1994 年 5 巻 2 号 p. 53-58
The sire effect and the fixed effects of sex, training center, type of course, specific racetrack, horse age and track condition were evaluated regarding best racing times of thoroughbred horses. The records of best racing times at the Japan Racing Association were collected from the electronic racing book. The data set was edited so that each sire was required to have at least twenty progeny in the racing distances of 1200m and 1800m. The total numbers o f sires and progeny were 34 and 1486 for 1200m and 34 and 1520 for 1800m. First, the least-squares analysis of variance was carried out by using the linear model which included the sire effect and the six fixed effects. Second, heritabilities for best racing times were calculated, and additionally sire evaluations were carried out by using the sire model of the BLUP method. As a result of analysis of variance, almost all effects included in the linear model were significant for both racing distances. The sire, type of course, horse age and interaction between the type of course and track condition had a highly significant effect on best racing times. Estimates of heritability were 0.11 and 0.09 for 1200 m and 1800 m, respectively. Rankings of sires' breeding values based on progeny records were quite different for the two racing distances. The Spearman rank correlation coefficient for these rankings was 0.350 (P<0.05).