ABSTRACT
Analyses of variance procedures are normally used to separate means in variety test performance data. One shortcoming of this procedure is that the interaction between a genotype and the environment is difficult to see. Yield and fiber quality of individual cotton varieties were regressed on test averages to more easily recognize how varieties perform across environments. A data set was constructed from the Louisiana Cotton Variety Test which included all varieties which had been produced at all locations in the test from 1991 to 1993. The genotype times environmental interaction was significant for all varieties for lint yield, micronaire, and fiber length, and all but one variety for fiber strength. Graphing the models provided ready comparisons of performance for individual varieties. Although not a new technique to crop science, using linear models depicting the genotype times environmental interaction appears to readily identify superior cotton varieties and may be useful to producers, consultants, and breeders.
|