Fifty-six individually penned pregnant ewes were offered silage ad libitum with wheat or barley in a replicated experiment. Each of these cereals was either unprocessed or processed by rolling or milling and was offered at 300 g/day from 6 weeks before lambing and at 600 g/day from 3 weeks before lambing. Five and 2 weeks before lambing, total faeces were collected over 6-day periods. On average proportionately 0·18 of the unprocessed wheat and 0·20 of the unprocessed barley consumed was recovered as intact grain in the faeces. Mean digestibility of the organic matter in the complete diet increased from 726 g/kg on whole to 748 g/kg on rolled and 759 g/kg on milled cereal (P < 0·01) at the 300 g/day level and from 721 g/kg on whole to 766 g/kg on rolled and 772 g/kg on milled cereals (P < 0·001) at the 600 g/day level. When the effect of egested whole grain was removed there was no evidence of an effect of cereal processing on organic matter digestibility and although digestibility of fibre was higher with whole than with processed grain the effect was not significant. Wheat-based diets were more digestible than barley-based diets and, compared with barley, wheat appeared to have a beneficial effect on fibre digestion.

Correlated responses in slaughter and carcass traits to divergent selection for high or low carcass lean content in Texel-Oxford sheep were measured in the crossbred progeny of 23 rams. Rams were selected using the selection index: -0·995 FATD + 0·206 WT20, where FATD and WT20 are the ultrasonic backfat depth and live weight at 20 weeks of age, with both traits standardized to have zero mean and unit phenotypic variance. A total of 329 crossbred lambs were slaughtered at 16 weeks of age and 250 lambs were slaughtered at fixed weight, 38 to 40 kg for castrated males and 36 to 38 kg for ewe lambs. Progeny from high-line rams grew faster, as indicated by the heavier slaughter weight of lambs slaughtered at fixed age (04 (s.e.d. 0·5) kg) and the lower age at slaughter for lambs slaughtered at fixed weight (-5 (s.e.d. 3) days). Subcutaneous and internal fat scores and the Meat and Livestock Commission carcass appraisal and conformation scores were all lower in the high line than in the low line and the magnitude of the selection line differences for lambs slaughtered at fixed weight was essentially double that of lambs slaughtered at fixed age (0·8 v. 0·5 score units).

Shoulder joints of 254 lambs, slaughtered at fixed age, were dissected and half carcasses were dissected on 66 of the lambs. Carcass and shoulder joint information was combined with double-sampling methodology, using multiple regression equations to predict carcass composition. For each tissue type, viz. lean, subcutaneous fat, internal fat and bone, the correlations between carcass content and shoulder joint content were at least 0·90 and the proportions of variation in carcass tissue content accounted for by the prediction equations were also at least 0·90. Progeny of high-line rams had higher carcass lean weight, (402 (s.e.d. 140) g) than progeny from the low line, with no difference in carcass fat weight, such that carcass lean content was higher, (10 (s.e.d. 5) g/kg), and carcass fat content was lower, (-10 (s.e.d. 7) g/kg). Based on the slaughter and carcass traits of these crossbred progeny, divergent selection for high and low carcass lean content has established lines of animals with different rates of lean growth but similar rates offat deposition.

The effect of the anabolic steroid trenbolone acetate (TEA) on growth and body composition was studied in Scottish Blackface mature female sheep weighing 45 kg. The sheep were blocked by weight and randomly allocated to be untreated controls (C) or subcutaneously implanted with 20 (Tl), 40 (T2) or 60 (T3) mg TBA 60 days before slaughter. The sheep were offered to appetite a loose-mix diet which contained an estimated 11 M] metabolizable energy and 120 g crude protein per kg dry matter. Live-weight gain and food intake, on average, were not significantly affected by the hormonal treatment which significantly reduced gut fill and significantly increased empty body and chilled carcass weights. Of the carcass tissues which were dissected only weights of total lean and intermuscular fat were significantly increased following TBA implantation. The relative development of fat, lean and bone in the carcass in the hormonally treated compared with control sheep was not significantly altered. The only effects on weights of organs and body parts associated with TBA implantation were increases in fat deposition in omental and perinephric and retroperitoneal depots. There was some evidence of an increasing response to TBA treatment with increasing dose level for weights of chilled carcass, carcass lean and the ratio of subcutaneous fat to intermuscular fat.

The response of finishing Blackface lambs (mean initial live weight 29·6 kg) to shearing and level of concentrate supplementation on a silage-based diet was examined in two experiments. Silage intake decreased as concentrate level increased at mean rate of 0·13 and 0·26 g silage dry matter (DM) per g concentrate DM, in experiments 1 and 2 respectively, with no significant differences between unshorn and shorn lambs. Lambs gained 36, 97, 135, 193 and 224 g/day (s.e. 9·1; P < 0·001) in experiment 1 and 97,133,170,185 and 222 g/day (s.e. 9·9; P < 0·001) in experiment 2 when 0, 200, 400, 600 g/day and concentrate ad libitum were offered respectively. Shearing increased silage DM intake from a mean of 0·50 to 0·56 kg/day (s.e. 0·016; P < 0·05) in experiment 1, and from 0·80 to 0·90 kg/day (s.e. 0·013; P < 0·001) in experiment 2. There was no interaction between shearing and the level of concentrate offered. Despite higher DM intakes by shorn lambs, their live-weight gains were lower than those of unshorn lambs. Gains of unshorn and shorn lambs averaged 153 and 120 g/day (s.e. 5·8; P < 0·001) in experiment 1 and 162 and 161 g/day (s.e. 5·7; P > 0·05) in experiment 2 respectively. This appears to have resulted because shorn lambs could not fully compensate for their increased energy requirements after shearing, by increasing their DM intake. Shearing did not significantly influence carcass characteristics.

It was calculated that with lambs gaining 150 g/day, concentrate requirement was increased from 286 to 335 g/day and total DM intake from 965 to 1107 g/day as a result of shearing.

Lactation curves for several dairy traits in sheep were studied. Data were collected from seven pure breeds, three synthetic breeds and twelve single breed crosses in a 5-year milking trial. Wood's lactation model (yn = anbe-cn) was used to describe the lactation curves. Actual and fitted lactation curves had the same shape within each trait except for proportion of lactose. Correlations between actual and fitted lactation curves ranged from 0·92 to 0·99 for all traits except lactose (0·24). There was no apparent explanation for the low correlation for lactose content. It was concluded that the algebraic formula can be used to obtain estimates of parameters for sheep lactation curves. However, some caution should be exercised in extending results to proportion of lactose in the milk.

Crossbred groups had higher total milk production than pure and synthetic breeds. Lactation curves for crossbred ewes, grouped by breed of sire, had a similar pattern. Finnsheep and Rambouillet-sired crossbred ewes produced proportionately 0·23 more milk than ewes sired by Dorset and Lincoln rams. Breed differences in total milk production were observed among the pure breeds. The Suffolk breed was superior in milk production. All breeds, except the Romanov, followed a similar pattern for lactation curves. The Romanov breed produced milk at a relatively low but consistent level throughout the lactation trial.

Pure Merinos d'Aries male lambs were reared indoors (IN), with hay and concentrates, or outdoor (OUT) on improved Mediterranean range (Garrigues) and slaughtered when they reached 25 (light) or 32 (heavy) kg live weight. The alimentary tract weight was significantly affected by slaughter weight and by rearing method, which also significantly modified rumen contents. IN lamb body composition was higher in fat and energy and lower in water and protein than OUT lambs but differences induced by the rearing method were reduced when slaughter weight was increased. Allometric equations established for each environment were used to determine body composition and body gain composition at 20 and 25 kg empty body weight. Changes were very limited for IN lambs but protein and fat content of OUT lambs changed greatly as a result of the tremendous increase of grass availability towards the end of the fattening period.

Two experiments are described in which growing lambs were offered ad libitum two concentrate pelleted diets containing 60 (diet C) or 230 (diet D) g/kg barley straw.

In the first experiment, eight male and eight female lambs were used to determine the outflow rate from the rumen of chromium-mordanted fish meal and soya-bean meal, estimated from faecal marker excretion. Diet D promoted a higher rumen outflow rate (0·0769 per h) than diet C (0·0486 per h); no differences were found either between sexes or between Cr-mordanted protein supplements.

In the second experiment, the rumen degradation of fish meal, soya-bean meal, sunflower meal, pea seeds and sweet lupin seeds was studied by incubating the supplements in polyester bags in the rumens of four lambs, following a change-over design. Increasing the proportion of straw to 230 g/kg increased the rate of nitrogen disappearance from all vegetable supplements, the differences being significant only for pea seeds.

When degradation kinetics and outflow rates were considered together, effective nitrogen degradability of protein supplements was reduced by the increased proportion of dietary straw, indicating a greater influence of rumen outflow rate than of the increased rate of microbial degradation.

About 8 weeks after parturition six multiparous British Saanen goats with permanent rumen fistulas were allocated to three dietary treatments in a random sequence for three successive 30-day periods. The treatments were perennial ryegrass hay ad libitum with 1000 glday of concentrates containing 112 (low-protein, LP), 182 (high-protein, HP) or 255 (very high protein, VHP) g crude protein (CP) per kg dry matter (DM) daily. The CP concentration was increased by substitution of soya-bean meal for cereals. From day 19 of each period, when measurements were made, concentrate allowance was reduced to 850 glday and hay intake was offered at proportionately 0·10 more than consumption in the previous 6 days. Hay intake and milk yield increased linearly with CP in the concentrates, but differences between HP and VHP were small and non-significant. Milk composition was unaffected by the treatments, so yields of milk solids followed the same trends as yields of milk, but these were only significant for solids-not-fat. The digestibility of neutral-detergent fibre, acid-detergent fibre and nitrogen (N) all increased with increasing CP, but significantly so only for N; the digestibility of DM and organic matter was unaffected.

The digestibility of the hay incubated in the rumen was unaffected by the treatments. With increasing N intake, faecal, urine and milk N increased linearly as did apparently digested N but N retention did not follow a consistent pattern. Rumen volatile fatty acids and pH were unaffected but concentration of rumen ammonia and blood urea increased with increasing CP as did rumen fluid dilution rate. The results suggest that, with lactating goats given diets of moderate quality grass hay and concentrates containing soya-bean meal, responses to concentrations of CP greater than about 180 g/kg DM in the concentrates or 140 g CP per kg DM in the total diet are likely to be small.

Computer tomography (CT) was evaluated as a method to predict body composition of live goats. Forty-one lactating goats were scanned. CT values (X-ray attenuation data) representing soft body tissues were summarized into 140 frequencies. By principal component analysis the 140 multicollinear frequencies were reduced to six orthogonal principal components at the loss of only 0·032 of the original variation.

Following scanning at four positions the goats were slaughtered, dissected and the data analysed. Seven compositional traits (water, protein and fat from the carcass and non-carcass parts, and total energy of the body) were regressed on live weight and the six principal components from each of five CT data sets (data from the four scan positions and pooled data) in a stepwise linear regression procedure. Models from the pooled data set were further evaluated in a cross-validation procedure. Compared with evaluation based on calibration only, a more simple and precise model was chosen as the best by this procedure.

After addition of CT variables in prediction equations already comprising live weight, cross-validation deviations were reduced by 0 for carcass water, 0 for carcass protein, 0·48 for carcass fat, 0·14 for non-carcass water, 0·23 for non-carcass protein, 0·73 for non-carcass fat, and 0·73 for total energy. These results show that CT is a valuable in vivo method for predicting body fat and energy, but less valuable for body water and protein.

Seventy-eight Beulah (BH) and 78 Welsh Mountain (WM) wether lambs (mature live weights (MLW) 60 and 70 kg respectively), were used to investigate genotypic differences in performance and compensatory growth during the post-weaning period. After an initial slaughter of 12 of each breed, a complete straw-based pelleted diet (122 g crude protein and 10·4 Mj metabolizable energy per kg dry matter (DM)) was offered ad libitum to a further 29 of each which were slaughtered at a range of live weights between proportionately 0·35 andl -00 of ML W. As well as these groups undergoing normal growth (NG), the remaining lambs were given the same diet at a level sufficient to maintain constant live weight (LW) for 3 months. Eight of each breed were then slaughtered and the remaining 29 of each offered the diet ad libitum for a period of post-restricted growth (PRG); these 29 were also slaughtered between 0·35 and 1·00 of MLW. During the periods of ad libitum food supply, voluntary food intake was greater in BH than WM lambs (BH: 38·3, WM: 35·3 g DM per kg LW, P < 0·01) and during PRG as compared with NG although this result was due to a response only in WM lambs (NG - BH: 38·0, PRG-BH: 38·5, NG - WM: 33·3, PRG - WM: 37·2 g DM per kg LW, breed × pattern of food supply interaction, P < 0·05). Daily live-weight gain (LWG) was greater during PRG than NG in both breeds (NG: 3·15, PRG: 4·80 g/kg LWper day, P < 0·05). This effect was greatest in the first 8 weeks of growth but persisted into the period 16 weeks to maturity. WM lambs had overall slightly leaner carcasses and a lower proportion of protein in the fleece-free empty body than BH when compared at the same proportion of MLW; particularly during PRG. The protein in the carcass-weight gain (CWG) was greater in WM lambs during PRG than during NG (NG: 98, PRG: 123 g/kg, P < 0·05), but did not differ significantly in BH (NG: 106, PRG: 118 g/kg). The proportion of fat in CWG was concomitantly smaller during PRG. There are thus breed differences in the performance and carcass composition of lambs during compensatory growth. Genotypes to be finished by continuous feeding or a regimen incorporating food restriction should be carefully chosen for date offinishing in relation to desired carcass composition.

Seven replicates of five littermate boar Landrace pigs were used to evaluate the effects of (a) reducing the food intake within the practical range or (b) reducing energy intake without reducing protein intake, between 33 and 90 kg live weight on growth, nitrogen balance and carcass characteristics. Three levels of food intake (ad libitum, 0·90 and 0·80 ad libitum) were used and the dietary protein contents ranged from 250 to 312 g crude protein per kg dry matter (DM) to equalize protein intake with reduced food intake. All the diets were of similar amino acid composition, and lysine per kg DM ranged from 14·3 to 17·9 g. The food was given twice daily. Balance trials were conducted at 45 and 65 kg live weight and all pigs were slaughtered at 90 kg live weight to assess carcass characteristics. Reducing food intake or only energy intake significantly reduced live-weight gain (P < 0·001). Reducing food intake decreased nitrogen retention (P < 0·01) during the 65-kg balance period. Reducing food intake had no effect on fat thickness measurements but measurements at shoulder (P < 0·001), mid back (P < 0·001), C(P< 0·01) and P2 (P < 0·05) decreased and lean: fat in rumpback increased (P < 0·05) when energy intake was reduced. The weights of small intestine (P < 0·001), large intestine (P < 0·05), liver (P < 0·05) and kidneys (P < 0·05) were significantly reduced as food intake was reduced. Energy is more limiting than protein when the intake of a diet providing for optimal growth of high potential boars is reduced between 33 and 90 kg live weight.

Twenty-four castrated male British Angora goats of Australasian origin aged 11 months and weighing 23·5 kg live weight were studied in a 112-day experiment. They were allocated in a 2 × 2 factorial design to receive diets containing, per kg dry matter, concentrations of estimated metabolizable energy (ME) of 10·2 M] (LE) or 11·9 MJ (HE) and crude protein concentrations of 108 g (LP) or 180 g (HP). The HE diets were offered to provide a dry-matter intake of 30 g/kg live weight and the LE diets to provide 0·85 of the ME intake of HE diets.

In comparison with LE diets, the HE diets, on average, increased live-weight gain, food conversion efficiency, carcass weight, killing-out proportion, empty body weight, cross-sectional area of m. longissimus dorsi and fat thickness over m. serratus dorsalis and weights in the carcass for dissected lean tissue, crude protein (N × 6·25) and dissected and chemically extracted fat. Weights of shoulder, hind barrel, and best-end joints and their weights of dissected lean and fat were increased on HE diets. No consistent effects due to dietary energy were observed for yield or diameter of mohair fibres.

The greater intake of dietary protein resulted, on average, in increased values for live-weight gain, efficiency of food conversion, carcass weight, weight of dissected carcass lean tissue and crude protein, killing-out proportion and cross-sectional area of m. longissimus dorsi. The greater intake of dietary protein increased significantly both yield and diameter of mohair fibre.

Significant interactions between protein and energy for empty body weight, carcass weight and carcass fat indicated that increasing dietary protein intake had a greater positive effect on the low than on the high energy diets. Similarly, increasing the dietary energy consumption on the LP diets produced a greater positive effect than that recorded on the HP diets.

The results indicated that growth and lean tissue deposition were affected by both protein and energy intakes but that mohair fibre characteristics were affected consistently only by protein nutrition.

The effects on reproductive performance offeeding supplements to ewes grazing predominantly perennial ryegrass swards before and during mating which started in mid November were examined with approximately 200 Scottish Blackface ewes in each of three experiments. In two experiments, ewes were offered no supplement (C), 600 g barley (S(B)) or 500 g barley + 100 g white-fish meal (S(B + FM)) per head per day from 3 weeks before until 3 weeks after the start of the mating period. Ewes in experiment 1 were grazed on different areas and in experiment 2 were rotated around these areas on a 12-day cycle. Initial herbage allowance in both experiments was 40 to 50 kg organic matter per ewe and stocking rates ranged from 10 to 14 ewes per ha. Ewes in experiment 3 were rotated around the same areas on a 10-day cycle at a stocking rate of 18 ewes per ha and offered either no supplement (C) or 600 g barley (S(B)) per head per day when live weight started to decrease or sward height declined to 3·5 cm, whichever occurred first. Both of these occurred shortly after the start of mating. Sward height fell from mid October and throughout November and was between 3 and 4 cm at the start of mating in all experiments.

There were no differences in response to the (S(B)) and (S(B + FM)) treatments in live weight, body condition or reproductive performance in experiments 1 and 2 and the effects of supplementation were examined with these treatments pooled.

Supplementation had very little effect on live weight, resulting in higher values only in experiment 1 but did produce higher body condition in all experiments. This occurred by the start of mating in experiment 1 but not until mid December in experiment 3 and early January in experiment 2.

A significant increase in the number of lambs born per ewe lambing was obtained from supplementation when the data from experiments 1 and 2 were combined (P < 0·05). There was no response in reproductive performance to supplementation in experiment 3. It was concluded that supplementation in the pre-mating period can improve reproductive performance when sward height declines from 6 to 3·5 cm but that supplementation in the post-mating period may not, even when the sward height is below 3·5 cm.

The effect of reducing energy intake within the practical range or reducing energy intake without reducing protein intake on chemical composition of gain, whole-body protein turn-over and energy metabolism was studied between 33 and 90 kg live weight with seven replicates of five littermate boar Landrace pigs. Three levels of food intake (ad libitum, 0·90 and 0·80 ad libitum) were used and the dietary protein contents ranged from 250 to 312 g crude protein (CP) per kg dry matter (DM) to equalize protein intake with reduced food intake. All the diets were of similar amino acid composition and were offered twice daily as pellets. There was no effect of dietary treatment on the DM, CP and fat contents (g/kg) of the empty body (EB), but fat content and fat: protein ratio in EB tended to decrease with reduction of food intake or energy intake. The relationship between energy intake and protein deposition was linear and the mean maximum protein retention was 187 g/day. Retention of DM (P < 0·001), protein (P < 0·001), fat (P < 0·001), energy (P < 0·001) and ash (P < 0·01) decreased linearly with reducing food intake or energy intake. The calculated residual heat production was 0·604 MJ metabolizable energy per kg M0·75 per day. The dietary treatments had no effects on composition of longissimus dorsi, fillet, liver, kidney, backfat or kidney fat. The nitrogen flux, nitrogen synthesis and breakdown tended to decrease with reduction in food intake or energy intake though the effects were not statistically significant. Nitrogen accretion decreased significantly (P < 0·05).

The effects on the production of indoles of dietary fibre, lactose and antibiotic were examined in a factorial design using 16 crossbred pigs (eight boars and eight gilts) from 40 to 90 kg live weight. The control diet was based on wheat and soya-bean meal which was partially replaced either by sugar-beet pulp (400 mg/g) or lactose (25 mg/g) or both. All diets were offered with or without the antibiotic tylosin phosphate (200 mg/kg diet). Animals were penned individually and the diet restricted to provide 1·3 M] digestible energy per kg M0·75. All faeces were collected for two 4-day periods at about 60 and 75 kg live weight. Animals were slaughtered on completion of the experiment and subcutaneous fat was sampled. Faeces and carcass fat were analysed for skatole and indole. There were no significant effects of treatments on growth rate but killing-out proportion was reduced (P < 0·05) on fibre or lactose diets with an additive effect (P < 0·001) of the combined ingredients. Dietary fibre significantly increased the daily elimination of skatole and indole and the concentration of indole in faeces but because of greater faecal bulk on the fibre diet the concentration of skatole in faecal dry matter was higher (P < 0·05) on the control diet. Dietary lactose had no effect on indole in faeces but significantly reduced the concentration and daily output of skatole. Levels of both skatole and indole in faeces tended to be reduced with dietary antibiotic which had a significant interaction with the fibre treatment on indole levels. The concentrations of skatole or indole in subcutaneous fat were neither affected by dietary treatment nor significantly correlated with concentrations or outputs in faeces.

The effects of feeding supplements in relation to sward height in the autumn on the reproductive performance of ewes grazing predominantly perennial ryegrass swards were studied in four experiments with three genotypes. Treatments in which either no supplements were given or supplements were offered according to the height of the sward were compared when ewes were grazed on swards which were managed to decline in height from > 5 cm to 2·5 cm either before or after mating. At a sward height of < 3·5 cm, 400 g ground and pelleted barley was offered per head per day; at < 3·0 cm, 700 g was offered; and, at < 2·5 cm, hay was offered in addition to the barley. Responses to these treatments were compared with that of a treatment in which a sward was maintained at > 5 cm. Sward heights were controlled by varying stocking rate. In experiment 1, with 305 Scottish Blackface ewes, supplementation was not necessary until after the start of the mating period. In experiment 2, with 200 Brecknock Cheviot ewes, the supplementation protocol was examined on swards declining in height either before or after a synchronized mating. Experiment 3 repeated experiment 2 with 200 Beulah Speckled-face ewes and experiment 4 repeated only the pre-mating phase of experiment 3 with 195 Beulah Speckled-face ewes.

In experiment 1, ovulation, ova loss and potential lambing rates were determined from the number of corpora lutea and viable embryos at slaughter at 8 to 10 weeks after the start of mating in 103 ewes and lambing rate was determined from the number of lambs at lambing in the remaining 202 ewes. In experiments 2, 3 and 4, ovulation, ova loss and potential lambing rates were determined from the number of corpora lutea and viable embryos at slaughter at return to service or at 4 weeks after a synchronized first mating.

Grazing a sward declining in height below 3·5 cm in the pre-mating period significantly reduced ewe live weight, body condition and reproductive performance, particularly in the relatively prolific Beulah Speckled-face breed, compared with grazing a sward maintained at > 5 cm. Supplementation at the levels offered overcame some of the effects of the lower sward height treatment in the pre-mating period but did not compensate fully. Grazing a sward declining in height below 3·5 cm in the post-mating period, although reducing ewe live weight and body condition, had little effect on reproductive performance, compared with that achieved on a sward maintained above 4 cm. Supplementation at the levels offered at the lower sward height during the post-mating period produced no increase in overall reproductive performance and had a negative effect on some reproductive parameters.

Food intake patterns of growing pigs given food ad libitum in individual and group housing were derived from food intake recordings with IVOG®-stations. The IVOG-station is a feeding station that records animal identification number, time, duration and amount of food intake during each visit of a pig to the food hopper. Data were collected in three testing batches of 90 Dutch Landrace pigs each, housed in 10 individual pens and in 10 group pens of eight pigs per batch.

Based on survival analysis theory, intervals between visits shorter than 5 min (which was used as meal criterion) were regarded as within-meal intervals and these visits were grouped into meals. In group housing, food intake per day and rate of food intake had no significantly non-normal distribution. In individual housing, rate of food intake, food intake per day and number of meals per day had no significantly non-normal distribution. In individual housing rate of food intake, food intake per day and number of meals per day had no significantly non-normal distribution. All traits were normally distributed after discarding extreme values, except food intake and eating time per visit and per meal. In group housing these traits were not significantly non-normally distributed after logarithmic transformation.

Pigs housed in groups ate faster, had a higher food intake per meal but less meals per day, less eating time per day and a lower daily food intake than pigs penned individually. During the day two peaks of feeding activity occurred, especially in group housing: one in the morning and one in the beginning of the afternoon.

A distinction was made between meals with a major contribution to daily food intake and meals of minor importance. In group housing, 69% of the daily number of meals accounted for proportionately 0·87 of daily food intake and 0·83 of daily food intake time. In individual housing 39% of the meals accounted for proportionately 0·90 of the daily food intake and 0·79 of the daily eating time.

Higher repeatabilities of day to day recordings of food intake traits, were found when estimated within 2-week periods compared with months or with the total fattening period.

An in vivo digestibility trial was conducted by feeding sheep the leaves, fruits and twigs of Atriplex barclayana in a proportion roughly equivalent to that eaten by sheep grazing freely in Atriplex plantations. Four treatments were imposed on each of four sheep in a 4 × 4 Latin-square experimental design: Atriplex offered alone or with 100, 200 or 300 g/day tapioca meal.

The mean apparent digestibility of the Atriplex dry matter (DM) and organic matter (OM) consumed were 0·59 and 0·56, respectively. Addition of tapioca to the Atriplex in the diet did not improve these digestibility coefficients. The low OM content of 760 g/kg together with its digestibility of 0·56 resulted in the Atriplex having a low metabolizable energy concentration of 6·28 MJ per kg DM. The in vitro apparent DM digestibility of Atriplex was approximately 0·09 higher than the in vivo apparent digestibility.

The mean nitrogen concentration of the Atriplex DM was 16·6 g/kg, and its apparent digestibility 0·73, which was not improved by the addition of tapioca to the diet. Nitrogen retention of the sheep eating only Atriplex was proportionately 0·17 of the nitrogen intake. The addition of 300 g tapioca improved nitrogen retention to 0·27 but was not significantly different from the other treatments.

Water intake and urine excreted were as high as 14 and 12 I/day respectively, for an Atriplex DM intake of about 1300 g/day. During the experiment the sheep only maintained live weight, despite daily intakes of up to 1200 g Atriplex DM and up to 300 g tapioca.

Meat quality was studied in pigs reared in groups kept as units during the fattening period and slaughter (unmixed). The results were compared with conventionally reared pigs, which were mixed during the fattening period, at transport to the abattoir and at lairage (mixed). It is concluded that the unmixed pigs in this study had a better meat quality measured as pH,, pHu, weight loss and also fewer skin lacerations compared with the mixed groups (i.e. less development of pale, soft exudative (PSE) and dark, firm dry (DFD) in the unmixed animals). There were no significant differences in meat colour or in lean content between the unmixed and mixed pigs.

Patterns of milk production and circulating blood metabolite and hormone concentrations were determined throughout a 14-week lactation in groups of 10 twin-rearing East Friesland EF) and Scottish Blackface (SBF) ewes. All ewes were offered, ad libitum, a pelleted diet throughout the experiment and milk yields were measured weekly. Pooled blood samples (six samples, 20-min intervals), collected before feeding, on 1 day of each week, were assayed for plasma glucose, non-esterfied fatty acids, 3-hydroxybutyrate (3-OHB), albumin, total protein, insulin, growth hormone (GH), cortisol, prolactin, thyroxine (T4), and triiodothyronine (T3). At weeks 2, 4,10, and 14, samples were collected at 20-min intervals for 8 h and assayed individually for plasma insulin, GH, cortisol, and prolactin. Mean daily intakes and milk yields of ewes were similar in the two breeds. Mean pre-feeding concentrations of most blood metabolites during the experiment were also similar in ewes of the two breeds but SBF ewes had higher plasma concentrations of 3-OHB (0·67 v. 0·43 mmol/l; P < 0·01) and total protein (73·2 v. 66·4 g/l; P < 0·001). SBF ewes had higher overall mean concentrations of insulin (9·53 v. 4·36 mU/l; P < 0·001), cortisol (7·59 v. 573 μg/l; P < 0·05), prolactin (457·1 v. 316·1 μg/l; P < 0·05), and T3 (1.32 v. 1·12 μg/l; P < 0·001). GH and T4 profiles were similar in the two breeds. Following the daily introduction of fresh food, there were significant increases in concentration of insulin (P < 0·001) and cortisol (P < 0·01) while there was a decrease in mean concentrations of GH (P < 0·001) and prolactin (P < 0·01). There were significant interactions between breed and feeding effects on insulin, GH, cortisol and prolactin concentrations. Patterns of milk production in ewes of these breeds were associated with changes in insulimGH ratios both before and after feeding which may be of a causal nature.