Implications of genetic-environmental interaction in animal breeding

Abstract

This report is an attempt to interpret evidence concerning geneticenvironmental interaction as it relates to the effectiveness of selection.Methods of estimating genetic-environmental interaction were investigated with the conclusion that the standard analysis of variance is a satisfactory tool provided (1) the interaction component of variance is adjusted for important variation between environments in the scale of genetic effects and (2) the variance component for the average effects of genetic groups is recognised as equivalent to the average covariance of the same genetic group in different environments (i.e.) to include the real possibility of negative genetic correlation.It is shown that (1) response to selection for improved average performance over varying environments (ΔGt) is proportional to the genetic correlation between expressions of the same genotype in different environments (rG), that (2) gain in ΔGi, from measuring performance inkenvironments is proportional toifktimes as many animals are tested, but toif total numbers per genotypic class (nk) are held constant, whereg2is heritability of individual variation within environments andrGis the correlation between the phenotypic expressions of the same genotype in different environments.Evidence is presented that (1) variations among poultry-farm environments in California cause important, but largely unpredictable, shifts in ranking of genetic stocks for egg production, and (2) improved accuracy in measuring genetic differences in average performance across environments probably justifies utilising a sample of 5 to 10 farms representing the range of environments for which the stock is bred.