Abstract
Most investigations of host-parasitoid interactions use solitary insects and hymenopteran parasitoids; however, non-hymenopterans are a significant share of parasitoid biodiversity. Intriguing questions arise when diverging from these model systems. Non-hymenopteran parasitoids may utilize novel strategies to locate, attack, and occupy their hosts than their more widely studied counterparts. In addition, in solitary insect systems, fitness losses are often direct when parasitoid hosts are eventually killed. However, among social insects, fitness costs may be partially absorbed by sterile castes, and fitness costs are better measured colony wide. The aim of this work was to further investigate parasitism/parasitoid infestation by conopid flies in primitively eusocial native bumble bees, building most questions upon known differential immunocompetence among species of Bombus in Virginia, and exploring host-parasitoid interactions in a system where the host is both beneficial and experiencing declines.
These studies included an investigation of land cover as a predictor of parasitism, a comparison at caste-level of susceptibility to parasitism in B. bimaculatus, and an experiment to test the prevalence of “overnighting behavior” -- the tendency remain outside of the safety of the nest through the nighttime hours-- among several bumble bee species for the prevalence of. Finally, this work describes a means to conduct behavioral tests in the laboratory setting using conopid fly parasitoids and bumble bee hosts. These questions were approached using numerous techniques: passive-trapping bees, geospatial analysis, dissections of passively trapped bees, dissections of net-collected bees and conopid flies, the use of radio-frequency identification tags, number tags, and video capture to monitor bees individually, and preliminary tests of behavioral experiments.
Based on these numerous efforts, this work identifies 1) high variability in peak parasitism rates among a variety of landscapes; 2) that nearly a third of B. bimaculatus queens are parasitized before they can overwinter and do not show evidence of having particular immunocompetence against this attack; 3) that overnighting is less common in a species (B. griseocollis) that mounts a stronger immune response and higher in B. impatiens relative to this species and that two other species (B. perplexus & B. fervidus) engage in this behavior as well, and 4) that conopid flies can be reliably induced to attack in captivity.
This work supplies information useful to understanding parasitism in Virginia, and more broadly, the final chapter opens up avenues of research related to the impacts of conopid flies on bumblebees by providing details about how successful conopid fly choice tests are achieved in the laboratory environment. Future work could investigate the mechanisms of host manipulation in these flies, and these methods may prove useful for other systems.
