AbstractTwenty varieties of pigeonpeas were evaluated at two locations for a 3-year period. Analysis of these data provided estimates on the nature and magnitude of the variance components for yield, date of flowering, plant height, and seed weight. The first- and second-order interactions for all characters studied were not as large as the variety component of variance. The variety x location x year interaction, although statistically significant, were of small magnitude, and equal to or smaller than the variety x year interaction. There are clear indications of much greater interaction of varieties x years in this study, which suggests that the number of years should not be less than three when testing for these characters. The variety x location interaction was of much smaller magnitude than the variety x year interaction, and statistically non-significant, except for date of flowering. The significance for date of flowering was explained on the basis of differences in date of planting at the two locations rather than to the effect of locations per se. The experimental data presented in this study reveal the necessity of testing for at least 3 years at one location.
Download data is not yet available.