Comparison of Relative Accuracy between Testing Semimonthly Composites or Single Random Samples of Herd Milk for Composition

Abstract

Summary The relative accuracy of testing composites or single random samples for composition was compared for milk samples from 30 herds. The bulk milk from each herd was sampled on alternate days for 2 wk during February, May, August, and November. Percentage milk fat in composite samples averaged 0.055 and 0.021% less than the average of the fresh milk aliquots, respectively, for Laboratories I and II. The average per cent SNF (TS gravimetric—F Babcock) in composite samples was 8.95%, compared with 8.91% in their fresh milk aliquots. For testing composite samples for per cent SNF, the plastic bead method of Golding was equal or superior to the gravimetric procedure. Two random tests for per cent milk fat each 2 wk for herds on alternate-day delivery estimated per cent milk fat as accurately as one test of a composite sample. Comparisons between testing single random samples and the average of the fresh milk aliquots over 2-wk periods showed variations from regression (S y · x ) of 0.13, 0.10, 0.13, 0.11, 0.08, and 0.12%, respectively, for milk fat, SNF (gravimetric), SNF (plastic beads-single equation), SNF (plastic bead-milk fat interval equations), milk protein, and lactose + mineral.