CHARACTERESTICS OF FREE FAT 0F DRY WHOLE MILK T519”: for the Dwru of M. S. MICHIGAN STATE UNIVERSETY Karin L. Lindquist 1962. WW WWWWWW WWWWWWW WWWWWWW u 3 1293 00796 . STAG HUG 2*;3 f inwflbmw “4"“. I PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or beware date due. DATE DUE DATE DUE DATE DUE r r‘ ,_ if“: V 7? -’>w;\ V 17...; MSU Is An Affirmative Action/Equal Opportunity Institution cztclrchna-ox ABSTRACT CHARACTERISTICS OF FREE FAT OF DRY WHOLE MILK by Karin L. Lindquist This study was undertaken to determine the compositional charac- teristics of the free fat fraction of dry whole milk. Adsorption chromatography on silicic acid and gas-liquid chromatography were employed to ascertain the major lipid components and the glyceride fatty acid composition, respectively, of the total lipid and free fat. Lipid analyses by silicic acid chromatography indicated that free fat contained lower concentrations of phospholipid and free fatty acids than the total fat. Fatty acid studies by gas chromatography revealed that the free fat contained slightly higher proportions of C10'018 saturated acids and lower amounts of C18 unsaturated acids than the total lipid. These differences in fatty acid distribution were enhanced during storage. Data obtained from stored powders were inconclusive in regard to changes in the relative proportions of the major lipid components. The results of this study indicate that the free fat extracted from dry whole milk with a mixture of petroleum and ethyl ethers is not substantially different from the total lipid. CHARACTERISTICS OF FREE FAT 0F DRY WHOLE MILK BY Karin L. Lindquist A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Food Science 1962 ACKNOWLEDGMENT Earnest gratitude is extended to Dr. J. R. Brunner for his guidance throughout the course of this study. His patience, understanding and helpfulness are sincerely appreciated. Thanks are also offered to Dr. L. R. Dugan and Dr. C. M. Stine for their professional advice. Gratefulness is also extended to the graduate students and staff members for their friendship and encouragement. ii TABLE OF CONTE NT 3 INTRODUCTION . . . . . . . . . . . . REVIEW OF LITERATURE. . . . . . . . . Definition of "Free" Fat , . , . Occurrence of Free Fat . . . . . Significance of Free Fat . . . . Composition of Free Fat . . . . EXPERIMENTAL PROCEDURE . . . . . . . Plan of Experiment . . . . . . . Analytical methods . . . . . . . Total fat . . . . . . . Free fat . . . . . . . Silicic acid chromatography Gas-liquid chromatography Preparation of sample Column selection and operating Qualitative analysis Quantitative analysis Phosphorus . . . . . . . . EXPERIMENTAL RESULTS . . . . . . . . Free Fat Content , . . . . . . . Neutral Glyceride Fatty Acids. . Major Lipid Components . . . . . iii conditions. page 10 10 10 10 11 ll 11 12 l3 l4 l7 l7 17 18 page DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . 36 Analytical MGthOdS. 0 O O O O O O O O O O O O O 0 0 O O 36 Methyl ester preparation and gas-liquid chromatography . . . . . . . . . . . . . . . . . . 36 Silicic acid chromatography. . . . . . . . . . . . 38 Free Fat Content 0 O O O O O O O O O O O O I O O O O O 39 Neutral Glyceride Fatty Acids . . . . . . . . . . . . . 40 Major Lipid Components. . . . . . . . . . . . . . . . . 42 SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . . . . 46 LITERATURE CITED . . . . . . . . . . . . . . . . . . . . . 48 iv I. II. III. IV. VI. VII. VIII. IX. XI. XII. XIII. XIV. TABLES Manufacturing conditions for spray-dried whole milk POWderS. O O O O O O O O O O O O O O O O I O 0 Total lipid and free fat contents of fresh whole milkPOWderseeeeeeoooeoooecoceeo Glyceride fatty acid composition of total free fat extracted from powder No. l. , , Glyceride fatty acid composition of total free fat extracted from powder No. 2. . . Glyceride fatty acid composition of total free fat extracted from powder No. 3. . . Glyceride fatty acid composition of total free fat extracted from powder No. 4. . . Glyceride fatty acid composition of total free fat extracted from powder No. 6. . . Glyceride fatty acid composition of total free fat extracted from powders No. 5 and lipid lipid lipid lipid lipid lipid No. 7 and and Comparison of the total lipid and free fat average fatty acid compositions from fresh and stored milk pOWderS. O O O O O O O O O O O O O O O O O O O O 0 O 0 Carbon numbers and identification of fatty acid peaks from the gas chromatogram of the total lipid extracted from powder No. 3 after storage for one month, see Figure 3 O O O O O O O O O O I O I O O O O O O O O O O The major lipid components of total lipid and free fat samples as determined by silicic acid chromato- graphy I O O O O O O O O O O O O O O O O The phospholipid content of the total lipid and free fat Sanlples. O C O O O O O O O O O O O O O O O I O O 0 Conversion of gas chromatographic peak area to micro- moles, mole percentage and comparisioh to percentage 0 O O O O O O O O O O O O O O I area .page , l6 , 24 , 25 , 26 . 27 . 28 . 29 . 3O .44 Comparison of selected total lipid and free fat average fatty acid compositions of fresh and stored milk pOWderS O O O O O O O O O O O O O O C O . 45 FIGURES Page A plot of the adjusted retention volumes of authentic straight chain saturated fatty acids versus chain length for the identification of unknown fatty acids by the method of carbon number . . . . . . . . . . . . .15 Gas chromatogram of fatty acid methyl esters obtained from the total lipid extracted from powder No. 3 after one month of storage . . . . . . . . . . . . . . .20 Gas chromatogram of fatty acid methyl esters obtained from the free fat extracted from powder No. 3 after onemonthOEStorageeeeeeeoc000.00.00.21 Silicic acid chromatogram of the total lipid extracted from powder No. 7 when fresh . . . . . . . . . . . . . .22 Silicic acid chromatogram of the free fat extracted from powder No. 7 when fresh . . . . . . . . . . . . . .23 vi INTRODUCTION Reconstituted dry whole milk has never gained wholehearted acceptance as a beverage. Both flavor and appearance are not characteristic of normal fresh pasteurized milk. Because these defects are found to a much lesser extent in non-fat dry milk, the fat of whole milk powder has been blamed for these adverse characteristics. In particular, the guilt has been placed on that portion of the fat, usually referred to as "free” fat, extracted from whole milk powder with non-polar organic solvents. Although researchers have hypothesized that the deleterious effects of free fat are on the flavor, dispersion and wettability of dry whole milk, the chemical composition of free fat has not been elucidated. This study was undertaken to determine if the composition of the free fat from a given whole milk powder is representative of the total fat or is somehow unique. Because off-flavors due to fat oxidation are important in dry whole milk, changes in the total lipid and free fat characteristics during storage of the powder under adverse conditions were studied. REVIEW OF LITERATURE Definition of “Free" Fat In a normal extraction of the lipid of milk, alcohol and/or ammonia are used to degrade the fat globule membrane and to liberate protein- bound lipid thus enabling the extraction of the fat with common fat solvents. However, in dry whole milk, a portion of the total fat can be extracted with non-polar organic solvents alone. Holm and Greenbank (1925), believing that this fat fraction was not protected by a protein fihm, used the term “free" fat to describe the lipid which they obtained from dry whole milk by extraction with carbon tetrachloride. Following this work, all fat fractions extracted from whole milk powder with benzene (Favstova and Boika, 1958), petroleum ether (Litman, 1955), ethyl ether (Lampitt and Bushill, 1931), carbon tetrachloride (Greenbank and Hufnagel, 1953), carbon disulfide (Lampitt and Bushill, 1931), or hexane (Janzen, McGugan and Swanson, 1953) have been referred to as free fat or surface fat. The extraction technique has also varied. Shaking the powder and solvent (Thomas, Holgren, Jokay and Bloch, 1957), allowing the powder and solvent to stand (Lampitt and Bushill, 1931) and extracting the fat by the Soxhlet method (Nickerson, Coulter and Jenness, 1952) represent techniques. Occurence of Free Fat To comprehend the existing theories concerning free fat formation, some knowledge of the physical structure of spray-dried milk is desirable. The particles are small spheres containing interior air cells with protein and small fat globules evenly dispersed in lactose, which is amorphous in fresh spray-dried milk (Coulter, Jenness and Geddes, 1951; Lampitt and Bushill, 1931; King, 1948; Choi, Tatter and O'Malley, 1951J At the present time, the reasons for the occurence of free fat are not completely understood. It has been hypothesized that the solids- not-fat portion of the powder particle and/or the fat globule membrane serve to protect the majority of the fat of dry whole milk from sol- vation (Lampitt and Bushill, 1931). The contribution of lactose to the occurence of free fat was first proposed by Lampitt and Bushill (1931). They reported a liberation of free fat and an exhibition of optical activity when optically inactive fresh, spray-dried milk attained specific moisture levels by absorbing water from humid atmosphere. From this evidence they deduced that some of the lactose changed from an amorphous to a crystalline state at a "critical moisture content" of 8.6% to 9.2%. The significant increase in free fat at this point was attributed to lactose crystals which mechanically reptured the solids-not-fat mass rendering the fat accessible by solvents. In further support of this hypothesis, they observed that fresh, untreated roller-dried milk, which notably has much higher amounts of free fat than spray-dried milk, is optically active, indicating the presence of lactose crystals. In agreement with Lampitt and Bushill (1931), King (1948) concluded, from optical activity studies, that fresh spray-dried milk contained only amorphous lactose, while both fresh roller-dried and stored spray-dried 'milk contained crystalline lactose. Also, reconstituted portions of these three powders exhibited surface fat which he called l"demulsified" fat, comparable to the free fat of Lampitt and Bushill (1931). King (1948) 4 observed increasing amounts of surface fat in fresh spray-dried milk, stored spray-dried milk and fresh roller-dried milk in that order. To account for the quantitative differences in free fat among the three samples, he considered both the physical state of lactose and the fat globule membrane. He hypothesized that the characteristically low amount of free fat obtainable from fresh spray-dried milk was due to the impermeability of the amorphous lactose by the solvent, while in roller-dried milk, a mosaic of tiny lactose crystals formed capillary spaces admitting solvent to the interior fat. However, even in this powder he would not eXpect to extract fat if the fat membrane was undamaged. Probably the shearing effect of lactose crystals present in roller-dried milk and the increased denaturing effect of the higher temperatures used in roller drying collaborate to "demulsify" the fat thereby allowing its extraction with non-polar organic solvents. These two factors, although operative, occur to a lesser extent in spray- dried milk than in roller-dried milk thus accounting for the lower amounts of free fat. By combining microscopy with special staining techniques to dis- tinguish particle constituents, King (1955) was able to observe the mode of fat distribution in dry milk. In Spray process powder which was fresh or stored under favorable conditions, he reported that tiny fat globules were distributed uniformly throughout the particle and that there was no apparent fusing of fat globules or fat patches on the surface of the particles. But under unfavorable, high-moisture storage conditions some of the fat was liberated as indicated by the appearance of coalesced fat around the interior air cells and on the surface of the powder particle. He also observed that particles of Spray-dried milk which contained free fat exhibited birefringeance in polarized light due to the crystallization of lactose, whereas powder possessing well dispersed fat showed no bire- fringeance. These observations support the reputed importance of the physical state of lactose in the occurence of free fat. Choi, Tatter and O'Malley (1951) also studied the effects of moisture and 95% ethanol on the crystallization of lactose in dry whole milk by determining the quantity of cK-lactose hydrate. In accord with Lampitt and Bushill (1931), they reported extensive lactose crystallization in milk powders subjected to a humid atmosphere. For spray-dried whole milk, they observed little lactose crystallization until a 6.5% moisture content was attained in the powder, which was comparable to the 8.6% to 9.2% "critical moisture content" reported by Lampitt and Bushill (1931). Choi 2551. (1951) found very little OC-laCtose hydrate in whole powder treated with 95% ethanol. However, the alcohol treatment caused a hundred-fold increase in powder solubility and increased the free fat value from 8.0% to 25.5%. These investigators considered then that fat liberation by alcohol treatment resulted from protein coagulation rather than lactose crystallization. Contrary to the obser- vations of Lampitt and Bushill (1931) and King (1948, 1955), Choi g£.gl. (1951) found only amorphous lactose in some samples of roller-dried milk. From this evidence they suggested that the sizeable difference in the amount of free fat found in spray- and roller-dried milk was due to the extent of protein coagulation rather than the degree of lactose crystalli- zation. Significance of Free Fat Several researchers have called attention to the deleterious effects of free fat on the reconstitutability of dry whole milk. Janzen, McGugan and Swanson (1953) observed that fresh powder samples were easily wet and dispersed when the free fat was removed. However, after two months of storage at 700 and llOOF. the treated samples no longer displayed the improvement in wettability and dispersibility. Favstova and Boiko (1958) also reported the adverse effects of free fat relative to the keeping quality of the milk powder and the stability of the reconstituted product. Stone, Conley and McIntire (1954) stated that the percentage of total fat present in a milk powder was inversely proportional to the ease of self-dispersion. They did not cite free fat values, but Litman (1955) reported that the percentage of total fat extractable as free fat increased with the fat content of the powder. Litman (1955) reported the appearance of an insoluble "skin" on the surface of reconstituted whole milk and flaky "scum“ specks on the sides of the container, which they believed were complexed free fat and protein. In powder stored at 85°F. they observed a definite increase in scum formation concurrent with a decrease in free fat content. He postu- lated that free fat in the scum would be stable to non-polar solvent extraction. King (1960) applied his staining techniques to this skin material and observed: 1) clusters, which were conglomeration of casein micelles, individual fat globules and coalesced fat and 2) clumps, which were intermingled free fat and protein with imbedded air bubbles and fat globules. Tamsma, Edmondson and Vettel (1958) reported that free fat, account- ing for up to 40% of the total lipid, had no effect on the dispersibility of stored milk powders, but that higher free fat values, of 40% to 95% decreased the dispersibility of the powder. Reinke (1959) found no statistical correlation between free fat content and ease of self- dispersion in spray-dried milk containing 10% to 69% free fat. In addition to the problem of imperfect reconstitution, flavor defects resulting in part from fat oxidation are sometimes encountered in dry whole milk. Some researchers have indicated that the rate of flavor deterioration by fat oxidation increased with an increase in free fat content. Holm and Greenbank (1925) found that the time re- W'quired for a whole milk powder to absorb oxygen decreased as the free fat content increased. Shipstead and Tarassuk (1953) obtained free fat from milk powder by extraction with petroleum ether and found it to be oxidezed to the point of tallowy flavor and complete loss of color. The fat extracted from the interbr of the particle had good flavor and normal color. Greenbank and Hufnagel (1953) reported that the keeping quality of milk powders was inversely proportional to the quantity of carbon tetrachloride extractable free fat. Greenbank and Palansch (1961) studied the progress of fat oxidation in puff-dried whole milk stored at 40°F. Four fat fractions: 1) free glycerides, 2) free lipids (glycerides and phosphatides), 3) total glycerides and 4) total lipids were obtained from the milk powders and were evaluated for oxidation by the peroxide value, the thiobarbituric acid value and taste panel score. They observed that the free glycerides and free lipids began to oxidize two to four weeks before the total glycerides and lipids. Since quantitative data on these fractions were not re- ported, it is not known if the powders contained relatively high or low amounts of free fat. Reinke (1960) found no correlation between the quantity of free fat and the flavor scores of fresh or stored spray- dried milk containing 10% to 69% free fat. Two related theories have been presented to explain why free fat is more easily oxidized than the remaining fat. Washburn (1922) supported Storch‘s hypothesis that free fat was more easily oxidized since it was not surrounded by a gelatinous envelope which protected the fat from oxidation as long as it remained unbroken. Shipstead and Tarassuk (1953) suggested that in addition to the fat globule membrane all the solids-not-fat aid in the protection of the fat on the interior of the particle from oxidation. Composition of Free Fat The chemical composition of this fat fraction extractable from ' dry whole milk with non-polar organic solvents has not been studied. Litman (1955) reported that the fat portion of the skin material, which he believed was complexed free fat, found on the surface of reconstituted milk, had a lower iodine number and a higher melting point than the fat of the dispersed milk. King (1960) believed that the fat of this complex was a high melting fraction for it exhibited low flourescence, due to a high content of crystalline fat which does not take up dye. EXPERIMENTAL PROCEDURE Plan of Experiment Seven lots of spray-dried whole milk powder were manufactured under commercial processing conditions from mixed herd milk obtained from the Michigan State University Dairy. Raw milk was preheated in a tubular heater at approximately 185°F. for five minutes. The milk was then condensed to approximately 40% total solids in a Rogers single- effect evaporator and dried in a Rogers spray drier at conditions listed in Table 1. Manufacturing procedure varied slightly for pow- ders No. 1, No. 5 and No. 7 in that the milk was pasteurized and homogenized prior to condensing. The fresh milk powders were assayed for total lipid and free fat content and aliquots of each fraction were subjected to glyceride fatty acid analysis by gas-liquid chromatography. The fatty acid analyses were repeated during the storage of powder lots 1, 2, 3, 4 and 6. Total lipid and free fat samples extracted from fresh powders l, 2, 3 and 7 and the total lipid of the condensed milk from which powder No. 2 was made were analyzed quantitatively for the major lipid components by adsorption chromatography on a silicic acid column. This analysis was repeated on the total lipid and free fat fractions extracted from powders No. 2 and No. 3 following one month of storage. The phosphorous content of the total lipid and free fat obtained from all seven whole milk powders was also determined. The powders were stored at atmospheric conditions in either green sample bottles (lots 1, 2 and 3) or poly- ethylene bags (lots 4, 5, 6 and 7). 10 Analytical Methods The solvent extraction method of Mojonnier and Troy (1925) was used to determine total fat in the dry milk powders. The lipid fraction representative of the free fat was extracted with a 50-50 (V/V) mixture of petroleum and ethyl ethers as described by Thomas 2E.§l. (1957). A 20 g sample of milk powder was weighed into a glass-stoppered 250 ml Erlenmeyer flask and shaken for one minute with 40 ml of solvent. The powder particles were allowed to settle for one minute prior to filtering the solvent supernatant into a suction flask fitted with a sintered-glass funnel of medium porosity. Following a second extraction, a third 40 ml portion of solvent was shaken with the milk powder for one minute and the entire flask contents were emptied into the funnel. TheErlenmeyer flask was rinsed twice with 20 and 10 ml portions of solvent and the suction filtering was continued until all the solvent was pulled into the flask. The filtrate was transferred to a weighed aluminum dish. The solvent was removed and the weight of the free fat residue was determined. Free fat values were expressed as the percentage of the total lipid of the powder. Silicic acid chromatography The components of the total lipid and free fat fractions were resolved by adsorption column chromatography on silicic acid according to the method of Hirsch and Ahrens (1958). The column was charged with approximately 300 mg of lipid. Petroleum ether (b.p. 600-7000.) and ethyl ether were used for gradient elution of the neutral fat followed by methanol to remove the phospholipids. The column eluate was collected 11 in weighed test tubes in 10 m1 portions by a Rinco fraction collector. The solvent was evaporated in an oven at temperatures near the boiling point of the solvent phase. The tubes were reweighed under standard conditions in an air conditioned room maintained at 72° to 74°F. and 40% to 50% relative humidity. The weight of lipid residue was plotted on the ordinate against test tube number. The peaks of the resulting chromatograms were tentatively identified by reference to elution volumes of authentic lipids fractionated under the same conditions. When enough residue was available the identification was confirmed by infrared analysis with reference to spectra of standard lipids. The samples, carried in spectr0photometric grade carbon tetrachloride, were analyzed in the 2 to 14,; range with a Beckman IR-5 infrared spectro- photometer. Ga s- 1 iqu id chromatography Preparation of samples. Methyl esters of the neutral glyceride fatty acids were prepared by base-catalyzed interesterification of the glycerides with methanol as described by Smith and Jack (1954). Approximately one gram of fat in 4 ml of pentane was treated with 3.5 ml of absolute methanol and 0.005 ml of l N alcoholic KOH. The reaction mixture was allowed to stand at room temperature for two days after which the reaction was stopped by washing with three 3 m1 portions of dilute hydrochloric acid. The methyl esters were extracted with pentane most of which was removed from the esters prior to gas chromatography by passing a gentle stream of nitrogen over the solvent-ester solution. Column selection and operatingjconditions. Five and ten foot one- quarter inch (0.D.) capper gas chromatographic columns packed with 12 butanediol succinate, polyoxyalkalene adipate (Reoplex 400), diethylene glycol adipate cross-linked with pentaerythritol (LAC-446) or diethylene glycol succinate (DECS) were prepared to find the liquid phase most suitable for the resolution of C4 to C13 fatty acid methyl esters. Column performance was estimated by the 1) separation factor (stearate- oleate) and 2) the height equivalent per theoretical plate (H.E.T.P.) as suggested by Keulemans (1957). A ten foot column of 20% diethylene glycol succinate polyester (DEGS) was found to give the best resolution. The separation factor for the unresolved stearate-oleate pair was 1.15 and the H.E.T.P. was .10 inch. During the course of this study several DEGS columns were employed having approximately these same column para- meters. An Aerograph Model A-90-C Gas Chromatograph equipped with tungsten hot wire thermal conductivity detectors was employed in conjunction with a Leeds and Northrup Type C Speedomax Recorder having a 5 my scale span and chart speed of 2 min/inch. The following operating conditions were found to be optimum for the resolution of fatty acid methyl esters (C4-C18) on the DECS column: Oven temperature 200°C. Injector temperature 240°C. Helium flow rate 50 cc/min Detector current 265 ma Qualitative analysis. The method of carbon numbers developed by Woodford (1960) was adopted for qualitative identification of fatty acid esters. A mixture of authentic, even—numbered, saturated acids from four to eighteen carbons and the unsaturated eighteen-carbon—acid esters was chromatographed under standard conditions. The adjusted retattion volume, VR', where 13 VR' a vR-vm VR = the volume of gas required to elute the com- pound under study Vm = the volume of gas required to elute a non- absorbed gas, was determined for each ester and a semi-log plot of adjusted retention volumes of the saturated esters versus the length of the carbon chain was made. From the straight line, the carbon number, plotted on the abscissa, for any ester could be determined if its adjusted retention volume was known, see Figure 1. Saturated, straight-chain esters have carbon numbers identical to their chain length while unsaturated esters have non-integral carbon numbers, which are characteristic for a parti- cular ester in a particular stationary phase. Quantitative analysis. Approximately 0.3 g of each standard ester was weighed accurately into a 5 m1 volumetric flask; a known weight of ethyl ether was added to volume. The weight, thus the number of micro- moles, of each ester was therefore known for any aliquote of the mixture. One, two, three, five and ten microliter portions of the standard mixture were chromatographed and the areas of the resulting peaks were measured with a planimeter. From these measurements, the quantity-area relation- ship for each fatty acid ester was established by a geometric plot of the number of micromoles against the corresponding area. The identifiable peaks on chromatograms of samples of unknown fatty acid composition were also measured with a planimeter. Unresolved peaks were separated for measurement by drawing the shortest possible perpendicular line from the recorder tracing to the baseline of the chromatogram. The area percentage for each known fatty acid, based on the total identifiable area, was computed. The unassigned area was not included in this calculation. l4 Phosphorus Approximately 0.5 g samples of total lipid and free fat were accurately weighed into ceramic crucibles and were treated with saturated alcoholic magnesium nitrate and ashed overnight in a muffle furnace at 600°C. according to the method of Horrall (1935). The ash was taken up in dilute hydrochloric acid and analyzed for phosphorus by the col- orimetric procedure of Fiske and Subbarow (1925). The concentration of phosphorus, estimated from a standard curve, was converted to lecithin by the factor 25.94 (Horrall, 1935). 0.000 (cu?) ADJUSTED RETENTION VOLUME ._ § CARBON NUMBER Figure 1. A plot of the adjusted retention volumes of authentic straight chain saturated fatty acids versus chain length for the identification of unknown fatty acids by the method of carbon number. nsocxon k mmcooom omuoq haouwawxoudmm momma o>onm oEHu ou uson>Hovo "Hwy: m .mooom hfioumsfixopdmm um moafloudfimfi mp3 owsuwuodsou has mafiaooaa oEH l6 N .momm haoumawxouamm um mououomo moz nonrandom> a mmH om\oo comm me ma ooow moa HmHm .w:< n m9. . ONE... 88 02 2 2. 2. as: 0 m2 02$ 8mm 03 2 88 m2 N 53. m maa om\mo ooom k in an .nmz q oom om\mo ooom omH m in in .amh m omm mH\No ooom omH m in in .>oz N 2: 02$ 83 a: 2 ooom 02 amp: .uamm H SE :2: 2e Ea 3.3 . SE mass was ouwm .moum mama oEHu wawmooonoo aaou mafia dump .oz uoH uwxm ofiuuoz mean wcwuoonoum ou Howum .msamz umpBom chfiwwo hwumm muwuwofioum soauouwcowoaom cowuouwunoumwm Hmaowuwuooo wcwmmoooum muonsom xHHE oHOLB componhmuam you mGOHuHocoo wsfiusuommsonz H mqm:._. 29thme 3 av es Q. »m A” .w, QWW‘ m WimFW m w o 1H9|3H W}! mi .wmmuoum mo nunoa woo nmumm m .02 ummaon Eoum mouuwuuxm new menu use Scum nonfieuno muwumo HmnucE mwom muumm mo Enumoumaoueu mmu .m mummwm «2.5 ms: zotzmtmm no 3. 0' on an on on an o. u r o v o «a n. u.. s on o . e WM 9 mm o 1. n .zmmnm Gm£3 5 .oz nonzon 50um nwuumpuxo UHQHH Hmuou otu mo sapwoumfiouto mwom oaoflafim .d musmwm mmmZDz ugh oem oo~ 00. ON. om 0v 0 o N _ m .¢ . m .3 .m. .oN .cm «20.»:4u rzu.o<¢ov rill cu:»u 4>z»u-x:uao¢»ua '11" mm 0 Id I1 :30 .LH913M (s...) W.) a...” n .oz worsen Scum mmuomuuxm umw menu mtu mo EmDmOumEouno mace oHoHHHm 0mm Om. 04083:: 6 mum—232 monk ON_ Om .Emmum not: c8326 panaccov 3:: at: ”733.85! .m musmflm ov o . fidddrlirlflo .v ,m “M 3 Nu H In. .N. O 4.. H H .2 o ,m .ON on mm 24 TABLE II Total lipid and free fat contents of fresh whole milk powders Total Free Percentage of total fat Powder lot lipid1 fat extractable as free fat2 (7.) <7.) 1 25.06 0.73 2.9 2 28.30 1.68 6.0 3 25.30 1.50 5.9 4 28.06 2.15 7.7 7 25.60 0.08 0.3 1 All percentages are based on the total weight of powder 2 Free fat values 25 m.am ~.om w.om m.om asses mac smuQsSuamcs m.am s.km m.~o m.am wHo-oHo N.m H.0 0.0 ~.NH 00:00 mwfiom woumuSDmm 0.H m.0 0.0 0.0 oflcoaoawq m.¢ H.¢ 0.N H.¢ oHoHoaHA 0.~m ¢.Hm m.n~ ¢.mm oHoao 0.0H 0.~H 0.0a m.NH owumoum 0.mm «.mm ¢.mm ¢.mm oHuHEme o.HH m.0H n.0H 0.0 owumauzz m.m m.m 0.0 H.¢ caused 0.m q.m «.0 m.m ogummu ¢.H m.N m.H ¢.~ oHHmumwu m.H m.m N.~ m.s oaouaso 0.0 0.0 m.~ 0.0 owuhuom raisin-unnuunnraiuraN mouH mamom uaom moo awoum Amnucofiv moaned owmuoum 0 .oz Henson scum wouoouuxo umm menu was mfimqa Hmuou mo nowufimomsoo 000m muuwm ovfiuoomaw > mqm<8 28 0.00 0.00 0.00 0.00 0.00 0.00 00000 000 00000000000 0.00 0.00 0.00 0.00 0.00 0.00 000-000 0.0 0.0 0.0 0.0 0.0 0.0 00-00 00000 000000000 0.0 0.0 0.0 0.0 0.0 0.0 000000000 0.0 0.0 0.0 0.0 0.0 0.0 00000000 0.00 0.00 0.00 0.00 0.00 0.00 00000 m.ma n.~0 0.00 N000 0.0 0.00 0000000 0.00 0.00 0.00 0.00 0.00 0.00 00000000 0.00 0.00 0.00 0.00 0.00 0.00 00000002 0.0 0.0 0.0 0.0 0.0 0.0 000000 0.0 0.0 0.0 0.0 0.0 0.0 000000 0.0 0.0 0.0 0.0 0.0 0.0 00000000 0.0 0.0 0.0 0.0 0.0 0.0 000.0000 0.0 0.0 0.0 0.0 0.0 0.0 0000000 rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr N mmudwllrrrrrrrrrrrrrrrrrlrllIrrlrrrrrlrr 000m 00008 000m HmuoH 000m kuoH 0000 50000 0>Hm 0:0 Smoum Amfiucoav 000000 ammuoum 0 .02 000300 500m 0ouowuux0 00m 000% 0:0 00000 kuou mo 00000000500 0000 huumw 000000500 H> mam¢H 29 0.00 0.00 0.00 0.00 00000 000 00000000000 0.00 0.00 . 0.00 0.00 000-000 0.0 0.0 0.0 0.0 00-00 00000 000000000 0.0 0.0 0.0 0.0 00a00oa00 o.~ 0.N m.m 0.0 00000000 w.m~ o.w~ 0.0m «.mm 00000 0.00 0.N0 ©.m o.m 0000mum n.0m «.0m m.m~ 0.0N 00005000 «.00 o.M0 o.m0 0.00 00000002 0.0 0.0 0.0 0.0 000000 0.0 m.m 0.0 0.0 000000 ¢.N o.~ m.m m.m 00050000 0.0 0.0 w.¢ m.~ 0000000 0.0 0.0 0.0 o.N o0uhusm u- -------- - uuuuu ---N 000< ----- - -------------- -- 0000 00009 0000 0000B 0000 muu0m 0300 50000 a0£uc05v 000009 0w0uoum 0 .02 000300 5000 00000uux0 000 0000 0:0 00000 00000 00 00000000500 0000 huu0m 000umom0u HH> m0m¢H 3O mvfiom wHo vmuNMSummCD 0.0m «.mm m.o~ m.mm .10 93 «.2 93 20-2.0 in N4 o.~ 0:6. wo-qo wwwow kumusumm 0.0 0.0 0.0 0.0 oafimaoawn H.~ m.N m.H m.m ofimaocwa m.m~ ~.om «.mm m.o~ cacao ¢.¢H m.¢H N.~H o.NH afiummum o.~m m.m~ o.mm ¢.¢m owuwaamm WE h .3 H .2 H .2 033;: n.¢ m.¢ o.¢ H.¢ ofiuamg ¢.m n.m n.m m.m afiummu H.~ N.H o.m w.H oHHhuamo o.o o.o 0.0 w.o oHoumwo 0.0 o.o 0.0 0.0 owu%usm ------ ...... ------N mmu¢ uuuuuuu --- ------ -- mmum Hmuoa mmum kuoy vfiom huumm m .02 umc3om m .02 “musom n .02 van m .02 muowzom scum vmuomuuxm umm menu can UHQHH kuou mo coaufimomsoo wean huumm «@Humohfiu HHH> MAm¢H 31 TABLE IX Comparison of the total lipid and free fat average fatty acid compositions of fresh and stored whole milk powders Percentage difference Fatty acid group Total Free (free-total) -------------- Area %----------------- Fresh powder Saturated acids C4-C8 6.5 6.5 0.0 C10'C18 64.0 65.5 +1.5 Unsaturated €18 acids 29.5 28.0 -1.5 Stored 1 to 4 months Saturated acids C4-C8 6.4 5.9 -005 Clo-C18 63.8 66.6 +2.8 Unsaturated C18 acids 29.8 27.4 ~2.4 Stored 5 to 9 months Saturated acids C4-C8 4.9 4.9 0.0 Unsaturated 018 acids 29.8 28.4 -1.4 32 TABLE X Carbon numbers and identification of fatty acid peaks from the gas chromatograms of the total lipid extracted from powder No. 3 after storage for one month, see Figure 3 Peak number Carbon number Fatty acid1 1 4.00 4:02 2 4.86 5:05 3 6.00 6:02 4 6.77 7:05 5 7.59 7:15 6 8.00 8:02 7 8.70 9:05 8 10.00 10:02 9 10.87 10:13 10 12.00 12:02 11 12.12 12:13 12' 12.83 13:03 13 13.42 14:04 * 14 14.00 14:02 15 14.82 14:13 16 14.89 15:03 17 15.45 15:04 18 16.00 16:02 19 16.41 16:14 20 16.60 17:04 21 16.88 17:03 22 17.37 18:04 23 18.00 18:02 24 18.45 18.12 25 19.00 18:22 1 The first figure designates the number of carbons and the second the number of double bonds 2 Identified in this laboratory 3 Identified by Smith (1961) 4 Preposed by Smith (1961) 5 Proposed by this researcher * Branched 33 TABLE XI Major lipid components of total lipid and free fat samples as determined by silicic acid column chromatography Peak number and identification l 2 3 Powder Hydro- Cholesteryl Trigly- lot carbons esters cerides - ------------------ weight Z ------------------- Fresh powder No. 1 Total lipid 0.1 0.2 95.8 Free fat 0.5 0.0 96 7 No. 2 Total lipid of condensed milk 0.4 0.1 91.4 Free fat 0.1 0.0 99 0 No. 3 Total lipid 0.6 0.0 88.1 Free fat 1.0 0.0 84.2 No. 7 Total lipid 0.1 0.1 82.0 Free fat 0.1 0.9 93 9 Powder stored 1 month Two Total lipid 0.1 0.0 83.5 Free fat 0.1 1.1 91.1 Three Total lipid 0.4 0.2 80.3 Free fat 0.4 0.0 77 4 Averages of fresh powders Total lipid 0.3 0.1 88.6 Free fat 0.4 0.2 93 5 34 TABLE XI (CONT.) Monogly- Phospho- cerides cerides lipids Unassigned Digly- Free fat ty acids none 0.11. 92 00 60 00 56 92-. 00 0.0 70 00 28 20 01 00 8O 20 [+0 20 09 09 73 10 5/4. 10 82 21 [+9 52 06 70 38 10 20 30 87 23 60 3O 68 07 90 02 10 00 66 20 41 36 16 38 30 12 80 10 20 70 39 22 35 10 71 O 10 CI... 21 07 30 35 TABLE XII The phospholipid1 content of the total lipid and free fat samples Powder lot Total lipid Free fat (7.) (7.) l 0.78 0.06 2 0.70 0.22 3 1.20 0.22 4 0.55 0.14 5 0.60 0.21 6 0.68 0.06 7 1.06 0.12 1 Calculated as lecithin DISCUSSION Analytical Methods Methyl ester preparation and gas-liquid chromatography Base-catalyzed methanolysis was used in this study because the mild conditions of this reaction are least destructive to the unsaturated fatty acids of milk fat. Compared to acid-catalyzed methanolysis, the base catalyzed reaction is faster, more complete and less likely to alter fatty acid structures for it is capable of being used at lower tempera- tures (Markley, 1947). Rmthylation employing diazomethane was avoided because reactions may occur at ethylene bonds to form pyrazolines (Gehrke, Goerlitz, Johnson and Richardson, 1960). Because the fatty acids of only the neutral glycerides were of interest in thissiudy, simple methods to remove phospholipids and free fatty acids were attempted. According to Choudhury and Arnold (1961), phospholipids can be separated from neutral fat by shaking the lipid sample with silicic acid and chloroform. However, this technique was abandoned after gravimetric examination of the two fractions indicated that materials other than phospholipids were adsorbed by the silicic acid. Removal of free fatty acids and their conversion to methyl esters was attempted by the anion exchange technique described by Hornstein, Alford, Elliott and Crowe (1959). Using this method no methyl esters were re- covered from the resin as indicated by gas chromatography. Either the free fatty acids were not adsorbed onto the resin or the esterification was not successful. Theoretical considerations of the reaction described by Smith and Jack (1954) for methyl ester preparation indicated that it was unnecessary 36 37 to remove free fatty acids and phospholipids for only the neutral gly- cerides were active in the methylation. Under the basic conditions of this reaction no free fatty acids are methylated because carboxylic acids will convert to carboxylate anions which, because of their negative charge, are not subject to nucleophilic attack by methanol (Gould, 1960). Phospholipids will not undergo interesterification. under the conditions of this reaction for the small amounts of KOH (5 x 10"5 moles) will be neutralized by the phospholipids (1.4 x 10'4 moles, based on phospholipid comprising 1% of 1 g of total lipid). The KOH catalyst would not be available for the formation of the methoxy ion. cholesteryl esters are methylated with more difficulty than the other esters because of steric retardation by the bulky cholesteryl group in forming the reaction intermediate. These hypotheses were found to be valid by experimentation. Even if free fatty acids, phospholipids and cholesteryl esters were esterified to a slight extent, esters contributed by them could be considered to be very negligible. The inherent error in this technique occurs in the solvent removal step prior to gas chromatography. The more volatile fatty acids, especially methyl butyrate, are lost to some extent. However, even poorer recoveries of fatty acid esters are obtained when either heat or vacuum are employed to remove solvent. Ideally, fatty acid compositional data should be expressed as mole percentage rather than area percentage for the thermal conductivity detector responds on a weight rather than a molar basis. For compara- tive purposes both methods were used to indicate the fatty acid composi- tion of free fat obtained from powder No. 6 after four months of storage, see Table XIII. The results obtained by expressing the raw data as mole 38 percentage and area percentage agree especially well for the fatty acids having ten or more carbons. For the shorter fatty acids, the results obtained by the area percentage method are about 1.5% lower than those arrived at by the mole percentage method of expression. Because comparisons of the relative fatty acid compositions of the fat samples was an objective of this study, the expression of fatty acid data as area percentage was adequate. Although it is erroneous to exclude the minor fatty acids in the calculation of either area percentage or mole percentage, only the chromatographic peak areas of the eleven major fatty acids were used to compute area percentages. The unassigned areas of the chromatograms, which included the minor fatty acids and a portion of the compounds resulting from oxidative deterioration, varied considerably among the samples. The inclusion of the unassigned area would confuse the comparison of individualiatty acid concentrations among samples. Silicic acid chromatography Accurate quantitative analysis of a complex lipid mixture by silicic acid chromatography is complicated by at least three factors. One dis- advantage is incomplete resolution of components having similar elution characteristics such as l) triglycerides and free fatty acids and 2) cholesterol and diglycerides. Also, peak identification by reference to elution volumes of authentic lipids is dependent on exact reproduction of the conditions of the analysis. The greatest limitation in achieving reproducible results is column preparation. Because identification using elution volumes cannot be trusted completely, confirmation by infrared analysis should be employed. However, these analyses are often either 39 not possible or inconclusive because of insufficient amounts of sample. Free Fat Content The free fat values of the fresh milk powders which ranged from 0.3% to 7.7% agree well with those reported in the literature for spray- dried milk containing approximately 26% total fat. Litman (1955) reported initial free fat values of 1% to 18% and Reinke (1959) ob- served free fat accounting for 14.4% to 20% of the total fat in milk powders manufactured under conditions similar to those in this study. The high percentages, 70% to 80%, of the total fat extractable as free fat from powders stored five or more months is surprising and unprece- dented. Reinke (1959) observed only small increases in free fat during storage in the order of two to three percent. Litman (1955) reported no changes in free fat content in powders stored at 45°F., and at higher temperatures, 850 and llOOF., a decrease in free fat value was observed. The general increase in free fat values during storage observed in this study may be attributed to the fat freeing action of lactose crystals which form, much as described by Lampitt and Bushill (1931), when milk powders absorb moisture. Although the extent of lactose crystallization was not determined in any of the powder samples, the stored powders developed a dry, hard, powdery texture and browned considerably. Similar observations were made by Lampitt and Bushill (1931) during water ab- sorption. The milk powders under study in this research did appear to have absorbed water for the total lipid percentage, based on the total weight of the powder, decreased during storage from 26% to 23% or 24% indicating an increase in moisture. In addition to physical destruction of the fat globule membrane by lactose crystals, protein degradation and 40 oxidative deterioration was probably a more important contributor to the freeing of the fat. The atmospheric storage conditions including exposure to light are unfavorable and are mainly responsible for the destruction of the fat globule membrane. Neutral Glyceride Fatty Acids By averaging the results of all the fatty acid analyses by gas chromatography, the free fat was noted slightly richer in ClO'C18 saturated acids and poorer in C18 unsaturated acids and C4-C8 saturated acids than the total lipid. These differences were enhanced by selecting and averaging only the data which followed this trend, see Table XIV. In these cases the free fat of fresh milk powder contained an average of 4.4% more ClO'ClB saturated acids, 2.5% fewer €18 unsaturated acids and 2.5% fewer C4'CB saturated fatty acids than the total lipid. The minor distinction found in the C4‘C8 group is meaningless because of the variable losses of volatile fatty acids, but the differences in concentrations of the longer fatty acids is probably significant. These findings are in agreement with those of Litman (1955) who con- cluded that complexed free fat found in insoluble particles of recon- stituted dry whole milk has a lower iodine number and a higher melting point than the fat of the dispersed milk. Bullock and Winder (1960), without direct experimental evidence, have postulated that pressures exerted on the fat globules during drying cause a redistribution of glycerides resulting in a thin film of high-melting glycerides on the surface of the spray-dried milk particle. If the free fat extraction technique removes mostly fat on the surface of the powder particle, this hypothesis of Bullock and Winder (1960) partially accounts for the 41 slightly higher proportion of high melting glycerides found in the free fat. Averages of the fatty acid data from all milk powders, Table XII, indicate that storage has no effect on the fatty acid composition of total lipid or free fat. However, the averaged selected data discussed above and listed in Table XIV indicates that the proportion of C10'Cl8 saturated acids increased after storage of five or more months at the expense of C18 unsaturated acids in both total lipid and free fat fractions. Oxidative deterioration is probably responsible for the loss of unsaturated acids. The averaged fatty acid compositional results listed in Table XIV indicates that the free fat is more repre- sentative of the total lipid in powder stored five or more months than was initially the case. This observation can be explained by the finding that at this age the free fat accounts for 80% to 90% of the total fat. The butyric acid concentrations of the fat fractions studied were much lower than values found by other researchers. Jenness and Patton (1959) reported that butyric acid has been found in proportions of 8.5 to 10.5 mole percentage in butterfat. The very low amounts or absence of butyric acid found in this study must be due to evaporation losses prior to gas chromatography. Linolenic acid was noted in four of seventeen total fat samples and five of sixteen free fat samples. The highest concentration of linolenic acid was found in powders manu- factured in September and January. No linolenic acid was observed in powders manufactured in April, May or August. Possibly linolenic acid, although present, is not detected in all cases. This acid occurs in low concentrations in milk and because of its large retention volume it diffuses considerably during passage through the column. 42 Major Lipid Components Some of the differences between total lipid and free fat elucidated by silicic acid chromatography can be attributed to the non-polar nature of the solvent used to extract free fat. The observation that lesser concentrations of free fatty acids and phospholipids are found in the free fat can be explained on this basis. Gurd (1960) reported binding of fatty acid anions to human serum albumin. A similar fatty acid association with milk albumin would certainly be stable to extraction with a 50-50 (V/V) mixture of petroleum and ethyl ethers. Even if the fatty acid anions were not associated with proteins they would be only slightly soluble in this solvent. The difference in free fatty acid concentrations between total lipid and free fat, which averaged 3.9% and 1.5% respectively, would be more outstanding if 75% of the fatty acids were not discarded as ammonium salts in the Mojonnier total fat extraction (Starr and Herrington, 1941). Assuming that the presence of free fat indicates fat globule meme brane damage for one thing, phospholipids, which are a component of the membrane, might be expected to occur at the same, if not higher, concen- trations in free fat than in total fat. The finding, by both chemical and physical assay, that phospholipids occur at lower levels in free fat than total fat can also be explained by protein-lipid association. Even though the membrane might be disrupted, the majority of the phospholipids would still be complexed to membrane protein. A more polar solvent than petroleum or ethyl ether is required to break the protein-phospholipid complex. Also free uncomplexed phospholipids, due to their polar nature, may have limited solubility in the petroleum-ethyl ether mixture. These reasons help to explain the observation that free fat contains less than 43 half as much phospholipid as total fat. Cholesterol, which Jenness and Patton (1959) report occurring in concentrations of 0.25 to 0.4% in butterfat, was not identified in any total lipid or free fat fractions. Cholesterol and diglycerides have similar elution characteristics on silicic acid (Hirsch and Ahrens, 1958) and may emerge from the column together as Peak 5, see Figure 4. The decrease in triglyceride concentrations in both total lipid and free fat at the expense of an increase in diglyceride during powder storage indicates that either lipolysis is occurring or that Peak 5, tentatively identified as diglyceride, is not diglyceride at all. This fraction may be triglyceride which has been chemically altered by oxida- tion. These changes would effect the elution characteristics of the lipid causing it to emerge from the column after the unoxidized triglyceride. The increase in the phospholipid concentration in free fat during storage indicated extensive damage to the membrane protein resulting in a higher proportion of uncomplexed free phospholipids. 44 TABLE XIII Conversion of gas chromatographic peak area to micromoles, mole percentage and comparison to area percentage Micro- Fatty acid Area moles Mole % Area % (cu-2) Caproic 2.3 0.34 6.4 4.7 Caprylic 1.2 0.16 3.1 2.4 Capric 2.4 0.32 6.1 4.9 Lauric 2.3 0.24 4.7 4.7 Myristic 6.0 0.63 . 11.9 12.2 Palmitic 15.4 1.63 30.8 31.3 Stearic 5.6 0.59 11.1 11.4 Oleic 12.7 1.26 23.8 25.8 Linoleic 1.3 0.14 2.6 2.6 1 Gas chromatographic peak area data of free fat obtained from powder No. 6 after four months of storage. 45 TABLE XIV Comparison of selected total lipid and free fat average fatty acid compositions of fresh and stored milk powders Percentage Fatty acid No. of samp- Total1 Free1 difference group les selected lipid fat (free-total) ----- Area %----- Fresh powder C4-C8 saturated 3 of 6 6.7 4.2 -2.5 Clo-C18saturated 5 of 6 63.3 67.7 +4.4 C18 unsaturated 5 of 6 29.9 28.0 -2.5 Stored 1 to 4 months C4-08 saturated 3 of 5 7.9 5.0 -3.2 Clo-C18saturated 5 of 6 61.2 66.9 +5.7 C18 unsaturated 5 of 6 30.8 28.1 -2.7 Stored 5 to 9 months C4-C8 saturated 4 of 5 5.8 3.8 -2.0 ClO-Clgsaturated 4 of 5 65.3 69.3 +4.0 C18 unsaturated 4 of 5 28.8 26.8 -2.0 1 Corrected to total 100 SUMMARY AND CONCLUSIONS The free fat fraction of whole milk powder, extractable with non- polar organic solvents, has been reported to have contributed adversely to the reconstitution and flavor of that product. The purpose of this study was to determine the compositional characteristics of free fat obtained from both fresh and stored whole milk powder. The total lipid and free fat of seven spray-dried whole milk powders manufactured under commercial conditions were analyzed. Adsorption chromatography on silicic acid columns and gas-liquid chromatography were employed to ascertain semi-quantitatively the major lipid components and the glyceride fatty acid composition, respectively, of the total lipid and free fat fractions. Lipid fractionation by adsorption chromatography indicated that free fat contained lower concentrations of phospholipid and free fatty acids than the total fat. Data obtained from stored powders were incon- clusive in regard to changes in the lipid during storage. No major differences in the distribution of individual fatty acids between total lipid and free fat fractions were noted by gas chromatographic fatty acid analyses. However, by grouping the fatty acids according to chain length, subtle dissimilarities were observed. Free fat was found to be slightly richer in Clo-018 saturated acids and poorer in C18 unp saturated acids than the total lipid. These differences were enhanced during storage. The results of this study lead to the conclusion that free fat ex- tracted from dry whole milk with a 50-50 (V/V) mixture of petroleum and ethyl ethers is not substantially different from the total lipid. Minor 46 47 dissimilarities were as follows: 1. Free fat contained lower concentrations of phospholipid and free fatty acids than the total lipid. 2. Free fat contained slightly lower concentrations of C18 unsaturated acids and higher concentrations of Clo-C18 saturated acids than the total lipid. LITERATURE CITED (1) Choudhury, B.R., and Arnold, L.K. 1960. The determination of the neutral oil content of crude vegetable oils. J. Amer. Oil Chem. Soc., 31:86. (2) Choi, R.P., Tatter, C.W. and O'Malley, C.M. 1951. Lactose crystallization in dry products of milk. 11. The effect of moisture and alcohol. J. Dairy Sci., 34: 850. (3) Coulter, S.T., Jenness, R. and Geddes, W.F. 1951. Advances in Food Research. pp 45-118. New York: Academic Press Inc. (4) Favstova, V. and Boiko, N. 1958. Changes in butterfat during the concentration and drying of milk. Mol. Prom., 12; 31. (Abst.) Dairy Sci. Absts., 29: 526. (Original not seen). (5) Fiske, C.H. and Subbarow, Y. 1925. The colorimetric determination of phosphorus. J. Biol. Chem.,.§§: 375. (6) Gehrke, C.W., Goerlitz, D.F., Johnson, H.D. and Richardson, C.W. 1960. Quantitative determination of the methyl esters of fatty acids by gas chromatography. Abst. material presented at Amer. Dairy Assos. 56th Ann. Meet., Logan, Utah (Mimeo). (7) Gould, E.S. 1960. Mechanism and Structure l2 Organic Chemistry. p. 315. New York: Henry Holt and Company. (8) Greenbank, G.R. and Hufnagel, C.F. 1953. The effect of the fat content of milk powder on the '” keeping quality of the dried powder. J. Dairy Sci., 36: 566. (9) Greenbank, G.P. and Pallansch, MgJ. 1961. The progress of oxidation in the milk powder granule. (Abst.) J. Dairy Sci. 44: 1165. (10) Curd, F.R.N. *‘ 1960. in Hananan, D.J. Ed., Lipid Chemistry. pp. 248. New York: John Wiley and Sons, Inc. (ll) Hirsch, J. and Ahrens, E.H. 1958. The separation of complex lipid mixtures by the use of silicic aCid chromatography. J. Biol. Chem.,_223: 311. 48 (12) (1'3) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) 49 Holm, G.E., Greenbank, G.R. and Deysher, E.F. 1925. The effect of homogenization, condensation and varia- tions in the fat content of a milk upon the keeping quality of its milk powder. J. Dairy Sci., 2: 512. Hornstein, 1., Alford, J.A., Elliott, L.E., and Crowe, P.F. 1960 Determination of free fatty acids in fat. Anal. Chem. 32: 540. Horrall, B.E. 1935. A study of the lecithin content of milk and its products. Purdue Univ. Agr. Expt. Sta. Bul. 401. Janzen, J.J., McGugan, W.A. and Swanson, A.M. 1953. Some factors involved in the wettability and dis- persibility of dried whole milk. J. Dairy Sci., ‘26: 566. Jenness, R. and Patton, S. 1959. Principles gf Dairy Chemistry. pp. 30-71. New York: John Wiley and Sons, Inc. Keulemans, A.I.M. 1957. Q Gas Chromatography. p. 113. New York: Reinhold Pub- lishing Corp. King, N. 1948. The microscopic examination of milk powders. Netherlands Milk and Dairy J., 2; 137 King, N. 1954. Physical structure of milk powder. Dairy Ind., 12: 39. King, N. 1955. Flourescence microscopy of fat in milk and milk powder. J. Dairy Research, 22: 205. King, N. 1961. Microstructure of some non-dispersible particles in milk powder. Aust. J. Dairy Tech.,.l§: 77. Lampitt, L.H. and Bushill, J.H. 1931. The physico-chemical constitution of spray-dried milk powder. Pat in spray-dried milk powder. J. Soc. Chem. Ind.,.29: 45. Litman, 1.1. 1955. A study of a fat-protein complex in powdered milk. Ph.D thesis. State College of Washington, Pullman. Markley, K.S. 1947. Fatty Acids. p. 299. New York: Interscience Publishers, Inc. 50 (25) Mojonnier, T. and Troy, H.C. 1925 Technical Control 2: Dairy Products. 2nd Ed. p. 936. Chicago: Mojonnier Bros. Co. (26) Nickerson, T.A., Coulter, S.T. and Jenness, R. 1952. Some properties of freeze-dried milk. J. Dairy Sci., 35: 77. (27) Reinke, E. 1959. The effect of processing variations on the free-fat self-dispersion, and flavor of whole milk powder. M.S. thesis. Michigan State University, East Lansing. (28) Shipstead, H. and Tarassuk, N.P. 1953. Chemical changes in dehydrated milk during storage. Agr. Food Chem.,.1: 613. (29) Smith, L.M. and Jack, E.L. 1954. The unsaturated fatty acids of milk fat. 1. Methyl ester fractionation and isolation of monoethenoid constituents. J. Dairy Sci.,lzl: 380. (30) Smith, L.M. 1961. Quantitative fatty acid analysis by gas-liquid chromatography. J. Dairy Sci., 44: 607. (31) Starr, M.P., and Herrington, B.L. 1941. The determination of fat in the presence of free fatty acids. 1. The Mojonnier test of mixtures of free fatty acids and butterfat. J. Dairy Sci., 24; 165. (32) Stone, W.K., Conley, T.F., and McIntire, J.H. 1954. The influence of lipids on self-dispersion and on ease of dispersion of milk powder. Food Tech.,.§: 367. (33) Tamsma, A., Edmondson, L.F. and Vettel, H.E. 1958. Free fat in foamrdried whole milk. J. Dairy Sci., 41: 710. (34) Thomas, R.J., Holgren, C.J., Jr., Jokay, L. and Bloch, I. 1957. A rapid method for determining free fat in dry milks. (Abst.) J. Dairy Sci., 42: 605. (3§)'Thompson, M.P. ' 1960. Structure of the milk fat globule membrane. Ph.D. thesis. _ ' Michigan State University, E. Lansing. (36) Washburn, R.M. 1922. The physical analysis of dry milk. J. Dairy Sci., 5: 388. (37) Woodford, F.P. and van Gent, C.M. 1960. Gas-liquid chromatography of fatty acid methyl esters, the carbon number as a parameter for comparisons of columns. J. of Lipid Research, 1: 188. N-"‘ .- in» $753,215! mi 1:1“- .- ca :31 05:1- mud "I71111111111111'1'“