In the present study, high-performance liquid chromatography analysis of butterfat allowed separation of 46 peaks at 32°C. Knowing the theoretical carbon number value of each triglyceride (TG), 32 peaks of the butterfat chromatogram were identified. These TGs were determined by extrapolation of their capacity factor values, and their identifications were confirmed with some standard TGs. Analysis of winter and summer butterfat from five different French areas showed significant seasonal and regional variation in the TG composition. However, the most important contribution to this variation was provided by TG groups represented by only four peaks. To approximately select the predominant TGs in these four peaks, a random distribution hypothesis was used to predict the amount of each TG. This hypothesis allowed the prediction of the TG components that seem to provide the most important contribution to both seasonal and regional variation.Many authors have investigated the seasonal fluctuation of the fatty acid (FA) composition of butter (1,2). The variation in FA composition explains the regular seasonal fluctuations of iodine values observed by Cox and McDowall (3) and of solid fat content of milk observed by Norris et aL (4). However, the triglyceride (TG) structure, i.a, arrangement and distribution of FAs in TGs, seems to have more influence on the physical characteristics of fats. In fact, Deman (5) and Pitas et at (6) showed that interesterification of milkfat, which transforms a highly selective arrangement of FAs into a random distribution, markedly increased hardness, solid fat content and proportion of high melting TGs.More accurate studies (7) showed that, in the case of milkfat, butyric acid (Bu) and caproic acid (Co) are mostly esterified in position 3, while myristic acid {My) and palmitic acid are in position 2 and in position 1 or 2, respectively. Stearic acid (S) and oleic acid (O) partitions depend on the TG molecular weight, mainly in positions 1 and 3 for highmolecular weight and in position 1 for low-molecular weight TG (8). To separate TGe. many scientists have exploited the advantages of reverse-phase high-performance liquid chr~ matography (HPLC). But because of animal fat complexity, few papers dealt with the TG separation.It was important to find the proper parameters, La, column, eluent, temperature or adequate temperature gradient, and detector to obtain the best peak separatiorL Several eluents were used to attempt to separate TGs, such as mixtures of methanol]water (9:D (9), methanol]chloroform (9:1) (10) and methanol]acetone (11). However, according to Deffense {12}, the most efficient way was to use an acetone] acetonitrile mobile phase at 50°C Frede and Thiele (13) con-*To whom correspondence should be addressed. firmed its efficiency by using the same mixture as the mobile phase (35:65), but they set the temperature of the column (Nucleosil C18-5 ~m, 15 cm; Macherey and Nagel, Duren, Germany, + Microspher C18-3 ~m, 10 cm, in series; chrompack, Mfilhein, Germany) at 30°C.In HPLC anal...