SAMPLE-SIZE FOR ESTIMATING THE WITHIN PLOT VARIANCE IN EXPERIMENTS DESIGNED TO EVALUATE COMMON BEAN PROGENIES

Ninety-three segregant progenies, of the F-3, F-4 aid F-5 generations, from the cross between common bean cultivars Jalo and Small White, we re assessed during tile dry season of 1991. Tile purpose was to calcul ate the within plot variability and to determine the minimum plot samp le size necessary to quantify this variation. For these experiments fi ve plants from each plot were harvested individually and the mean and variance of yield, number of pods and number of seeds per plant were c alculated. Different sample sizes were simulated, varying from 5% to 5 0% of die number of plots. Ten samples of the same size were used for each simulation. By comparing the estimates obtained per sample with t hose of the whole population, it was shown that the samples were repre sentative of the population for all characteristics and generations ev en when the smallest size, 5% of the plots, was used.