Abstract
Parametric relationships between the genotype x environment interaction and the genetic correlation of the same attribute measured in two different environments are derived. It is shown that the criticism by Fernando et al. (1984) of Yamada's method (1962) in the case of unbalanced data is irrelevant.