Research on host plant use in insect herbivores has largely focused on developing our understanding of the chemical and nutritional basis of dietary specialization. Both theory and empirical studies have emphasized the importance of bottom-up factors as determinants of herbivore host plant preference and use, but more recent studies have highlighted the important role that natural enemeis can play in shaping feeding niches. Using two species of highly polyphagous forest caterpillars, Acharia stimulea and Euclea delphinii (Lepidoptera: Limacodidae), we experimentally quantified herbivore performance on six co-occurring tree species in both the presence and absence of natural enemies. In each of two years, 150 caterpillars (5 caterpillar/tree x 5 trees per tree species x 6 tree species) were placed in sleeve cages on each host plants in the field and we measured survival, development time, and pupal mass. In a similar manner we assessed survival and parasitism of sentinel larvae on each host plant in a series of short-term (1-2 wk) expsosure experiments. Exposed caterpillars were reared in the lab on foliage from their respective host plants and assessed for parasitism by fly and wasp parasitoids.  

Results/Conclusions

All measures of herbivore performance (survival, development time, relative growth rate, and pupal mass) varied significantly among host plants in the absence of enemies. Black cherry was the highest quality host plant and black gum and red oak were the lowest quality host plants and the rankings of host plant quality was consistent across years. Predation, as measured by disappearance of larvae in the field also varied significantly among host plants and was consistently high on red oak in both years. Across exposure experiments, parasitism levels varied among host plants in the majority (>60%) of field trials but the rank order of host plants varied across the season and among years. In addition, parasitism by Hymenopterans predominated in early instars and was replaced by parasitism by tachinids in later instars. Because bottom-up fitness effects were consistent and top-down pressure was highly variable (but routinely high), the net effects of host plant on herbivore fitness for both caterpillar species favors the use of different species during different parts of the season and life cycle of the caterpillars. We conclude that unpredictabiliy in mortality from the third trophic level is consitent with a bet-hedging strategy maintaining broad diets in these forest caterpillars.