Some of the methodology used to infer the magnitude of bird predation on caterpillars in forest sites that vary in area (degree of fragmentation): bird exclosure on witch hazel (left), clay caterpillar on red maple (right) (photos by Michael S. Singer).
In collaboration with Drs. Robert Bagchi, David Wagner, and Christopher Elphick at U. Connecticut, I extended the tri-trophic work in Connecticut forests into the realm of landscape fragmentation. This NSF-funded project was the focus of my research from 2016 to 2020. As has been seen more generally with habitat specialist species, we observed that small forest patches in the landscape harbor fewer dietary specialist caterpillars relative to dietary generalist caterpillars (Mickley et al. 2025). Our project tested several alternative hypothetical mechanisms for this pattern. As discussed in our review paper on this topic (Bagchi et al. 2018), these mechanisms include neutral sampling effects on the plant community due to limited habitat area, disrupted top-down control of dietary generalist caterpillars by bird predation due to area effects and/or altered habitat quality for songbirds, and altered deer-browsing effects on the plant community. Other work has shown that intensive deer browsing can eliminate from the community plant species that host the highest caterpillar densities. It is also possible that intraspecific differences in host-plant quality for caterpillars somehow leads to greater fitness in large forest fragments versus small ones. We considered these mechanisms in testing the hypothesis that the fitness costs incurred by dietary specialist caterpillars are more severe than those incurred by dietary generalist caterpillars on woody host plants occurring in fragmented forest stands. Surprisingly, none of these previously hypothesized mechanisms explained the underrepresentation of dietary specialist caterpillar species in small forest fragments (Anderson et al. 2019, Mickley et al. 2025).