Over time, the stimulative effect of elevated CO2 on the photosynthesis of rice crops is likely to be reduced with increasing duration of CO2 exposure, but the resultant effects on crop productivity remain unclear. To investigate seasonal changes in the effect of elevated CO2 on the growth of rice (Oryza sativa L.) crops, a free air CO2 enrichment (FACE) experiment was conducted at Shizukuishi, Iwate, Japan in 1998–2000. The target CO2 concentration of the FACE plots was 200 µmol mol−1 above that of ambient. Three levels of nitrogen (N) were supplied: low (LN, 4 g N m−2), medium [MN, 8 (1998) and 9 (1999, 2000) g N m−2] and high N (HN, 12 and 15 g N m−2). For MN and HN but not for LN, elevated CO2 increased tiller number at panicle initiation (PI) but this positive response decreased with crop development. As a result, the response of green leaf area index (GLAI) to elevated CO2 greatly varied with development, showing positive responses during vegetative stages and negative responses after PI. Elevated CO2 decreased leaf N concentration over the season, except during early stage of development. For MN crops, total biomass increased with elevated CO2, but the response declined linearly with development, with average increases of 32, 28, 21, 15 and 12% at tillering, PI, anthesis, mid-ripening and grain maturity, respectively. This decline is likely to be due to decreases in the positive effects of elevated CO2 on canopy photosynthesis because of reductions in both GLAI and leaf N. Up to PI, LN-crops tended to have a lower response to elevated CO2 than MN- and HN-crops, though by final harvest the total biomass response was similar for all N levels. For MN- and HN-crops, the positive response of grain yield (ca. 15%) to elevated CO2 was slightly greater than the response of final total biomass while for LN-crops it was less. We conclude that most of the seasonal changes in crop response to elevated CO2 are directly or indirectly associated with N uptake.