Efforts to control the invasive gypsy moth in North America have been ongoing since its accidental release in 1869. Introductions of non-native natural enemies of gypsy moth have been largely ineffective, but several parasitoids have acquired native non-target insects as hosts. One species, the tachinid fly, Compsilura concinnata, has been implicated in the precipitous decline of giant silk moths (Saturniidae) in New England. However, large regional disparities in the severity of silk moth decline, despite the universal presence of the fly, suggests a more complex relationship. We hypothesized that forests with higher oak basal area, the preferred gypsy moth host species, may exacerbate parasitism of silk moths by Compsilura. In 2017, at 16 sites distributed within 4300 km2 of forest in Adirondack Park, New York, we recorded oak basal area, elevation, and degree of forest connectivity. At each site, relative abundance of gypsy moth and forest tent caterpillar were estimated and hourly temperatures were recorded on data-loggers. Thirty captive-reared Antheraea polyphemus silk moth larvae were deployed for 6 days on understory saplings at each site, re-collected, and reared in the laboratory to assess parasitism. We used a LASSO regression model to assess which site variables best predicted Compsilura parasitism.
Results/Conclusions
Recovery of sentinel larvae averaged 52%, with at least 10 caterpillars recovered from each site. Parasitism was highly variable, ranging from 0-74% (mean ± SE: 17.7% ± 5.4%). More interior forest sites (F1,13=26.1, p<0.0001) with fewer accumulated degree days (F1,13=10.1, p<0.007) had greater incidence of parasitism by Compsilura, regardless of oak basal area. These results suggest that Compsilura parasitism is dependent more on local climatic conditions than biotic variables such as forest composition and defoliating caterpillar abundance. Synchrony between the deployment of sentinel larvae and Compsilura phenology was tested using PRISM-derived degree day accumulation estimates for each site. Overlap between predicted Compsilura flight and sentinel caterpillar exposure captured both parasitism presence/absence (F1,14=13.4, p<0.003) and the proportion parasitized (F1,14=33.24, p<0.0001). The significant effects of temperature in predicting Compsilura parasitism suggests the potential for complex multi-trophic thermally-responsive interactions among host Lepidoptera, trees, and the parasitoid itself. While differences in forest type did not predict Compsilura parasitism in early summer, forest effects on later season interactions between Compsilura and lepidopteran communities are unknown. Compsilura has multiple generations and relies on different suites of host species throughout the growing season, which complicates understanding of the factors driving its impact on mid-season silk moth populations.