Samples of foremilk, composite milk and strippings were obtained from 13 cows for progesterone analysis to determine the effectiveness of the different sampling methods to detect open and pregnant cows. The milk samples were taken during the afternoon milking of the day of breeding (day 0), day 14, day 24 and after day 35 (from pregnant animals). Progesterone concentrations, determined by radioimmunoassay, were 1.4 ±.4, 2.1 ±.8 and 2.8 ± 1.1 at day 0; 7.9 ± 5.0, 19.1 ± 9.1 and 30.1 ± 6.2 at day 14; 8.6 ± 6.5, 21.7 ± 7.4 and 29.4 ± 7.3 at day 24; 7.2 ± 2.8, 21.8 ± 4.8 and 33.5 ± 1.4 at day 35+ for foremilk, composite and strippings samples, respectively. It was determined by discriminate analysis that all three methods would correctly place samples obtained on day 0 into the correct group. Only in the case of strippings samples did there appear to be no danger of classifying pregnant animals in the day 0 (estrus) group. It is suggested that milk strippings represent the preferred sample to take for milk progesterone diagnosis of pregancy or non-pregnancy due to ease of sampling and to a reduced chance of making a faulty diagnosis.