Many insect herbivores feed in concealed locations and thus limit their susceptibility to foraging natural enemies. However, some herbivores move into and out of concealment many times over the course of their development, creating windows of greater vulnerability to attack. Parasitoids that specialize on such hosts face an interesting foraging decision: having located an inaccessible host, how long should the parasitoid wait for the host to become accessible? We explored this question with a simple foraging model, and found that parasitoids should take one of two strategies, depending on three parameters. Parasitoids should leave immediately upon discovering a host is inaccessible if 1) the probability of locating another host is high, 2) the proportion of hosts that are accessible at any point in time is high, and 3) the probability that an inaccessible host will emerge from hiding is low. Conversely, the parasitoid should wait for hosts to emerge from concealment when the opposite conditions hold. We then estimated these parameter values for the European corn borer, Ostrinia nubilalis, and its specialist parasitoid, Macrocentrus grandii. From these values, we predicted that the optimal strategy for M. grandii should vary with host density, with a leaving strategy favored by high host density, and a waiting strategy by low host density. These results suggest that individual parasitoid foraging experience and learning should play an important role in the foraging behavior expressed by M. grandii.