I examined how elevated carbon dioxide (CO2) concentration alters carbon (C) and nitrogen (N) cycling in two annual grasslands, a sandstone grassland dominated by invasive mediterranean grasses, and a serpentine grassland dominated by native grasses and forbs. In both grasslands, elevated CO2 stimulated N mineralization and plant N uptake. Increased soil moisture in elevated CO2 caused by decreased plant transpiration best explained these changes, in contrast to other studies that have concluded that increased C input to soil is the mechanism altering N cycling under elevated CO2. I also examined differences among plant species in response to elevated CO2 in monocultures on serpentine soil. Elevated CO2 increased plant N acquisition for three invasive grasses, but decreased N acquisition for three native species; increasing CO2 may favor invasion of serpentine by mediterranean grasses. Elevated CO2 altered microbial N dynamics, but the direction and magnitude of changes were species-dependent, suggesting that feedbacks to plant growth through altered N cycling under elevated CO2 will depend on plant species composition. I grew sandstone grassland communities under factorial CO2 and nutrient enrichment for one growing season, allowed them to senesce over the summer, and then investigated changes in C and N cycling following simulated rain after the summer drought. Elevated CO2 increased C input to soil, causing increased heterotrophic N immobilization; under nutrient enrichment, these changes caused decreased nitrification and decreased NO efflux from soil, but did not alter N20 efflux. These results suggest that elevated CO2 could mitigate the effects of soil NO production on tropospheric ozone and acid rain, but will not affect soil N2O production and associated radiative forcing. Finally, I determined the effects of elevated CO2 on net ecosystem productivity after 4 growing seasons. Elevated CO2 increased total ecosystem C stocks in both serpentine and sandstone grasslands. We did not find changes in soil organic C, however, where long-term C storage must occur in these grassland ecosystems. Nevertheless, increased net ecosystem productivity in elevated CO2, if sustained, will eventually lead to increased soil organic C, suggesting that grassland ecosystems could act as a negative feedback to rising CO2.