J Sci Food Agric 1998, 76, 546È552 Eþect of Naked Oats in the Dairy Cow’s Diet on the Oxidative Stability of the Milk Fat Anna M Fearon,1* C Sinclair Mayne2 and Colin T Charlton3 1 Department of Food Science, The Queen’s University of Belfast, and the Food Science Division, Department of Agriculture for Northern Ireland, Newforge Lane, Belfast, BT9 5PX, UK 2 Agricultural Research Institute for Northern Ireland, Hillsborough, Co. Down. BT26 6DR, UK 3 Food Science Division, Department of Agriculture for Northern Ireland, Newforge Lane, Belfast, BT9 5PX, UK (Received 15 January 1997 ; revised version received 26 June 1997 ; accepted 7 August 1997) Abstract : Oxidative stability of milk fats from cows o†ered naked oats- and barley-based diets were compared in shelf-life tests using the Schaal Oven Test at 63¡C, and determination of peroxide and thiobarbituric acid values. Milk fat from cows o†ered the diet containing naked oats, although containing a higher proportion of monounsaturated fatty acids, had a signiÐcantly longer oxidation induction period (13 days) than milk fat from cows o†ered the control (barleybased) diet (9 days). However, sensory testing of the milk fats indicated that a perceptible odour di†erence was apparent between stored (63¡C) and unstored milk fats after 3 days of storage irrespective of the diet fed. It is possible that the odours perceived may have originated from the cowsÏ diets or from hydrolytic rancidity reactions. The difficulties of relating chemical and sensory testing of the oxidative deterioration of milk fat or butter are underlined. ( 1998 SCI. J Sci Food Agric 76, 546È552 (1998) Key words : milk fat ; dietary modiÐcation ; oxidative stability INTRODUCTION acid composition of the dietary material. The presence of oleic acid (C ) in the milk fat results primarily 18 > 1 from the conversion of C to C by the bovine 18 > 0 18 > 1 desaturase enzyme system located in the intestine and mammary gland (Kinsella 1972). In order to produce a milk fat with a higher proportion of unsaturated fatty acids, it is necessary either to increase the content of unsaturated fatty acids entering the blood stream from the rumen and supplying the mammary gland, or, to optimise the desaturating ability of the intestine and mammary gland (Banks and Christie 1990). The Ðrst approach was adopted by Australian scientists in the early 1970s, with the feeding of unsaturated lipid encapsulated within a formaldehydeÈprotein matrix (Scott et al 1970). The technique not only protected the dietary lipid from hydrogenation in the rumen but avoided interference with normal rumen activity. Early problems with the highly unsaturated nature of the resulting milk fat ([20% linoleic acid, There has been considerable interest in dietary modiÐcation of the fatty acid composition of cowsÏ milk fat to increase the content of unsaturated acids present and hence improve the physical properties of the fat. The short- and medium-chain fatty acids (C to C and a 4 14 proportion of C ) in cowsÏ milk fat are synthesised de 16 novo in the mammary gland. The longer chain acids (some C and the C acids) are derived directly from 16 18 the diet and it is their proportions in the milk fat that may be altered by modifying the lipid portion of the cowsÏ diet (Hawke and Taylor 1983). Dietary lipid normally undergoes enzymatic hydrolysis and biohydrogenation in the cowsÏ rumen, with the consequence that stearic acid (C ) is the principal 18 > 0 fatty acid leaving the rumen, irrespective of the fatty * To whom correspondence should be addressed. 546 ( 1998 SCI. J Sci Food Agric 0022È5142/98/$17.50. Printed in Great Britain Oxidative stability of dietary modiÐed milk fat C ) led to rapid oxidative deterioration, but these 18 > 2 have largely been resolved with improved technology, the inclusion of substantial quantities of a-tocopherol in the diet and the use of rapeseed (canola) as the lipid source (Goering et al 1976 ; Ashes et al 1992). The alternative approach has been to o†er dairy cows high lipid-containing cereals or oilseeds in order to provide a rich supply of C fatty acids to the 18 mammary gland for subsequent desaturation and incorporation into milk fat as C . In this study, 18 > 1 naked oats were used as the high-lipid source, resulting in a milk fat having an increased proportion of C 18 > 0 and C fatty acids and producing a softer, more 18 > 1 spreadable butter (Fearon et al 1996). This present work aimed to determine whether the increased content of monounsaturated fatty acids adversely a†ected the oxidative stability of the milk fat. EXPERIMENTAL Sample preparation Twenty Friesian dairy cows in early lactation were divided into two groups and o†ered either naked oatsor barley-based (control) concentrates and silage ad libitum in a continuous design experiment of 10 weeks duration. Full details of the experimental design and diet formulation have been described by Fearon et al (1996). Milk was collected from each group weekly ; 547 separated into cream and churned into butter. Fatty acid composition of the milk fats, as determined by methylation and subsequent analysis by gasÈliquid chromatography (Fearon et al 1996), is given in Table 1. Milk fat compositions were found not to di†er signiÐcantly between weeks during the 10 weeks of the feeding trial period (Fearon et al 1996) ; therefore, for the purposes of this study the weekly butter samples were identiÐed as replicates. Storage stability tests Portions of butter (approximately 250 g), representative of both diets, were randomly selected from the butters produced over the trial period to give three replicates per diet. These were placed in 1000 cm3 opaque white plastic food-grade containers (Polar-cup, Portadown, Northern Ireland) in an incubator at 20 ^ 1¡C. The Schaal Oven Test was conducted using clariÐed butter oils prepared from similar butter sources (BS 769 : 1961). Aliquots of the butter oil (800 cm3) were placed in 1000 cm3 amber glass beakers in a laboratory oven at 63 ^ 1¡C (Rossell 1983). Peroxide value measurements Peroxide value (PV) measurements were carried out in triplicate at 3-day intervals on the stored butter and butter oil samples. These were determined by the AOCS TABLE 1 Fatty acid compositiona (g kg~1) of milk fats from dairy cows o†ered barley-based (control) and naked oats-based concentrates and a-tocopherol content of concentrates (Values are means of Ðve observations per week over 10 weeks) Fatty acid Barley-based concentrates Naked oats-based concentrates SEM SigniÐcance of treatment e†ect : Concentrates type 4:0 6:0 8:0 10 : 0 12 : 0 14 : 0 14 : 1 16 : 0 16 : 1 18 : 0 18 : 1 18 : 2 a-Tocopherol content of concentratesb (kg g~1) 29É4 17É1 12É9 29É1 33É0 114É8 13É7 329É8 19É2 110É5 224É6 13É0 33É9 16É9 11É7 24É3 26É2 98É1 11É4 279É3 17É0 139É3 272É9 13É3 0É53 0É14 0É15 0É58 0É87 1É85 0É73 5É09 0É12 2É08 4É01 0É17 * NS ** ** ** ** NS ** ** *** *** NS 15É6 15É6 a Fearon et al (1996). b Base level 7 mg kg~1 (BlanchÑower, personal communication). *** P \ 0É001, ** P \ 0É01, * P \ 0É05, NS, not signiÐcant. 548 Official Method Cd 8-53 (AOCS 1990) and results expressed as milliequivalents peroxide per kg sample. 2-Thiobarbituric acid value measurements Thiobarbituric acid value (TBA) measurements were carried out in triplicate at 3-day intervals on the stored butter and butter oil samples. TBA values were determined by the method of Tarladgis et al (1960) using the modiÐed distillation apparatus described by Hoyland and Taylor (1989), with 50 cm3 of distillate being collected within 10 min. Absorbance readings at 538 nm were converted to mg malonaldehyde per kg sample by multiplying by 7É8 (Pearson 1976). A M Fearon, C S Mayne, C T Charlton Duo-trio test The duo-trio test (BS 5929 : Part 2, 1992) was employed to assess if an odour di†erence could be perceived between unstored standard butter oils and butter oils stored for di†erent periods at 63¡C. Thirty-two panellists were presented with an identiÐed reference sample followed by two coded samples, one of which was identical to the reference sample and the other di†erent. The panellists were asked to smell all three samples and identify which of the coded samples di†ered from the reference sample. Each comparison was balanced so that 16 panellists received the standard oil (unstored) as the reference sample and 16 panellists received a treated oil (stored at 63¡C) as the reference sample. Assessments with butter oils from each diet were carried out separately 7 days apart. a-Tocopherol content Statistical analysis The barley- and naked oats-based concentrates were analysed for a-tocopherol content using the extraction, saponiÐcation and chromatographic methods described by McMurray and BlanchÑower (1979). The highperformance liquid chromatography system (HPLC) was supplied by Merck Ltd (Poole, Dorset) and comprised a Hitachi L6000 pump, F1050 Ñuorescence detector, AS2000 autosampler and Hewlett Packard 3392 integrator. A Lichrosorb RP18 (5l) 250 ] 4 mm column (Merck Ltd, Poole, Dorset, UK) was used with 98% methanol as the mobile phase and a Ñow rate of 1É8 cm3 min~1. BS 5929 : Part 2 (1992) was applied to the sensory results to determine signiÐcance of di†erence between treatments and standards. Analysis of variance was carried out on peroxide value and TBA value results. Gompertz curves (Bliss 1970) were Ðtted to the PV and TBA results plotted against time (days of storage). The Gompertz curve is an asymmetrical sigmoidal curve which has the form y \ a ] c* exp([exp(exp([b(t [ m)))) (1) where y is the PV or TBA value at time t, and a, c, b and m are parameters to be estimated. Sensory analysis RESULTS AND DISCUSSION Samples A large volume (4000 cm3) of clariÐed butter oil from each dietary treatment was prepared (BS 769 : 1961), sub-divided into four portions per diet in amber 2000 cm3 glass beakers and held at 63 ^ 1¡C in laboratory ovens for 3, 8, 13 and 21 days to give a range of peroxide values. The samples were dispensed as 10 cm3 aliquots into capped amber glass vials and presented at a temperature [40¡C to the panellists for assessment. Unstored clariÐed butter oils from both diets were heated to 63¡C as standards for assessment. Sensory unit The sensory evaluation of the butter oils was undertaken in a sensory evaluation suite with each panellist in an individual booth with positive air Ñow vented to the outside of the building. Samples were presented in amber glass vials and evaluated under red lighting adjusted so that appearance di†erences between treatments were masked. The questions were delivered and responses analysed on a computerised system using PSA software, version 1.636 (Oliemans, Punter and Partners BV, Burgemeester, Utrecht, Netherlands). The high-lipid dietary material used in this study was naked oats, a cereal commonly containing twice as much lipid as other cereals grown in the UK, with a lipid content up to 80È100 g kg~1 dry matter (Valentine 1987) and approximately half of the fatty acids present are unsaturated (Morrison 1977). Other dietary materials used as rich sources of C fatty acids 18 have been soya beans and whole and ground rapeseed (Banks et al 1980 ; Murphy et al 1990). These high-lipid cereal sources not only provide the necessary substrate (C ) for the bovine desaturase enzyme but retention 18> 0 of the seed coat serves to minimise any adverse e†ect of the dietary oil on the activity of rumen microorganism (Banks and Christie 1990). The milk fats from cows o†ered high-lipid cereal sources have contained decreased contents of C ,C and C fatty acids 12> 0 14> 0 16> 0 and increased contents of C and C fatty acids, 18> 0 18>1 resulting in softer milk fats with lower solid fat contents at low temperatures (Murphy et al 1990 ; Fearon et al 1996). It is recognised that fats or oils containing a higher proportion of unsaturated fatty acids have a shorter shelf-life due to the more rapid oxidation of Oxidative stability of dietary modiÐed milk fat unsaturated fatty acids (Rossell 1989). Sonntag (1979) described the relative oxidation rates of linolenic (C ) : linoleic (C ) : oleic (C ) : stearic (C ) 18> 3 18> 2 18>1 18> 0 fatty acids as 150 : 100 : 10 : 1. It was hoped in the present work that improvement of the functional characteristics of milk fat, by increasing the proportion of monounsaturated fatty acids, would be achieved without a serious deterioration in oxidative stability of the fat. The resistance of the butters and butter oil samples to oxidative deterioration was measured by shelf-life tests at ambient and high temperatures. Stability testing at room temperature has the disadvantage of being very slow ; in this study no peroxides were detected and only low TBA values, between 0É01 and 0É2 mg malonaldehyde per kg, were determined in the butters stored at 20¡C during the storage period (Table 2). These results were in agreement with those of Farag et al (1990) who reported little change in PV and TBA values of butters stored at room temperature (unspeciÐed) for 18 days. The butter samples held at 20¡C had not been clariÐed and, as they contained O16% moisture, it must be acknowledged that deterioration, other than oxidation may have taken place. For example, hydrolytic rancidity with the formation of free fatty acids may have occurred. Accelerated methods of testing the shelf-life of lipidcontaining products have been reviewed by Rossell (1983) and Frankel (1993). The Schaal Oven Test employed here was originally devised by the cracker and biscuit industry. It is a simple, uncomplicated test using heat as an accelerant of oxidation and can be expected to provide a reasonable comparison between samples subjected to the same test conditions (Rossell 1983 ; Frankel 1993). The course of oxidation was followed in the butter oil samples by testing aliquots of oil at regular intervals using PV and TBA determinations. The initial stage of very slow oxidation (the induction period) can be measured as the time required to reach 549 an end point of oxidation corresponding either to a level of detectable rancidity (a deÐned peroxide value) or, as in this study (Figs 1 and 2), to a sudden change in oxidation rate. Graphs of peroxide values and TBA values for butter oils held at 63¡C are shown in Figs 1 and 2, respectively. Gompertz curves were Ðtted to the TBA and PV results for the butter oils and the conclusion of the induction phase was identiÐed by drawing a line tangential to the point of inÑexion to cut the lower asymptote. The signiÐcance of the di†erence between the duration of the initial lag or induction phases of the clariÐed milk fats (butter oils) from the di†erent dietary treatments was tested. Results are presented in Table 3 and show that there was a highly signiÐcant (P \ 0É01) difference in the length of the induction phase between milk fats obtained from cows o†ered barley-based or naked oats-based concentrates, with the milk fat from the naked oats diet having a longer induction period. The TBA values showed a similar trend with a very highly (P \ 0É001) signiÐcant longer induction phase for milk fat originating from cows receiving a naked oatsbased diet. For both sets of data the actual values for the duration of the induction phase as identiÐed from the PV and TBA graphs were similar, approximately 9 days for milk fats from cows o†ered the control barleybased diet, and 13 days for milk fats from cows o†ered the diet containing naked oats. In the early stages of oxidation any antioxidants present in a lipid-containing product will play a role in inhibiting or delaying the onset and progress of oxidation. By the end of the induction period, however, the antioxidant molecules will all have been involved in reactions. Although antioxidants were not added to the milk fats, it is acknowledged that tocopherols, in particular a-tocopherol, will be transferred to milk fat from the cowsÏ diet (Hidiroglou 1989 ; Nicholson and St Laurent 1991). An inverse relationship has been reported between susceptibility of milk to oxidised Ñavour TABLE 2 Peroxide values (PV) (mEq peroxide kg~1) and thiobarbituric acid (TBA) values (mg malonaldehyde kg~1) of butters prepared from milk from cows o†ered barley or naked oats-based diets, and stored at 20¡C Storage period (days) 0 1 8 15 22 a ND, not detected. Barley-based (control) concentrates Naked oats-based concentrates PV T BA PV T BA NDa ND ND ND ND 0É00 0É00 0É078 0É078 0É156 ND ND ND ND ND 0É00 0É00 0É078 0É078 0É156 A M Fearon, C S Mayne, C T Charlton 550 Fig 1. Gompertz curves Ðtted to peroxide values (PV) for clariÐed milk fats stored at 63¡C from cows o†ered barley (È = È) or naked oat-based (È È…È È) concentrates. Fig 2. Gompertz curves Ðtted to thiobarbituric acid (TBA) values for clariÐed milk fats stored at 63¡C from cows o†ered barley (È = È) or naked oats-based (È È…È È) concentrates. and tocopherol content (Schingoethe et al 1978). Therefore, it was necessary to eliminate the possibility of the dietary cereal sources contributing di†ering amounts of a-tocopherol by analysing the prepared concentrates. Both the barley- and the naked oats-based concentrates contained 15É6 kg g~1 a-tocopherol (Table 1). However, it has been reported that oats contain other antioxidant compounds in the form of esters of phenolic acids, which were found to be e†ective at increasing the oxidative stability of a vegetable oil system (Duve and White 1991). Nevertheless, although this may be one possible explanation for the di†erences observed in induction phase duration of the milk fats, it was not determined in this experiment if these phenolic compounds were transferred into the milk fat. Results from the sensory assessment of the butter oils are presented in Table 4. Assessors were asked to identify which of two coded samples smelled di†erently from a reference sample. Statistical analysis of the results tested if the number of correct replies was statistically signiÐcant (Table 4). It was found, irrespective of the cereal content of cowsÏ diet, that following storage at 63¡C for 3 days, a perceptible di†erence in odour could be detected between stored and unstored milk fats. This period of storage (3 days) corresponded to PVs \4É0 meq kg~1 and is in agreement with the work of Downey (1974) who observed perceptible odour development in milk fats having PVs as low as O2 meq kg~1. (In the sensory assessment the milk fat originating from cows on the barley diet was not compared directly with those milk fats from cows on the naked oats diet. Instead the storage periods at 63¡C which produced a perceptible change in odour for each milk fat were identiÐed.) The 3 days storage period however, contrasts markedly with the end of the induction phases as determined by TBA and PV analyses, which were approximately 9 days and 13 days for milk fats from barley- and naked oats-based concentrates, respectively. Nevertheless, the lower level of signiÐcance TABLE 3 Duration and statistical signiÐcance of di†erence between induction phases identiÐed from Gompertz curves Ðtted to peroxide values (PV) and thiobarbituric acid values (TBA) for clariÐed milk fats from dairy cows o†ered barley (control) or naked oats-based concentrates Diet treatment Barley (control) Naked oats Statistical signiÐcance of di†erencea PV (mEq peroxide kg~1) T BA (mg malonaldehyde kg~1) Induction phase (day) SE Induction phase (day) SE 9É56 12É85 0É006 0É607 0É653 (P \ 0É01) 8É84 13É26 0É0002 0É489 0É470 (P \ 0É001) a t value \ (induction [ induction /J(SE )2 ] (SE )2. 1 2 1 2 Oxidative stability of dietary modiÐed milk fat 551 TABLE 4 Results of duo-trio di†erence test for sensory assessment of odour di†erence of clariÐed milk fats obtained from cows o†ered barley or naked oats-based concentrates and stored at 63¡C T reatment (days at 63¡C) 3 (PV \ 4) 8 (PV 6È8) 13 (PV 10È15) 21 (PV [ 20) Barley vs standard Naked oats vs standard Correct repliesa SigniÐcanceb Correct repliesa SigniÐcanceb 29/32 24/32 26/32 25/32 P \ 0É001 P \ 0É01 P [ 0É001 P \ 0É001 22/32 21/32 26/32 28/32 P \ 0É05 NS P \ 0É001 P \ 0É001 a Panellists could identify odour di†erence between treated (stored at 63¡C) and standard (unstored) milk fats. b Level of signiÐcance (BS 5929, 1992) was stated as 26/32 (P \ 0É001), 24/32 (P \ 0É01), 22/32 (P \ 0É05) and less than 22/32 (not signiÐcant). c PV, peroxide value, mEq peroxide kg~1. at day 3 (P \ 0É05) and day 8 (not signiÐcant) for milk fat from the naked oats diet is in keeping with the order observed with the chemical indices of oxidation (Table 2). There are several factors which should be considered when interpreting these results. In the duo-trio test, the assessors were presented with an identiÐed reference, thus reducing the need for expert or selected assessors (Kilcast 1995). However, the assessors, although receiving some training, were not experienced in identifying or describing oxidative o†-Ñavours, and it was therefore difficult to interpret whether the odour di†erence they perceived was due to oxidative o†-Ñavour development or to odour contributions from other sources. For example, volatiles originating from the cereal (Urbach 1990) or from hydrolytic rancidity reactions (Downey 1974 ; Deeth, Fitzgerald and Wood 1979) which may become more dominant or pronounced with prolonged heating at 63¡C. Nevertheless, from this study the inclusion of naked oats in the diet of dairy cows would seem to be an e†ective means of modifying the composition of milk fat, enhancing its functional properties (Fearon et al 1996) whilst not adversely a†ecting the oxidative stability of the fat. Apparent discrepancies between the sensory testing and chemical indices of oxidation underlined problems associated with correlating such tests, although both were in agreement with the order of stability of the milk fats. ACKNOWLEDGEMENTS The authors wish to thank Dr Linda Farmer and the sta† of the Sensory Evaluation Unit for their assistance with this work. REFERENCES AOCS 1990 Sampling and Analysis of Commercial Fats and Oils. Peroxide V alue (Official Method Cd 8-53). American Oil Chemists Society, Chicago, USA. Ashes J R, St Vincent Welch P, Gulati S K, Scott T W, Brown, G H 1992 Manipulation of fatty acid composition of milk by feeding protected canola seeds. J Dairy Sci 75 1090È1096. Banks W, Christie W W 1990 Feeding cows for the production of butter with good spreadability at refrigeration temperatures. Outlook Agric 19 43È47. Banks W, Clapperton J L, Kelly M E 1980 E†ect of oil enriched diets on the milk yield and composition, and on the composition and properties of the milk fat, of dairy cows receiving a basal ration of grass silage. J Dairy Res 47 277È285. Bliss C I 1970 Statistics in Biology (Vol 2), Mcgraw-Hill, New York, pp 194È205. Deeth H C, Fitzgerald C H, Wood A F 1979 Lipolysis and butter quality. Aust J Dairy T echnol 34 146È149. Downey W K 1974 Butter Quality. An Foras Taluntis, Dublin, Republic of Ireland. Duve K J, White P J 1991 Extraction and identiÐcation of antioxidants in oats. J Am Oil Chem Soc 68 365È370. Farag R S, Ali M N, Taha S H 1990 Use of some essential oils as natural preservatives for butter. J Am Oil Chem Soc 68 188È191. Fearon A M, Mayne C S, Marsden S 1996 The e†ect of inclusion of naked oats in the concentrate o†ered to dairy cows on milk production, milk fat composition and properties. J Sci Food Agric 72 273È282. Frankel E N 1993 In search of better methods to evaluate natural antioxidants and oxidative stability in food lipids. T rends Food Sci T echnol 4 220È225. Goering H K, Gordon C H, Wrenn T R, Bitman J, King R L, Douglas Jr F W 1976 E†ect of feeding protected safflower oil on yield, composition, Ñavour and oxidative stability of milk. J Dairy Sci 59 (3) 416È425. Hawke J C, Taylor M W 1983 Developments in Dairy ChemistryÈ2 : L ipids, ed Fox P F. Applied Science Publishers, London, UK, pp 37È81. 552 Hidiroglou M 1989 Mammary transfer of Vitamin E in dairy cows. J Dairy Sci 72 1067È1071. Hoyland D V, Taylor A J 1989 A modiÐed distillation method for the detection of oxidation in foods. Int J Food Sci T echnol 24 153È161. Kilcast D 1995 Food T aints and O†-Flavours, ed Saxby M J. Blackie Academic and Professional, London, UK, pp 1È40. Kinsella J 1972 Stearyl CoA as a precursor of oleic acid and glycerolipids in mammary microsomes from lactating bovine : Possible regulatory step in milk triglyceride synthesis. L ipids 7 349È353. McMurray C H, BlanchÑower W J 1979 Determination of a-tocopherol in animal feedstu†s using high-performance liquid chromatography with spectroÑuorescence detection. J Chromatogr 176 488È497. Morrison W R 1977 Cereal lipids. Proc Nutr Soc 36 143È148. Murphy J J, McNeill G P, Connolly J F, Gleeson P A 1990 E†ect on cow performance and milk fat composition of including full-fat soya beans and rapeseeds in the concentrate mixture for lactating dairy cows. J Dairy Res 57 295È 306. Nicholson J W G, St-Laurent A M 1991 E†ect of forage type and supplemental dietary Vitamin E on milk oxidative stability. Can J Anim Sci 71 1181È1186. Pearson D 1976 T he Chemical Analysis of Foods (7th Edn), Churchill Livingstone, Edinburgh, UK, pp 494È495. A M Fearon, C S Mayne, C T Charlton Rossell J B 1983 Rancidity in Foods, ed Allen J C & Hamilton R J. Applied Science Publishers, London, UK, pp 21È46. Rossell J B 1989 Intermediate shelf life products as illustrated by fats and fatty foods. Food Sci T echnol T oday 3 235È240. Schingoethe D J, Parsons J G, Ludens F C, Tucker W L, Shave H J 1978 Vitamin E status of dairy cows fed stored feed continuously or pastured during summer. J Dairy Sci 61 1582È1589. Scott T W, Cook L J, Ferguson K A, McDonald I W, Buchanan R A, Loftus Hills G 1970 Production of polyunsaturated milk fat in domestic ruminants. Aust J Sci 32 291È293. Sonntag, N O V 1979 BaileyÏs Industrial Oil and Fat Products (Vol 1, 4th edn), ed Swern D. J Wiley and Sons, Chichester, UK, pp 99È176. Tarladgis B G, Watts B M, Younathan M T, Dugan L 1960 A distillation method for the quantitative determination of malonaldehyde in rancid foods. J Am Oil Chem Soc 37 44È 48. Urbach G 1990 E†ect of feed on Ñavour in dairy foods. J Dairy Sci 73 3639È3650. Valentine J 1987 Breeding cereals of high nutritional quality with special reference to oats and naked oats. Asp Appl Biol 15 541È548.