Abstract Elevated CO 2 and defoliation effects on nitrogen (N) cycling in rangeland soils remain poorly understood. Here we tested whether effects of elevated CO 2 (720 μl L −1) and defoliation (clipping to 2.5 cm height) on N cycling depended on soil N availability (addition of 1 vs. 11 g N m −2) in intact mesocosms extracted from a semiarid grassland. Mesocosms were kept inside growth chambers for one growing season, and the experiment was repeated the next year. We added 15N (1 g m −2) to all mesocosms at the start of the growing season. We measured total N and 15N in plant, soil inorganic, microbial and soil organic pools at different times of the growing season. We combined the plant, soil inorganic, and microbial N pools into one pool (PIM-N pool) to separate biotic + inorganic from abiotic N residing in soil organic matter (SOM). With the 15N measurements we were then able to calculate transfer rates of N from the active PIM-N pool into SOM (soil N immobilization) and vice versa (soil N mobilization) throughout the growing season. We observed significant interactive effects of elevated CO 2 with N addition and defoliation with N addition on soil N mobilization and immobilization. However, no interactive effects were observed for net transfer rates. Net N transfer from the PIM-N pool into SOM increased under elevated CO 2, but was unaffected by defoliation. Elevated CO 2 and defoliation effects on the net transfer of N into SOM may not depend on soil N availability in semiarid grasslands, but may depend on the balance of root litter production affecting soil N immobilization and root exudation affecting soil N mobilization. We observed no interactive effects of elevated CO 2 with defoliation. We conclude that elevated CO 2, but not defoliation, may limit plant productivity in the long-term through increased soil N immobilization.