B. Langlois et al., NEW RESULTS ON THE ESTIMATION OF THE HERI TABILITIES OF EARNINGS OF THOROUGHBREDS IN FRANCE, Genetics selection evolution, 28(3), 1996, pp. 275-283
Logarithms of yearly earnings of 232 893 horses in flat races and 85 1
45 horses in steeplechases and hurdle races in France from 1950 to 199
0 were analysed according to an animal model. REML methodology was app
lied to chronological subsamples of this enormous file to estimate gen
etic parameters. In flat races, from the birth year 1960 to 1987, heri
tability increased from 0.22 to 0.28 according to a statistically high
ly significant linear trend, and the repeatability between 2 and 3 yea
rs changed in a similary significant manner from 0.35 to 0.40. Over th
e jumps, from 1962 to 1985, fluctuations appeared probably due to a lo
wer precision of the estimates and no significant evolution could be s
een for heritability (near 0.25). Repeatability between 4 and 5 years,
however, increased significantly. The reasons for the evolution of th
ese parameters over time are discussed. Two main, hypotheses are menti
oned: (i) introduction of external genetic variability in French pedig
rees through American blood (this hypothesis however is not in agreeme
nt with the very small genetic distance between the two populations);
and (ii) an equilibrium between two opposite sources generating linkag
e disequilibrium, selection and assortative mating. The intensity of t
hese two forces observed in each subsample allows us to explain fairly
well the variations of observed parameters.