The Spatial Distribution of Two Pine Sawflies and Methods of Sampling for the Study of Population Dynamics

Abstract

The spatial distribution of egg clusters and cocoons of Neodiprion swainei and N. sertifer is described, and methods are developed for estimating the density of different stages on the same numerical basis.Egg-cluster density varies directly with height in the tree, branch length, and tree size, evidently due to the females’ preference for illuminated sites. The overdispersion of egg-cluster counts reflects the heterogeneity of distribution of suitable oviposition sites rather than an inherent aggregative behaviour of parent females. Cocoon density within stands is not strongly related to the distance from the nearest tree, due to the wide dispersal of cocoon-spinning larvae. Some of the overdispersion of cocoon counts is due to differences in cocoon density between sections of a stand. Cocoons of males are more highly aggregated than those of females, and cocoons attacked by predators are more highly aggregated than ones not attacked.The sampling unit selected for each life-history stage is one that permits density estimates of desired precision at minimum cost. In some stands, egg density can be estimated at minimum cost by direct counts per quadrat, in most stands, however, the whole tree is the most economical unit, despite the fact that per-tree density must be converted to absolute density in order to express all stages on a common basis. Sampling error can be reduced by stratification according to tree size. Gradients in cocoon density are seldom steep enough to permit stratification, but the cost of sampling can be minimized by adjusting unit size. Optimal unit size varies inversely with cocoon density and degree of aggregation, and is affected by soil type and the method used to extract cocoons.The role of spatial distribution and sampling programs in studies of population dynamics is discussed.