Over 3 years a total of 204 bucket-reared calves (from Holstein/Friesian dams) and 187 suckled calves (from Hereford × Friesian dams) were monitored during the first 9 or 6 months of life respectively. Within each rearing system there were approximately equal numbers of calves sired by Charolais and Hereford bulls and also about equal numbers of heifer and steer calves. All breeds and sexes within each rearing system were managed as a single group.

Differences in daily live-weight gain from arrival (January) for the bucket-reared calves and from birth (April) for the suckled calves through to the autumn were significantly different at 0·66 and 1·01 kg/day respectively (P < 0·001). By comparison, differences in growth rate between sex and breed within each rearing system were small. Hereford cross calves from both systems however carried significantly more condition in the autumn compared with Charolais crosses (P < 0·01).

The results suggest that when animals are reared on the same plane of nutrition and given similar management, differences in initial growth rate between early and late maturing breeds and sexes may be less than those perceived by commercial producers the views of whom are influenced by the fact that the types are often managed differently in practice.

Broiler chickens often make limited use of the free-range area. Range use is influenced by type of shelter available. Range use may possibly be improved by a more gradual transition from the house to the range and by using dark brooders (secluded warm, dark areas in the home pen) that mimic aspects of a broody hen and possibly reduce fearfulness. The aim of this study was to assess effects of dark brooders on fearfulness, free-range use and behaviour later in life. Another aim was to test the chickens’ preference for shelter type and the effects of overhangs outside of the pop holes to provide a gradual transition to the range. Three production rounds, each with 440 Sasso broiler chickens (110/group), were completed. Chicks were housed indoors from days 0 to 25; per round, two groups had access to a dark brooder, whereas the other two groups had conventional IR lamps. Fearfulness was assessed by the open field (OF) and tonic immobility (TI) tests on days 22 to 24 on 25 chicks/group per round. Birds were then moved to four mobile houses from which they could access both grassland with artificial shelter (AS) and short rotation coppice (SRC). Two of the houses had overhangs extending from the pop holes; these were switched between the four houses weekly. Free-range use and behaviour were observed three times daily from Monday to Friday. Dark brooders did not affect results from the OF or TI test, except for jumps in the OF test which tended to occur less often in brooded chicks. Neither dark brooders (34.9% without v. 31.7% with brooder) nor overhangs (32.5% without v. 34.1% with overhangs) influenced the percentage of chickens outside. Chickens showed a clear preference for SRC, range use increased over time in SRC, and more birds ranged farther from the house in SRC. Behaviours of chickens observed outside were mainly influenced by shelter type, age of the birds and distance from the house. Locomotion tended to occur more in the presence of overhangs. Overall, these results could not confirm the hypothesis that dark brooders would decrease fearfulness and thereby increase free-range use. Overhangs also did not improve free-range use, and neither brooders nor overhangs had considerable impact on behaviour of chickens outside. Chickens clearly preferred dense natural vegetation over AS and ranged farther in it, indicating that this type of shelter is more suitable for slow-growing free-range broilers.

Current techniques for measuring the dry matter intake (DMI) of grazing lactating beef cows are invasive, time consuming and expensive making them impractical for use on commercial farms. This study was undertaken to explore the potential to develop and validate a model to predict DMI of grazing lactating beef cows, which could be applied in a commercial farm setting, using non-invasive animal measurements. The calibration dataset used to develop the model was comprised of 94 measurements recorded on 106 beef or beef–dairy crossbred cows (maternal origin). The potential of body measurements, linear type scoring, grazing behaviour and thermal imaging to predict DMI in combination with known biologically plausible adjustment variables and energy sinks was investigated. Multivariable regression models were constructed for each independent variable using SAS PROC REG and contained milk yield, BW, parity, calving day and maternal origin (dairy or beef). Of the 94 variables tested, 32 showed an association with DMI (P < 0.25) upon multivariable analysis. These variables were incorporated into a backwards linear regression model using SAS PROC REG. Variables were retained in this model if P < 0.05. Five variables; width at pins, full body depth, ruminating mastications, central ligament and rump width score, were retained in the model in addition to milk yield, BW, parity, calving day and maternal origin. The inclusion of these variables in the model increased the predictability of DMI by 0.23 (R2 = 0.68) when compared to a model containing milk yield, BW, parity, calving day and maternal origin only. This model was applied to data recorded on an independent dataset; a herd of 60 lactating beef cows two years after the calibration study. The R2 for the validation was 0.59. Estimates of DMI are required for measuring feed efficiency. While acknowledging challenges in applicability, the findings suggest a model such as that developed in this study may be used as a tool to more easily and less invasively estimate DMI on large populations of commercial beef cows, and therefore measure feed efficiency.

The objective was to study the effect of maternal supplementation with a yeast cell wall-based product containing a mannan-rich fraction (MRF) during gestation and lactation on piglet intestinal gene expression. First parity sows were fed experimental gestation and lactation diets with or without MRF (900 mg/kg). After farrowing, piglets were fostered within treatment, as necessary. Sow and litter production performance data were collected until weaning. On day 10 post farrowing, jejunum samples from piglets were collected for gene expression analysis using the Affymetrix Porcine GeneChip array. Most performance parameters did not differ between the treatments. However, protein (P<0.01), total solids less fat (P<0.03) and the concentration of immunoglobulin G (IgG) in milk were greater (P<0.05) in the MRF-supplemented group. Gene expression results using hierarchical clustering revealed an overall dietary effect. Further analysis elucidated activation of pathways involved in tissue development, functioning and immunity, as well as greater cell proliferation and less migration of cells in the jejunum tissue. In conclusion, feeding the sow MRF during pregnancy and lactation was an effective nutritional strategy to bolster colostrum and milk IgG that are essential for development of piglet immune system and gut. In addition, the gene expression patterns affected by the passive immunity transfer showed indicators that could benefit animal performance long term.

During a series of routine experiments to evaluate 16 foods offered to wether sheep, conventional measurements of digestibility were augmented by measurement of fractional outflow rates (FOR) of solids (ks) and liquid (k1) from the rumen and urinary allantoin nitrogen excretion (UAN) as an index of microbial yield. Fixed (1 kg) amounts of ‘dry’ foods were offered to sheep ranging in weight from 42 to 89 kg which permitted a retrospective analysis of plane of nutrition effects on diet digestibility, ks, k1 and UAN. Each diet was simultaneously offered to three sheep (one each of large, medium and small size). Smaller sheep had significantly higher FOR, significantly lower apparent digestibilities of dry matter and significantly higher UAN (g/kg digestible organic matter (DOM)). Higher UAN (g/kg DOM) indicated higher microbial energetic efficiencies with the smaller sheep.

Fertility can be defined as the natural capability of giving life. It is an important factor both for human medicine, where ~10% of the couples call for the services of assisted reproductive technologies, and for species of economic interest. In particular, in dairy cows, the recent years have seen a kind of competition between milk production and fertility, and genes improving fertility are now considered as parameters to be selected for. The study of fertility pathways is nevertheless made difficult by the strong impact of environmental factors on this parameter, as well as by the number of genes potentially involved (as shown by systematic transcriptome analysis studies in the recent years). One additional level of complexity is given by the fact that factors modulating fertility will probably be sex specific. The usage of mouse models has been one of the solutions exploited for tackling with these difficulties. Here, we review three different approaches using mice for identifying genes modulating fertility in mammals: gene invalidation, positional cloning and in vitro mutagenesis. These three approaches exploit specific characteristics of the mouse, such as the possibility of controlling precisely the environment, an excellent genetic characterization and the existence of genomic and molecular tools equalled only in humans. Many indications suggest that at least some of the results obtained in mice could be easily transposed to the species of interest.

Evaluations of the energy requirements of working animals have been confounded by the constraints of indirect calorimetric techniques (Cal). This study sought to investigate a non-restrictive methodology, the doubly labelled water (DLW; 2H218O) technique, for the measurement of energy expenditure in free ranging equids. Six pony geldings were intravenously injected with DLW in two sequential studies that permitted first, isotope equilibration and half-lives to be determined and then second, heat production (HP) to be measured simultaneously by DLW and traditional (Cal) techniques.

In study 1, three animals were injected with DLW, three animals were untreated controls. Blood samples were collected every 30 min for 12 h and thereafter at 24 h intervals for 14 days. Isotopes equilibrated throughout the body water pool within 300 (2H) and 240 (18O) min and half-lives were 6.3 ± 0.6 days (2H) and 5.6 ± 0.4 days (18O).

In study 2, HP was simultaneously determined by Cal and DLW over a 4-day period. Animals (no. = 6) were assigned to pairs and in successive weeks two ponies were injected with DLW and confined to metabolism chambers 12 h later. Cal HP was 0.51 ± 0.02 MJ/kg M0.75 per day compared with 0.48 ± 0.29 MJ/kg M0.75 per day estimated by DLW. Maintenance metabolizable energy intake was 0.53 ± 0.01 MJ/kg M0.75 per day (Cal) and 0.50 ± 0.01 MJ/kg M0.75 per day (DLW). Validation of the DLW technique may empower essential, systematic appraisal of energy requirements in unrestrained working horses.

A bacterial cocktail of living strains of Clostridium perfringens type A (CPA) without β2-toxin gene and non-pathogenic Escherichia coli was administered orally to newborn piglets before first colostrum intake and on 2 consecutive days on a farm with a high incidence of diarrhoea and antibiotic treatment in suckling piglets associated with E. coli and CPA. This clinical field study was driven by the hypothetic principle of competitive exclusion of pathogenic bacteria due to prior colonization of the gut mucosal surface by non-pathogenic strains of the same bacterial species with the aim of preventing disease. Although CPA strains used in this study did not produce toxins in vitro, their lack of pathogenicity cannot be conclusively confirmed. The health status of the herd was impaired by a high incidence of postpartum dysgalactia syndrome in sows (70%) and a high incidence of neonatal diarrhoea caused by enterotoxigenic E. coli and CPA during the study. No obvious adverse effect of the bacterial treatment occurred. On average, more piglets were weaned in litters treated (P=0.009). Visual pathological alterations in the small intestinal wall were more frequent in dead piglets of the control group (P=0.004) and necrotizing enteritis was only found in that group. A higher average daily weight gain of piglets in the control group (P<0.001) may be due to an increased milk uptake due to less competition in the smaller litters. The bacterial cocktail was tested under field conditions for its potential to stabilize gut health status in suckling piglets before disease development due to colibacillosis and clostridial infections; however, the gut flora stabilizing effect of the bacterial cocktail was not clearly discernible in this study. Further basic research is needed to confirm the positive effects of the bacterial treatment used and to identify additional potential bacterial candidates for competitive exclusion.

With still limited information on vitamin requirements and considering that many commercial practices adopt dietary vitamin levels above the values suggested by nutritional tables, this study aimed to assess the effect of administering vitamin supplementation to sows in gestation and lactation and to their litters on the reproductive performance and body condition of the sows and on the performance and immune profile of the litters until slaughter. The trial was split into two phases. The first phase used 104 sows, assigned to be randomized to blocks according to parity, submitted until 21 days of lactation to two treatments: control–standard (standard levels of vitamins) and test–elevated (elevated levels of vitamins). Each sow and its respective farrow were considered an experimental unit. The sows underwent evaluations of body condition score, back fat thickness and reproductive performance. In the second phase, 60 barrows and 60 gilts at 21 days of age and mean initial weight of 5.33 ± 1.5 kg until slaughter at 164 days of age. The piglets were assigned to randomized blocks according to the weight and sex of the animals in a 2 × 2 factorial model, with 10 replicates per treatment, where a pen with three animals represented the experimental unit. Following the same treatments of the first phase, the piglets were evaluated for daily weight gain, daily feed intake, feed conversion, mortality rate and humoral immune response. Vitamin supplementation had no positive effects on the reproductive parameters or body composition of sows. However, it positively impacted the performance of the litters in the early nursery stage, but did not lead to superior effects on the immune responses to vaccination against circovirus or mycoplasma.

With the objective of determining genetic × environmental interaction for beef production under direct grazing conditions, measured as both individual average daily gain (ADG) and per ha (PROD), 160 steers were utilized through 2 years, from two breeds of different growth potential, and four stocking rates (SR) tending to establish different nutritional environments. The breeds were Aberdeen Angus (A) and high-grade Limousin crossbred (L) and the SRs established were 2·25, 2·87, 3-50 and 4·13 steers per ha. The pasture where the steers grazed was Festuca arundinacea. Response variables were analysed by least-squares using a fixed model of year, breed, SR and their two-way interactions. There were highly significant effects (P < 0·01) of the interaction of breed × SR for ADG and PROD. Quadratic and linear regressions (P < 0·01) were adjusted for PROD and ADG on SR respectively using the least-squares means. The proportional superiority of L on A at the lowest SR was 0·27, while at the highest SR the situation was reversed, A gained proportionately 0·32 more weight than L. These results indicated that under limiting conditions of nutrition as generated at the highest SR in this experiment, the small-sized individuals tended to produce meat more efficiently both individually and per ha, the opposite situation being true when nutritional conditions were not restrictive.

The effect on animal performance of the inclusion in the diet of avoparcin at one or more of the following levels 0, 150, 300 and 450 mg per head daily and 200 mg per head daily monensin was studied i n a series of four trials. Finishing cattle of 428 (s.e. 3·9) kg live weight were used in three trials and yearling heifers (218 (s.e. 4·9) kg live weight) in the fourth. All animals received grass silage ad libitum with 0·5 to 6·0 kg per head daily of a cereal-based supplement. Relative to the control diets with no additive, proportionate increases in live-weight gain of between 0·07 and 0·38 were obtained with the inclusion of avoparcin at 150 mg per head daily in the diet with a small depression in silage dry-matter intake. The effects at higher levels of inclusion of avoparcin and of monensin were more variable. Alteration in the proportion of propionate to acetate in the rumen fluid of the animals was observed only at the highest levels of additive inclusion. There was no effect of additive treatment on carcass weight or dressing proportion.

Streptozotocin-induced diabetes mellitus (STZ-DM) in rats is a model of type 1 diabetes, which is commonly used to study diabetes, but differs from human diabetic pathophysiology in its insulin resistance. An STZ-DM rat can be administered five times the dose of insulin compared to that of a diabetic patient. Thus, attaining normoglycaemia in STZ-DM rats with insulin injections is complicated, and it involves an obvious risk of overdosing before getting a response. This study was designed to investigate whether suboptimal treatment with long-acting insulin restores insulin sensitivity in the STZ-DM rat, and thus an approach to more closely mimic the human condition. Male Sprague-Dawley rats were made diabetic by means of a single intravenous injection of STZ (55 mg/kg body weight (BW)), resulting in an increase in blood glucose (BG) from 6.5 ± 0.2 to 22.5 ± 1.0 mmol/l (P ⩽ 0.05) within 24 h. After treating the STZ-DM rats with vehicle for 14 days, BG was 26.1 ± 1.1 mmol/l, and the response to a single injection of fast-acting insulin (Humalog, 5 IE/kg BW) was a 23% reduction in BG. Thereafter, the rats were treated daily with a suboptimal dose of long-acting insulin for a total of 7 days (Insulatard, 5 IE/kg per day), which resulted in a BG level of 19.4 ± 2.7. The response to fast-acting insulin after the suboptimal treatment was a 61% reduction in BG. Thereafter, the animals were vehicle-treated for another 7 days, which resulted in a response to fast-acting insulin similar to the initial values (−34%). Furthermore, the group treated with suboptimal doses of long-acting insulin had a longer duration of the reduction in BG (150 min, as opposed to 90 min in the vehicle-treated groups). We conclude that the development of a decreased insulin response occurs rapidly within the first 2 weeks after the onset of diabetes in STZ-DM rats. This leads to a brief and significantly reduced decrease in BG when fast-acting insulin is administered. The insulin response is increased by treatment with suboptimal doses of long-acting insulin, but rapidly decreases again when treatment is withdrawn. Regular administration of suboptimal insulin doses may provide an approach to eliminate the effects of a lowered insulin response.

Prediction of carbohydrate fractions using equations from the Cornell Net Carbohydrate and Protein System (CNCPS) is a valuable tool to assess the nutritional value of forages. In this paper, these carbohydrate fractions were predicted using data from three sunflower (Helianthus annuus L.) cultivars, fresh or as silage. The CNCPS equations for fractions B2 and C include measurement of ash and protein-free neutral detergent fibre (NDF) as one of their components. However, NDF lacks pectin and other non-starch polysaccharides that are found in the cell wall (CW) matrix, so this work compared the use of a crude CW preparation instead of NDF in the CNCPS equations. There were no differences in the estimates of fractions B1 and C when CW replaced NDF; however, there were differences in fractions A and B2. Some of the CNCPS equations could be simplified when using CW instead of NDF. Notably, lignin could be expressed as a proportion of DM, rather than on the basis of ash and protein-free NDF, when predicting CNCPS fraction C. The CNCPS fraction B1 (starch + pectin) values were lower than pectin determined through wet chemistry. This finding, along with the results obtained by the substitution of CW for NDF in the CNCPS equations, suggests that pectin was not part of fraction B1 but present in fraction A. We suggest that pectin and other non-starch polysaccharides that are dissolved by the neutral detergent solution be allocated to a specific fraction (B2) and that another fraction (B3) be adopted for the digestible cell wall carbohydrates.

In the first of two experiments, steers which received a single implant of 60 mg hexoestrol grew 24% faster than non-implanted controls and 7% faster than steers given 30 mg. The growth rate of the steers given a single implant of 60 mg did not differ significantly from that of animals implanted twice with 30 mg or 5 times with 15 mg. In the second experiment, steers given 60 mg grew 23% faster than controls and their feed conversion efficiency was 13% better; the hexoestrol treatment resulted in a saving of 131 kg feed per animal. The tenth-rib joints of steers given 60 mg had a higher crude-protein content and lower fat content than those from non-implanted animals.

Physicochemical characteristics were determined in the longissimus lumborum muscle, after 8 days of ageing of steers (n=12) and bulls (n=12) from Mertolenga breed slaughtered directly from pasture (day 0) or after a finishing period of 50, 100 and 150 days in a feed-lot facility. Bulls and steers presented similar live weight (averaging 388 kg), carcass weight (CW; averaging 213 kg), dressing percentage (averaging 60%), carcass fatness (11.9% CW) and carcass fat thickness (averaging 3.03 mm). Live weight, CW, carcass fatness and fat thickness increased along time-on-feed. Gender only had a negligible effect on meat characteristics, with b* and h* being the only parameters of colour affected by gender, also presenting a significant interaction gender×time-on-feed. Nevertheless, both the genders presented a high-quality grade concerning tenderness (Warner–Bratzler shear force (WBSF)). L* increased until 50 days on feed and decreased afterwards, whereas a* and C* values increased along time-on-feed. Pigment content was also affected by time-on-feed and showed a gender×time-on-feed interaction. Beef colour became darker and redder along time-on-feed, but still in a colour range highly acceptable by Portuguese consumers. Despite the increase in intramuscular fat and myofibrillar fragmentation index, as well as the decrease in collagen content of steers and bulls along time-on-feed, it did not affect the tenderness/hardness, indicating a small effect of time-on-feed in meat characteristics. Despite only small differences in carcass characteristics and meat-quality parameters that have been noticed along time-on-feed, those differences were only significant after 100 days on feed. Principal component analysis (PCA) was performed. The first PC axis (39.6% of the total variance) included colour variables a*, b* and C*, and carcass fatness, fat thickness, CW and live weight, whereas the second one (12.7% of the total variance) included h*, cooking losses and dressing-out. The principal component (PC) analysis confirmed the lack of differences between bulls and steers and indicates a differentiation of the first two periods of feeding (0 and 50 days on feed) from the two latter (100 and 150 days on feed) periods of feeding.

There are few published studies on the genetics of body growth to maturity in sheep. Such information is required for choosing an optimal time of selection for body weight and to predict correlated responses to selection at all stages of an animal's lifetime. The data were derived from a randomly selected control flock of Scottish Blackface sheep on a hard heather hill research farm in Peeblesshire, Scotland. The objectives of the analysis were to estimate the heritability of body weight at various ages between birth and maturity, and the genetic correlations between these weights.

Rabbit commercial maternal lines are usually selected for litter size (LS) and paternal lines for growth rate (GR). Line OR_LS was selected by ovulation rate (OR) and LS to improve LS more efficiently. In this study, growth traits of line OR_LS were evaluated by estimating the correlated response on weaning weight (WW), slaughter weight (SW) and GR during fattening period as well as their variability (DWW, DSW and DGR, respectively). Data were analyzed using Bayesian inference methods. Heritability estimates were low for growth traits (0.09, 0.13 and 0.14 for WW, SW and GR, respectively) and negligible for growth traits variability (0.01, 0.004 and 0.01 for DWW, DSW and DGR, respectively). Moderate common litter effect ratio (c2; 0.35, 0.28 and 0.27) and low maternal effect ratio (m2; 0.11, 0.05 and 0.01) were obtained for WW, SW and GR, respectively. Both c2 and m2 were lower at slaughter than at weaning. In addition, low common litter effect and negligible maternal effect were observed for growth traits variability. Genetic correlations between LS and both growth traits and their variability were close to zero. Positive genetic correlations were observed between OR and growth traits (0.19, 0.38 and 0.36 for WW, SW and GR, respectively) as well as between OR and growth traits variability (0.35, 0.62 and 0.20 for DWW, DSW and DGR, respectively). Positive correlated responses in both periods were obtained for growth traits, WW, SW and GR (0.037, 0.156 and 0.110 kg, respectively). The correlated response found in growth traits might be due to the positive genetic correlations between OR and these traits. However, selection for OR and LS using independent culling levels did not modify the growth traits variability. Therefore, no negative consequences on growth traits can be expected in current commercial maternal lines.

Management of intestinal microbiota of monogastric animals has increased in importance since the ban of growth promoting antibiotics in many countries. Organic acids have been used as alternatives to antibiotics by many feed manufacturers. Regardless of the wide usage, the effect, dose response and mode of action of acids on intestinal microbes is poorly understood. In this study, we investigated the effects of dietary supplementation of three commonly used products, namely formic acid (FA) (90%), dl-methionine (DLM) (99%) and liquid methionine hydroxy analogue-free acid (88%), on ileal microbiota of pigs. Laboratory simulation system, mimicking swine ileum, was used to study the products at various concentrations and combinations. Furthermore, selected combinations were tested in a piglet trial to confirm the findings made in in vitro studies. FA turned out to have a dual effect on ileal microbiota. At concentrations below 0.5%, it significantly stimulated bacteria, but at higher inclusion rates it was highly inhibitory. This finding, which was consistent in in vitro and in vivo studies, implies that reducing the dose of FA does not lead to a diluted inhibitory effect, but in fact, an opposite, stimulatory effect on intestinal microbiota. It is highly important that feed compounders acknowledge this finding. Unlike FA, the inhibitory effect of methionine hydroxy analogue on ileal bacteria was linearly dose dependent and significant at inclusion levels above 0.2%, in vitro. Partial replacement of methionine hydroxy analogue by FA, or FA by methionine hydroxy analogue, led to an unpredictable outcome due to the dual effects of FA; e.g., a minor inclusion of added FA changed the inhibitory effect of methionine hydroxy analogue into microbial stimulation by FA. Inhibition of ileal microbiota by methionine hydroxy analogue was detected only in in vitro studies, suggesting that intact methionine hydroxy analogue may not have reached the ileum, in live animals. Therefore, if the target is to ensure the inhibitory effect of FA, the FA level in feed should be kept above 0.6%, and not reduced, if methionine hydroxy analogue is used as a methionine source instead of DLM. DLM was totally inert with regard to bacterial growth and metabolism, both in vitro and in vivo. The results of these studies reveal the importance of knowing how each acid product works. Inconsistent results in animal trials may have been partly due to quadratic dose–response effects of products, and unpredictable product combination effects.

To investigate the extent to which the level of androstenone and skatole decreases with a decrease in live weight and/or age at slaughter of entire male pigs produced under organic standards, 1174 entire male pigs were raised in parallel in five organic herds, distributed across four batches in summer and winter. The median androstenone level was high for organic entire male pigs (1.9 µg/g), but varied greatly both within and between herds. Median skatole level was 0.05 µg/g, also with a wide range both within and between herds. Decreasing live weight over the range of 110±15.6 kg s.d. was found to decrease androstenone as well as skatole concentration, however, with different patterns of association. Age did not have significant direct effect on either androstenone or skatole levels. Androstenone levels were higher during winter than summer (P<0.0001), but no difference in skatole was found between seasons. The study concludes that decreasing live weight at slaughter could be an applicable management tool to reduce risk of boar taint and the level of tainted carcasses for a future production of entire male pigs within the organic pig production system, although further studies are needed as great variation in boar taint was found also for low weight animals.

In two experiments, pigs were weaned at 4 to 5 days of age and offered dry diets based on milk proteins and tallow. Live-weight gains between 7 and 28 days of age were 220 g/day with feed conversion ratios of 0·80 when the optimum diets were given.

Expt 1 was a 2 × 2 × 2 factorial with eight pigs per treatment. There were two sources of carbohydrate, lactose and wheat starch, two levels of fat, 40 and 250 g/kg and two levels of crude protein, 140 and 250 g/kg of the diet. There was no significant difference in the performance of pigs when they were fed lactose or wheat starch as the source of carbohydrate. However, the feed intake of the pigs and their weight gains and apparent digestibilities of dry matter and nitrogen were lower when the diets contained 250 g fat/kg than when the diets contained 40 g fat/kg.

In Expt 2 six diets were fed to 12 pigs each. The diets contained 150, 180, 210, 240, 270 or 300 g/kg of crude protein. Weight gains and nitrogen retention showed a positive correlation (r = 0·997) with the dietary nitrogen intake but there was no statistical increase in these parameters above the 270 g crude protein/kg level. The apparent digestibilities of dry matter and nitrogen showed a similar trend. A value of 141 mg nitrogen/day per kg M0·75 was estimated as the metabolic endogenous loss of nitrogen.