1. An experiment is described which lasted for 5 years and in which a comparison was made of twelve grasses or grass mixtures under different fertilizer nitrogen treatments and also when grown with clover. The herbage was cut 4–6 times in each season. Adequate amounts of mineral fertilizers (280–340 lb. K2O and about 100 lb. P2O5 per acre per annum), and the following nitrogen treatments were applied: (1) no nitrogen, no clover, (2) grass sown with clover, (3) 140–208 lb. nitrogen per acre per annum in four to six equal dressings, (4) 350–416 lb. nitrogen per acre per annum in five and six equal dressings.
In 1951, 1952 and 1953 the clover dominant swards (treatment 2) were split between the following treatments; (X) as (3) above, (Y) 35 lb. nitrogen per acre in spring and again in late summer, (Z) no nitrogen as (2) above.
2. The average yields for the 4 years were 2180, 5940 and 8300 lb. dry matter per acre, and 290, 850 and 1460 lb. crude protein per acre for treatments 1, 3 and 4. With treatment 2 the average yields were 2830 lb. dry matter and 400 lb. crude protein in 1949 and 4270 lb. dry matter and 820 lb. crude protein in 1950. An approximate average yield for the 4 years from treatment 2 was 4630 lb. dry matter and 860 lb. crude protein. In 1951–3 average yields for treatments 2X, 2Y and 2Z were, 7240, 6340 and 5750 lb. dry matter and 1240, 1180 and 1100 lb. crude protein per acre.
3. There were considerable differences between grasses in nitrogen response and compatibility with clover. The highest yields with fertilizer nitrogen were given by cocksfoot strains, but, in the presence of clover, ryegrass and timothy strains gave the highest yields. There were also differences between strains within each species.
4. Mean crude protein contents were, for treatments 1, 3 and 4, 13·3, 14·3 and 17·6%, and for treatments 2X, 2Y and 2Z in 1951–3, 17·2, 18·6 and 19·1%. Differences between species were significant in only a few instances.
5. The distribution of yield over the season was most regular with treatment 4. Cocksfoot species gave the least variable yields from cut to cut, while those from timothy and ryegrass swards were the most variable.
6. Treatments 3 and 4 maintained a high proportion of sown grasses in the swards. In treatment 2 the clover percentage rose to a high level by 1950. A high percentage was maintained under treatment 2Z in 1951–3. Treatment 2Y depressed the clover content in some grasses, and treatment 2X further depressed it in those grasses. A fairly high clover content was maintained, however, even with treatment 2X with some timothy strains and meadow fescue.
7. The mineral fertilizers applied maintained the soil analysis at a satisfactory level.
8. The results are discussed with special reference to the relative merits of fertilizer nitrogen and clover nitrogen and to the differences between species and strains.