Atmospheric CO 2 is rising rapidly, and options for slowing the CO 2 rise are politically charged as they largely require reductions in industrial CO 2 emissions for most developed countries. As forests cover some 43% of the Earth's surface, account for some 70% of terrestrial net primary production (NPP), and are being bartered for carbon mitigation, it is critically important that we continue to reduce the uncertainties about the impacts of elevated atmospheric CO 2 on forest tree growth, productivity, and forest ecosystem function. In this paper, I review knowledge gaps and research needs on the effects of elevated atmospheric CO 2 on forest above-and below-ground growth and productivity, carbon sequestration, nutrient cycling, water relations, wood quality, phenology, community dynamics and biodiversity, antioxidants and stress tolerance, interactions with air pollutants, heterotrophic interactions, and ecosystem functioning. Finally, I discuss research needs regarding modeling of the impacts of elevated atmospheric CO 2 on forests.Even though there has been a tremendous amount of research done with elevated CO 2 and forest trees, it remains difficult to predict future forest growth and productivity under elevated atmospheric CO 2 . Likewise, it is not easy to predict how forest ecosystem processes will respond to enriched CO 2 . The more we study the impacts of increasing CO 2 , the more we realize that tree and forest responses are yet largely uncertain due to differences in responsiveness by species, genotype, and functional group, and the complex interactions of elevated atmospheric CO 2 with soil fertility, drought, pests, and co-occurring atmospheric pollutants such as nitrogen deposition and O 3 . Furthermore, it is impossible to predict ecosystem-level responses based on short-term studies of young trees grown without interacting stresses and in small spaces without the element of competition. Long-term studies using free-air CO 2 enrichment (FACE) technologies or forest stands around natural CO 2 vents are needed to increase the knowledge base on forest ecosystem responses to elevated atmospheric CO 2 . In addition, new experimental protocols need to continue to be developed that will allow for mature trees to be examined in natural ecosystems. These studies should be closely linked to modeling efforts so that the inference capacity from these expensive and long-term studies can be maximized. D