In deterministic comparisons of breeding schemes it is necessary to take account of variance reduction due to selection. This can take place as multi-stage selection within generations and it takes place across generations when offspring of selected parents are selected. A standard way to deal with this is to set up selection index equations where the parameters are altered as a consequence of selection. It is shown that if the breeding schemes use a univariate or multivariate best linear unbiased prediction (BLUP) animal model for prediction of breeding values this procedure can be simplified. This is done by modelling the distribution of estimated breeding value (EBV) utilizing that changes in EBV of an individual are independent of selection. In the univariate case the variance reduction and the resulting genetic gain can be calculated from the selection intensities and the accuracies in the unselected population. An expression is given for the response in each generation when selection is started in a base population with complete pedigree. This shows that a limiting value is obtained within three to four generations. The asymptotic response for several traits is described in the case where selection is on multitrait BLUP.

Methods which estimate the effects on farm profit of a change in a genetic trait (economic weight) are compared. A neoclassical economic model based on a Cobb Douglas type production function, is extended to the long run and compared with a conventional profit equation for two types of trait changes. With this model, an optimal production plan can be derived for the farm which depends on input prices and the production function parameters. In the calculation of economic weights by animal breeders, costs are usually linear with respect to farm output, and so input levels and output are determined arbitrarily. Differences between methods for the estimation of both absolute and relative economic weights are shown to be small for proportional trait changes of 0·01 to 0·05. However, for proportional trait changes of 0·3, proportional differences were from 0·01 to 0·75 and from 0 to 0·6 for absolute and relative economic weights respectively, depending on the production function parameters. This dependence on the size of the trait change is attributed to the non-linearity of the production function. The effects of ‘rescaling’, which implicitly discounts economic weights for output increasing traits, are also evaluated. Absolute and relative ‘rescaled’ economic weights for 0·01 proportional trait changes are shown to differ proportionally from those estimated by the neoclassical economic model by from 0·14 and 0·12 to 0·50 respectively. For 0·3 trait changes, proportional differences were from 0·16 to 0·79 and from 0·8 to 0·43 for absolute and relative trait changes respectively. The results have important implications for conventional breeding programmes, and the economic evaluation of breeds and major genes.

Forty-four Borana × Friesian F1 crossbred calves born to Borana (Bos indicus) cows were randomly assigned to either suck their dams hvice daily or bucket-feeding with 1341 milk over a 57-day pre-weaning period. Their dams were also allocated either to a supplement group with grass hayad libitumand 5 kg per head per day of a concentrate ration or to an unsupplemented group withad libitumgrass hay only. Calves were fed in individual boxes in a calf house. The mean birth weight of calves was 24·8 (s.e. 1·8) kg. Suckled calves were heavier (P < 0·01) at weaning (47·7 (s.e. 1·9) v.37·1 (s.e. 1·8) kg) and had faster pre-weaning growth rate (411 (s.e. 30)v.215 (s.e. 29) g/day) than bucket-fed calves. Calves which sucked supplemented cows were heavier (P < 0·001) at weaning (51·2 (s.e. 2·6)v. 44·2 (s.e. 2·6) kg), had faster (P < 0·01) pre-weaning growth rate (477 (s.e. 41)v.346 (s.e. 41) glday) and higher (P<0·01) daily milk intake (2·6 (s.e. 0·2)v.1·6 (s.e. 0·2) I/day) than those which sucked unsupplemented cows. Mean concentrate intake was 302 (s.e. 30) g/day for bucket-fed calves and 341 (s.e. 33) and 408 (s.e. 33) g/day for calves which sucked supplemented and unsupplemented cows. Cow milk production was higher (P < 0·01) in suckled than in milked cows (3·3 (s.e. 0·2)v.1·0 (s.e. 0·2) I/day) and in supplemented than unsupplemented cows (3·9 (s.e. 0·2)v.2·7 (s.e. 0·2) I/day). Concentrations of milk protein, fat and total solids averaged 39 (s.e. 15), 49 (s.e. 16) and 143 (s.e. 21) g/kg, respectively and were not significantly influenced by either concentrate supplementation or suckling. Restricted suckling and early weaning coupled with appropriate calf and cow feeding regimes could be one of the alternatives for calf rearing under smallholder and small scale peri-urban dairy production systems. However, further studies are warranted to evaluate the effects of such a strategy on both calf and cow performances under these production systems.

Breeding programmes for Wakwa and Ngaundere Gudali cattle breeds were initiated in Wakwa (Cameroon) in 1952 and 1965, respectively, to improve growth performance of the indigenous breeds. Animals were selected on 240-day weaning weight (WWT). Data analysed covered the period 1971 through 1985 and consisted of 2211 calf birth weight (BWT) and 1409 WWT records on Gudali and 1196 BWT and 763 WWT records on Wakwa. Selection responses in BWT and WWT were estimated as regressions of average sire estimated transmitting abilities (ETAs) and average dam estimated breeding values (EBVs) on year of calving. ETAs and EBVs were predicted using best linear unbiased prediction procedures. Estimated annual changes in sire ETAs for WWT were 0·67 (s.e. 0·17) and 1·69 (s.e. 1·13) kg/year for Gudali and Wakwa, respectively. Corresponding trends in dam EBVs were –0·03 (s.e. 0·03) and –0·24 (s.e. 0·19) kg/year. Estimated correlated trends in BWT among Gudali and Wakwa sires were 0·09 (s.e. 0·05) and –0·14 (s.e. 0·04) kg/year, respectively. Corresponding correlated trends in BWT associated with dam selection were essentially zero, being –0·001 (s.e. 0·001) and –0·01 (s.e. 0·01) kg/year. Environmental trends were negative for all traits except Gudali WWT. Phenotypic trends were all positive, except that of Wakwa BWT. Overall, selection on WWT yielded moderate selection reponse despite substantial infrastructural constraints in the breeding programme. Trends in sire ETAs and dam EBVs for WWT jointly accounted for estimated genetic gains of 8·60 kg in Gudali and 20·4 kg in Wakwa over the study period. Overall correlated response in BWT during the study period was significant (P < 0·001) and negative for Wakwa. It was positive but not significant for Gudali.

Forty Charolais and 40 Limousin × Holstein Friesian bulls, 8 months old and weighing 315 kg live weight at the start of the experiment, were given grass silage to appetite throughout.

From 315 to 425 kg live weight they received: no supplement (S0), 1 kg (S1), 2 kg (S2) or 4 kg (S4) of a rolled barley supplement per head per day. After a mean treatment live weight of 425 kg had been achieved for each breed, the daily supplement of all treatments was changed to 3 kg of rolled barley per head per day until individual animals were adjudged to have attained a European Community (EC) external fatness score of 3 to 4L, at which point they were slaughtered.

Daily growth rates from 315 to 425 kg live weight were 0·67, 0·82, 0·94 and 1·25 (s.e.d. 0·042) kg for treatments S0, S1, S2 and S4 respectively. Following the change of supplement rate at 425 kg live weight, daily gains were 1·46, 1·36, 1·27 and 1·16 (s.e.d. 0·051) kg/day, giving overall gains of 0·99, 1·05, 1·08 and 1·19 (s.e.d. 0·033) kg/day respectively. The numbers of days taken to achieve slaughter condition from the start of the trial at 315 kg were 277, 251, 243 and 207 (s.e.d. 5·3) days respectively. Bulls were slaughtered at 17·0, 16·2, 15·9 and 14·7 months of age and produced carcass weights of 330, 324, 322 and 305 (s.e.d. 6·5) kg respectively.

Silage dry-matter intakes were 5·9, 5·1, 4·8 and 4·2 kg/day during the initial phase when differential supplement levels were being given and 7·0, 6·8, 6·2 and 5·8 respectively during the finishing period when all groups were being given 3 kg barley supplement daily. Food conversion ratios (kg dry matter intake per kg gain) were 9·1, 7·3, 6·9 and 5·8 during the initial phase and 6·4, 6·9, 6·9 and 7·2 respectively after 425 kg live weight.

It is concluded that barley supplements of 0 to 2 kg/day offered with high-quality grass silage during the growing stage (8 to 12 months of age), followed by at least 3 kg/day will produce finished bulls at an average of 325 kg carcass weight at 16 to 17 months of age. A higher level of 4 kg barley during the growing stage will reduce the slaughter period to approximately 15 months of age but carcass weight will also be reduced. Further studies are required to define the precise level of barley supplement during this earlier period.

Bulls (½ Blonde d'Aquitaine ⅜ Charolais) were used in a 2 × 3 factorial design experiment with two planes of nutrition in the finishing period and three slaughter weights. High and low planes of nutrition were based on diets with similar forage to concentrate ratio (0·40 of dry matter (DM)) offered ad libitum or at 0·78 of ad libitum DM intake at equal live weight respectively. The target slaughter live weights were 550, 625 and 700 kg. Twelve bulls were taken to each slaughter point and an additional five animals were killed as a pre-experimental slaughter group. Mean initial live weight was 412 (s.e. 5·3) kg at a mean age of 342 (s.e. 2·5) days. No significant interactions were found between the main factors. For the high and low planes, live-weight and estimated carcass gains were 1251 and 989 (s.e. 47·7), and 816 and 668 (s.e. 35·3) g/day respectively, the reduction in gains being similar to the proportional degree of nutritional restriction. Plane of nutrition had no effect on live-weight or carcass gain per unit of energy intake. The low plane of nutrition produced significant decreases in body cavity fat depots, subcutaneous fat in the sample joint and increased the proportions of both saleable beef and high-priced joints in the carcass.

With increase in slaughter weight, energy intake per unit of live weight0·75 and rates of both live-weight and carcass gain tended to decline. The dressing proportions were 583,579 and 609 (s.e. 9·1) g/kg for the slaughter live weights of 550, 625 and 700 kg. Carcass conformation improved while fat depots in the body cavity and estimated concentration of separable fat in the carcass increased with increase in slaughter weight, and both estimated lean and bone concentrations decreased. Forequarter as a proportion of total side tended to increase with increase in slaughter weight. Slaughter weight had no effect on concentration of saleable meat or ultimate pH of carcass muscle. It is concluded that bulls of this genotype can be taken to high slaughter weights on diets having a relatively high proportion of forage as grass silage, with high rates of growth and acceptable carcass leanness.

Thirty-six individually penned lambs (mean live weight 32·4 (s.d. 2·27) kg) were offered maize silagead libitumand one of three concentrate mixes, two of which contained extracted rapeseed meal (control and HR) and the other fish meal (FM). The concentrates were given according to live weight and in sufficient quantities to provide proportionately about 0·4 of the dry matter (DM) intake of the lambs. The dietary concentrations of nitrogen (N) g/kg DM were 22·4, 27·4 and 27·5 and of rumen undegradable N 6·6, 7·3 and 11·6 for the control, HR and FM diets respectively.

Both the HR and FM diets depressed maize silage intakes compared with the control during the first 21 days (P < 0·05) and lambs given the FM diet continued to have lower intakes than control lambs (P < 0·05) throughout the experiment. The live-weight gain of HR lambs was considerably depressed in comparison with the control and FM lambs during the first 21 days of the experiment (P < 0·05). Overall HR lambs gained weight more slowly than control and FM lambs up to 45 kg live weight but the difference was not statistically significant. Food conversion ratio was better for FM than HR (P < 0·01). There were no treatment differences in wool growth.

The results obtained provide little evidence that fish meal had any beneficial effects upon lamb growth compared with the control diet except a possible increase in the efficiency of metabolizable energy utilization.

Thirty-six individually penned lambs (mean live weight 32·4 (s.d. 2·27) kg) were offered maize silage ad libitum and one of three concentrate mixes, two of which contained extracted rapeseed meal (control and HR) and the other fish meal (FM). The concentrates were given according to live weight and in sufficient quantities to provide proportionately about 0·4 of the dry matter (DM) intake of the lambs. The dietary concentrations of the nitrogen (N) g/kg DM were 22·4, 27·4 and 27·5 and of the rumen undegradable N 6·6, 7·3 and 11·6 for the control, HR and FM diets, respectively. All lambs were slaughtered at 45 kg live weight and chemical composition of the empty body and some of the component parts determined. A further 12 lambs were slaughtered at the beginning of the experiment to establish body composition before the dietary treatments were imposed.

The HR lambs had lower fleece.free empty body (FFEB) gains than either control or FM lambs (P < 0·05). This reduced gain of HR lambs was particularly associated with a reduction in fat deposition (P < 0·01) such that the FFEBs contained less fat than control and FM lambs (P < 0·01). The efficiency of conversion of metabolizable energy for growth (kg) was worse than both the control (P < 0·01) and the FM (P < 0·001) diets. The FFEBs of HR lambs also contained more ash (P < 0·05) than the lambs receiving the other diets.

The FM diet was associated with greater gains of fat and energy in the guts compared with the control diet (P < 0·05) and FM lambs had a better kg, value than control lambs (P < 0·05). There was no evidence that FM lambs had better N retention than lambs on the other two treatments.

Although breed substitution is a very common method of rapidly altering the characteristics of a livestock population, within-breed selection for lamb carcass characters is becoming more widely used in the terminal sire breeds in Britain. Such selection schemes require accurate and applicable genetic parameters for the carcass traits of interest. This paper reports a genetic analysis of 11 traits measured on 3592 lamb carcasses using a derivative-free restricted maximum likelihood technique. Because of the trial design it was possible to estimate the genetic parameters of the traits at comparable fat cover, carcass weight and age at slaughter, using appropriate within-subclass covariates.

Heritability estimates for muscle depth (0·23, s.e. 0·04), muscle width (0·32, s.e. 0·05), fat depth (0·31, s.e. 0·05), caliper fat (0·23, s.e. 0·04) and conformation (0·29, s.e. 0·05) were similar in the weight and age-corrected results. Heritabilities for fat cover (0·35, s.e. 0·05), daily carcass weight gain (0·63, s.e. 0·07), and leg conformation (0·44, s.e. 0·05) were higher with slaughter at fixed weight than at fixed age. The heritabilities of traits at comparable fat cover were similar to those at comparable age at slaughter with the exception of daily carcass weight gain, fat depth, caliper fat and leg conformation.

The different measures of fat were highly correlated with each other. Conformation was moderately correlated with external fat score, carcass length and muscle depth. Muscle depth was moderately correlated with carcass length. Differences were found between genetic correlations calculated at equal weight, a common point of comparison in breeding schemes, and equal fat cover, a common end-point for slaughter lambs.

A total of 252 lambs was subjected to two contrasting growth rates (high and low) for 7 weeks commencing at 6 weeks of age, factorially arranged with two contrasting growth rates (high and low) from 13 weeks to slaughter. Growth rate was controlled either by weaning lambs at 6 weeks and feeding indoors throughout on differing levels of a concentrate diet (84 lambs in experiment 1) or by grazing at differing sward heights up to weaning at 13 weeks after which weaned lambs were grazed at differing sward heights and where necessary housed from early October and offered differing levels of concentrate and silage (168 lambs in experiment 2). Lamb slaughter weights on each growth rate treatment covered the range 40 to 48 kg live weight.

Increasing the growth rate between 6 and 13 weeks from 213 to 320 glday in experiment 1 and from 191 to 291 g/day in experiment 2 did not significantly affect the fat depth measured at various locations throughout the carcass, the weight of perinephric and retroperitoneal fat, fat grade or the total lipid content of soft tissue dissected from the shoulder joint when lambs were taken to slaughter weights. Increasing the growth rate between 13 weeks and slaughter from 136 to 338 g/day in experiment 1 and from 99 to 146 g/day in experiment 2 significantly increased the depth of carcass fat and fat grade but, though lipid content of the tissue in the shoulder also increased, this difference was not significant.

Carcass fat depth, perinephric and retroperitoneal fat, fat grade and lipid content of the shoulder tissue increased as slaughter weight increased. The effect of carcass weight on carcass fat was greater on the high plane of feeding from 13 weeks than on the low plane. For example, mean fat depth increased by 0·52 mm/kg increase in carcass weight on the high plane compared with 0·13 mm/kg on the low plane.

Comparing the effect of plane of feeding and carcass weight on carcass fat depth the high finishing plane of feeding had an effect equivalent to an increase of 3·8 kg carcass weight. The equivalent figure for the depth of tissue over the 12th rib was 2·8 kg while in relation to fat grade the effect was smaller at 1·0 kg carcass weight.

Twenty-five primigravid ewes were used in a preliminary study to investigate the effect of bovine somatotropin (bST) between 97 and 124 days of gestation, on maternaland foetal growthand development. Five ewes (reference group) were slaughteredat 96 days of gestation, and the remaining ewes were injected daily with saline (control group: no = 10) or 0·1 mg/kgbody weight of bST (bST group: no. = 10). Following bST treatment, five control- and five bST groupewes were slaughtered (slaughter group). The remaining ewes (production group) were allowed to lamb, after which the lambs were slaughtered within 2 h of birth. Bovine somatotropin reducedthe proportion of maternal body fat, but had no major impacton wool growth, maternal gain and gross skeletal measurements, and foetal growth and development. Maternal bST treatment tended to affect the number of placentomes. It should, however, be emphasized that these results are based on small numbers of animals and that further research in this area is required.

Poultry litter has been used as an economical nitrogen (N) source in ruminant diets. However, litter must be processed to eliminate pathogens. Broiler poultry litter was processed by three different deep-stack methods: (1) uncovered, (2) covered with 0.1524 mm clear polyethylene, and (3) aerated by placing perforated pipe 1 m apart within the stack. Stack temperature was measured for 28 days. The litter was used to formulate experimental diets (186 g litter per kg), while the control diet contained urea as the N source. All diets contained 100 g/kg cottonseed hulls, 16 to 17 g/kg limestone, varying amounts of cracked maize and added retinol. The control diet contained 131 g/kg crude protein (CP) while the litter diets contained 125 g/kg CP. The diets were offered to 16 crossbred wether lambs (37 (s.e. 6·1) kg) in individual metabolism crates for 16 days. Blood was collected at the initiation and termination of the trial. Covering litter with plastic decreased temperature ivithin the stack (P < 0·05). Maximum temperature for the uncovered, covered and aerated stacks was: 68°C, 57°C and 72°C, respectively. Dry-matter intake and dry matter, energy and neutral-detergent fibre apparent digestibilities were not different (P > 0·05) among the diets. Apparent digestibility for N was less (P < 0·05) in lambs given uncovered and aerated litter than in those given the control diet. However, N apparent digestibility in lambs given covered litter was not different from that of those given the control diet. Gross energy and N retention of the four diets were not different (P > 0·05). Covering deep-stacked broiler poultry litter improved N digestibility by proportionately 0·15 compared with uncovered litter. Serum chloride was decreased while serum urea was increased (P < 0·05) in sheep consuming the litter-containing diets. Deep-stacked broiler litter should be covered in an air-tight manner in order to maximize its nutritive value.

Energy and nitrogen balance partitioning was studied in sheep given whole cottonseed (WCS), vetch hay and commercial concentrates. Sheep were offered diets containing 0 (CON) and 250 g/kg WCS at maintenance (LF) and twice maintenance (HF) levels. Digested energy, metabolizable energy (ME) and heat production (HP) were measured in a balance trial combined with indirect calorimetry. Dietary ingredients were also incubated in sacco in sheep rumens. Feeding WCS resulted in decreased energy digestibility at HF (0·718 v. 0·751) but not at LF. Loss of energy as methane expressed as a proportion of gross energy intake for CON and WCS respectively was 0·078 and 0·057 at LF and 0·061 and 0·051 at HF. In both diets, HP was similar at LF. In contrast HP amounted to 0·657 of the ME for CON and only 0·61 for WCS at HF.

Apparent nitrogen digestibility was not affected by WCS consumption. Urinary nitrogen was similar for both diets at LF, but at HF it was higher in the WCS diet. Thus at HF, retained nitrogen was 045 and 0·12 of apparently digested nitrogen for CON and WCS respectively.

Ruminal degradation of organic matter was proportionately 0·87 and 0·77 of apparent digested organic matter in the total tract for the CON diet, at LF and HF respectively and 0·80 and 0·73 for the WCS diet, at LF and HF respectively. In the four groups ruminal nitrogen degradation was equal to apparent whole-tract nitrogen digestion. It was concluded that feeding WCS may decrease metabolic HP in ruminants. Utilization of WCS could be improved by taking measures to decrease urinary nitrogen losses.

In three experiments, four grass silages and four grass hays were offeredad libitum to a total of 164 Finn Dorset ewes rearing twins in the 1st month of lactation. In addition, one of three constant daily levels (300, 650 or 1000 g fresh weight per ewe) of a barley-based concentrate was offered. In the first experiment, replacement rate of silage with in vitro organic matter digestibility (OMD) of 0·65 was less than for hay (OMD 0·59), at –0·09 v. –0·62 g forage OM per g concentrates OM. Consequently, lamb growth rate was greater and ewe weight loss was less with increased concentrate level for the silage than for the hay. In the second experiment, the replacement rates of three silages (OMD 0·56, 0·60 and 0·67) were not significantly different (mean –0·29) and, once again, increased concentrate level gave significantly better ewe and lamb performance. Finally, the replacement rates for three hays (OMD 0·51, 0·59 and 0·72) were also not significantly different (mean –0·30). Equations are presented to predict the intakes of hays and silages by ewes in early lactation.

Three experiments were conducted to investigate the effect of genotype on the morphological characteristics, chemical composition and feeding value of nine chopped barley straws, and responses of the feeding value to a supplement of soya-bean meal (SBM). In experiment 1, the straws were offered ad libitum to 12 wethers and measurements were made of straw intake and dry (DMD) and organic (OMD) matter digestibility using a balanced lattice design with three periods. Experiment 2 used the same experimental design but the straws were offered ad libitum with 200 g SBM. Experiment 3 measured the dry matter losses (DML) of ground straws from nylon bags placed in the rumen of three wethers. The extent (b) and rate (c) of DML were estimated using the equation DML = a + b(l – ect).

There were significant differences in the stem height (P < 0·001), leaf proportion (P < 0·001), and protein (P < 0·001) and modified acid detergent fibre (P < 0·01) concentrations between the straws. Significant (P < 0·05) effects of genotype on intake were seen but the variation in intake between them decreased when the protein supplement was added. Effects of genotype on DMD (P < 0·05) and OMD (P < 0·05) were seen in unsupplemented straws and differences of OMD (P < 0·05) were still apparent when SBM was given with straw. Although there were significant (P < 0·05) differences in DML after 48 h incubation, the constants b and c were similar (P > 0·05) across the genotypes. Protein concentration in the straw, although closely correlated with intake, appears to have less value as a predictor of feeding value than previously thought.

Eight non-pregnant, non-lactating adult ewes weighing 56 kg were fed for 26 days on a meadow hay (organic matter 937 g, crude protein 107 g, metabolizable energy (ME) 8·3 MJ per kg dry matter) to satisfy half of their daily energy and protein maintenance requirements i.e. 200 kj ME and 1·3 g truly digestible protein per kg M0·75. They were then allotted to two groups of four each. Control ewes continued to receive the initial diet for another 74 days while the others were supplemented with 60 g/day fish meal to receive 200 kj ME and 2·7 g truly digestible protein per kg initial M0·75 per day.

In spite of its high nitrogen content, the digestibility of the hay was increased by the protein supplementation, e.g. from 0·56 to 0·60 for neutral-detergent fibre and from 0·54 to 0·59 for energy.

The supplemented ewes lost significantly less body weight than the controls (96 v. 134 g/day) essentially because of their higher ME intake (272 v. 252 kj/kg average M0·75 during the experimental period). The measurements of blood metabolites suggest that the daily amounts of mobilized body protein increased with duration of underfeeding. The additional amino acids provided by fish meal were principally used as a source of energy and for wool growth.

Cashmere production was evaluated on Scottish feral (F) goats, on goats imported from Iceland (I), Tasmania (T), New Zealand (N) and Siberia (S), and on two- and three-way crosses between feral and imported lines. Evaluations were based on the weight of cashmere in 10 cm2 mid-side patch samples taken at 5 months of age, with annual cashmere production being predicted from sample cashmere and body weights. Data were collected on 121 purebred and 706 crossbred kids of both sexes. Mean fibre diameter for the F, I, T, N and S lines was 13-75 (s.e. = 0.82), 14·04 (s.e. 0·41), 16·13 (s.e. 0·35), 16·63 (s.e. 0·49) and 17·97 (s.e. 0·50) μm, mean estimated annual cashmere production was 37·3 (s.e. 71·3), 914 (s.e. 31·7), 227·1 (s.e. 28·1), 275·1 (s.e. 42·5) and 579·8 (s.e. 44·7) g, and mean live weight was 15·71 (s.e. 2·16), 17·65 (s.e. 0·93), 16·39 (s.e. 0·83), 16·53 (s.e. 1·28) and 21·9 (s.e. 1·28) kg, respectively. Significant positive heterosis existed between some lines for body weight and cashmere production, with the I line goats consistently showing the largest effects. Combining fibre diameter and cashmere production by their relative economic importance into an index designed to indicate the total value of the fibre produced by each genotype, the cashmere production index, reduced the large production differences between the lines, although the S line was still superior to all other lines. When the cashmere production index was adjusted to account for the economic importance of fibre colour, however, the T and N lines, the only lines which produced white fibre, were comparable to the S line. The cashmere production index for the S line was very sensitive to changes in the relative economic weight for fibre diameter, and if the price differential for high quality (i.e. fine) fibre was increased by a factor of 1·36, or greater, then the T and N lines were superior to the S line. Three-way cross means were estimated from line means and heterosis effects. No cross was consistently superior to all other genotypes, but several of the crosses showed the advantages of potentially producing white fibre as well as having high cashmere production indexes, with their indexes being insensitive to changes in the relative economic weights. Future selection for cashmere production in this population should concentrate on individuals of outstanding genetic merit, regardless of their line or cross.

Electronic feeding systems for progeny testing pigs, allow selection to occur under conditions found on commercial farms. This paper reports on the feeding behaviour of six groups of 10 pigs, balanced for sex and initial body weight (mean starting and finishing weight: 32·1 kg v. 68·5 kg), monitored on such a system, for an average of 38 days. These data represent 26 542 feeder visits for which the pen, identity of the pig, feeder entry and exit times, and the amount of food consumed are known. The results show significant time-of-day effects for frequency of feeder visits, feeding rates, mean feeder occupation time, mean food intake per feeder visit, total food intake and total feeder duration (all P < 0·001). A single peak in feeding behaviour was observed between 13·00 and 16·00 h. Significant pen effects were observed on all feeding variables (e.g. feeder occupation times P < 0·05) with the exception of total food intake (P > 0·05). Individual pigs were found to possess different types of meal regulation, as measured by prandial correlations. Pens were found to have significantly non-random sequences of feeder entries (e.g. yen 1, P < 0·001). The range of feeder visits (three to 69) observed, was higher than reported in any previous study. Physical performance on the system, in terms of gain in body weight and total food intake was best predicted by total feeder occupation time, suggesting that individual pigs could adapt to the physical and social constraints imposed by the system, by altering aspects of their feeding behaviour. The results suggest that feeder access competition resulting from social synchrony and facilitation strongly influenced the feeding behaviour of pigs on this computerized food intake recording system.

Forty-five New Zealand White and 35 Californian doe rabbits, 8 months old and in the third parity of their 1st year of production were used in this study. Thirty-five does were concurrently pregnant and lactating, the remaining 45 does were non-pregnant. Milk samples were collected in colostrum and at the 7th, 14th, 21st, 28th and 35th days of lactation. Neither breed nor pregnancy status affected the does milk composition except for total phosphorus which was significantly lower (P < 0·05) in the pregnant does' milk. However the major components of milk except lactose, were significantly affected by stage of lactation. Both moisture: total solids and calcium: phosphorus ratios were the same (2:1) throughout the lactation. The ratio between retinol concentration in colostrum and in the 5th week was about 10:1. Specific gravity was significantly (P < 0·05) correlated with protein, lactose (positive) and fat (negative).