Five graziiig experiments each lasting 2 or 3 years were made between 1955 and 1967, all starting in the first year of ryegrass/cocksfoot/clover or ry^rass/clover leys. A high and a low rate of N, 235 and 45 Ib/ac on average (263 and 51 kg/ha) were compared for beef production. Higb-and low-N treatments gave mean clover contents for the grazing season of 8 and 24 % on a dry-weight basis, respectively. High N consistently gave a smaller liveweigbt gain/animal than low N, on average 192 and 208 Ib/day (0 87 and 094 kg/day), respectively. Liveweigbt gain/ac was 20 % greater for hig^ N than for low N, and in terms of net energy tbe production from bigh-and low-N, respectively, was 18,500 and 15,000 MJ/ac (45,700 and 37,100 MJ/ha). Data from these experiments, together with published results, were used to calculate a regression of liveweight gain response on N rate and an equation was derived from this to express tbe output in terms of profit. At 1971 prices profit was maximal at £0 9/ac (f2/ha) with 112 Ib N/ac (125 kg N/ha); it was considerably greater at 1973 prices wben bigber rates of N were justified.