Over a 3-year period, productivity characteristics and criteria of trypanosomiasis incidence and severity have been monitored by monthly examination of individual N'Dama cattle in villages in The Gambia. From this database, 60 lactating cows in which Trypanosoma congolense or T. vivax had been detected on blood examination (group 1) were compared with 50 cows which had not been found infected with trypanosomes during the monitoring period (group 2). The latter were selected on the basis of comparability of age and stage of lactation to those of group 1 for examining the effect of trypanosome infections on the quantity of milk extracted for human consumption, and on the growth of their sucking calves. Data from a 6- to 7-month period were examined in the analysis.

The quantity of daily milk extracted during the 1st month ot intection (group 1) decreased by proportionately 0·25 in comparison to the amount extracted during the preceding month when parasites were not detected. The corresponding figure in the uninfected controls (group 2) was 0·02. The mean daily milk extracted for human consumption from uninfected cows during a 6-month period was proportionately 0·26 higher than the mean for the infected cows. Growth rates of calves sucking infected and uninfected dams were similar.

These observations indicate that infection with pathogenic trypanosomes of lactating N'Dama cattle causes a reduction in milk production.

In economic terms, it was estimated that the decline in milk extracted for human consumption due to trypanosome infections amounted to an average of £1 per month per cow.

Five randomized-block experiments were carried out over 2 years using British Friesian cows managed as three separate herds. The effects of offering cows different levels of concentrates, ranging from 0·8 to 7·2 kg/day, in addition to grass silage ad libitum during the winter period on reproductive performance was investigated. The effects of substituting 4 kg/day maize gluten, or 0·8 kg/day fish meal, for part or all of the standard concentrate were also examined.

Neither level of concentrates nor the inclusion of maize-gluten meal significantly affected reproductive performance even where milk production and quality was considerably influenced. The inclusion of fish meal improved conception rates to all services (0·64 v. 0·44; P < 0·05) and reduced the number of services required per conception (1·62 v. 2·31; P < 0·01).

Plasma urea levels were raised consistently by the feeding of fish meal and, with the exception of weeks 2 and 6 of lactation, by the feeding of higher levels of concentrates. From the 6th week of lactation levels of beta-hydroxybutyrate (BHB) in the blood were significantly higher in the group of cows receiving the highest level of concentrates. The BHB level at week 6 was correlated with the number of services required per conception.

A number of relationships between production factors and fertility are also presented.

The published theoretical estimates of equilibrium genetic change possible by progeny testing in dairy cattle are too high. Much of the overestimation is because no account has been taken of the temporary reduction in genetic variation due to selection of parents and ancestors. A particular efficient progeny testing scheme was modelled theoretically and simulated on computer using a continuous infinitesimal additive model. To make the simulation feasible an elite breeding herd of females was used, and progeny testing was done outside the herd. The proportional reduction in genetic variance due to previous selection was 0·25, 0·27 and 0·29 for original heritabilities of 0·05, 0·25 and 0·50, with the theoretical and simulated results being very similar. The theoretical and simulated proportional losses in equilibrium selection responses were 0·22 and 0·26, 0·21 and 0·22 and 0·21 and 0·25, respectively, for the three levels of heritability. Thus, most of the loss in response in the simulation results came from the reduction in genetic variance due to selection.

A long-term experiment (first 24 weeks of lactation) with Danish Black and White cows and a short-term experiment with Jersey cows (experiment periods of 6 weeks) were carried out. In the long-term experiment, the effect of supplementing a normal dairy ration with 0·5 kg calcium (Ca) soap, 10 kg Ca-soap or 2 kg rolled barley was investigated in two herds, where silage was given ad libitum. The daily intakes of fatty acids in the diets were approximately 625, 1000, 1400 or 650 g respectively. Energy intake increased markedly by the supplementation, but not differently between soap or barley supplementation. The effect on milk production was not consistent in the two herds. In one herd, the soap supplement tended to increase milk fat yield (80 g/day) but not more than the barley supplementation which, in contrast to the fat supplementation, also increased milk protein yield. The effect of the fat was consistent over the first 6 months of lactation. In the other herd, only small differences in milk production were obtained with either supplements. In the short-term experiment carried out in mid lactation, a supplement of 0·8 kg Ca-soap increased daily intake of fatty acids from 730 to 1280 g per cow daily. Total food intake, milk and milk fat yield were not affected by the supplement but milk protein yield and live-weight gain were significantly reduced.

Twenty-four lactating British Friesian type cows were used in a four period, partially balanced, changeover design experiment to evaluate eight treatments consisting of four protein sources in the supplement (soya-bean meal, formaldehyde treated soya-bean meal, fish meal and a mixture supplying equal crude protein levels from fish and soya-bean meal (respective degradabilities of total supplement nitrogen (N) were 0·56, 0·37, 0·32 and 0·44), each offered in supplements formulated to contain 150 and 200 g crude protein per kg fresh weight in a 4 × 2 factorial design. All supplements were offered at 8·0 kg/day in addition to ad libitum access to a medium digestibility grass silage (digestible organic matter as proportion of dry matter (DM), 0·66); with a relatively poor quality fermentation (pH, 4·5; ammonia-N proportion of total N, 0·21).

There were no significant interactions between source and level of protein on any of the animal performance variables. Source of protein did not influence milk output or milk composition but increasing protein intake significantly increased milk yield (26·0 and 26·7 (s.e. 0·19) for the low and high protein supplements respectively). Blood urea levels were significantly influenced by both protein source and level.

Total diet digestibility and energy and N utilization studies were undertaken with six replicates per treatment and the full data are presented. Neither protein source nor level had any significant effect on DM or energy digestibility. N digestibility was significantly reduced following formaldehyde treatment of soya-bean meal (0·69 v. 0·72 (s.e. 0·01)) and most N utilization variables were influenced by the level of protein intake but not by protein source.

Theoretical rates of genetic response are calculated for adult multiple ovulation and embryo transfer (MOET) nucleus breeding schemes and conventional breeding systems in dairy cattle. For a single trait breeding objective, the relative responses with MOET schemes decline as the heritability falls and the number of progeny in the progeny test scheme increases. Selection for an aggregate breeding objective of milk yield and another trait, for different values of heritability, correlation and relative economic importance, is also examined. Generally, the overall response of efficient MOET schemes exceeds the progress of an efficient conventional breeding scheme with 50 to 100 daughters per bull. Only if the relative economic weight for the lowly heritable trait exceeds 0·60 and the progeny test is on 200 offspring are conventional systems superior. Testing additional paternal half-sib groups from a MOET scheme gives a considerable extra response, especially when lowly heritable traits have a dominating economic importance.

Six steers, each fitted with a ruminal cannula and T-type duodenal and ileal cannulae, were used to investigate protein and fibre utilization from lucerne, birdsfoot trefoil and sainfoin preserved both as hay and silage. Steers were fed at 2-h intervals during six 12-day periods in an experiment with a 6 × 6 Latin-square design. Animals were fed all-forage diets of six treatments (lucerne hay and silage, birdsfoot trefoil hay and silage, sainfoin hay and silage). Samples were collected from all alimentary sites twice daily during the last 3 days of each period. Acid insoluble ash was used as a solid marker and Cr-ethylene diamine tetra-acetic acid as a liquid marker. Organic matter (OM), hemicellulose (HC) and cellulose (CL) total tract digestibilities of lucerne, birdsfoot trefoil and sainfoin were similar (OM: 599, 608, 580 g/kg; HC: 499, 497, 480 g/kg; CL: 590, 618, 608 g/kg). However, crude protein (CP) digestibilities were lower (P < 0·001) for sainfoin (582 g/kg) than for lucerne (732 g/kg) or birdsfoot trefoil (693 g/kg). Nitrogen (N) flows at the duodenum or ileum were similar among forages despite lower N intake for sainfoin. Presence of tannins in sainfoin may have been responsible for low protein degradation in the rumen, and reduced N digestion in the small intestine. Forage preserved as hay or silage had similar CL, HC and OM digestibilities, while CP digestibility was higher (P < 0·05) for silages.

One hundred and sixty-two spring born steers comprising 54 Charolais × Friesian (CH), 54 Hereford × Friesian (HE) and 54 Friesians (FR) were reared together from shortly after birth until the end of their second grazing season at a mean age of 573 days. The animals were then blocked on weight within genotype, housed in a slatted shed and allocated to either a high (12·6 MJ metabolizable energy (ME) per kg) or a medium (10·4 MJ ME per kg) energy diet offered ad libitum until slaughter at one of three target carcass weights · light (260 kg (HE, FR and CH)), normal (300 kg (HE and FR) and 320 kg (CH)) or heavy (340 kg (HE and FR) and 380 kg (CH)) to give a 3 (breed type) · 2 (energy level) · 3 (slaughter weight) factorial arrangement during the finishing winter. The trial was repeated in 3 years consecutively.

CH were significantly heavier than HE at all times up to 18 months of age when they were 43 kg heavier, at which stage they were also 24 kg heavier than FR. CH grew significantly faster at pasture than HE in both seasons and faster than FR in the second season. HE had the best daily gain in the finishing period.

Overall slaughter weights per day of age, carcass weights per day of age and carcass weights of CH, HE and FR were 812, 771 and 761 (s.e. 9·2) g, 448, 418 and 405 (s.e. 4·8) g and 318. 299 and 298 (s.e. 2·2) kg, respectively. CH had significantly better killing-out proportions (lOg/kg) than HE which in turn were 10 g/kg better than FR. CH and FR had similar carcass length, carcass depth, leg length and leg widths and both were significantly larger than HE. CH had significantly larger longissimus dorsi areas than either HE or FR (80·9 v. 67·0 and 66·4 cm-); they also had significantly better conformation scores than HE which in turn were significantly better than FR.

Increasing the energy level during the finishing period increased daily gain by proportionately 0·52 (1089 v. 684 g/day), reduced slaughter age by 51 days (711 v. 762 days) and significantly increased killing-out proportion (551 v. 534 g/kg) and conformation score (4·86 v. 4·64). Offal as a proportion of empty live weight tended to be higher for FR than for HE or CH, the exception being the hide, which was significantly lower for FR.

Crude protein concentration in diets based on high-moisture maize grain was increased using rumen degradable protein (RDP) or rumen undegradable protein (UDP). Treatment diets (numbers in parentheses indicate approximate RDP and UDP concentrations, g/kg dry matter (DM)) were: low-RDP (59, 80), high-RDP (79, 92), low-UDP (66, 67), high-UDP (69, 107, with maize gluten meal as chief UDP source) and high-UDP (71, 91, with fish, meat and maize gluten meals as UDP sources). Rumen pH and concentrations of volatile fatty acids, measured hourly for 12 h after feeding in four rumen fistulated Holstein steers, were unaffected by treatment (P > 0·05). Rumen ammonia-N concentration was greater for high-RDP (193 mg/1) than low-RDP (68 mg/1; P < 0·001). Within 12 h of feeding, ammonia-N concentration declined from 169 to 58 mg/1 for low-RDP and remained stable for high-RDP.

In a 56-day feeding trial using Holstein heifers (live weight 128 kg) and the first four treatments, RDP concentration had no effect on DM intake (P > 0·05). Lower DM intake (P < 0·01) for high-UDP (4·40 kg/day) than low-UDP (5·13 kg/day) was attributed to low palatability of maize gluten meal. Live-weight gain and food conversion efficiency were unaffected by treatment.

In a 56-day feeding trial using Holstein bulls (live weight 110 kg) and all treatments, RDP concentration had no effect on animal performance (P > 0·05). Food conversion efficiency increased (P < 0·001) when UDP concentration was increased only if UDP was supplied by a mix of fish, meat and maize gluten meals.

When RDP supply was not limiting to DM intake, protein source (amino acid composition) was better than protein degradability as an index of the feeding value of dietary protein.

Two groups of six bulls were offered a fattening diet based on barley or sugar-beet pulp in the proportions of either 500: 200 g/kg or 200: 500 g/kg to compare the effects of starch and degradable fibre on rumen fermentations, plasma hormones and metabolites. The bulls were fed twice daily at 08.00 and 16.00 h.

The large proportion of sugar-beet pulp in the diet induced higher levels of acetic and butyric acids and less propionic acid than the high proportion of barley. The diurnal patterns over a 24-h period of rumen pH, volatile fatty acids, glucose and ammonia were characterized by two cycles. Sugar-beet pulp in large proportions induced more even fermentations as indicated by flatter curves with less extreme values than those produced by a large inclusion of barley in the diet. Ammonia concentration was high before feeding and decreased subsequently during 4 to 6 h after the meal.

No typical patterns were observed in plasma concentrations of glucose and alpha-amino nitrogen. Plasma urea concentration rose 2h after feeding and dropped during the following 6 to 8 h. There were no effects of dietary carbohydrate composition on growth hormone concentration but plasma insulin was significantly reduced with the diet high in sugar-beet pulp. Furthermore, the insulin profile showed peaks after each meal with the barley diet but only after the afternoon feeding with the diet high in sugar-beet pulp.

The carcass composition of Hereford × Friesian (HE), Friesian (FR) and Charolais × Friesian (CH) steers finished on diets of high (H) and medium (M) dietary metabolizable energy (ME) concentrations and slaughtered at low (L), normal (N) and heavy (W) carcass weights was examined in a 3 × 2 × 3 (no. = 9 per individual treatment) factorial experiment. A pre-finishing slaughter group of nine animals of each breed type was also included, giving a total 189 animals in 21 experimental groups. ME concentrations of the H and M diets which were offered ad libitum were 12·6 and 10·4 MJ/kg dry matter. Target carcass weights for L, N and W were 260, 300 and 340 kg for HE and FR and 260, 320 and 380 kg for CH.

Carcass side weights (before tissue separation) of the HE, FR and CH pre-finishing slaughter groups were 90·8, 970 and 101·0 (s.e.d. 3·9) kg. Corresponding tissue proportions were 188, 199 and 200 (s.e.d. 4·3) g/kg bone, 663, 686 and 690 (s.e.d. 5·3) g/kg muscle and 135, 99 and 96 (s.e.d. 5·7) g/kg fat. Main effect side weights of the finished groups were 152·3, 151·4 and 162·2 (s.e.d. 1·4) kg for HE, FR and CH. 158·5 and 152·2 (s.e.d. 1·1) kg for H and M and 131·2, 155·2 and 179·6 (s.e.d. 1·4) kg for L, N and W, respectively. Tissue proportions in the same order were 146, 160 and 157 (s.e.d. 2·0), 149 and 159 (s.e.d. 1·6) and 163, 154 and 146 (s.e.d. 2·0) g/kg bone, 579, 601 and 635 (s.e.d. 5·5), 600 and 610 (s.e.d. 4·5) and 637, 599 and 574 (s.e.d. 5·5) g/kg muscle and 264, 228 and 195 (s.e.d. 6·4), 240 and 219 (s.e.d. 5·2) and 188, 235 and 264 (s.e.d. 6·4) g/kg fat. CH had more (P < 0·001) of their muscle in the pelvic limb and less (P < 0001) in the thorax than HE and FR. HE had more (P < 0·001) of their carcass fat in the subcutaneous depot and less (P < 0·001) in the intermuscular depot than FR and CH. The allometric regression coefficients for the main joint and tissue weights on side weight were <1·0 for both limbs, loin, bone and muscle. Coefficients were >1·0 for the thorax, ribs, flank and fat. The regression coefficients for the main muscles of the pelvic limb and loin and total thoracic limb muscle on total side muscle were <l·0, while the coefficients for the flank, ribs and thorax muscles were >1·0. Similarly the regression coefficients for the bones of the two limbs on total side bone were <10, while the coefficients for the loin, ribs and thorax bones were >10. It was calculated (for the H diet) that at a carcass weight of 300 kg, HE, FR and CH would have carcass tissue proportions of 576, 600 and 642 g/kg muscle and 261, 227 and 180 g/kg fat. The three breed types would have similar carcass fat contents at carcass weights of 264, 300 and 376 kg for HE, FR and CH, respectively.

The extra energy for walking compared with standing still (EW) (J/m per kg live weight) was measured in three Brahman cattle and two water buffalo. Ew was not affected by species or speed within the most comfortable range of speeds (V = 0·6 to 1·0 m/s) but over the whole range tested, Ew = 0·947F + 1·99 (r = 0·66, no. = 61) with average Ew = 2·1 (s.e. 0·06).

The extra energy cost of carrying loads while walking (Ec) (J/m per kg carried) was measured using two Brahman cattle, two water buffalo and a pony. Ec was independent of load (up to 70 kg) and speed but was generally lower when loads were placed over the animals' shoulders instead of on their backs. Average values for the cattle, buffaloes and the pony were 2·6, 4·2 and 3·3, respectively.

The efficiency of doing work defined as: work done/energy expended was measured in two Brahman cattle and two water buffalo and gave average values of 0·30 and 0·37 respectively for the two species. Efficiency was proportionately about 0·03 higher for animals wearing a collar than when wearing a single yoke but was unaffected by whether the animals wore single or double yokes, by the speed of travel, the size of the load or whether the load was steady or variable.

Along with appropriate values for the energetic efficiency of raising body weight when walking uphill, these data are used to derive a factorial equation for estimating the energy expenditure of animals working in the field.

Breeding data from 1543 artificial inseminations, performed on 763 purebred and crossbred zebu cattle reared on 73 small farms, in the Amazon basin of Peru from 1980 to 1986, were evaluated. Inseminations were performed during morning hours using Holstein or Brown Swiss frozen semen. The calving to first oestrus and calving to conception intervals, number of services per conception, intervals between services and conception rate were evaluated. A total of 64-4% of first services and 61-4% of conceptions occurred between the middle of the dry season and the middle of the rainy season. Farm effect proved to be the most important source of variation for all reproductive traits. Other important factors influencing the interval from parturition to first service were breed of cow, parity and season of service. Parity influenced the interval from parturition to conception while parity, time of oestrous detection and inseminator influenced conception rate. The long post-partum anoestrous interval was found to be the most important factor limiting reproductive efficiency on small farms involved in the artificial insemination programme. The reproductive performance of the crossbred cattle tended to be better than that of the purebred zebu cattle.

For the Mpwapwa cattle breed in Tanzania, the efficiency of various breeding schemes including an open nucleus was investigated by model calculations. Artificial insemination and intensive recording of production are assumed to be applied in a nucleus which is the main breeding unit. As a pre-nucleus, associated herds with less intense data recording serve as the basis to screen superior cows for nucleus replacements, provide the capacity for progeny testing, and operate as bull multipliers for commercial herds.

The criteria of efficiency were genetic gain and profit from selection for a dual purpose breeding objective (milk and beef) in a total population of 10 000 cows. Introducing a two breeding tier scheme through separating all recorded cows into a nucleus and pre-nucleus leads to an increase of the genetic gain rather than the profit. Further improvement is obtained by introduction of artificial insemination in pre-nucleus herds. The nucleus size should not exceed about 5% of the cow population and an optimum size of the pre-nucleus is about 15%. Opening the nucleus to replacement cows coming from the pre-nucleus affects the aggregate genetic gain very little although it can be recommended if milk yield is to be mainly improved or if the total profit is taken into account. Furthermore, the nucleus should be opened if there is little difference between the heritabilities in the nucleus and pre-nucleus and also in order to avoid detrimental effects of inbreeding and genotype × environment interaction.

Correlated responses to selection in a tropical beef herd for high (H) and low (L) estimated breeding value for pregnancy rate were measured over three calf crops. The H line calves were proportionately 0·04 lighter at weaning but during the ensuing dry season they lost less weight than L line animals, so that between the ages of 12 and 18 months there were no significant line differences in live weight. It was concluded that selection for cow fertility had not directly altered the progeny's growth rate. Tick resistance and heat tolerance also did not differ between the lines. However, resistance to worms was lower (P < 0·05) in the H line indicating genetic antagonism between cow fertility and worm resistance. Scrotal circumference and scrotal circumference adjusted for live weight were significantly higher in H line bulls between 9 and 18 months of age. Pregnancy rates of heifers in the H line were 0·12 higher than in the L line despite similar average live weights at mating. It is suggested that accelerated sexual maturity in both heifers and bulls has occurred as a result of selection for lifelong cow fertility.

Intake, rumen digesta pool sizes, ruminal digestion rates, passage rates of liquid and particle pools and faecal particle-size distrubutions of grass, browse and lucerne diets were compared in ruminally fistulated moose, wapiti and cattle. For each diet, ruminal digestion rates (nylon bag technique) were similar for the three ruminants. Intakes, rumen digesta pool sizes and passage rates varied among species and higher rates were associated with passage of larger faecal particles. However, it was not possible to simply rank the three ruminants as representative grazers (cattle), browsers (moose) and mixed feeders (wapiti) since intakes, passage rates and rumen pool sizes interacted with diet and season. Contrary to expectation, digestible dry matter intakes of each species were not greatest for their respective typical diets.

To separate ingestive factors involving palatability from post-ingestive factors responsible for forage intake being depressed by ensiling, silages were evaluated using sham-fed animals. Protocol for all experiments was to formulate low dry matter (DM) silages from a high DM lucerne silage reconstituted with either water, solutions containing constituents commonly found in silage, or extracts of low DM silage. Solutions of silage constituents were adjusted before reconstitution to a pH typical of silage. Tests were performed by offering silages one at a time once daily to oesophageal fistulated sheep, maintained on lucerne hay, but starved for 5 h prior to the test. Intake was measured over 30 min. Sheep were adapted to intermediate levels of each test ingredient(s) before experiments were done to evaluate effects of four to five levels of each ingredient using Latin-square designs. Animals were sham fed during both periods of adaptation and of testing ingredient levels. Mean of 30-min intakes for control silages was 573 g DM. Lactic and acetic acids added together, to levels of 53·2 and 35·4 g/kg, respectively, enhanced intake with a linear response due to level (P < 0·05). Acetic acid added by itself to 88·0 g/kg depressed intake with a linear response due to level (P < 0·05) in which 10 g/kg added acetic acid depressed 30-min intake by 13·5 g. Acetic and butyric acids added together to levels of 36·2 and 54·4 g/kg respectively, were without effect as was ammonia added to 3·8 g/kg as ammonia nitrogen (N). A mixture of free amino acids added to a level of 9·2 g amino-N per kg was without effect, but at a level of 13·8 g/kg, intake was depressed to proportionately 0·75 of the control. A mixture of two amines and gamma amino butyric acid added to a level of 4·6 g N per kg caused a quadratic response (P < 005) in which intermediate levels enhanced intake but the highest level was without effect. An extract from good-quality silage had no effect on intake and that from a poor-quality silage enhanced intake in a linear manner (P < 0·05). These data indicate that elevation of acetic acid in silage without increases in amounts of other constituents decreased intake of silage through an effect on palatability. This conclusion modifies interpretation on the significance of palatability, determined from studies of ingestive behaviour of ruminants, as a factor to explain reduced intake of sheep given silage.

Data on individually recorded silage dry-matter intake (SDMI), concentrate dry-matter intake (CDMI) and live weight of steers and data on silage composition including toluene dry matter (TDM), pH, total nitrogen (N), ammonia nitrogen (NH3-N), volatile fatty acids (VFAs), digestible organic matter in the dry matter (DOMD) and neutral-detergent fibre (NDF) obtained from experiments conducted at three sites were used to obtain simple and multiple linear regressions of SDMI on other variables.

Live weight accounted for a high proportion of the variation in intake but this effect could generally be removed by scaling intake by live weight raised to the power of 0·75 (M0·75). CDMI was the most important factor affecting scaled intake in mixed diets. TDM, NH,-N and VFAs all had important effects on SDMI. The relationship of SDMI with TDM was curvilinear suggesting that there is little to be gained in intake terms from wilting to TDM above 250 g/kg. The effect of NH3-N appeared to be related more to its correlation with VFAs than with any other nitrogenous constituent while the VFAs appeared to have a direct effect on SDMI. The effects of N and pH on SDMI were generally small. DOMD and NDF had relatively little effect on SDMI. Significant differences in intercepts between sites were found for most relationships although common slopes were often found.

Data on individually recorded silage intake (SDMI), concentrate intake and live weight of steers and data on silage composition including toluene dry matter, pH, total nitrogen, ammonia nitrogen, volatile fatty acids, digestibility and fibre measures obtained from experiments at three sites were used.

Correlation and principal component analyses indicated that there was severe collinearity among a number of the variables, particularly among various fermentation characteristics, between different measures of digestibility and between different measures of fibre. This collinearity was shown to have caused instability in least-squares multiple regression coefficients of silage intake on other variables previously obtained from the same data (Rook and Gill, 1990). Principal component regression and ridge regression allowed the derivation of new coefficients which were more stable and more in line with a priori expectations from experimental results. Linear functions of the original variables and the use of models containing fewer variables also proved effective in overcoming the problems of collinearity.

Fermentation characteristics were shown to be the most important silage factors affecting intake and butyric acid was shown to be more important than other volatile fatty acids or ammonia nitrogen. Neutral-detergent fibre was found to be better related to intake than other fibre or digestibility measures.

Nine male and nine female lambs each from the Karagouniko (K), Chios (C) and East Friesland (F) breeds were weaned at a mean age of 50 days and were killed over a commercial range of live weight from 16 to 44 kg (carcass weights 6 to 22 kg).

Carcass weight was linearly related to fasted live weight prior to slaughter (r = 0·994). There were significant but small differences between breeds in mean weight of lungs, pelt and internal fat. The carcass components of bone, muscle and fat tissues were related to carcass weight by linear or quadratic regression. Adjusted to mean carcass weight, the mean weight of bone did not differ between breeds (2·82 kg, s.e. = 0·056 at 12·8 kg carcass). The K and C lambs had significantly less muscle than the F lambs (6·77, 6·63 and 8·20 kg, respectively, s.e.d. = 0·094) and more fat (2·87, 2·95 and 1·29 kg, respectively, s.e.d. = 0·017) when compared at mean carcass weight. The breed differences were significant over the whole range whether compared at equal weight or at an equal proportion of estimated mature weight. At the same proportion of carcass fat, corresponding to the United Kingdom market average (250 g/kg) the calculated carcass weights were 16, 17 and 30 kg, respectively, for K, C and F lambs.

It is concluded that the Friesland breed differs from many others in the proportions of muscle or fat at corresponding carcass weights but this difference is not a general characteristic of dairy sheep. Comparison of breeds at defined proportions of carcass components, as determined by market standards, can be effectively achieved by serial slaughter over a suitable weight range.