Effect of age and stage of lactation on the simple and partial correlations between the yield and the percentages of fat and protein of cow’s milk

Abstract

Coefficients of simple and partial correlations between the milk yield and the percentages of fat and protein in milk were calculated separately for the first-calvers and older cows during the early period of lactation (5—39 days post partum), during a later period (40—360 days p.p.), and during the entire period of lactation, from the data of 1028 milk samples obtained during the indoor-feeding period from 70 different Ayrshire cows belonging to the Viik Experimental Farm dairy herd. Statistical treatment of the data revealed that the milk yield affected the protein content of the milk more than its fat content, especially in the first-calvers. The protein-fat correlation, which was somewhat affected by the milk yield, was generally less close during the early stage of lactation than during its later stage. The percentage of variation in a) the milk yield that was explained by the variations in the percentage of fat and protein, was affected more by age than by the stage of lactation, b) the fat content of milk that was explained by the variations in the milk yield and the protein content of milk, was influenced more by the stage of lactation than by the age of the cow, c) the protein content of milk that was explained by the variations in milk yield and the fat content, was influenced by age as well as by stage of lactation. The equations for estimating theoretically the daily milk yield, the percentage of fat, or the percentage of protein in milk, when two out of the three variables are known, have been presented separately for the first-calvers and for the older cows at early and at later stages as well as during the entire period of lactation. The results of the study indicated that the milk yields and the percentages of fat and protein in milk vary to such an extent independently of each other, even when the age of cows, the stage of lactation, and the season of the year have been taken into consideration, that none of the dependent variables in question could be very closely estimated on the basis of knowing two of the independent variables.