1. Records used to compile the Contemporary Comparisons of 57 Friesian, 8 English Ayrshire and 11 Scottish Ayrshire A.I. bulls, each with at least 100 ‘effective daughters’ were analysed.

2. For each bull, the herd-years were divided into three equal groups on the basis of the average heifer yield of both daughters and contemporaries (high-, medium-, and low-producing herd-years) and three independent Contemporary Comparisons were calculated for each bull, one at each of the three yield levels.

3. In the data from England and Wales, the mean Contemporary Comparison declined with increasing mean level of production. This decline was such as to imply that some 20% of the differences in production between herds were genetic in origin. A possible explanation lies in the gradual change from Dairy Shorthorn to Friesian and Ayrshire which has taken place in England and Wales, but not in Scotland, during the last 15 years.

4. The variance within progeny groups within herd-years increased from the low level to the high but the coefficient of variation decreased slightly. The variance between sires also increased with the mean level of production but almost exactly in parallel with that within sires so that the heritability and consequently the accuracy of the progeny test for milk yield was effectively the same at all production levels.

5. The correlation between the true breeding values of the bulls at the different levels was very close to one.

6. From these results it is concluded that there is no need to provide special strains within breeds to suit particular management levels or to concentrate progeny testing in the higher-producing herds and that daughter records from all herds, irrespective of level of production, can be used with equal confidence.