Performance of a ryegrass-white clover sward subjected to a range of nitrogen treatments was recorded over three years. Herbage was removed either by periodic mob-stocking with sheep or cutting with a shearing handpiece. Nitrogen was applied as lime-ammonium nitrate in split dressings, over the period late autumn-winter-spring-early summer. A cloverless sward receiving no fertiliser nitrogen was included to gain information on the nitrogen status of the experimental site. An unbalanced design was used, with 7 nitrogen treatments represented under grazing but only 4 under cutting. The performance of pastures subjected to relatively frequent, close defoliation by mob-stocked sheep was reasonably reproduced in cut swards. The cutting technique allowed better simulation of clover herbage removal by sheep than is possible with mowing, and it was not until the third year that a difference was recorded in yearly clover dry herbage yield between grazed and cut swards. Both the cutting technique and avoidance of nutrient depletion (especially of potassium) in cut treatments are considered important in determining the pattern of results. Under these conditions, grazing increased total dry herbage (herbage nitrogen) yield each year by from 13 to 24% (18 to 22%). Annual nitrogen turnover in urine and dung averaged approximately 650 kg N/ha, with apparent recovery in herbage no more than 20% on average over the three years. Annual total dry herbage (herbage nitrogen) yield from this ryegrass-white clover association was substantial, averaging 15.6 t DM/ha (560 kg N/ha) each year without fertiliser nitrogen. The heaviest fertiliser input, 448 kg N/ha/an, increased total dry herbage (herbage nitrogen) yield each year by from 20 to 25% (30 to 43%). Quite clearly, nitrogen availability limited total herbage production from this well managed grass-clover association. On average over the three years, total herbage responses were approximately linear, and apparent recovery of fertiliser nitrogen averaged 44%. This figure would be higher if it were possible to correct for enhanced soil nitrogen uptake by clover in the nitrogen-treated swards. N2(C2H2) fixation under grazing totalled 263, 165 and 53 kg N/ha/an in grass-clover pastures receiving 0, 112 and 448 kg N/ha/an, respectively. Fertiliser nitrogen reduced symbiotic fixation through the combined influence of clover suppression and reduced fixation efficiency by clover in the nitrogen-treated swards. Main treatment (defoliation method x nitrogen treatment) interactions were largely absent from annual production data. Comparison of main treatment effects indicated that sheep-grazing and fertiliser nitrogen exerted similar but certainly not the same effects. Nitrogen removal in animal produce was estimated to have averaged about 20 kg N/ha/an, while average removal in herbage from cut grass-clover associations was almost 600 kg N/ha/an. Yet the contrast between grazing and cutting in the several parameters of nitrogen relationships measured in the soil-plant complex was unspectacular. The soil total nitrogen pool throughout the sampled profile declined over the course of the experiment. The average annual reduction under grazing (60 kg N/ha-45.7 cm) was less than under cutting (150 kg N/ha-45.7cm). Fairly complete utilisation of herbage, coupled with a close C/N ratio (c 10) for soil at the site, are considered to provide the most likely explanation for the observed pattern of change. No influence of nitrogen treatments was apparent, indicating that factors regulating soil carbon accumulation are inseparable from the size of the annual input of nitrogen, in determining nitrogen accumulation within a developed grass-clover system. It is concluded from nitrogen balances that unaccounted-for nitrogen outgoings of several hundred kg N/ha occurred each year from the grazed grass-clover systems. Results from this and associated research indicate that grazing animals cause substantial, but hitherto largely unrecognised nitrogen losses from developed grass-clover ecosystems. Most of the loss stems from aggregation of excess dietary nitrogen into urine patches.