Local variation in host plant quality affects spatial distribution of fall webworms (Hyphantria cunea)

Background/Question/Methods

Fall webworm (Hyphantria cunea) is a generalist moth species that is polyphagous over its entire range, but populations appear to have become more specialized within specific regions.  For instance, in Colorado fall webworm feeds on ~17 different host plants whereas populations farther east can be found feeding on dozens of hosts.  One host plant for fall webworm in Colorado is plains cottonwood (Populus deltoides).  Fall webworm larvae are not as likely to feed on plains cottonwood located on the canyons in the foothills of Colorado, as they are to feed on plains cottonwood located on the adjacent plains.  Furthermore, plains cottonwood is never as commonly used as a host as its congener, narrowleaf cottonwood (Populus angustifolia), which is a relatively high-quality host for fall webworm in Colorado.  We investigated this intra- and inter-specific variation in host use and tested whether fall webworm larval performance varied among cottonwood populations and species.  Using a split-brood design, we reared fall webworm larvae on plains cottonwood from the canyons, plains cottonwood from the plains, and narrowleaf cottonwood from canyons to determine the relative performance of larvae on these host plants.  We recorded development time and pupal mass as measures of fitness.

Results/Conclusions

We found significant differences in survival of fall webworm larvae on the three different host plants. Host plant species, dam, and sex of the pupa all had significant effects on pupal weight, but none of the interactions were significant.  Larvae reared on plains cottonwood from the plains achieved the highest pupal weights while pupal weights did not differ significantly between larvae reared on narrowleaf cottonwood or plains cottonwood from the canyons.  Host plant species, dam, and sex of the pupa again all had significant effects on larval development time and there was also a significant interaction between dam and host plant.  Development time was shortest on narrowleaf cottonwood, followed by plains cottonwood from the plains and then plains cottonwood from the canyons. Our two measures of larval performance, pupal weight and development time, both demonstrated that plains cottonwood from the plains is a superior host to plains cottonwood from the canyons for fall webworms in Colorado.  Our results may explain spatial distribution of fall webworms in Colorado.