Genetic and maternal influence on the first three lactations of holstein cows.

Milk records of up to the first three lactations of artificially sired Holstein daughters and their dams were used to estimate heritability from both daughter-dam regression and paternal half-sib correlation. The within-herd estimates h'om records expressed as deviations from herd-mate averages were :37, .30, and .24 from daughterdam regression for the first three lactations ; and the corresponding estimates h'om halfsib correlation were .24, .21, and .23. These results suggest a large maternal effect in the first lactation, a small amount in the second, and little in the third lactation. Analysis of the records not as deviations gave the same pattern for the daughterdam regressions, but confounding of sires with year-seasons apparently biased the half-sib correlations upward. The consequences of considering unequal heritability for different lactations if the apparent difference is due to maternal effects is illustrated for selection for first-lactation breeding value from up to three records on the cow herself and on her dam. Some recent reports (2, 7) have suggested that the first-lactation record of a dairy cow is more important than any later records for evaluating her breeding value. Little will be added to the accm~ey of prediction by using more than the first record. Evidence from other reports [see, for example, Rende] et al. (8)] which have found apparently higher heritability estimates from first-lactation records than for later records would confirm this opinion. A question of major importance in determining the relative value of various lactation records is concerned with selecting cows which will be mated to selected bulls to produce the next generation of young bulls to be sampled in artificial insemination (A,I.). I f the common opinion that one must know at least three or four records of a cow before making her Received for publication June 16, 1965. Department of Animal Husbandry, University of California, Davis. eligible to produce a son for A.I. is really folklore, then this must be proved, so that the generation interval can be shortened through use of cows with only one known record. The same reasoning will also be true for within-herd culling, but the effects of selection on the whole population will be determined much more through A.I. than through herd culling. This study was made to determine estimates of heritability for the first three lactations and the genetic correlations among these records, so that the importance of later lactations for genetic selection can be evaluated.