The Design and Analysis of Genetic Experiments to Modify Genotype-Environment Interaction. I. Single Pair Matings and Selection Using One Simple Quantitative Attribute Measured in Two Environments

Most breeding programs are aimed at producing quantitative changes in the genetic structure of the population in question. Available theory and designed experiments have failed to show how to modify the expression of genotype-environment interaction and assume that it is negligible or is not present.This paper considers the design of an experiment to test the feasibility of modifying this interaction and gives the necessary formulae to evaluate the results. It suggests that a number of genetic groups, say each with 2n full-sibs, should be equally divided into two random sub-groups and placed in different environments, e.g., two nutritional or climatic levels. One environment is where the parents and the first group of n randomly chosen offspring are reared continuously and the other environment is the one in which the second group of offspring is reared from birth. A criterion is then established for a selection program based on the performance differential of the same simple quantitative attribute measured in full-sibs reared in the two environments. This scheme can be employed for selecting for this criterion in three directions. Extensions of these theoretical considerations for the cases of more than one simple quantitative attribute, part-whole correlated attributes, indirect selection and more complicated designs will follow.