A STUDY OF THE VARIATION IN THE FAT CONTENT OF HERD BULK MILKS

1. The mixed milks of twenty‐eight herds were analysed daily for fat content over a period of thirteen days. 2. Daily fat figures are given together with maximum and minimum and mean daily fat variations. 3. Coefficients of variation have been calculated for individual herds. 4. The permissable variation between the fat contents of random samples of mixed herd milks have been calculated for individual herds and for herds grouped on the basis of the daily gallonage produced. 5. Estimates have been made of the accuracy of the fat percentage of a single sample of mixed herd milk, and of a mean fat percentage derived from the fat contents of four random samples of such milk, as a measure of the true mean fat percentages for individual herds and for the grouped herds. 6. Day‐to‐day variations in the fat percentages of mixed herd milks are rather greater than were indicated by previous workers. Generally larger herds show less variation than smaller herds but this does not hold good for individual herds. No evidence was obtained in the present investigation to support the claims of previous workers of a correlation between fat content and variation, or that the variations were not of a purely random character. The fat contents of two random samples of a mixed herd milk taken within a few days of each other can show considerable differences. Single samples of a mixed herd milk are of little value in assessing the compositional quality of the milk being produced by the herd over a period of several weeks. To obtain a reasonably accurate assessment of the compositional quality of a monthly supply, a mean fat percentage derived from the analysis of at least four random samples should be used. It is suggested that two weekly samples should be bulked for the single weekly analysis to obtain a more valid monthly mean based on eight samples.