One of the major problems confronting breeders evaluating the yield of genotypes is the presence of large genotype X environment interactions (G X E). From a theoretical viewpoint, heterozygous and heterogenous populations are expected to be more stable in their producing ability and thus have smaller G X E than homozygous and homogenous populations. To test this hypothesis in cotton, isogenic populations of normal, okra leaf, and their F₁ and F₂ hybrids from eight cultivars were grown in 12 environmental regimes. The 32 isogenic populations were grown in two years, 1979 and 1980; two planting dates, April and May; and three seeding rates, 10, 20 and 30 seed/m- on 1m row.

The lint yield response of the four isogenic genotypes interacted with years, planting dates, years X planting dates, years X cultivars, and years X planting dates X seeding rates. The greatest interacting factors were years and years X planting dates with F = 12.2 and 5.7, respectively. Okra leaf produced significantly higher lint yields (77 kg/ha) than normal for the May 1979 plantings, but produced significantly less for the April 1979 and May 1980 plantings, yielding 57 and 88 kg/ha less than normal for the respective regimes. Significant positive dominance (F -midparent) of 68 kg/ha was detected for the April 1980 planting but significant negative dominance of 44 kg/ha was detected for the May 1979 plantings. The F₂ yield was significantly lower, -40 kg/ha, than expected on the basis of the performance of its individual genotypic components [F₂-(F₁ + midparent)/2] for both 1979 plantings. In 1980 the F₂ yielded 29, and significantly, 43 kg/ha more than expected for the April and May plantings, respectively. In addition, the F₂ yield regression for environmental yield index was 0.85 and significantly lower than the regression coefficient for the other three genotypes which averaged about 1.05. Thus, F₂ populations yield better than expected in the low yielding environments and less than expected in high yielding environments. This study indicated F₂ populations were more stable in yield performance than their isogenic homozygous and homogenous normal and okra leaf parents, and also more stable than their heterozygous and homogenous F₁ parents.