Heritability of racing ability in thoroughbreds — A review

A survey is made of heritability estimates of timing (h² < 0.20), weight handicaps, performance rates and earnings (0.30 < h² < 0.40). The problem of genetic progress is then examined because, the progression of record times being slow or inexistent, it has been recognised for many years that the population has reached its selection plateau. Several hypotheses for explaining this phenomenon are reviewed. They can be ranked into two groups: the first includes those hypotheses which assume that the heritability values found correspond to a true genetic variability; and the second includes those hypotheses enumerating the causes of overestimations.In the first analysis, the theory of a reduction of the variance parallel to an increase of the mean, according to which high heritabilities might be explained without any observation of progress during important races, is not confirmed. In addition, it also does not seem to be justifiable to assume that the breed fights against fitness. However, it is to be noticed that there is no contradiction between the slow progress of record times and low heritability values of timings, especially as selection is not made directly on timing measures, but on race performance criteria with which they are not necessarily strongly correlated (a correlation of about ?0.50 only).In the second analysis, many sources of heritability overestimation can be pointed out: assortative mating, high homogeneity of the harems and distribution of environmental effects according to the quality of the parents.These considerations lead to the question of the genotype-environment correlation, which should not deliberately be discarded in the case of thoroughbreds.