Herbage from the first regrowth of perennial ryegrass based swards was direct-ensiled following treatment with either an inoculant of Lactobacillus plantarum (Ecosyl, Imperial Chemical Industries pic) at 3·0 1/t, formic acid (850 g/kg) at 2·9 1/t, or no additive (control). During harvesting, alternate loads of inoculant material were treated with an absorbent polymer (ammonium polyacrylamide) at the rate of 1 kg/t herbage and ensiled in separate 100-t capacity silos. The mean dry matter (DM) and water soluble carbohydrate concentrations of herbage used for the four treatments was 157 and 120 g/kg respectively. Lactic acid levels post ensiling increased more rapidly in inoculant-treated herbage than with the other treatments. Formic acid and inoculant-treated silages were well preserved whereas control and inoculant-plus-polymer silages were only moderately well preserved. Losses of DM during ensilage were greater with the formic acid treatment with DM recovery values of 0·78, 0·72, 0·76 and 0·73 for the control, formic acid, inoculant and inoculant-plus-polymer silages respectively. Treatment of herbage with an absorbent polymer prior to ensiling resulted in a proportional reduction in effluent volume of 0·2 whereas formic acid treatment increased effluent flow by 0·28. The silages were evaluated in a changeover design experiment with two periods each of 4 weeks duration, using 24 British Friesian dairy cows. Animals were housed in individual stalls and in addition to the treatment silages, received 5 kg/day of supplement containing 193 g crude protein per kg DM. Silage intakes were increased by proportionately 0·10, 0·14 and 0·05 respectively with the formic acid, inoculant and inoculant-plus-polymer treatments compared with the control. The increased silage intakes with the inoculant treatment were reflected in an increased milk yield of 1·1 kg milk per day whereas formic acid and inoculant-plus-polymer treatments had no significant effect, although formic acid treatment did result in a significant increase in milk fat concentration. There were no major differences between treatments in energy or nitrogen digestibility, when determined on a complete ration basis. In conclusion, a large milk yield response was obtained as a result of treatment of herbage with inoculant prior to ensiling and this resulted from increased silage and hence energy intake. Treatment with formic acid increased silage and energy intake but had no effect on milk energy output.

Two experiments were carried out in the springs of 1986 and 1987 to investigate the effect of energy concentration in complete diets on the performance of heifers and mature cows. In 1986, all animals were on trial for an 8-week indoor period. In 1987, 14 heifers were on trial for 8 weeks, four heifers were on trial for 7 weeks and six heifers were on trial for 4 weeks. Carry-over effects were compared while animals were on pasture. Two diets offered ad libitum were compared. Diet H contained 0·7 concentrates and 0·3 grass silage. Diet L contained 0·5 concentrates and 0·5 grass silage. In 1986, the concentrate portion of diet H had a higher metabolizable energy (ME) content than diet L whereas in 1987 the same concentrate mixture was used in both diets.

Increasing the energy concentration of the diet from 10·66 MJ ME per kg dry matter (DM) to 11·21 MJ ME per kg DM resulted in a significant proportional increase of 0·36 in DM intake for heifers and mature cows in 1986. The proportional increase in DM intake in 1987 for heifers was 0·17. DM intake (kg DM per day) for heifers and mature cows in 1986 and for heifers in 1987 for diets H and L was 11·8, 8·7, 15·0, 11·0, 14·9, 12·7 respectively. The energy concentration of the complete diet did have a significant effect on average milk yield during the treatment period in the 1986 trial for mature cows but not for heifers in either year. Milk yields were 19·1, 17·5, 26·3, 240, 19·8, 18·9 kg/day respectively for heifers and mature cows in 1986 and for heifers only in 1987. In vivo digestibility coefficients in 1986 for energy, DM, organic matter, modified acid-detergent fibre and protein for diet H (0·708, 0·702, 0·724, 0·678, 0·671) were significantly higher than those for diet L (0·677, 0·666, 0·694, 0·653, 0·625).

The total lactation yield for the heifers and mature cows was not significantly influenced by the energy concentration of the diet given in early lactation.

Selection of bulls for rate and efficiency of lean gain was studied in a herd of Hereford cattle. There were two selection lines, one selected for lean growth rate (LGR) from birth to 400 days and the other for lean food conversion ratio (LFCR) from 200 to 400 days of age, for a period of 8 years. A control line bred by frozen semen from foundation bulls was also maintained. Generation interval was about 2·4 years and average male selection differentials, per generation were 1·2 and — 1·1 phenotypic standard deviation units for LGR and LFCR respectively.

Genetic parameters and responses to selection were estimated from the deviation of the selected lines from a control line and by restricted maximum likelihood (REML) techniques on the same material. Realized heritabilities were 0·40 (s.e. 0·12) for LGR and 0·40 (s.e. 0·13) for LFCR using the control line. Corresponding estimates from REML were 0·42 (s.e. 0·10) and 0·37 (s.e. 0·14). The estimate of the genetic correlation between LGR and LFCR was about — 0·69 (s.e. 0·12) using REML.

The estimates of direct annual genetic change using deviations from the control were 3·6 (s.e. 1·3) g/day for LGR and — 0·14 (s.e. 0·07) kg food per kg lean gain for LFCR. Corrsponding estimates from REML were similar but more precisely estimated. The correlated responses for LFCR in the LGR line was higher than the direct response for LFCR.

Correlated responses in two lines of Hereford cattle selected for lean growth rate (LGR) from birth to 400 days of age and lean food conversion ratio (LFCR) from 200 to 400 days of age for a period of 8 years were evaluated. Correlated changes were estimated by two methods: deviation of selected lines from a control line and restricted maximum likelihood. Generally, estimates from the two methods were similar but tended to be more precise for the latter. Statistically significant correlated responses occurred in growth rate in the LGR line and in lean proportion and food conversion ratio in both selected lines. Selection for LGR was accompanied by increases in body weight at various ages in both male and female progeny. In the LFCR line there were little or no changes in body weight for male calves but some increases at certain ages for female progeny. There were no adverse correlated responses detected in reproductive traits such as calving difficulty and calving and pre-weaning mortality.

Three trials were conducted to evaluate the effects of treating soya-bean straw with various alkalis plus ensiling on digestibility of various nutrients and on nitrogen retention in ruminants. Soya-bean straw was treated with sodium hydroxide, calcium hydroxide or ammonium hydroxide and ensiled at 650 g/kg moisture. In trial 1, the effects of level and type of alkali on in vitro digestibility of ensiled soya-bean straw were evaluated. Concentrations of alkalis used were 0 (control), 20, 30, 40 and 50 g/kg dry matter. Both in vitro digestibilities of dry matter and organic matter of soya-bean straw were significantly (P < 0·05) increased with alkali treatment at all levels plus ensiling compared with untreated or treated unensiled straw. The 40 g/kg alkali concentration level in combination with ensiling appeared to be the most effective overall. In trial 2, the effects of type of alkali on nutrient digestibility and nitrogen retention in wether lambs were evaluated. Sixteen wether Suffolk lambs were given untreated or alkali-treated (40 g/kg) ensiled soya-bean straw. Treatments were: (1) water-treated straw; (2) NaOH-treated straw: (3) Ca(OH)1-treated straw; and (4) NH4OH-treated straw. Lambs given NH4OH-treated ensiled straw digested the dry matter, organic matter, and crude protein better than those given water-treated ensiled straw (P < 0·05). NH4OH improved nitrogen intake by proportionately 0·181 but not nitrogen retention. For trial 3, eight Angus steers were used. Chopped soya-bean straw was treated with NH4OH and ensiled in Silopress bags. Treatments were: (1) untreated dry straw; (2) water-treated straw; (3) 30 g/kg NH4OH-treated straw; and (4) 40 g/kg NH4OH-treated straw. The trial consisted of a 7-day collection period which was replicated once. Apparent digestibilities of dry matter, crude protein and various fibre constituents were increased by ensiling or ensiling and NH4 OH treatment combinations (P < 0·05). Alkali treatments in combination with ensiling appeared to improve digestibility of soya-bean straw, but ammoniation did not improve nitrogen utilization. Ensiling appeared to have a more significant effect than alkali treatments per se on the degradability of soya-bean straw.

A theoretical analysis of the potential benefits of marker-assisted selection (MAS) of candidate bulls prior to entry into a young sire progeny testing programme was carried out. It is assumed that quantitative trait loci (QTL) affecting milk production have been mapped with respect to known genetic markers, and MAS is based on evaluation of elite sires in order to identify marker alleles in coupling to favourable or unfavourable QTL alleles. Candidate bulls, descendants of the elite sire will then be selected, prior to conventional progeny testing, on the basis of the marker alleles derived from the elite-sire ancestor.

The analysis considers recombination between marker and QTL, the difficulty of tracing specific marker alleles from sire to progeny, and the expectation that MAS, in practice, will be implemented in the grandsons, rather than in the sons of elite sires. It is shown that MAS of candidate bulls, based on the use of a single diallelic marker in linkage to a QTL will have only a negligible effect on the rate of genetic progress. Increases of 15 to 20% in the rate of genetic gain, however, can be obtained by the use of single polyallelic markers, and increases of 20 to 30% can be obtained by utilizing haplotypes of diallelic or polyallelic markers.

Three experiments were conducted to determine the metabolizable energy (ME) concentration of dry maize gluten feed (DMGF) in starter (experiment 1), growing (experiment 2) or finishing (experiment 3) diets for steers. Seventy-two weanling Holstein-Friesian steer calves (initially 52 kg live weight); 32 Aberdeen Angus and 24 Shorthorn steer calves (initially 250 kg live weight) and 32 Aberdeen Angus, 24 Shorthorn and 68 crossbred steers (initially 367 kg live weight) were used in experiments 1, 2 and 3, respectively. In experiment 1, calves were given one of three total mixed rations in which energy and protein ingredients on a dry matter (DM) basis were (g/kg): (1) 590 ground maize grain (MG), 260 rolled oats (RO) and 150 soya-bean meal (SBM); (2) 210 DMGF, 400 MG, 260 RO and 130 SBM; or (3) 440 DMGF, 210 MG, 250 RO and 100 SBM. Daily live-weight gains (LWG) and food-to-gain (F/G) ratios were similar across dietary treatments (P > 0·05). Calves given the diet containing 440 g DMGF per kg had higher DM intakes (DMI) than those given no DMGF (P < 0·05). ME concentration of the diet without DMGF was greater than ME concentrations of diets containing DMGF (P < 0·05). In experiment 2, calves were given one of four diets containing, on a DM basis (g/kg): (1) 800 maize silage (MS) and 200 MG; (2) 800 MS and 200 DMGF; (3) 500 MS and 500 DMGF; or (4) 200 MS and 800 DMGF. LWG and DMI of calves given 500 or 800 g DMGF per kg diets were higher than those of calves given 0 or 200 g DMGF per kg diets (P < 0·05). F/G ratios tended to increase and dietary ME concentrations tended to decrease with increasing content of DMGF in diets. In experiment 3, steers were given one of four diets, which on a DM basis were (g/kg): (1) 850 MG and 150 MS; (2) 300 DMGF, 150 MS and 550 MG; (3) 500 DMGF and 500 MG; or (4) 750 DMGF and 250 MG. LWG was not altered when DMGF replaced MG only or MG and MS (diet 2 v. 1 or diets 3 and 4 v. 1, respectively (P > 0·05)) in the diet. Steers given the 300 or 750 g DMGF per kg diets had higher DMI and lower F/G ratios than those fed diets without DMGF (P < 0·05). ME concentration of the diet without DMGF was greater (P < 0·05) than the ME concentration of 300 or 750 g DMGF per kg diets. ME concentration of DMGF, calculated by regression, was proportionately about 0·9 of the value of MG in all experiments and averaged 11·46 MJ/kg DM.

Thirty-nine Hereford × Friesian autumn-calving cows (mean calving date 17 September) and their Charolais-cross calves were used in a 2 × 2 × 2 factorial experiment to assess the effects of feeding level in winter and sward height in summer on intake and performance. Mean estimated metabolizable energy intakes during winter (30 October to 6 May) were 63 (L) or 92 (H) MJ/day. Two sward heights, 4·8 (SI) or 7·6 (Tl) cm, were imposed from turn-out to weaning on 3 July, and a further two sward heights, 4·7 (S2) or 8·0 (T2) cm, from weaning to housing on 1 October. At turn-out the H cows were 17 kg heavier (P > 0·05) than the L cows and their calves were 30 kg heavier (P < 0·01). Cow herbage intakes were significantly higher on the Tl and T2 treatments compared with the SI and S2 treatments as were live-weight changes: 0·94 v. –0·01 kg/day (P < 0·001) for Tl and SI treatments and 0·96 v. –0·18 kg/day (P < 0·001) for T2 and S2 treatments respectively. Calves on the Tl treatment had higher live-weight gains than those on the SI treatment (1·40 v. 0·63 kg/day; P < 0·001). After weaning the SI cows had significantly higher herbage organic matter intakes than the Tl cows (P < 0·05) and had significantly higher live-weight gains (0·51 v. 0·28 kg/day; P < 0·05). Calving difficulty and calf birth weight were not affected by any of the treatments. The results demonstrate that both previous and current level of nutrition have an effect on cow performance. When cows are turned out to pasture at body condition scores of 2·25 to 2·5, the sward heights recommended for spring-calving cows are equally appropriate for autumn-calving cows.

The effect of body condition on hypothalamic and pituitary function in post-partum suckled beef cows was studied. Thirty-eight Blue-Grey cows were calved at a body condition score of either 2·35 (L) or 2·80 (H), fed to maintain live weight and body condition thereafter and ovariectomized at either 5 or 9 weeks post partum. The H cows had higher LH pulse frequencies before ovariectomy (0·36 v. 0·28 pulses per h; P < 0·05) and 15 days after ovariectomy (0·86 v. 0·72 pulses per h; P < 0·01), indicating a direct effect of body condition on GnRH release from the hypothalamus. Body condition had no effect on the ability of the pituitary to release LH in response to a physiological dose (2·5 u.g) of GnRH in entire or ovariectomized cows at 5 or 9 weeks post partum. The magnitude of the increase in circulating levels of LH following GnRH injection was greater at 9 weeks compared with 5 weeks post partum (2·23 v. 1·73 u.g/1; P < 0·001) in entire cows, but there was no difference between 5 and 9 weeks in ovariectomized cows. Following ovariectomy there was a 2·5-fold increase in LH pulse frequency within 15 days and LH pulse frequency approached 1 pulse per h in all treatments groups. It is concluded that both the hypothalamus and the pituitary are capable of functioning normally by 5 weeks post partum in suckled beef cows, that the pulsatile release of GnRH from the hypothalamus is inhibited, possibly due to a very high sensitivity to oestradiol, and that the sensitivity of the pituitary to the negative feedback effect of oestradiol may decrease with time post partum.

The effect of body condition at calving on ovarian follicle characteristics during post-partum anoestrus was studied in 38 suckling Blue-Grey cows with mean body-condition scores of 2·80 (s.e. 0·05) (H) or 2·35 (s.e. 0·05) (L) at calving and fed so that they maintained live weight and body condition during lactation. Cows of each level of body condition were ovariectomized at either 5 or 9 weeks after calving and follicles ≥3 mm in diameter were dissected out and incubated in a medium (Medium 199) for 2 h. Follicle incubates were assayed for progesterone, testosterone and oestradiol. L cows had fewer follicles ≥3 mm in diameter than H cows (24·1 v. 44·9; P < 0·05) at 5 weeks post partum due to a lower number of small (3 to 4 mm) follicles (12·9 v. 30·6; P < 0·05). At 9 weeks there was no effect of body condition on mean numbers of follicles (38·1 v. 40·4; P > 0·05). The proportions of healthy, intermediate and atretic follicles in all cows were 0·35, 0·20 and 0·45 respectively; these proportions were not affected by body condition or time of ovariectomy. Body condition did not affect the overall mean capacity of the cultured follicles to produce oestradiol or progesterone in vitro but testosterone production was higher in follicles from H than L cows (6406 v. 5206 pg/h; P < 0·05). However, at 9 weeks post partum a greater proportion of H than L cows had a large (>8 mm), healthy, highly oestrogenic follicle (4/8 v. 1/7; P = 0·18).

At 9 weeks post partum, the four large, healthy, oestrogen-active follicles (those secreting >250 pg/h) from H cows were approximately 11-fold more oestrogenic than the single comparable follicle from L cows (7100 v. 635 pg/h; P > 0·05). There was little difference at week 5 post partum: 3288 and 2785 pg/h (P > 0·05) for H and L cows respectively. Oestrogen-active follicles from H cows were significantly more oestrogenic at 9 than 5 weeks post partum (5807 v. 3021 pg/h; P < 0·05). The results suggest that the body condition-related reduction in the number and steroidogenic capacity of large oestrogenic follicles in post-partum cows may be associated with a reduction in LH pulsatility.

Five diets of identical metabolizable energy (ME) content were formulated, in which wheat silage, made from a variety bred for a high proportion of ears to roughage, was incorporated at the rate of 480 g/kg dry matter (DM). Their calculated ME density was 11-4 MJ/kg DM, the protein concentration of four of the five diets was 150 g/kg DM, and that of the fifth was 87 g/kg DM. The effectiveness of protein supplementation of such diets by non-protein nitrogen (NPN), fish meal (FM) or soya-bean and cottonseed oil meals (SBM and CSM) was studied in a digestibility and nitrogen balance trial, and in a feeding trial using Friesian and beef breed young bulls. One diet served as a negative control (NC). The apparent digestibility of organic matter (kg/kg DM) was 0·709, 0·708, 0·704, 0·680 and 0·667 (s.e. 0·0125) and nitrogen retention (g/kg M0·75 per day) was 0·20, 0·65, 0·65, 0·64 and 0·68 (s.e. 0·059; P < 0·01) for the NC, NPN, FM, CSM and SBM supplemented diets, respectively. The feeding trial was divided into two periods; the first 100 days, and from day 101 to slaughter. The NC groups of both breeds gained significantly less in both periods of the experiment than did the supplemented groups. There were no significant differences among the supplemented groups.

Dairy ewes (Prealpes breed) were subjected to differential photoperiod treatments: either long (15·5 light (L): 8·5 dark (D) or short (8·5 L: 15·5 D). Treatments started 15 days before lambing and continued until the 5th week of lactation. After lambing the 22 ewes of each photoperiodic treatment were divided in two groups and fed on either a high or low plane (19 v. 10·5 MJ/day and 370 v. 237 g crude protein per day). Milk yield was not affected by photoperiod but significantly modified by feeding levels (1154 v. 789 ml/day). Protein content of the milk was reduced under long days (—4·9 g/1) compared with short days, while no effect was observed for fat content. Body chemical composition changes, estimated by the deuterium oxide (heavy water) dilution technique between weeks 1 and 4, were affected by both treatments, with a significant interaction between photoperiod and diet for body protein changes: under short days lipid and protein changes were closely related while under long days they were independent.

Blood samples were taken on day 23 of lactation from six ewes per group, every 30 min from 07.30 to 15.30 h for growth hormone (GH) and prolactin (PRL) measurements. The timing of GH and PRL pulses were different, but there was evidence that secretory patterns of each hormone were synchronized across ewes within the four treatments. Long photoperiod increased plasma PRL baseline rather than pulse amplitude, with no effect of diet. Undernutrition mainly increased GH pulse amplitude rather than baseline concentration, with no significant effects of photoperiod.

High spontaneous concentrations of GH were associated with body lipid mobilization (R2 = 0·51) and were also involved in the control of protein changes (R2 = 0·71). These results indicate that GH (and PRL) are involved in nutrient partitioning in the lactating ewe.

Forty Menz rams weighing 20 to 22 kg and aged 13 to 18 months were individually offered 1·0 kg/day chopped maize stover and one of five supplements over a 70-day growth trial followed by a 10-day metabolism trial. The amounts of supplement offered were designed to supply 4·4 g nitrogen per day and with stover, sufficient nutrients to allow growth rates of 50 g/day. The control supplement, 80 g dry matter (DM) per day extracted noug (Guizotia abyssinica) meal, was compared with 194 to 212 g DM per day of fruits (pods (carpels) with seeds) of Acacia tortilis, A. albida, A. nilotica and A. sieberiana. For the control and four Acacia supplements respectively, mean growth rates were 32, 32, 21, 16, 4 (s.e.6) g/day and mean stover intakes 483, 430, 401, 347, 320 (s.e. 74) g DM per day. The low growth rates and intakes with A. sieberiana and to a lesser extent A. nilotica were probably related to their content of phenolic compounds including insoluble proanthocyanidins which may have lowered the digestibility of neutral-detergent fibre and nitrogen. A. tortilis and A. albida would appear to be of comparable nutritive value to noug meal as supplements to maize stover. Unlike noug meal Acacia fruits are widely available in rural areas of Africa. Further research is required to identify other tree species suitable for use as supplements and how to overcome the anti-nutritional factors.

Four groups of eight lambs (four male + four female) of about 25 kg live weight were given a concentrate diet which was supplemented with either calcium (Ca) or Ca and phosphorus (P). The levels of Ca and P in the diets were set in relation to estimated requirements (R) as determined by the Technical Committee on Response to Nutrients (TCORN, 1990) and the four treatments were as follows: treatment A Ca and P = 1R; treatment B Ca = 2R P = 1R; treatment C Ca = 4R P = 1R; treatment D Ca = 4R P = 2R.

At about 50 kg live weight the lambs were slaughtered and their body composition was determined.

The composition of their gains was also determined using information obtained from a fifth group of lambs killed at the start of the trial. There were significant treatment effects in relation to body mineral composition with Ca and P levels being higher in lambs on treatments C and D compared with those on treatments A and B. Magnesium levels were also higher in lambs on treatment D. Lambs on treatments A and B grew at the same rate (0·t25 kg/day) and retained Ca and P in relation to live-weight gain at rates close to those on which TCORN (1990) based their estimates of requirements. Lambs on treatments C and D, however, retained these minerals at somewhat higher rates though this improvement in mineralization was accompanied by a slight reduction in growth rate and in some lambs on these treatments there were health problems due to the formation of kidney stones.

The levels of Ca and P recommended by TCORN (1990) for lambs are adequate for normal growth and mineralization. Some improvement in mineralization may be achieved by feeding these minerals at above recommended levels but this improvement would appear to be at the expense of a reduction in growth rate and an increased risk of urolith formation.

Finn Dorset ewes were offered forage ad libitum from week 16 of pregnancy until lambing and were either unsupplemented (U) or given 600 g/day of concentrates (S). Grass silage and white clover silage, ensiled separately and having similar dry matter concentrations, were offered in four mixtures with white clover proportions of 0, 0·20, 0·40 and 0·60 on a fresh basis. Forage intake was significantly higher for mixtures containing higher proportions of clover (9·1, 10·5, 12·3 and 13·6 g organic matter (OM) per kg live weight) and the ewes gained more weight (33, 65, 178 and 174 g/day) and had smaller losses in body condition score (–0·59, –0·49, –0·39 and –0·17). Higher proportions of clover in the diet in late pregnancy also resulted in significantly higher growth rates of lambs during lactation, when a common diet was offered.

The number of foetuses carried in pregnancy had significant effects on intake and some aspects of performance. For ewes carrying singles, twins and multiples respectively, mean daily forage intakes were 12·2, 11·9 and 10·4 g OM per kg live weight and losses in body condition score were –0·05, –0·40 and –0·62.

Offering the supplement reduced forage intake and for treatments U and S respectively, mean daily values were 12·6 and 10·2 g OM per kg live weight. Whilst the supplemented ewes had smaller losses in body condition score in late pregnancy (–0·54 v. –0·28), lamb birth weights and growth rates were significantly increased only for ewes which had carried three or more foetuses in pregnancy and reared their two heaviest lambs.

The consistency of differences between varieties in nutritive value of cereal straws was tested over 3 years consecutively for nine varieties of spring barley and 10 varieties of winter wheat straw and over 2 years consecutively for 12 varieties of winter barley and six varieties of oats. For all varieties there were large year-to-year differences in nutritive value measured by ruminal degradation characteristics using the nylon bag technique. For spring barley there were large differences between varieties and the correlations between varieties in different years were mostly in excess of r = 0·90 for 48-h degradability and maximum potential degradability. For winter wheat the correlations between varieties for different years were lower but mostly significant. For the winter barley differences between varieties were less, but even so differences between varieties in different years were significant. For oats differences between the varieties were small and there was no significant correlation within varieties between years. The results suggest that the nutritive value of cereal straws is generally consistent between years and is a characteristic that could be improved by selection.

The effects on reproductive performance of different strategies of pasture management designed to conserve herbage in situ during the late growing season for use during the pre-mating and mating period were studied in three experiments over 3 years with Greyface ewes. Two experiments with 341 ewes compared set-stocking of pasture at 12 ewes per ha on two initial levels of herbage mass with a conservation strategy which left half of the area ungrazed and grazed the other half of the area at 24 ewes per ha from late August until mid October (2 weeks before the start of mating) after which the total area was grazed at 12 ewes per ha. In a third experiment with 124 ewes, the conservation strategy left a quarter of the area ungrazed and varied stocking rate to a minimum of 12 ewes per ha to maintain two initial levels of herbage mass on the other three-quarters of the area and then compared set-stocking at six ewes per ha on this area with set-stocking at 18 ewes per ha on the previously ungrazed quarter of the area during the pre-mating and mating period.

Initial levels of herbage mass were within the range of 2000 to 2100 kg dry matter (DM) per ha (7 cm sward height; high) and 1500 to 1700 kg DM per ha (4 to 5 cm; low). Mean herbage accumulation rates between August and October were estimated to be 25 and 44 kg DM per ha per day in the first two experiments. Initial live weight and body condition score also varied considerably between years and subsequent response was influenced by herbage growth rate. Where mean ewe body condition was within the score 2·75 to 3·00, range over the mating period and herbage mass on set-stocked areas was not higher than 2200 kg DM per ha (8 cm sward height) in the pre-mating period or not lower than 1300 kg DM per ha (3·5 cm) at about 3 weeks after mating, the strategy of management did not influence reproductive performance in terms of the number of lambs born. Within these limits, however, reproductive performance was positively related to herbage mass in late August (low = 1·60, high = 1·94).

When herbage mass fell below the 1300 kg DM per ha level before or during mating, reproductive performance was improved by a herbage-conservation strategy which maintained ewe body-condition score within the 2·75 to 3·00 range. When herbage mass and growth rate were high, reproductive performance was also improved by a herbage-conservation strategy which restricted the development of excessive body condition before mating by avoiding the increase in barrenness shown to derive from very high levels of body condition.

The heritability of growth traits from birth to 6 months of age and the genetic and phenotypic correlations among them were determined from paternal half-sib analyses of 1689 records of D'man lambs. Heritability estimates were 0·34 (s.e. 0·08), 0·23 (s.e. 0·06), 0·52 (s.e. 0·10) and 0·43 (s.e. 0·09) for weights at birth, 30 days, 90 days and 180 days, respectively. Heritabilities for average daily gains from birth to 30 days and from 30 to 90 days were estimated to be 0·12 (s.e. 0·05) and 0·56 (s.e. 0·11), respectively. Genetic and phenotypic correlations among growth traits were all positive and of high magnitude. The former estimates varied from 0·66 to 0·98, the latter estimates ranged from 0·23 to 0·92. Selection for high weight at 90 days would be expected to have large effect on other growth traits as a result of their large positive genetic correlations.

Six replicates (five Landrace and one Duroc) of four littermate entire male pigs were used to evaluate the effect of protein level in diets of constant amino acid composition on live-weight gain, efficiency of gain, nitrogen balance and carcass characteristics. The diet contents ranged from 151 to 282 g crude protein (CP) and 8·5 to 16·4 g lysine per kg dry matter (DM) and were studied over two live-weight ranges, 33 to 55 kg and 55 to 88 kg. All the diets had similar metabolizable energy (ME) values (15·0 MJ/kg DM). Pigs were given food twice daily to appetite. Sixteen pigs (four replicates) were used for metabolism trials at 35 to 43, 58 to 65 and 78 to 88 kg live weight and after attaining a live weight of 88 kg they were slaughtered to assess carcass characteristics. Between 33 and 55 kg live weight significantly linear improvements (P < 0·01) in the daily live-weight gain and in the ratio ME: live-weight gain were observed for daily intakes in excess of 300 g ideal protein and 21 g lysine (212 g CP and 12·5 g lysine per kg DM). Between 55 and 88 kg live weight, linear improvements in the same two variables were observed for daily intakes of 395 g ideal protein and 28 g lysine (212 g CP and 12·5 g lysine per kg DM). For the 33 to 88 kg live-weight period the average ME intake was 30 MJ/day and the predicted intakes for maximum growth and food conversion efficiency were in excess of 370 g ideal protein and 26 g lysine. Nitrogen retention improved significantly with increasing dietary crude protein content in all three balance periods. Fat thickness measurements decreased linearly (P < 0·05) and the proportions of lean and fat in rumpback increased and decreased respectively (P < 0·01) as the dietary protein level increased.

To test the proposition that growing pigs, when given a choice between two foods, are able to select a diet that meets their requirements, and to investigate the rules of diet selection, four foods (L, A, B and H) with similar energy yields, but different concentrations of crude protein (CP) (125, 174, 213 and 267 g CP per kg fresh food respectively) were formulated. The four foods were offered ad libitum either singly, or as a two-way choice using all the six possible pairs, to 40 individually caged pigs from 12 to 30 kg live weight. On the single foods the rate of food intake fell from 1001 to 971 to 961 to 868 (s.e.d. 40) g/day (F < 0·05) as the protein concentration of the foods increased from L to H; the growth rate followed an opposite trend (492, 627, 743 and 693 (s.e.d. 31) g/day respectively; P < 0·01). When the pigs had to select between two foods limiting in protein (L and A) the less limiting one was preferred (710 (s.e. 200) g A per kg total food intake; the protein concentration of the selected diet was 160 (s.e. 10) g CP per kg). On the choice between B and H (a choice between a food with protein concentration close to requirements and a food with protein excess) the lower food was markedly preferred (928 (s.e. 4) g B per kg total food intake; the protein concentration of the selected diet was 218 (s.e. 1) g CP per kg). When the animals were given a choice between two foods, a combination of which was non-limiting (pairs LB, LH, AB and AH), the protein concentrations of the selected diets were not different between treatments (208, 204, 202 and 205 (s.e.d. 13) g CP per kg respectively) and they also declined systematically with time and weight. The growth rate of the animals on these pairs were 752, 768, 769 and 763 (s.e.d. 54) g/day (P > 0·05), which were not significantly different from the highest growth rate achieved on a single food. The results suggest that pigs, when given a choice between a suitable pair of foods, are able to choose a balanced diet and to change its composition to reflect their changing requirements. The choice-feeding method may well be useful as an effective and economic way of estimating and meeting requirements, and of measuring the growth potential of pigs.