A field experiment was conducted over six years between 1990-1991 and 1995-1996 at R.S. Pura (32.5°N, 74.8°E, 308 m above mean sea level) and Kanpur (25.26°N, 80.34°E, 180 m above mean sea level) to work out fertilizer N application practices when crop residues were recycled in a rice-wheat cropping system. From the recommended levels of N (i.e. 100 kg/ha at R.S. Pura and 120 kg/ha at Kanpur for rice, and 120 kg/ha at both the locations for wheat), when 20 kg of the nitrogen was applied at the time of residue incorporation and the remainder during the crop growth period, yields of the crops were significantly lower than those obtained under other N scheduling practices tried in the experiment. However, when an extra 20 kg/ha nitrogen was applied at the time of residue incorporation over and above the recommended N doses at R.S. Pura, and N levels were enhanced by 20 kg/ha over the recommended rates at Kanpur, the crops yield was the highest. On the average of six crop cycles, these practices produced an additional 150 and 510 kg grains/ha at R.S. Pura, and 570 and 810 kg grains/ha at Kanpur in rice and wheat crops, respectively. Every rupee invested in the purchase of additional nitrogen for application in the rice-wheat system thus earned rupees 8.58 at R.S. Pura and rupees 17.76 at Kanpur. The time-trend equations indicated a gradual increase in the productivity of rice-wheat system at Kanpur, while at R.S. Pura yields started to decline after the third year of the experiment. These trends were associated with the organic-C content of the soil.