Host Quality and Sawfly Populations: A New Approach to Life Table Analysis

Abstract

In the field, population density of a shoot—galling sawfly Euura lasiolepis was on eighth as great and mortality was five times as great in dry habitats as in wet habitats. The effect of host water stress on Euura populations was experimentally tested by initiating Euura populations on potted Salix lasiolepis willow plants given various amounts of water. As in wild populations Euura mortality was much higher (2.5 times) on waterstressed Salix. The life table analyses of experimental Euura cohorts began with the potential number of eggs in the adult females of the preceding generation. This enabled us to describe the effects of female reproductive responses to host quality on Euura densities. These responses, failure to initiate a gall or to release an egg during gall initiation, resulted in a 27.5% reduction of the Euura cohort on hosts given a high—water treatment and an 85.1% reduction on low—water hosts. Excluding behavioral limits on the potential number of eggs laid, first—instar larval death was the largest and most variable Euura mortality factor of the experimental populations. Host water treatment accounted for 51.9% of the variance in first—instar larval mortality (P ° .001), with 13.6% of the first instars dying on hosts given high—water treatments and 54.5% dying on the low—water hosts. The threefold decrease in gall initiation and egg release on water—stressed hosts has probably been selected for by the fourfold increase in first—star larval mortality on these water—stressed hosts. From an evolutionary perspective the reproductive behaviors have amplified the effects of the first—instar mortality. From an ecological perspective the reproductive responses of the adult females were the critical factors regulating Euurra densities. Understanding the effect of variation in host quality on herbivore survivorship and behavior is essential to understanding the densities of local herbivore populations and should be routinely incorporated into life table analyses.