1. A small-scale plot experiment which had been carried out from 1947 to 1949 (Holmes, 1951) to study the effect of massive dressings of nitrogen, with and without phosphate and potash, on the yield of a ryegrass dominant sward was continued in 1950–2. A 4 × 4 Graeco-Latin square was used.

The nitrogen treatments applied each year were:

(1) no nitrogenous fertilizer, (2) 260 lb., (3) 520 lb. and (4) 416 lb. (312 lb. in 1951) nitrogen per acre per annum applied in four or five equal dressings, one for each cut. Treatments 1, 2 and 3 were cut each time they reached the long leafy stage (8–11 in. in height), treatment 4 was cut when 13–16 in. in height.

The mineral treatments were (A) no mineral fertilizer, (B) 336–538 lb. K2O per acre per annum depending on nitrogen treatment, (C) 120–180 lb. P2O5 per acre per annum, (D) treatments B and C combined. Mineral applications were applied in four or five dressings each year, one for each cut.

2. Applications of phosphate did not affect the yield or protein content of the herbage, but yields were severely restricted in the absence of potash. Where potash was applied the yields under each nitrogen treatment were maintained or increased over the 6-year period. Average yields of dry matter for the 6-year period when potash was present were 4760, 8050, 9620 and 9320 lb. per acre per annum for treatments 1, 2, 3 and 4. Without potash the corresponding average yields were 3980, 5610, 5190 and 5100 lb. Average crude protein yields with potash were 710, 1410, 1990 and 1640 lb. per acre per annum and without potash 550, 1090, 1190 and 1020 lb.

3. The presence of potash resulted in earlier growth in each season through the maintenance of the earlier vigorous grasses in the sward. Although the growth curve was variable with treatment 1, treatments 2, 3 and 4 gave nearly uniform distribution of herbage production over the season.

4. The weighted mean contents of crude protein for each year ranged from 13·9% for treatment 1 to 20·6% for treatment 3 when potash was given and from 12·9% for treatment 1 to 23·6% for treatment 3 when potash was absent. There was a gradual increase in protein content at the later cuts in each season, but the range was less where nitrogen was applied.

5. The efficiency of utilization of fertilizer nitrogen was calculated. When the yield was compared with that of a no-clover sward the average response was 15·6, 10·8 and 11·8 lb. dry matter per lb. of nitrogen applied for treatments 2, 3 and 4 respectively. In terms of crude protein the percentage recovery was 53, 44 and 42 respectively. When the yields were compared with those of the clovery swards the nitrogen recovery figures were reduced by about one-third.

6. The botanical composition of the plots was determined by the nitrogen and potash treatments. Where both were adequate a vigorous sward of ryegrass and timothy was maintained. Where nitrogen was absent but potash present a clovery sward developed. In the absence of potash with or without nitrogen the better grasses declined and were replaced by poor grasses.

7. Provided potash was applied there were no marked changes in the soil analysis.

8. The results are discussed with particular reference to the maintenance of high grass yields and the relative roles of clover and fertilizer nitrogen.