To understand the CO2 emission from soil-crop system as influenced by crop growth and tissue N content, pot and field experiments were carried out during 2001-02 wheat and rice growing seasons. Black chambers were used to take gas samples within a closed soil-crop system. The CO2 emission rate was detected by a gas chromatograph. Seasonal change of the CO2 emission was observed from the soil-crop system. Respiration from the soil-rice system was higher than that from the soil-wheat system. Dark respiration of the crop shoot was positively correlated to the shoot biomass. The respiration coefficient Rd, defined as the amount of CO2-C respired by per unit biomass C within one day under a reference temperature of 25 degrees C, can be well quantitatively expressed by shoot N content for either wheat or rice crop. Relationship between the Rd and the N content can be described as a linear regression of Rd = 0.0124N - 0.0076 (R2=0.9879, p<0.001) for the wheat crop and as a quadratic equation of Rd = 0.0085N2 - 0.0049N (R2=0.9776, p<0.001) for the rice crop, respectively. The crop roots promoted the soil respiration greatly, which increased by 178% for the wheat and 338% for the rice in comparison with the respiration from root-free soil. A further calculation of the root respiration, including root autotrophic respiration and rhizosphere respiration, suggested that the contribution of crop rhizosphere respiration to the total soil respiration was greater in the upland soil than that in the irrigated paddy soil.