A stratified sampling plan demonstrated that pattern of distribution of insects among nine regions of a bin (center and edge or midway in four compass directions) differed between bins of wheat. Within a bin, variation between two samples taken at a site was generally the largest, followed by variation between regions and then variation within a region between two sites 30 cm apart. The five most common species were most abundant in the center of the bin and another four species were most common along the bin wall. The negative binomial best described distribution of all species and variance increased more rapidly than the mean. Probability of detection increased with density, although progressively more slowly as density increased. Increase in number of samples with insects as a function of increasing density was apparently reduced by a disproportionate increase in number of insects per sample as density increased. Sampling effort required for detection or estimation of abundance was modeled as a function of insect density, sample-to-sample variation, and acceptable accuracy or confidence levels.