|
|
||||||||
University of Maine,2 Orono 04473
Abstract
Two separate heritability studies, using a newly developed rating system called Performance Rates based on multiple comparisons of horses with one another, were performed. The first study using 39 pairs of maternal full sibs estimated 0.36 for heritability. The second study utilized 75 sets of three paternal half brothers. This gave a heritability of 0.68 as an estimate. So a weighted average of the two gave a heritability estimate of 0.55.
A repeatability study utilizing the average earnings per start for 256 horses for three separate racing years resulted in an estimate of 0.37.
A selection index using multiple linear regression was developed to select young sires: Y=1.209+.097 (X1)–.144 (X2)+.027 (X3) where Y is the estimated progeny average earnings index (A.E.I.) of the horse, X1 the progeny A.E.I, of the sire, X2 the progeny A.E.I. of the broodmare's sire, and X3 the horse's Average Earnings Per Start Index (A.E.P.S.I.). This index predicted within .949 A.E.I. for 67% of the horses in this study. The squared multiple correlation coefficient was .345 which gave a correlation coefficient of .59 and consequently a reasonable accuracy. The greatest indicator by far of a sire's breeding potential was found to be his phenotype, or in this case his Average Earnings Per Start Index.
1 Gratitude is expressed to Raymond H. Gillespie of The Jockey Club Statistical Bureau for data furnished for the repeatability study.
2 Department of Animal and Veterinary Sciences.
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |