ABSTRACT
Estimates of population parameters are required in developing selection indices. Estimated indices will not give as great an overall response to selection as would an index based on true population parameters. It is important to know how large a sample is required to develop precise estimates of population parameters, as well as knowing how serious it is to proceed with estimates of questionable precision. In practice, the genotypic effects of alleles at different loci are distributed according to a multinomial distribution. However, the genotypic value of an individual or a line will be the overall sum of the effects at the individual loci. The distribution of the overall genotypic values will approach a normal distribution if the number of loci is large. In most applications of selection indices, it has been assumed that, for a particular trait, genotypic worth either increases or decreases linearly with changes in genotypic value.
