Thursday, August 5, 2010

PS 73-74: Evidence for non-selective feeding strategies by root-feeding insects in a longleaf pine-wiregrass system

Matthew J. Dittler and Robert H. Jones. Virginia Tech

Background/Question/Methods

The nutritional ecology of root-feeding insects is largely unknown, but may have implications for plant community dynamics and ecosystem function. We asked whether root-feeding insect larvae of family Scarabaeidae (white grubs) and families Elateridea and Cebrionidae (wireworms) foraged for a nutritionally optimal mixture of fine roots from C4 grass (higher C:N) and C3 non-grass species (lower C:N). Optimal foraging theory suggests a tradeoff between nutritional benefits of selective foraging and additional time spent searching for an ideal diet. Because the mobility of root-feeding insects is limited by soil, we hypothesized that benefits of selective foraging might be outweighed by search time, resulting in random feeding. To test this prediction we collected random samples (0-25 cm depth) of fine roots and root-feeding beetle larvae within randomly established plots in a longleaf pine-wiregrass system in southern Georgia. Grass dominance had been experimentally manipulated using presence or absence of prescribed fires, leading to higher and lower dominance, respectively (p=0.0002). Because C3 and C4 plants differ in discrimination against 13C, we were able to estimate the proportion of C4 grass in the diets of the herbivores and in the roots sampled from each plot.

Results/Conclusions

During the summer of 2009, two white grub and three wireworm taxa were collected. Carbon isotope ratios of all but one insect taxon (Maladera, p=0.043) did not differ significantly from the average isotope ratios of root samples from each burn treatment, satisfying our prediction for random feeding with respect to the proportion of grass consumed across the manipulated range of dominance.  Most taxa, including the dominant grub and wireworm (Phylophaga and Cebrionidae, respectively), tracked the δ13C values of roots from each plot from which individuals were collected, though Agriotes and the Meladera deviated slightly from predicted values, with proportion of grass slightly below predicted values (Maladera), or slightly above (Agriotes). We also observed a significantly higher median 15N enrichment of approximately 4‰ for wireworms, relative to 2‰ for white grubs (p<0.0001), suggesting that the wireworms are omnivorous. We conclude that the herbivores in our study system are nonselective generalists, or that selective foraging may be physically constrained by the soil environment.