An experiment was conducted during the wet seasons of 1989 and 1990 to study the performance of rice (Oryza sativa L)-based intercropping systems in rainfed uplands of Chhotanagpur plateau. 'Kalinga 3' upland rice intercropped with 'BR 65' pigeonpea [Cajanus cajan (L.) Millsp.] in 4: 1 row ratio proved most advantageous. with average yield of 2.06 tonnes/ha of rice and 0.6 tonne/ha of pigeonpea, and a net return of Rs 4 020/ha compared with total variable costs and return of 2.72/Re invested. Rice + 'A 404' finger millet [Eleusine coracana (L.) Gaertn.] intercrop system with 2: 1 row ratio was the second best, with net return of Rs 2 754/ha and benefit of 2.32/Re invested; giving 1.72 tonnes/ha grain yield of rice and 1.01 tonnes/ha grain yield of finger millet. Another ideal intercrop system for red sandy soils of uplands was rice + 'GG 2' groundnut (Arachis hypogaea L), with 2: 1 row ratio, giving 1.7 tonnes/ha rice and 0.37 tonne/ha groundnut, with a net return of Rs 2 462/ha and benefit of 1.97/Re invested. Intercropping with 'T 9' blackgram (Phaseolus mungo L.), 'Sunaina' greengram (Phaseolus radiatus L.) and 'PK 327' soybean [Glycine max (L.) Merr.] proved uneconomical.