Milk collected at retail markets and in schools in Connecticut in 1970–71, 1974, and 1975 was examined for flavor quality, temperature at collection, code date (last day product is intended to be offered for sale), and age of sample (days from bottling). The relationships between quality and age of sample and the processor's code date were tested. Regression analysis of 1975 samples showed that, on the average, processors overestimated shelf life by about 2 days, but this overestimate varied from 0 to 7 days for individual dairies. Although the temperature of samples collected in 1974 and 1975 was lower than in the 1970–71 sampling, there was no diminution in the percentage of samples of unsatisfactory quality.
1Department of Biochemistry
3Department of Ecology and Climatology.