Sample sizes were determined for three adult corn rootworm sampling techniques. Sources of variation in the sampling plan used were analyzed with a nested analysis of variance. Plant-to-plant, ear-to-ear, and trap-to-trap variation accounted for the greatest percentage a variability, on a within-field basis, in their respective analyses. Optimum allocations of sample units were two plants per site, five ears per site, and one trap per site. Comparisons of precision and cost of the three methods revealed that plant counts provide the most precise estimate of beetle populations for the least cost.