Yield response and isotopic N-fertiliser recovery of rainfed wheat were assessed as influenced by fertiliser rate and timing. A popular bread wheat cultivar, Seri 82, was planted in a 4-year experiment from 1994/1995 to 1997/1998. Urea fertiliser was applied at rates of 0-240 N ha(-1) in two split applications. Fertiliser-N recovery and residual N remaining in the soil after wheat harvest were measured using (15)N-labelled fertilisers. The highest wheat grain yield ranged from 4.9 to 6.9 t ha(-1) with 240 kg N ha(-1) fertiliser. The 4-year results showed that wheat benefited least from the fertiliser applied near planting. N-fertiliser recovery was higher from fertiliser applied during tillering compared with application at emergence. The results suggest that applying one-third or less of the total N at planting and applying the remainder at tillering can minimise leaching risks, Another benefit of this strategy would bean overall increase in N-fertiliser recovery. Residual fertiliser-N left in soil after wheat harvest was proportional to N application rates and mainly confined to the upper 40 cm depth. (15)N recovery by wheat at maturity was 50-60%, indicating that 40-50% of fertiliser-N remained in the soil or was lost. Over 95% of total fertiliser application to wheat could be accounted for in the wheat crop or soil after harvest at the 240 kg N ha(-1) rate. The results, therefore, suggest that leaching losses of fertiliser-N below 90 cm were not likely during the growing season for rainfed wheat grown on these heavy-textured soils (Palexerollic Chromoxeret) of the Mediterranean region. (C) 2001 Elsevier Science B.V. All rights reserved.