Randomly breeding populations

Abstract

So far we have been concerned with the analysis of data obtained from true-breeding lines and the descendants of crosses made between them. Following such a cross, a multiplicity of genersations and types of family can be raised experimentally — a multiplicity limited only by the biological properties of the material (whether, for example, it can be selfed as well as crossed, whether individuals can be kept alive for crossing to their own offspring and so on) and by the time and facilities available for the experimental programme. Each generation and type of family will have its own mean and variance, and its own covariances with other related families. Thus a large number of statistics can be obtained from which we can estimate the genetical and environmental components of both means and variances. The specification of these components of variation is simpler because by starting with true-breeding lines we can, in the absence of selective elimination, specify the relative frequencies of the types of zygotes and gametes that we expect in and from any given type of family.