Using superparasitism by a stem borer parasitoid to infer a host refuge

1. Macrocentrus cingulum Reinhard (Hymenoptera: Braconidae), a parasitoid o f the European corn borer, Ostrinia nubilalis, and the Asian corn borer, O. furnacalis (Lepidoptera: Pyralidae), has high fecundity but has been repor ted to parasitize a low proportion of host larvae. This was corroborated in field collections: in Hebei (China) and Delaware (U.S.A.), M. cingulum par asitized only 15 and 25%, respectively, of hosts collected. 2. Because M. cingulum females cannot oviposit through plant tissue, they m ust parasitize hosts either before they have bored into stalks or while the y are near entrance holes, so that at any one time, many Ostrinia larvae ma y be unavailable to M. cingulum. This refuge, together with fluctuations in abundance of foraging M. cingulum females, may explain why M. cingulum par asitizes relatively few Ostrinia larvae. 3. To test this hypothesis, levels of superparasitism were measured in the field. Low parasitism resulting from a refuge for host larvae should cause high rates of superparasitism in hosts outside the refuge. 4. Because M. cingulum is polyembryonic, the number of parasitoids per host does not indicate the level of superparasitism. Random amplified polymorph ic DNA markers were used to determine the number of different genotypes eme rging from each host. The resulting frequency distributions were fitted to those expected under random oviposition to estimate the proportion of Ostri nia larvae unavailable to M. cingulum. 5. In the samples from Hebei and Delaware, the level of superparasitism was much higher than expected by chance if all hosts were available. Fitting t he frequencies of genotypes per host to a Poisson distribution, the authors estimated that 74-82% and 69-74% of host larvae were unavailable to M. cin gulum in these collections, respectively. This means that M. cingulum paras itized 60-84% and 82-95% of available hosts in these collections, respectiv ely. These levels of parasitism contrast strongly with the 15-25% found whe n all hosts were assumed available for oviposition. 6. Genetic distances of M. cingulum within and between hosts did not differ , allowing rejection of the hypothesis that high levels of superparasitism resulted from a female laying several eggs in the same host. 7. The hypothesis that M. cingulum parasitizes few Ostrinia larvae because many larvae are in a refuge explains these data and previously published in formation better than other hypotheses that have been suggested.