IMPLICATIONS OF CONFOUNDING OF GENOTYPE X YEAR AND GENOTYPE X CROP EFFECTS IN SUGARCANE

In sugarcane (Saccharum spp.) field experiments, genotype .times. year (G .times. Y) and genotype .times. crop (G .times. C) interaction effects are usually confounded. In this study, the two effects were separated. Forty-seven random genotypes from a heterogeneous population and a check cultivar in 1984 as ratoon crop. Variance components and broad-sense heritability estimates were calculated for the following: plant crop 83 and ratoon crop 84 (Analysis I); plant crop 84 and ratoon crop 84 (Analysis II); and plant crops 83 and 84 (Analysis III). The Analysis II gave the "true" G .times. C interaction which was higher for sugar concentrations (SC) and tons per ha of sugar (THS) than that detected in Analysis I (confounded G .times. Y and G .times. C). For stalk weight and tons per ha of cane (THC), the G .times. C component was negligible in both analyses. The effect of years on plant crop (Analysis III) was noted for stalk number, THC and THS. Broad-sense heritability estimates increased, for almost all traits studied, in Analysis III as compared with Analyses I and II, indicating that the genetic component was maximized by growing genotypes in the plant crop in two years. The effect of confounding G .times. C with G .times. Y was not considered substantial. We feel that the usual analysis (Analysis I) would be adequate for calculating heritability.