Effect of sample size and structure on the bootstrap estimate of variance of heritability

The bootstrap estimates of heritability using half-sib analysis was obtained by drawing samples of size-200, 500 and 800 for different family structures to study their effect on the variance estimate. It was found that stable bootstrap estimate of standard error of heritability required about 200 bootstrap replications for low heritability except for small sample size 200 for which at least 400 bootstrap replications were required. The optimum family sizes and structure for low heritability was 10 sires and 50 progenies per sire. The optimum number of bootstrap replications required to get the stable estimate of standard error of moderate and high heritability was 200 for all sample sizes. The optimum sample size and structure for getting the accurate estimate of standard error of heritability were 25 sires and 20 progenies per sire and 40 sires and 10 progenies per sire for moderate and high heritability respectively.