Increasing concentrations of atmospheric CO₂ will interact with other environmental factors to influence the physiology and ecology of trees. This research evaluated how plant phytochemical responses to enriched atmospheric CO₂ are affected by the availability of soil nitrate (NO₃⁻) and how these chemical changes, in turn, alter the performance of a tree‐feeding folivore. Seedlings of three deciduous tree species—quaking aspen (Populus tremuloides), red oak (Quercus rubra), and sugar maple (Acer saccharum)—were grown in ambient (355 μL/L) or elevated (650 μL/L) CO₂ in combination with low (1.25 mmol/L) or high (7.5 mmol/L) soil NO₃⁻ availability. After 60 d, foliage was analyzed for changes in nutrients and allelochemicals likely to be influenced by the availability of CO₂ and NO₃⁻. Penultimate gypsy moth larvae (Lymantria dispar) were reared on foliage (aspen and maple) to determine how performance would be affected by host chemical changes.

Using the framework of carbon–nutrient balance (CNB) theory, we tested three hypotheses regarding the impact of CO₂ and NO₃⁻ availability on plant chemistry and insect performance: (1) nitrogen‐based compounds will decrease, and carbon‐based compounds will increase in response to elevated CO₂ and/or low NO₃⁻; (2) aspen will exhibit the greatest change in C:N ratios, and maple the least; and (3) phytochemical changes will influence gypsy moth performance, with larvae fed aspen being affected more than those fed maple.

Concentrations of nitrogen and soluble protein decreased, whereas concentrations of starch, condensed tannins, and ellagitannins increased, in response to elevated CO₂ and/or low NO₃⁻. Responses of simple carbohydrates and phenolic glycosides were variable, however, suggesting that foliar accumulations of “dynamic metabolites” do not follow the predictions of CNB theory as well as do those of stable end products. With respect to Hypothesis 2, we found that absolute (net) changes in foliar C:N ratios were greatest for aspen and least for oak, whereas relative (proportional) changes were greatest for maple and least for aspen. Thus, Hypothesis 2 was only partially supported by the data. Considering Hypothesis 3, we found that elevated CO₂ treatments had little effect on gypsy moth development time, growth rate, or larval mass. Larvae reared on aspen foliage grown under elevated CO₂ exhibited increased consumption but decreased conversion efficiencies. Gypsy moth responses to NO₃⁻ were strongly host specific: the highest consumption and food digestibility occurred in larvae on high‐NO₃⁻ aspen, whereas the fastest growth rates occurred in larvae on high‐NO₃⁻ maple. In short, our results again only partially supported the predicted pattern. They indicate, however, that the magnitude of insect response elicited by resource‐mediated shifts in host chemistry will depend on how levels of compounds with specific importance to insect fitness (e.g., phenolic glycosides in aspen) are affected.

Overall, we observed relatively few true interactions (i.e., nonadditive) between carbon and nitrogen availability vis à vis foliar chemistry and insect performance. Tree species, however, frequently interacted with CO₂ and/or NO₃⁻ availability to affect both sets of parameters. These results suggest that the effects of elevated atmospheric CO₂ on terrestrial plant communities will not be homogeneous, but will depend on species composition and soil nutrient availability.