Search

A selection experiment with Scottish Blackface sheep was used to compare predicted and realized responses to selection. Three closed lines, of approximate annual size of 270 ewes and 10 rams, were maintained between 1956 and 1974, in which selection was at random, or for high and low values of an index of cannon-bone length at 8 weeks of age adjusted for body weight at the same age. An unselected base flock (1954-55) and the randomly selected line were used to estimate base population parameters, while the selected lines were used to estimate realized responses to selection.Heritabilities and genetic correlations were obtained in the unselected lines from a variety of collateral and ancestral relationships. The important components of phenotypic variance were estimated and likely responses to selection predicted for the index and its two component traits. Realized responses to selection were estimated from the regression of response on selection differential. The expected variance-covariance matrix of observed responses was included in generalized least-squares estimates of these regressions.The realized heritability of the index under selection, estimated from the divergence of selected lines, was 0·52 (s.e. 0·02). After allowing for the expected reduction in heritability arising from linkage disequilibrium, this was very similar to the base population estimate of 0·56 (s.e. 0·04). The responses in the component traits of the index were also very close to those expected from base population parameters

Reproductive performance, body growth, wool production and ewe survival were examined in Scottish Blackface sheep. The data set was from a large randomly selected population maintained under hill conditions between 1954 and 1974. Paternal half-sib analyses of variance were used to estimate heritabilities and genetic correlations, while maternal half-sib analyses of variance indicated the relative importance of maternal effects (both genetic and environmental).Among reproductive components, the heritability of fertility (ewes lambing per ewe mated) and lamb survival (lambs weaned per lamb born) were about zero. Litter size (lambs born per ewe lambing) had an estimated heritability of 0·12 (s.e. 0·04) and a repeatability of 0·22 (s.e. 0·02), indicating some scope for improvement. Additionally, the strong genetic correlation between adult body weight and litter size (0·66, s.e. 0·18) suggested greater response from a combination of indirect selection (on body weight) and direct selection (on dam's litter size). Prior to weaning, direct genetic effects on body weight were very small (heritability less than 0·1) with a large influence of maternal environment. After weaning, these maternal influences gradually diminished and the heritability of weight increased. Genetic correlations between body weights showed a similar increasing trend with age. Genetic improvement of early growth rate would be more effective by selecting on a post-weaning rather than a pre-weaning body weight. Fleece weight had a high heritability (about 0·5) but was genetically uncorrelated with other measures of performance. Ewe survival had a small genetic component although estimates of heritability were very imprecise.

A selection experiment with Scottish Blackface sheep was used to compare predicted and realized correlated responses to selection. Three lines were maintained between 1956 and 1974 in which selection was either at random or for high and low values of an index of cannon-bone length at 8 weeks of age adjusted for body weight at the same age.There was no evidence of asymmetrical responses in any trait. Selection for increased cannon-bone length, adjusted for body weight, resulted in (i) increased body weights at all ages between birth and maturity, (ii) increased reproduction rate, principally from increased litter size but also from small responses in ewe fertility and lamb survival and (iii) decreased survival of adult ewes. The realized genetic correlations were relatively small (up to 0·3) but mostly significantly different from zero, whereas base population genetic correlation estimates were very imprecise. The realized responses in reproduction rate were probably a consequence of the genetic response in body weight.