Abstract
During 1977, a detailed study was undertaken to develop efficient sampling procedures for estimating egg and larval populations of the alfalfa blotch leafminer, Agromyza frontella (Rond.), in Vernal alfalfa. Distributions of the stages within and between stems were contagious and stem totals tended to follow negative binomial distributions. Within stems, the density of eggs and larval mines increased exponentially from base to tip. Taking these patterns and the costs of sampling into account, and using the relevant components of variance, four leaves taken systematically from a single stem is proposed as the appropriate sample unit for both eggs and mines. On this basis, 200 stems per field will give adequate precision (less than 10% of the mean with confidence probability 90%) for population estimates. However, the propensity of the leafminer to prefer the middle leaflet to the basal leaflets in the approximate ratio 3:4:3 suggests that sampling costs may be further reduced by restricting counts to the middle leaflet and raising them by a factor of 2.5.