Heritability and repeatability of speed for two- and three-year- old standardbred racehorses

Repeatabilities (t) and heritabilities (h²) for racing time were estimated from data on 7206 2- and 3-year-old standardbred racehorses competing in charted races at six tracks between 1975 and 1976. There were a total of 38,487 records from 2367 sire progeny groups. The data were divided according to gait into two subsets. Trotters only were used to test alternative models proposed to characterize horse’s racing time. Initially, class of race was recognized as a subjective categorization reflective of non-random assignment of horses to races. After extensive investigation, it was concluded that adjustment of horse’s time should include regression on pace (time of the winner). Data were analyzed within track-age-gait subclasses. Intra-horse regression coefficients of time on pace (sec/sec) ranged from .61 ± .08 to .79 ± .05. When records were adjusted for pace, repeatability and heritability estimates ranged from .32 ± .08 to .55 ± .04 and from -.25 ± .22 to .66 ± .15, respectively. Pooled estimates of t and h² were .44 ± .01 and .29 ± .03, respectively. When records were not adjusted for pace, sire, horse and error variance components were inflated by environmental bias (t_pooled = .62: h² pooled = .36). When records were pre-adjusted for pace, the permanent environmental variance appeared smaller compared with the total phenotypic variance than if there was no preadjustrnent for pace. Since genetic progress is based on selection of genetically superior sires and dams, time adjusted for pace should be used for estimation of breeding values and for objective evaluation of an individual horse's real ability to compete against horses having similar abilities. Thus, resulting breeding values should be more accurate dnd biased to a lesser degree than if unadjusted racing times were utilized.