The richness, abundance, and biomass of phytophagous arthropods like lepidopteran larvae is highly uneven among sympatric tree taxa. Optimal foraging theory predicts that predation pressure will be greatest on foraging substrates that support the highest abundance and/or diversity of prey, thus offering the greatest reward and maximizing fitness. Predation pressure can also vary with the nutritional or energetic needs of predators across the annual cycle. For insectivorous birds, prioritizing foraging effort in trees that support the most insect prey can benefit individuals by improving their foraging efficiency, condition, and ultimately fitness. However, we lack an understanding of how trees vary in their support of bird foraging activity across seasons and among plant taxa. We used plasticine caterpillar models to measure avian predation rates on 9 native North American tree species that vary in caterpillar-hosting potential. We measured avian predation rates during May, June, and October to compare caterpillar mortality in seasons that vary in life-history needs, abundance, and diversity of avian predators. We modeled daily survivorship and total mortality using Cox-proportional hazard models and logistic regression. We found that, across seasons, caterpillars had significantly higher predation rates on trees that are predicted by literature host records to support the most species of caterpillars (β = 0.22 ± 0.05, 95% CI = [0.13,0.32], z = 4.73, P < 0.0001). Caterpillars had the highest mortality in June, coinciding with avian breeding seasons, and the lowest rates in October, coinciding with fall migration and dispersal. Our study suggests that birds disproportionately forage on trees that have the highest potential to support caterpillar richness and presumably prey biomass. The observed pattern of non-random foraging has many implications; for example, the utility of using informed tree selection to improve bird foraging in managed ecosystems or potential negative implications to bird populations of forest-composition shifts due to climate change. Applying this information to habitat restoration will enable land managers to better support avian populations by planting trees that best support foraging substrates for insectivorous birds in managed ecosystems.