan cattle contaminated with poly— brominated biphenyls. J. Dairy Sci. 58(5):76A. Reynolds, L. M. 1969. Polychlorinated biphenyls (PCBs) and their interference with pesticide residue analysis. Bull. Environ. Contam. Toxicol. A:l28. Rhee, K. S., and F. W. Plapp, Jr. 1973. Polychlorinated biphenyls (PCBs) as inducers of microsomal enzyme activity in the house fly. Arch. Environ. Contam. Toxicol. 1:182. Risebrough, R. W. 1968. Polychlorinated biphenyls in the global ecosystem. Nature 220:1098. Ruzicka, J. H. A., J. H. Simmons, and J. O' G. Tatton. 1967. Pesticide residues in food stuffs in Great Britain. IV. Organochlorine pesticide residues in welfare foods. J. Sci. Food Agric. 18:579. Ruzo, L. 0., and M. J. Zabik. 1975. Polybrominated biphen- yls: Photolysis of hexabromo and hexachlorobiphenyl in methanol solution. Bull. Environ. Contam. Toxicol. 13:181. Safe, 8., O. Hutzinger, and D. Jones. 1975. The mechanism of chlorobiphenyl metabolism. J. Agric. Food Chem. 23:851. Savage, E. P., J. D. Tessari, J. W. Malberg, H. W. Wheeler, and J. R. Bagby. 1973. A search for polychlorinated biphenyls in human milk in rural Colorado. Bull. Environ. Contam. Toxicol. 9:222. Savage, E. P., J. D. Tessari, and J. W. Malberg. 1973. The occurrence of polychlorinated biphenyls (PCBs) in silage stored in pit and upright silos. Bull. Environ. Contam. Toxicol. 10:97. Schmidt, T. T., R. W. Risebrough, and F. Gress. 1971. Input of polychlorinated biphenyls into California coastal waters from urban sewage outfalls. Bull. Environ. Contam. Toxicol. 6:235. Skrentny, R. F., R. W. Hemken, and H. W. Dorough. 1971. Silo sealants as a source of polychlorobiphenyl (PCB) contamination of animal feed. Bull. Environ. Contam. Toxicol. 6:A09. 71 Snedecor, G. W., and W. G. Cochran. 1967. "Statistical Methods." 6th Ed. pp. 59-62. The Iowa State University Press, Ames, Iowa. Stanovick, R. P., S. I. Shahied, and E. Missaghi. 1973. Determination of polychlorinated biphenyl (Aroclor 12A2) migration into food types. Bull. Environ. Contam. Toxicol. 10:101. Stemp, A. R., and B. J. Liska. 1966. Effects of processing and storage of dairy products on telodrin and methoxy- chlor residues. J. Dairy Sci. A9:1006. Strik, J. J. T. N. A. 1973. Toxicity of hexachlorobenzene (HCB) and hexabromobiphenyl (HBB). Ghent Rijksuniver- zsteit Faculteit Landbouwwenten Schappen Mededelungem, 38(1):709. Sullivan, W. N., and I. Hornstein. 1953. Chlorinated poly- phenyls to improve lindane residues. J. Econ. Entomol. A6:158. Turner, J. C. and R. S. Green. 197A. The effect of a poly- chlorinated biphenyl (AroclorR 125A) on liver micro- somal enzymes in the male rat. Bull. Environ. Contam. Toxicol. 12:669. United States Department of H.E.W. 1953. "Milk Ordinance and Code." F.D.A., Washington, D. C. United States Department of H.E.W. 1969. Report of the Secretary's Commission on pesticides and their rela- tionship to environmental health. Part I and II. Van Patten, K., 1975. Michigan Department of Agriculture. Private Communication. Villeneuve, D. C., D. L. Grant, W. E. J. Phillips, M. L. Clark, and D. J. Clegg. 1971. Effects of PCB administration on microsomal enzyme activity in pregnant rabbits. Bull. Environ. Contam. Toxicol. 6(2):120. Yoshimura, H., and H. Yamamoto. 1975. A novel route of excretion of 2,A,3',A'-tetrachlorobiphenyl in rats. Bull. Environ. Contam. Toxicol. 13:681. Willett, L. B., and H. A. Irving. 1975. Distribution and clearance of polybrominated biphenyls by cows. J. Dairy Sci. 58(5):76A. influkfli I... wow .4. Eh I -. W. .H. 1|:iflhl'lfiumw "I7'111111111111131Tl“