The positive effect of administration of growth hormone (GH) on milk production and its presence in selected dairy cattle lines of higher GH concentrations prompted an examination of the presence of restriction fragment length polymorphism (RFLP) in the growth hormone gene using the enzyme Taq/ and to investigate associations between this polymorphism and milk production traits. Blood was sampled from 251 Italian Holstein-Friesian cows. Three fragment bands, arbitrarily denoted A, B and E, of 6·2, 5·2 and 1·9 kilobase (kb), respectively, were observed. Their combinations exhibited six patterns, AA, AB, ABE, AE, BB and BE with frequencies of 64·5, 24·3, 2·4, 6·8, 1·4 and 0·4%, respectively. The statistical analysis was performed using linear mixed animal models. The results indicated an effect of the GHgene polymorphic Taq/restriction fragment on 305-day productions of milk, fat and protein; the low frequency pattern AE showed productions inferior to those for AA or AB patterns. Effect estimates of AA, AB and AE were, respectively, 200 (s.e. 215), 218 (s.e. 267) and -910 (s.e. 380) kg for milk production, 7·75 (s.e. 7·98), 16·10 (s.e. 9·79) and -22·14 (s.e. 14·42) kg for fat production, and 6·78 (s.e. 6·21), 8·57 (s.e. 7·58) and -20·74 (s.e. 11·14) kg for protein production. The average substitution effect estimates of E were -891 (s.e. 278), -26·56 (s.e. 10·16) and -24·50 (s.e. 9·43) kg for milk, fat and protein yields, respectively. In conclusion, these results suggest that the E fragment deserves further designed and specific study.

Acetone concentration in milk is correlated with subclinical and clinical ketosis and also with milk yield and reproductive efficiency. The purpose of this study was to assess a threshold of milk acetone at which cows tend to be adversely affected. The critical values were estimated on data from 11690 lactations where the change in daily milk yield from weeks 1 to 2 to weeks 5 to 6 of lactation was evaluated in relation to different milk acetone concentrations. Daily milk yield tended to be reduced in cows with acetone concentrations between 0·7 and 1·4 mmol/l, but responses were not entirely consistent. At concentrations >l·4 mmol/l, daily milk yield was clearly reduced. Long-term milk yield (day 0 to 100 and day 0 to 200) was about 190 kg of 40 g/kg fat-corrected milk (FCM) less in cows with acetone concentrations >l·4 vis-à-vis <0·7 mmol/l. Reproductive efficiency was also impaired in cows with acetone concentrations >l·4 mmol/l; 4·9 days longer interval from calving to first service (first parity and parity 2+ cows) and 5·7 times a greater risk of cystic ovaries (parity 1), compared with cows with milk acetone <0·7 mmol/l. However, no significant effects on long-term milk yield or on reproductive efficiency were found in cows with acetone concentrations 0·7 to 1·4 vis-à-vis <0·7 mmol/l. The results suggest that 1·4 mmol/l acetone in milk may be used as the most important critical value, as higher concentrations are detrimental to productivity. The interval 0·7 to 1·4 mmol acetone per I milk may be used as a warning class, since early lactation yield may be reduced.

Five published equations were compared on their ability to adjust different patterns of lactation curves. The equation y(t) = a exp(b1t'2/2 + b2/t – c(l + t'/2)t') with t' = (t – 21·4)/l00 (i) was retained because of the quality of the adjustment and the absence of convergence problems when applied on individual curves. Including an effect of season (SE) and an effect of pregnancy (PE) improved the quality of individual adjustments (no. = 339). The final equation was equal to (i) + SE + PE with SE = a + b cos (2π(wc + w - l)/52) + c sin(2n(wc + w - 1)/52), with wc= week of the year at calving, w = week after calving and a = 0·0065, b = -1·26 and c = 0·374, and PE = a (wp –18) e-bwp, with w p = week of pregnancy and a, b fixed parameters. Individual cow characteristics were used to analyse equation (i) parameters. The predictions obtained with this equation and several predictive functions of the equation (i) parameters, which differed in the way they included or not the potential of production, were compared. With no indication of this potential, the prediction was very poor. With the initial production (mean production of the 4th, 5th and 6th days of lactation) as an estimate of this potential, 75% of the lactations had the median of absolute values of errors less than 2·95 kg/day. The results were highly improved by using the yield during the 5th week of lactation. The threshold of 2·95 was reduced to about 2 kg/day. The quality of the individual prediction was better for primiparous than for multiparous cows, and for French Friesian and Montbeliarde cows than for pure or crossbred Holstein cows. Although individual predictions were not always satisfactory, they provided excellent agreement when averaged per group (20 cows).

Effects of feeding a protein supplement to dairy cows during the dry period on performance during the following lactation were investigated in two experiments. Holstein-Friesian cows were paired towards the end of lactation, and, after drying off, one of each pair received a typical dry cow management regime of ad libitum grass silage (experiment 1), or a mix of grass silage and distillers' grains or pressed beet pulp (experiment 2). The other cows were offered restricted access to the same basal diet, together with ad libitum access to barley straw and 0·5 kg/day high protein maize gluten meal. During the following lactation, animals from both groups were treated without reference to dry period treatment, and were offered equal access to the same lactation diet. Data were analysed by analysis of variance of experiment means and by parallel curve analysis using sample means. In experiment 1, milk yields were similar (27·2 v. 27·9 (s.e.d. 2·12) kg/day for control and supplemented animals respectively) but milk protein yields, and hence concentrations, were significantly higher (P < 0·001) from supplemented animals (28·9 v. 31·8 (s.e.d. 0·58) g/kg). In experiment 2, milk yields were significantly higher (P < 0·001) from supplemented animals (mean 33·3 v. 35·4 (s.e.d. 1·66) kg/day; however, milk protein yields were also significantly increased (P < 0·001) and the change in milk protein concentration was small. No difference in dry-matter intake was recorded in a subset of animals during early lactation in experiment 2. It is hypothesized that the maternal labile body protein pool was maintained or replenished during the dry period by the provision of the protein supplement, and that this had a significant effect on subsequent lactation performance.

In a lifetime study of spring born cattle managed in a 20-month beef system, four feeding treatments were imposed during the 5-month winter period and two grass heights (low 6 to 8 cm, high 8 to 10 cm) during the subsequent grazing period. Three slaughter weights were imposed over the grazing period, early, mid and late, at an average of 67,110 and 154 days post turn-out. A multi-factorial design was used with three animal factors — maturity (early maturing Hereford crosses v. late maturing Charolais crosses), sex (heifer v. steer) and method of rearing (suckled calves v. bucket-reared calves). There were significant differences in growth rate for both sex and maturity (P < 0·001) and a highly significant negative effect of winter food level on summer growth rate (P < 0·001), the growth rate of food treatment 4 being proportionately 0·61 of that treatment 1.

Sward height significantly influenced summer growth rate (P < 0·001) but showed no interaction with winter food level in any of the three grazing periods. Growth rates increased over the summer but differences between winter food treatments decreased with daily gains for food treatment 4 being proportionately 0·44, 0·81 and 0·84 of food treatment 1 as the grazing season progressed.

Eliminating winter feeding treatment as a factor and including condition score at turn-out as a co-variate improved the variation explained and reduced differences in growth rate for the main effects with only the main effect of grass height remaining significant. This suggests that the condition of animals at turn-out in conjunction with the subsequent grazing sward height provides a simple practical guide to subsequent animal performance.

In each of three experiments, 60 Limousin × Holstein Friesian bulls, initially of 191, (s..e. 5·0), 177 (s.e. 5·5) and 210 (s.e. 7·2) kg mean live weight for experiments 1, 2 and 3 respectively, were offered maize silage ad libitum plus one of three barley-based, isonitrogenous compound diets containing 200 g rapeseed meal, 575 g maize-gluten feed or 100 g fish meal per kg fresh weight. Compound diets were given at the rate of 3·0 kg (fresh weight) per head per day up to 320 kg live weight and at 3·3 kg per head per day from 320 kg to 420 kg live weight. In experiment 3, when bulls were taken through to slaughter, the compound diet was further increased to 3·5 kg per head per day from 420 kg live weight to slaughter. Experiments 1 and 2 were conducted over 128 and 194 days, with mean live weights at the end of the experiments of 347 (s.e. 17·2) and 414 (s.e. 24·2) kg respectively. In experiment 3, bulls were slaughtered at 16 months of age at a mean live weight of 544 (s.e. 25·0) kg.

In experiment 1, the fish-meal diet gave the highest rate of daily live-weight gain at 1·34 compared with 1·17 and 1·16 (s.e.d. 0·041) kg for rapeseed-meal and maize-gluten diets, respectively. In experiments 2 and 3, daily gains were similar on all treatments at 1·21, 1·19 and 1·26 (s.e.d. 0·036) kg and 1·06, 1·09 and 1·11 (s.e.d. 0·035) kg for the rapeseed-meal, maize-gluten and fish-meal diets respectively. Total daily dry-matter intakes did not show consistent trends: in experiment 1, intakes were 5·8, 5·6 and 5·8 kg/day, in experiment 2, 6·2, 6·2 and 6·3 kg/day; and in experiment 3, 7·1, 7·3 and 7·1 kg/day for rapeseed-meal, maize-gluten and fish meal, respectively. The higher intakes of dry matter in experiment 3 were due to the greater intakes of maize silage by bulls taken to heavier weights in this experiment.

Jugular and ovarian venous profiles of progesterone and endometrial progesterone concentrations were measured in pregnant and non-pregnant ewes. Mean jugular concentrations were not significantly different (3·12v.2·39 μg/l). The incidence of progesterone concentrations above a chosen threshold was significantly greater in the ovarian vein than in the jugular vein (P < 0·05) and was greater (P > 0·05) in the ovarian venous plasma of pregnant than of non-pregnant ewes. Mean endometrial progesterone concentrations were significantly higher in pregnant than in non-pregnant ewes (T < 0·05). It is suggested that the lower progesterone concentrations in the ovarian veins of non-pregnant ewes may have a rôle in pregnancy failure and that ovarian venous concentrations may influence endometrial progesterone concentrations.

The responses in traits measured at the start of the mating period and at farrowing were examined after seven generations of divergent selection for daily food intake (DFI), lean food conversion (LFC), lean growth rate (LGA) on ad-\ibitum feeding and lean growth on scale feeding (LGS). Weight and backfat depth at mating were measured on 330 gilts and 74 boars, which were selected on the basis of performance test traits. Farrowing information was available on 259 gilts. At the start of the mating period, gilts selected for high DFI, LGA or LGS had similar live weights (135, 137 and 137 (s.e.d. 4·5) kg) but different backfat depths (20·3, 14·0 and 11·3 (s.e.d. 1·3) mm) while the corresponding low lines had different live weights (129, 117 and 124 kg), but similar backfat depths (17·5, 17·8 and 17·8 mm). Gilts selected for high LFC had lower mean live weight and backfat depth (114 kg and 10·7 mm) than gilts in the other selection lines. Conception rates of gilts selected for low DFI or high LGS were similar (0·62 and 0·64, s.e.d. 0·12) and lower than for the alternative selection strategies (0-78), but the low DFI gilts were significantly older at farrowing than gilts selected for high LGS (424 v. 408 (s.e.d. 5·5) days). Responses in live weight, backfat depth, age at mating and particular reproduction traits were selection strategy dependent, such that identification of relationships between growth and reproduction traits will require measurement of additional growth traits at an earlier age than in the current study

Data from 4150 Landrace pigs tested during the period 1989-94 for backfat thickness and age at 100 kg in an open selection nucleus were analysed with the standard restricted maximum likelihood/best linear unbiased prediction method and with a Bayesian approach based on the marginal posterior distributions of parameters of interest achieved via Gibbs sampling. Breeding values and fixed effects were sampled from normal distributions and (co)variance components from inverted Wishart distributions. The Bayesian analysis indicated that the selection was effective for both traits. Assuming flat priors for the (co)variance components, the posterior means of the annual rates of response to selection for both traits were −0·473 days and −0·212 mm. The influence of informative priors constructed from (co)variances estimated in the French Landrace breed on inferences about genetic and common environmental parameters, genetic group effects and total and annual responses was also examined.

Predictions of two models were compared. The models relate ovulation rate (OR) and prenatal survival (PS) to litter size (LS): the uterine capacity model (UCM), where maximum LS is limited by uterine capacity (UC), and the threshold model (TM) whereby PS is modelled as a binary threshold trait. Records were simulated according to both models using statistics from Trench Large White gilts. Both models were able to reproduce closely the observed curvilinear relationship between OR and LS, with LS reaching a plateau at high OR. Several genetic correlations (ρg) fulfilling the conditions h20R = 0·34 and h2LS = 0·12 (the residual maximum likelihood estimates in the population) were studied by means of stochastic computer simulation. The genetic correlation between OR and LS was very sensitive to changes in h2uc, whereas ρgLS,PS was always positive, and ρgOR,PS was always negative. The correlation between PS and UC was larger than 0·90, except for very small h2UC This suggests that genes affecting PS have a strong influence on UC and that PS can be a good indirect criterion to select for UC. Both models predicted that the advantage of an index combining OR and LS with respect to direct selection on LS diminishes in successive generations of selection and that the size of the experiment needed to detect significant differences is very large. Records were also simulated by halving the mean and variance of UC, so as to mimic unilateral hysterectomy-ovariectomy (UHO). If the UHO treatment results in halving UC, LS of UHO females should behave very much as half the UC of intact females.

Protein accretion curves were derived using food intake, growth and body composition data from a total of 320 pigs, which comprised castrated males and gilts from five genotypes. The 32 pigs from each genotype/sex subclass were offered ad libitum a series of isoenergetic diets (13·8 MJ metabolizable energy per kg) designed to provide a non-limiting intake of nutrients and to allow maximum protein deposition rates. Four pigs from each subclass weren slaughtered at each of the live weights (kg) 25, 44, 65, 85, 100, 115, 130 and 150. Daily food intake (dF/dt) was described as an increasing exponential function of age (t), live weight (W) as an increasing exponential function of cumulative food consumed (F), and protein weight (Pr) as an allometric function of live weight (W). The rate of protein accretion in the body of pigs [d(Pr)/dt] was calculated as d(Pr)/dt = (dF/dt).(dW/dF).(dPr)/dW). Bootstrap procedures were used to estimate standard errors for the food intake, growth and compositional parameters and to obtain the confidence bands for the dependent variables (dF/dt, W, Pr and d(Pt)/dt).

Protein accretion rate as a function of live weight was curvilinear, increasing to a maximum, then decreasing with increasing live weight. There were significant differences between subclasses in the maximum rate of protein accretion, although there was no relationship between this rate and the live weight, stage of maturity, or age at which maximum protein accretion occurred. Describing protein accretion as a multiplicative function of food intake, food efficiency and the partitioning of nutrients in the body allowed changes in the magnitude and shape of the protein accretion curve to be ascribed to one, or a combination, of the above mechanisms.

Between 45 and 100 kg body weight, Large White castrated males (cLW), and crossbred Pietrain × Large White castrated males fcPP×) and boars (bPP×) were either given food ad libitum in experiment 1 or allocated to four energy levels (0·70, 0·80, 0·90, 1·00 ad libitum) in experiment 2. Daily protein and amino acid supplies were calculated to be non-limiting for growth. Protein (PD) and lipid (LD) deposition rates were measured according to the comparative slaughter technique over the whole experimental period (experiments 1 and 2) and according to the nitrogen and energy balance technique at four stages of growth: 48, 64, 79 and 94 kg mean body weight (experiment 2). Average daily gain increased linearly with metabolizable energy (ME) intake, the slope of the relationship being significantly different between types of pig (from −28 to +36 g per additional MJ ME). Daily PD increased with ME intake above maintenance (MEp) according to a linear-plateau relationship in the three types of pig: the slope was significantly affected by type of pig (from +3·4 to 6·0 g per extra MJ MEp) but not by stage of growth. Daily LD increased linearly with MEp intake; neither the type of pig nor the stage of growth affected its slope (+13-5 g per extra MJ MEp). The LD/PD ratio in the extra body-weight gain associated with increased energy intake was affected only by the type of pig.

The effect of energy supply on physical composition of body weight gain between 45 and 100 kg was studied in Large White castrated males (cLW), crossbred Pietrain × Large White castrated males (cPPX) and boars (bPPX). The pigs were either given food ad libitum and kept in individual pens in experiment 1, or allocated to four energy levels (0·70, 0·80, 0·90, and 1·00 ad libitum) and kept in metabolism cages in experiment 2. Daily protein supplies were calculated to be the same at the four energy levels within each type of pig and non-limiting for growth. Five additional animals for each type of pig were slaughtered at 45 kg. Daily tissue gain was measured according to the comparative slaughter technique. The daily lean gain increased with metabolizable energy (ME) intake according to a linear-plateau relationship whereas the daily fat gain increased linearly. The type of pig significantly affected the slope of the relationship between lean gain and ME intake (from 15 to 22 g per extra MJ ME) but not the slope of the relationship between fat gain and ME intake (10 g per extra MJ ME on average). Increased energy intake was associated with increased fatness of body-weight gain, which was higher in cLW and cPP× than in bPP×.

The objective of this study was to investigate the effect of low food intake on post-weaning fertility and hormone concentrations in primiparous sows. One hundred and forty primiparous sows (Landrace × Large White) were randomly assigned to one of four treatment groups: (A) restricted to 3 kg food per day during lactation and 2 kg food per day post weaning (no. = 36); (B) restricted to 3 kg food per day during lactation and ad libitum feeding post weaning (no. = 33); (C) ad libitum feeding throughout with a mean voluntary food intake of 3·7 kg/day during lactation (no. = 36); and (D) ad libitum feeding throughout with a mean voluntary food intake of 4·9 kg/ day during lactation (no. = 35). Litters were standardized at eight to 10 pigs at 2 days post partum and were weaned at 27·2 (s.e. 0·5) days post partum. After weaning, sows were observed twice daily for signs of oestrus and were mated on the day(s) of standing oestrus. Six sows from each group were fitted with indwelling jugular cannulae 2 days before weaning. Blood samples were collected 1 day before weaning (W−l), 1 day after weaning (W+l) and 4 days after weaning (W+4), and these sows were slaughtered on W+4. Ovaries were recovered, follicles and corpora lutea were counted, and folUcular fluid collected. Sows given food at a restricted level (groups A and B) lost more backfat during lactation than sows given food ad libitum (groups C and D) (P < 0·01). Piglet weaning weights, weaning to conception intervals and the proportion of sows ovulating by W+4 were not significantly different among groups. In sows that did not ovulate, numbers of large follicles were not significantly different among groups but the low food intake (groups A, B and C) sows had more medium-sized follicles than high food intake (group D) sows (P < 0·01). FolUcular fluid concentrations of oestradiol, progesterone and insulin-like growth factor-1 (IGF-1) did not differ among groups (P > 0·05). Plasma non-esterified fatty acids concentrations were greater in groups C and D than in groups A and B on W−l, W+l and W+4. Serum insulin and plasma glucose concentrations were lower in groups C and D than in groups A and B on W−l and W+l, but were similar among groups on W+4. Serum growth hormone concentrations were lower in group D than group A and B on W−l and W+l but not W+4. Concentrations of IGF-1 in serum were greater in groups C and D than groups A and B on W−l and W+l but not W+4. It is concluded that low food intakes during lactation and post weaning affected body condition, serum hormones and metabolites and ovarian function of primiparous sows without dramatically affecting ovarian hormones or reproductive performance.

An experiment was carried out to study the effect of changes in either the quality or the quantity of dietary protein intake on the excretion of nitrogen (N) compounds in the chicken. Thirty-two White Rock male broilers (1 day old) were raised in batteries and fed a commercial starter diet for 9 days. Then they were randomly divided into eight groups each of four birds, of similar body weight (mean live weight: 178 (s.e. 1·9) g), and individually housed in metabolism cages. Following a paired-feeding design based on metabolic body weight (kg M0·75), each group of birds was given, for an experimental period of20 days, each of four levels of protein (60, 120, 180 or 240 g/kg: 5 days each) in two groups of isoenergetic (14·5 kj metabolizable energy per g dry matter) and semisynthetic diets based on soya-bean meal, known to be first limiting in methionine, either unsupplemented (diets S) or supplemented with 2 g/kg DL-methionine (diets SM). Excreta were collected every 3 days for 48 h, frozen and stored at −20°C. The excreta samples were subjected to chemical analysis for uric acid by a rapid high-performance liquid chromatographic method, for urea and ammonia by a colorimetric method, and for total N by the Kjeldahl procedure. In general, the excretion of major N compounds was markedly affected by either the quality or the content of dietary protein. Overall, the excretion of total N, uric acid-N, ammonia-N and urea-N significantly (P < 0·05) decreased with improvement in dietary protein quality and significantly (P < 0·05) increased with increase in protein intake. Regression equations were obtained relating the excretion of uric acid, urea, ammonia and total N on protein supply. For the partition ofN compounds output, the ratios of uric acid-N and ammonia-N to total N significantly (P < 0·05) decreased on improving dietary protein quality and increased or remained unchanged, respectively, with the increase in dietary protein content. The use of the ratio of ammonia-N to total N is recommended as a rapid, easy and accurate indicator of dietary protein adequacy, as an alternative to measures based on total N balance, without the need for separation of urine and faeces.

Wild red deer (Cervus elaphus) hinds were subjected to one of five post-capture management strategies during the period from capture in late pregnancy in March, until weaning in September. The treatments were: remained enclosed at the capture site; relocated to a remote site; relocated to a conventional deer farm on improved pasture; relocated to a conventional deer farm and grazed with farmed hinds; housed prior to calving on a conventional deer farm site. For a further year, all hinds were managed under conventional deer farm conditions. There was considerable evidence, based on mortality and behaviour, to suggest that initial housing of the hinds following capture compromised their welfare. Deaths of hinds (seven out of 20) occurred in this group, associated with bullying during the housed period. Hinds on this treatment also showed a higher frequency of aggressive interactions compared with the groups directly moved to the deer farm (11% v. 2% respectively of scans conducted at 10-min intervals over 6h; P < 0·001) and a reduced level of lying behaviour (15% v. 34 to 47% of scans; P < 0·001) during the gestation period. Similarly, during the calving period, this group lay the least (27% v. 43 to 72% of scans; P < 0·001) and the mean number of interactions between hinds reached 11·7 compared with 1·5 to 4·9 (F < 0·001) in the other groups over a continuous 6-h recording period. The two groups of hinds relocated directly onto sown swards were generally similar to one another in terms of behaviour and performance. However, the group mixed with farmed hinds suffered from considerable calf losses in year 1 due to disease (four out of 12). Losses of hinds over winter in year 1 (nine out of 90) occurred principally amongst those animals which had not become habituated to human presence or supplementary feeding in the summer, i.e. groups remaining enclosed at the capture site or relocated to a remote site. When all hinds were treated similarly in year 2 the hinds from these same two groups, together with those initially housed indoors, showed more hind-hind interactions overall than the two groups located directly onto the deer farm (7·2 v. 3·1 interactions per hind over a 4-h period; P < 0·02) and it is suggested that these hinds had not yet adapted to their new environment. An ACTH stimulation test conducted during year 2 supported the view that the two groups of hinds most recently introduced into the managed system were subject to a chronic stressor at the time of testing, since administration of ACTH did not elicit a significant increase in plasma cortisol concentrations (mean values pre- and post-ACTH 188 and 217 nmol/l respectively). In contrast, the mean plasma cortisol concentrations of the two groups managed under extensive farm conditions from the outset, showed a significant rise (pre- and post-ACTH, 261 and 376 nmol/l respectively; P < 0·01). From this it is concluded that their adaptation to the farm environment had already occurred. The live-weight gains of the wild hinds calves (229 g/day) on the improved pastures in the 1st year of the study were below that for farmed hinds calves (282 g/day; P < 0·05), suggesting that they were not habituated to the management system. However, by the end of year 2 animal performance was comparable with that of farmed hinds and calf growth rates reached 276 g/day. Thus while housing wild red deer immediately after capture is associated with poor welfare, analysis of behaviour, adrenal response and animal production over a longer period suggests that by the end of the study few important differences remained between the groups.

The farming of ostriches is well established in South Africa and is gaining in popularity in agriculture around the world but few scientific data are available for rates of egg production and the factors affecting incubation and hatching. Production and hatchability of over 1200 ostrich eggs on a farm in the United Kingdom were recorded during the 1995 season. Females of breeding ages (no. = 43) were kept in 13 enclosures in sex ratios of 1: 1 or 1: 2 (M: F). Productivity of all ostriches over the laying season was 25·2% although when results from individual enclosures, and from individual birds, were examined higher rates of production were achieved although only one bird achieved her full laying potential. As group size increased productivity decreased. Overall fertility was 78·7% although fertility greater than 90% was achieved by females kept in pairs or trios. Larger group sizes had lower fertility. Mean weekly hatchability was only 24·1% with hatch of fertile eggs being 31·9%. Maximum hatchability of eggs produced from an individual female was 73% of fertile eggs. Microbial contamination of eggs was 32·6% and this was the biggest cause of failure to hatch, although rates of contamination varied between females. These results for productivity are similar to those published for ostriches in Israel and South Africa. Hatchability is lower, and microbial contamination is higher than published information although this is the most comprehensive study of ostrich production yet published. Future development of ostrich breeding programmes needs to concentrate on small breeding groups, preferably pairs, and on improvements in nest site hygiene and post-laying egg management.

Four bales of hay were chosen from a stack of hay, barn-stored for 6 months. Each bale was subdivided into four 2·5 kg sections and one section from each bale was subjected to one of four water soaking treatments: 0, 0·5, 3 or 12 h. Post treatment, the samples were shaken for 9 min under a Negretti LS45F personal dust sampler which sampled the surrounding air. The samples were than analysed for levels of respirable particules < 5 μm in size and for amounts of sodium (Na), potassium (K), phosphorus (P), magnesium (Mg), calcium (Ca) and nitrogen (N). Soaking for 30 min reduced the number of respirable particles by about 90% but also reduced the Na, K and P content of the hay. Soaking for a longer period, at least up to 12 h, did not give a significant further reduction in number of respirable particles but there was loss of Mg and further losses in Na, K, and P. These results clearly indicate that soaking for 30 min is sufficient to reduce the respirable challenge with minimum nutrient loss.

Five Chilean crossbred draught horses were used to study variation in work performance and biochemical and physiological parameters whilst working with a mould board plough (36 kg) for 6 h/day. The draught force, distance travelled, work done and the estimated extra energy for work were determined for each horse. The results showed that the horses used only 68·0 (s.e. 1·9)% of the total available working time ploughing a mean area of 3283 (s.e. 168) m2 with a mean depth and width of furrow of 12·6 and 22·3 cm, respectively. Estimated daily energy expenditure by the horses during work was 2·24 (s.e. 0·02) times maintenance requirements. The small changes observed in the physiological variables: heart rate, respiratory rate, packed cell volume, haemoglobin concentration and enzyme activity of creatine kinase, lactate dehydrogenase and aspartate amino transferase, showed that the horses were exercised under submaximal conditions. However, the significant increase in blood cortisol concentrations showed that horses experienced some degree of stress during ploughing work. The significant increase in blood triglycerides concentration observed after each period of work seems to demonstrate that fat mobilization is an important metabolic pathway as an energy supply for the working muscles during prolonged and low speed draught work.