In a field experiment conducted during the wet season of 1994 and 1995 on a deep black soil (Vertisol), the differential response of recently released rice (Oryza sativa L.) hybrids in terms of their yield components and yield to NH,-N and NH, + NO3-N sources [through urea and calcium ammonium nitrat, (CAN) respectively] and split application of N (as 3 and 4 equal splits) was studied. Four rice hyrbids, 2 of early (MGR 1 and KRH 1) and and the other 2 of mid-duration (APRH 1 and APRH 2) were tested using 2 checks (Rasi and Jaya). More stable and easily available NH,-N was found to be superior to unstable and leachable NO3-N fraction from CAN which resulted in higher number of tillers and panicles and was mainly responsible for higher grain yields (6.40 and 5.44 tonneslha with NH,-N and 5.73 and 4.59 tonneslha with NH, + NO,-N in1994 and 1995 respectively). Application of N in 4 splits, coinciding the last with flowering, improved filled grain percentage, 1,000- grain weight and finally resulted in higher grain yields owing to increased photosyn- thetic rate and delayed leaf senescence. Among the hybrids. MGR 1 in the short- duration group emerged as the most promising by outyielding the check Rasi to the extent of 18-20% more. Grain yield showed significant correlatidn during 2 years with tiller number (r = 0.48* and 0.68*), panicle number (r = 0.47* and 0.66**), filled grain percentage (r = 0.36 and 0.53**) and 1,000-grain weight (r = 0.48* and 0.39). The multiple linear regression analysis revealed significant contribution of 67% by tillers, panicles, spikelets, filled grain percentage and 1,000-grain weight towards grain yield.