Environmental correlations in sire evaluation.

Correlations were computed among averages of first lactation deviations from herdmate averages of groups of artificially sired daughters of the same Holstein bulls. The groups were constructed by specifying the time interval between animals in the different groups. Comparison of the actual correlations with the expected correlations suggests that environmental correlations are small or nil among artificially sired halfsibs in New York. The correlations between initial and later groups of 20 or 40 daughter records at different time intervals are so close to the expected correlations that there is no support for the view that atcuracy of prediction in sire evaluation is being over-estimated due to failure to properly account for environmental correlations among contiguous half-sibs in artificial insemination. Lush (3, 4) and Lush and McGilliard (5) have demonstrated the effect of environmental correlations among daughters of the same sire on sire evaluation. Bereskin and Lush (1) have discussed this problem and have presented evidence from Iowa D H I A data which indicates a significant effect of environmental correlaReceived for publication July 16, 1965. tions on predicting the production of future daughters. Heidhues et al. (2) and McDaniel and Corley (6), however, reported no significant departure from the expected correlations among the first ten, second ten, etc., daughters of a sire. Results presented in this paper suggest that under conditions of artificial insemination in New York, environmental correlations are not important in evaluation of sires used artificially. Data, Methods, and Results First-lactation milk production records (305day, 2 × , M.E.) of artificially sired Itolstein cows were expressed as deviations from their herd-mate averages and used in the analyses, providing certain conditions were met. The records were sorted into order according to the birth date of the cow. Sire groups for analysis were deternlined in several ways. These methods are described in Table 1. For example, analysis (20:12, 12, ~2) included 44 sire groups, all having at least 80 daughters distributed according to birth date into four equal groups, with a minimum time difference of 12 months for birth dates between individuals in first and second groups, between individuals in the second and third groups, and between individuals in the third and fourth groups. No restriction was placed on animals in a group except that if records of the first 20 daughters born were included in the first group, then records of all