Gypsy moth is a non-native forest defoliator that is spreading 10 km/yr in western Virginia while the range is static or retracting in eastern Virginia. Populations in eastern Virginia may be limited by the warming climate. Alternatively, predation on pupae is a major mortality factor in gypsy moth, thus, may affect expansion rate at the invasion front. Here we ask whether pupal mortality factors vary across Virginia. We examined three mortality catagories: small predators, large predators and other mortality. Six study sites were established east-west across Virginia spanning the coastal plain, Piedmont, and the Appalachian Mountains. Three plots were established at each site with ten trees per plot. Gypsy moth pupae were distributed by and on trees weekly in two treatments: caged-exposed and ground-tree. The caged treatment measured effects of small predators (e.g., invertebrates). Mortality by large predators, (e.g., rodents and birds) was estimated using a formula that discounts small predator effects on exposed pupae. Pupae were distributed for 4-5 weeks in the spring/summer, collected after 7 days, and scored for predation. Intact pupae were returned to the lab and scored for other mortality. Data were analyzed with mixed-effects models. Male moths were trapped weekly to measure site-specific developmental phenology.
Results/Conclusions
Predation of exposed pupae was 200-500% greater than caged pupae. Predation was 5-30% higher on the ground than in trees. Mortality probability by small and large predators ranged from 0.15-0.40 and 0.83-1.0, respectively. Predation rates generally increased across the study period then decreased in the last 1-2 sampling periods across site. No geographic differences were found in time-specific predation rates across Julian days. Male moth trapping coupled with phenological model predictions estimated that the pupal period in the coastal plain was 7 and 16 days earlier than the Piedmont and mountains, respectively. After accounting for these phenological differences, we found that mortality due to small predators was greater in the mountain region than in the Piedmont (p < 0.05) and coastal plain (p < 0.05), counter to our expectation. There was no geographic difference in the effect of large predators (p > 0.05). Other mortality was marginally lower in the mountains (0.43-0.58) than in the coastal plain (0.58-0.61) and Piedmont (0.68-0.72), although the plot-specific variation was high. These data provide weak evidence that suggests climatic differences could be a causal factor of geographic variation in gypsy moth spread across Virginia. No evidence supported the pupal predation hypothesis.