One of the primary challenges animals face is consuming enough nutrients of sufficient quality that they might realize their reproductive potential. Many strategies have evolved, but robust explanations for the causes and consequences of variation in diet breadth remain elusive. I’ve used field and comparative lab studies of a specialist parasitoid wasp (Pteromalus cassotis) with a closely-related generalist (Pteromalus puparum) across several potential butterfly host species to better understand how physiological and behavioral trade-offs might limit parasitoids' host ranges. When paired with eleven host species, both parasitoids preferred those hosts on which they performed best, but I found little support for the idea that physiological trade-offs drive the evolution of specialized foraging strategies. Intriguingly, the generalist was more likely than the specialist to accept unsuitable hosts, while the specialist was more likely than the generalist to reject suitable hosts. I also reared monarch butterflies (Danaus plexippus) to be more or less chemically defended and found that host toxicity influenced the behavior and performance of the specialist. The generalist often killed monarch hosts, but was incapable of developing in them. Finally, I investigated the role of behavior in avoiding natural enemy-induced mortality of insect pupae, which are relatively immobile and are often presumed to be vulnerable to natural enemies. I review many examples to show that pupae likely benefit from a variety of behaviors performed before pupation (by the larva or pre-pupa), behaviors of the pupa itself, and behaviors of conspecific and heterospecific individuals. Taken together, my dissertation provides the first details on the natural history and chemical ecology of interactions between P. cassotis and the monarch butterfly. My findings suggest that, in some cases, behavioral decisions may be more important than physiological trade-offs in shaping animal diet breadth. These findings have implications for organisms in rapidly changing environments and biological control efforts because they demonstrate that specialist foragers may retain the evolutionary potential for host switches, and generalists can cause difficult-to-measure non-target mortality.