Genotype by Environment Interaction Study of Cotton in Oklahoma1

Abstract

Eleven cotton (Gossypinm hirsutum L.) varieties were grown at three locations in Oklahoma over a three‐year period. Lint yield, fiber length, fiber coarseness, and two measures of fiber strength were studied using the analysis described by Comstock and Moll.Significant varieties by years mean squares were obtained for fiber length and one measure of fiber strength. However, those mean squares were so small relative to their respective varietal mean squares that the evaluation of length and strength over years was considered unnecessary. A very large and significant varieties by locations mean square for yield suggested that the state should be subdivided in some manner for varietal testing and breeding purposes. A division into dryland and irrigated production may be adequate. The second‐order interaction mean squares for yield and fiber coarseness were large and significant. Therefore, a sizable proportion of the interactions exhibited by those two traits could not be attributed to locations or to years. Inferences made from comparisons of variance components were the same as those obtained from an examination of mean squares.