Although rising levels of atmospheric carbon dioxide are expected to directly affect forest ecosystems, little is known of how specific ecological interactions will be modified. This research evaluated the effects of enriched CO2 on the productivity and phytochemistry of forest trees and performance of associated insects. Our experimental system consisted of three tree species (quaking aspen [Populus tremuloides], red oak [Quercus rubra], sugar maple [Acer saccharum]) that span a range from fast to slow growing, and two species of leaf—feeding insects (gypsy moth [Lymantria dispar] and forest tent caterpillar [Malacosoma disstria]). Carbon—nutrient balance theory provided a framework for tests of three hypotheses; in response to enriched CO2: (1) relative increases in tree growth rates will be greatest for aspen and least for maple, (2) relative decreases in protein and increases in carbon—based compounds will be greatest for aspen and least for maple, and (3) relative reductions in performance will be greatest for insects fed aspen and least for insects fed maple. We grew 1—yr—old seedlings for 60 d under ambient (385 ± 5 μL/L) or elevated (642 ± 2 μL/L) CO2 regimes at the University of Wisconsin Biotron. After 50 d, we conducted feeding trials with penultimate—instar gypsy moth and forest tent caterpillars. After 60 d, a second set of trees was harvested and partitioned into root, stem, and leaf tissues. We subsequently analyzed leaf material for a variety of compounds known to affect performance of insect herbivores. In terms of actual dry—matter production, aspen responded the most to enriched CO2 atmopheres whereas maple responded the least. Proportional growth increases (relative to ambient plants), however were highest for oak and lest for maple. Effects of elevated CO2 on biomass allocation patterns differed among the three species; root—to—shoot ratios increased in aspen, decreased in oak, and did not change in maple. Enriched CO2 altered concentrations of primary and secondary metabolites in leaves, but the magnitude and direction of effects were species—specific. Aspen showed the largest change in storage carbon compounds (starch), whereas maple experienced the largest change in defensive carbon compounds (condensed and hydrolyzable tannins). Consumption rates of insects fed high—CO2 aspen increased dramatically, but growth rates declined. The two species of insects differed in response to oak and maple grown under enriched CO2. Gypsy moths grew better on high—CO2 oak, whereas forest tent caterpillars were unaffected; tent caterpillars tended to grow less on high—CO2 maple, whereas gypsy moths were unaffected. Changes in insect performance parameters were related to changes in foliar chemistry. Responses of plants and insects agreed with some, but not all, of the predictions of carbon—nutrient balance theory. This study illustrates that tree productivity and chemistry, and the performance of associated insects, will change under CO2 atmospheres predicted for the next century. Changes in higher level ecological processes, such as community structure and nutrient cycling, are also implicated.