A mixed sward of grasses and clovers, previously established with adequate Ca, N, P and K on deep acid peat, was used for a factorial experiment to determine the amounts of nutrients required to produce herbage for conservation. The treatments consisted of all combinations of 0, 117and 234 kg N, 0, 29 and 58 kg P and 0, 156 and 312 kg K/ha, and their effects on the botanical composition of the sward, the yield and mineral composition of the herbage, and the nutrient status of the peat were measured over a period of five years. Three cuts were taken annually, one third of the N dressings being applied for each cut.
Tall fescue became the dominant grass species while cocksfoot persisted much better than either ryegrass or timothy. The proportion of clover in the sward was reduced to practically zero by applying N, but without N it was maintained at about 20% by applying K.
Yields of dry matter were greatly increased both by N and by K and there was a marked positive interaction between these two nutrients. Due to the marked suppression of clover by applied N there was no response to the first increment of N in 1959, but with this exception, 1 kg N/ha with adequate P and K produced between 15 and 30 kg dry matter/ha. The effect of K increased from 1958 to 1962, the first increment consistently producing most of the response. Applying P had little effect on yield until the fifth year when a significant response was obtained.
The mineral composition of the herbage varied during the season; the percentages of N, Ca, Na and Mg were generally higher in autumn than in early summer whereas the percentage of K showed the opposite effect.
The N, P and K treatments had large effects on the percentages of N, P, K, Na, Ca and Mg in the herbage and on the amounts removed in the crop. Without applied N, but with adequate P and K, percentage of N varied according to the amount of clover in the sward. With applied N there was less variation and the average percentage of N was about 1·7 with 39 kg N and 2·0 with 78 kg N/ha/cut. Percentage of P ranged from 0·14 to 0·54 and percentage of K from 0·5 to 3·5. Apparent recoveries up to 74, 65 and 96 % were obtained from the N, P and K treatments, respectively.
For continuous production by cutting and removing the herbage, a mixed sward containing clovers appears to require, without applied N, about 120 kg K/ha/year. With applied N, an additional 0·3 to 0·5 kg K/kg N appears to be needed. The amount of P required will depend mainly on the concentration desired in the herbage and on the rate of N used. To produce herbage with 0·3 % P in the dry matter about 15 kg P/ha/year, plus approximately 0·1 kg P/kg N applied seems to be needed; for 0·4 %P the corresponding quantities are about 40 kg P/ha plus 0·3 kg P/kg N applied.