Factors affecting the grazing habits of ruminants are discussed in relation to the attractiveness and nutritive value of herbages and the observed preferences of cattle and sheep at different times.Methods of measuring stock preferences are considered and it is concluded that, while frequent eye estimates of the quantity of each species appear to provide an adequate method, continuous observations and herbage sampling are useful adjuncts to it.The results of five experiments are given. Four were carried out in 3 years' winter grazing and one on spring, summer and autumn grazing over 2 years. In the latter experiment herbage samples were chemically analysed and a relationship with the observed preferences was calculated.In the winter grazing trials the greener grasses tended to be selected first, i.e. timothy, white clover, rough-stalked meadow grass, perennial and Italian rye-grass. Cocksfoot, Phalaris, red fescue and meadow fescue were not generally relished. Meadow foxtail, smooth-stalked meadow grass, tall fescue and sometimes the different varieties of rye-grass and timothy varied in rank from year to year as a result of differential frost damage and fungal attacks.In the summer grazing trials lucerne and white clover were found palatable in all the 1956 trials; in 1957 lucerne remained so, but white clover was less well liked, especially in June and July. The amount of grass growing with the clover seemed to affect attractiveness, mixtures being preferred to either species by itself. Meadow fescue and timothy were always ranked high, perennial rye-grass and cocksfoot usually next. Agrostis and red fescue were lowest. For the grasses, preference rank was positively correlated in 1957 with water-soluble dry matter, water-soluble ash, water-soluble carbohydrate and negatively correlated with lignin content.Results are compared with those of other investigators and they show a marked degree of similarity. Variability in the data for sheep is possibly due to limitations in technique. The results for sheep and for cattle are compared; and it was generally found that cattle were fonder of meadow fescue than sheep, and less fond of cocksfoot.Stock appear to graze those plants which will most readily supply their requirements for salts and energy (carbohydrates). Factors like dung, fungal attack, accessibility, density and toughness may interfere with this relationship.

The treatments in this summer-grazing experiment were designed as a 3 x 3 factorial; 3 levels of nitrogenous fertilizer, supplying 0, 336 and 672 kg N/ha per annum, and 3 stocking rates. The fertilizer was applied in 6 to 8 dressings at intervals of 3 to 4 weeks. The animals used were fattening Hereford x Friesian steers. The stocking rates were calculated on the basis of data available on herbage production under N and irrigation treatments and varied according to the level of N input. The experiment was laid out in 6 randomized blocks, which were grazed in rotation. Blocks were removed from the grazing cycle in spring and early summer of each of the 3 years (1962-4) over which the experiment was conducted and cut for conservation; yields were recorded. Details are given of: grazing management; results of studies of soil-N levels; the quantities of herbage removed for conservation; the quantity, chemical and botanical composition of the herbage available for grazing; and herbage consumption.

In situations where yield is variable but in which it is not desirable to take too large a cut sample, e.g. on grazed plots, the pasture‐yield estimator can be used to improve the accuracy of yield estimation. An outline of the theory and examples of the use of the method are given.

In a 3 X 3 factorial experiment, the effects of three levels of nitrogenous fertilizer (0, 336 and 672 kg/N per ha) and three grazing intensities on live-weight gains and carcass attributes were measured. The live-weight gains of the animals (per head) over the season were greater at the low and moderate grazing intensities than at the high; this effect occurred mainly from July to October each year. Gains were low on the no-nitrogen swards from July 1962 to the end of the grazing season and throughout 1963: this was related to the very high clover content of the swards. In all years, as grazing intensities increased, the carcass weights per animal became lower. Carcass balance, carcass length, depth of longissimus dorsi muscle, fat thickness over the longissimus dorsi, weight of kidney suet and channel fat, and weight of the alimentary tract also reflected the effect of treatment on weight gains. The results are discussed in relation to the botanical composition and consumption of the herbage, presented in Part I.

The treatments in this summer-grazing experiment were designed as a 2 x 3 factorial, there being two levels of nitrogenous fertilizer (nitro-chalk), supplying 0 and 336 kg of N/ha per annum (NO and N3, respectively) and three stocking rates. The fertilizer was applied in 7 dressings at approximately three-weekly intervals. The animals used were Hereford x Friesian steers; these were slaughtered at the end of the grazing season. The stocking rates were calculated on the basis of expected herbage production. The experiment was laid out in 6 randomized blocks which were grazed in rotation. Two blocks were cut for conservation before grazing each year; up to the end of June only 4 blocks were used for grazing. There was little contamination of the swards with unsown species. With the NO swards the clover contents declined each year, but remained fairly high at over 20% in 1965. Clover contents also declined each year in the N3 swards and by 1965 were only 2 to 4%. The quantities of herbage cut for conservation were higher from the swards receiving N. The live-weight gains of the cattle (per head) were higher at the lower stocking rates, the effect of stocking rate being more noticeable in the July to Oct. period. Live-weight gains per ha increased at the higher stocking rates. In 1963 and 1965, as stocking rates increased carcass weights per animal decreased and carcass quality, as indicated by the commercial grades, was poorer. In 1964, a dry spell from July to Oct. necessitated the removal of cattle from the experiment and the effects of treatment on carcass weight and quality were not apparent.

1. From 1955 to 1957 a cattle grazing experiment was carried out on lucerne and cocksfoot pastures. Hereford-cross steers were used. Results obtained on the assessment of herbage intake either by the use of chromic oxide and faecal index methods or by herbage sampling are presented here.2. Except in 1955, when techniques were being developed, the digestibility of the herbage on offer to the bullocks was estimated by cold storing herbage, cut with an Allen Autoscythe and later feeding it to sheep in indoor digestion trials. The in vitro digestibility of herbage samples cut in 1955 and 1956 was also determined.3. In 1956 the mean intake of organic matter (lb.) per 100 lb. live weight by faecal methods was 2·38, 2·07, 2·20 and 2·29 on pastures of lucerne, cocksfoot, lucerne/cocksfoot (broadcast) and lucerne/cocksfoot (alternate 1 ft. drills), respectively, and 1·99, 1·76, 2·02 and 1·97, respectively, in 1957.4. In 1956 the mean intake of organic matter (lb.) per 100 lb. live weight by herbage sampling methods was 2·14, 1·70, 2·11 and 1·99 on the same pastures of lucerne, cocksfoot, lucerne/cocksfoot (broadcast) and lucerne/cocksfoot (drills), respectively, and 2·49, 1·32, 2·35 and 2·03, respectively, in 1957. The modified Tarpen trimmer, cutting to within 1 in. of the ground was used for sampling in 1956 and the Allen Autoscythe cutting to within 2 or 3 in. of ground level in 1957.5. Detailed botanical separation of herbage samples showed that the bullocks were selecting the more digestible parts of the plants; the top 4 in. of lucerne and the ends of leaves and stems of cocksfoot. This and other factors which could affect the estimates of digestibility of herbage eaten by the cattle are discussed together with possible sources of error in the herbage sampling methods.

Silage was made from wilted and unwilted grass in May 1965, 1966 and 1967. The silage was made in sealed silos, but these were not evacuated and additives were not used. The silages were fed as follows: in 1965 to milking dairy cows and 1‐year‐old beef steers; in 1966 to steers coming into the experiment at 3, 9 and 15 months of age; in 1967 to weaned steer calves, with or without rolled barley (1 or 2 kg/head per day). In all years the silages were of high in vivo digestibility. The intake of wilted silage was higher than that of unwilted silage. There was no significant difference between the milk yields of cows fed on wilted and unwilted silage only. Liveweight gains were higher with wilted silage in 1965 and 1966, but not in 1967. In 1966 gains increased with age of animal and in 1967 gains were higher at successively higher levels of concentrate feeding.

1. Perennial ryegrass (S. 23 and S. 24 sown together) and cocksfoot (S. 37 and S. 143 sown together) were compared in grazing and individual feeding experiments during the summer months of 1960, 1961 and 1962, using spring-born calves.2. Herbage samples indicated that more grass was available to the animals grazing on ryegrass than on cocksfoot in 1961 but no differences were found in 1962. The ad lib. method of feeding should have ensured that quantity did not affect the results in the feeding experiments.