Abstract
Results are presented for the final two years of a previously described experiment.The return of dung and urine by sheep to a ryegrass/white‐clover ley was controlled by suitable harnesses to give four treatments (no dung or urine, dung, urine, dung and urine) which were combined in a replicated factorial design with four levels of nitrogenous fertilizer application (0, 52, 182, 312 1b. N per acre).Applied nitrogen and urine were the dominant factors affecting botanical composition. The percentage of ryegrass increased and that of clover decreased with the progressive increases in nitrogen application. Volunteer species (mainly Poa spp.) contributed up to 20 per cent by the final year, the maximum occurring under the medium‐high nitrogen treatment.Urine restricted the incursion of weed grasses.Combined with urine or the full return of excreta, high levels of applied nitrogen increased herbage production by up to 120 per cent. There was little response to dung except at the highest nitrogen level.The yield response to applied nitrogen was almost linear. In the absence of animal returns response was poor, partly due to shortage of potash. When both excreta were withheld the light nitrogen dressing depressed the annual production compared with the control; where both excreta were returned together with this dressing no reduction occurred in annual yield and the spring yield was improved (p <0.05).