In plant breeding programs, the knowledge about the appropriate sample size for the evaluation of populations is very important. A small sample reduces the chance of selecting superior genotypes, whereas a very large sample may lead to unnecessary increases in cost and labor. A population consisting of 192 soybean lines was divided in groups of 24 lines, which were assessed for grain yield in eight randomized complete block experiments. Analyses of variance were performed for each experiment as well as for groups of experiments, resulting in analyses of variance consisting of 24, 48, 72, 96, 120, 144, 168, and 192 lines. As the sample size increased, the width of confidence intervals of parameter estimates decreased, stabilizing with samples of 144 lines. Therefore, an appropriate sample size for the evaluation of soybean inbred populations should contain about 150 lines.
Genetic variance; heritability; parameter estimation accuracy; confidence interval
