This document specifies a reference method for the determination of milk fat purity using gas chromatographic analysis of triglycerides. The method utilizes the differences in triglyceride fingerprint of milk fat from the individual triglyceride fingerprints of other fats and oils to determine samples which are outside the range normally observed for milk fat. This is achieved by using the defined triglyceride formulae based on the normalized weighted sum of individual triglyceride peaks which are sensitive to the integrity of the milk[6][7]. The integrity of the milk fat can be determined by comparing the result of these formulae with those previously observed for a range of pure milk fat samples[12]. Both vegetable fats and animal fats such as beef tallow and lard can be detected. The method is applicable to bulk milk@ or products made thereof@ irrespective of the variation in common feeding practices@ breed or lactation conditions. In particular@ the method is applicable to fat extracted from milk products purporting to contain pure milk fat with unchanged composition@ such as butter@ cream@ milk and milk powder. Because a false-positive result can occur@ the method does not apply to milk fat related to these circumstances: a) obtained from bovine milk other than cow??s milk; b) obtained from single cows; c) obtained from cows whose diet contained a particularly high proportion of vegetable oils such as rapeseed@ cotton or palm oil@ etc.; d) obtained from cows suffering from serious underfeeding (strong energy deficit); e) obtained from colostrum; f) subjected to technological treatment such as removal of cholesterol or fractionation; g) obtained from skim milk@ buttermilk or whey; h) obtained from cheeses showing increased lipolysis; i) extracted using the Gerber@ Weibull?CBerntrop or Schmid?CBondzynski?CRatzlaff methods@ or that has been isolated using detergents (e.g. the Bureau of Dairy Industries method). With the extraction methods specified in i)@ substantial quantities of partial glycerides or phospholipids can pass into the fat phase. NOTE 1 In nature@ butyric (n-butanoic) acid (C4) occurs exclusively in milk fat and enables quantitative estimations of low to moderate amounts of milk fat in vegetable and animal fats to be made. Due to the large variation of C4@ for which the approximate content ranges from 3@1 % fat mass fraction to 3@8 % fat mass fraction@ it is difficult to provide qualitative and quantitative information for foreign fat to pure milk fat ratios of up to 20 % mass fraction[11]. NOTE 2 In practice@ quantitative results cannot be derived from the sterol content of vegetable fats@ because they depend on production and processing conditions. Furthermore@ the qualitative determination of foreign fat using sterols is ambiguous. NOTE 3 Due to special feeding practices such as those related to c) and d)@ false-positive results have sometimes been reported for milk from certain Asian regions[15]. Moreover@ grass-only diets such as mountain and@ in particular@ highland pasture feeding sometimes cause false-positive results@ which can be substantiated by a content of conjugated linoleic acid (C18:2 c9t11) of ??1@3 % fatty acid mass fraction[16][17]. Nevertheless@ results conforming to the criteria of milk fat purity specified in this document are accepted@ even if samples were undoubtedly produced under conditions reported in this note@ including those described in h). NOTE 4 In cases where a positive result is suspected to be caused by circumstances related to c) or d)@ another analytical method@ such as fatty acid or sterol analysis@ can be applied to confirm the finding. Due to similar or increased limitations (e.g. as described in NOTE 1 and NOTE 2)@ a negative result obtained by another method is not appropriate to contrastingly confirm milk fat purity.