1. The potential use of an extrinsic label to measure iron absorption from a ferric orthophosphate-fortified malted cocoa drink was examined by measuring the solubility of the FePO4 in 0·1 M-hydrochloric acid.

2. The validity of using the stable isotope 58Fe as an extrinsic label was tested by comparing Fe absorption by rats from wheat flour extrinsically-labelled with 58Fe or 59Fe.

3. Fe absorption from a malted cocoa drink (20 g powder made up with hot water) fortified with FePO4 (0·5 mg Fe/g powder) was measured in human subjects using 58Fe as an extrinsic label. Absorption was calculated by measuring unabsorbed 58Fe in faeces. Absorptions of Fe from the drink fortified with either FePO4 or ferrous sulphate were compared. The effect of the addition of ascorbic acid to the drink (1 mg/g powder) on Fe availability was also examined.

4. The effect of fasting on Fe absorption from the drink was determined in rats by giving the drink to fasting animals or shortly after they had consumed a small meal.

5. The FePO4 was totally soluble in 0·1 M-HCl and there were no differences in absorption between 58Fe- and 59Fe-labelled wheat flour. In the human experiment the proportion of Fe absorbed from the drink plus FePO4 and ascorbic acid was (mean with SE) 0·25 (0·02), from the drink plus FePO4 0·24 (0·02) and from the drink plus FeSO4 0·23 (0·03). Fasting had a significant effect on Fe availability; rats given the drink shortly after a small meal absorbed less than half as much Fe as those given the drink on a fasted stomach.

6. It was concluded that the FePO4 used to fortify the malted cocoa drink was as well absorbed as FeSO4 but that the high levels of absorption were a reflection of the fasting condition of the subjects. The level of ascorbic acid was not great enough to enhance the availability of the FePO4 any further.

1. Twelve young adult male volunteers were given a low-fibre white bread diet (9 g neutral-detergent fibre (NDF)/d) and a medium-fibre coarse-bran bread diet (22 g NDF/d), each lasting 20 d. In a third period of 20 d the volunteers were subdivided in groups of four, consuming a high-fibre coarse-bran bread diet (35 g NDF/d), a medium-fibre fine-bran diet (22 g NDF/d, bran particle size > 0.35 mm) or a wholemeal bread diet (22 g NDF/d). Digestion of dietary fibre and its components hemicellulose, cellulose and lignin were determined as well as colonic function.

2. An increase of the amount of dietary fibre (through bran in bread) from 9 to 22 g NDF/d resulted in the following significant changes (P < 0.01): increase in faecal wet weight of 63 g/d, decrease in the percentage of faecal dry weight from 27 to 24, increase in defaecation frequency of 0.2 stools/d and reduction of the intestinal transit time of 36 h.

3. Further significant changes with regard to all factors mentioned were observed during the high-fibre diet. Faecal wet weight was significantly (P < 0.05) lower with the fine-bran bread diet than with the coarse-bran bread on a similar fibre intake of 22 g NDF/d. Results obtained in the wholemeal-bread period did not show significant differences compared with those from the coarse-bran bread period of 22 g NDF/d.

4. Mean digestibilities for the fibre from bread were: for NDF 0.34, for hemicellulose 0.46, for cellulose 0.20 and for lignin 0.04.

5. The results obtained suggest that the theory of sponge activity of the fibre matrix structure is the predominant factor accounting for the water binding capacity of fibre in the colon.

1. Continuous infusions of [14C]glucose and [14C]lactate on separate days, and measurements of blood flow-rate, were used to obtain values for rates of unidirectional metabolism and of interconversion of glucose and lactate in the portal-drained viscera, liver and hind-quarters of ewes during late pregnancy and early lactation. All infusions were made within 5 h after the morning meal, when steady-state conditions appeared to exist.

2. Use was made of ewes that had been appropriately catherized during pregnancy, and whose catheter remained patent through into lactation.

3. The liver was the main source of glucose production (67–70%) during both pregnancy and lactation. Other sources were the portal-drained viscera (absorbed glucose) and, presumably, the kidneys. Over 80% of the glucose was utilized by the peripheral tissues with approximately 35–40% of utilization being attributable to the hind-quarters.

4. Of the total lactate production, 76% occurred in the peripheral tissues during pregnancy but only 36% during lactation. While the liver utilized 73% of lactate during pregnancy, this value fell to only 42% during lactation, at which time the portal-drained viscera utilized 26% of the lactate.

5. During pregnancy, approximately 80% of the lactate arose from glucose, chiefly in peripheral tissues, while at least 12% of the glucose arose from lactate, chiefly in the liver. During lactation the extent of these interconversions was decreased.

6. Despite the interconversions, whole-body turnover rates for glucose and lactate were under- or overestimated by only 4–10% and 2–5% respectively. Furthermore, a comparison of turnover rates obtained with [U-14C]- and [6−3H]glucose indicated that there was only 6 and 2% recycling of glucose-C during pregnancy and lactation respectively.

7. Under the conditions employed in this study, lactate does not appear to be a major precursor of glucose in the ruminant, and most of the lactate taken up by the liver must be used for purposes other than gluconeogenesis, such as oxidation or alternative anabolic pathways.

1. Twelve grain mixtures, one lucerne (Medicago sativa) hay and one maize silage which had been used in mixed diets for which dietary nitrogen undegraded in the rumen (UDN) had been estimated with duodenally-cannulated cows, were studied. Total N in the feeds was fractionated into pool A (N soluble in borate–phosphate buffer), pool B (total N–(pool A + pool C)) and pool C (acid-detergent-insoluble N or residual N after 24 h incubation in protease solution).

2. N solubilization in protease solution containing 6·6 units/ml (substrate-saturating enzyme concentration) indicated the presence of subfractions in pool B, with different rates of solubilization. Such subfractions were not detectable from in situ, Dacron bag, estimates of N solubilization.

3. UDN was estimated using a dynamic mathematical model and rate-constants obtained from N solubilization in protease solution or in situ.For three grain mixtures tested using the protease technique the model predicted UDN values of 7, 10 and 12% compared with values of 47, 66 and 59% estimated in vivo. The full range of experimental feeds was tested using the in situtechnique and UDN values predicted by the model were used to derive UDN values for twelve mixed diets. The latter values were significantly but not closely correlated with those determined in vivo (r2 0·41, P < 0·05).

4. An attempt was made to simulate rumen proteolysis in vitro by choosing a protease enzyme concentration (0·066 units/ml) providing a proteolytic activity similar to that of whole rumen fluid. The experimental samples of feed were subjected to simulated rumen proteolysis for 18 or 48 h to resemble the mean retention times in the rumen for grain mixtures and roughages respectively. The residual N at the end of incubation was considered as an estimate of UDN. The UDN values estimated from simulated rumen proteolysis and those determined in vivo for twelve mixed diets were in close agreement (r2 0·61, P < 0·01).

5. Simulated rumen proteolysis can serve as a simple, rapid and sensitive method to estimate UDN in a variety of feedstuffs.

1. A method is described for the determination of protein degradation based on measurements of ammonia concentration and gas production (Menke et al. 1979) when a feedingstuff was incubated with rumen fluid in vitro.

2. NH3 liberated during incubation is in part used for microbial protein synthesis. Production of carbon dioxide and methane can be regarded as a measure of energy available for protein synthesis. The ratio, gas production: incorporation of NH3-nitrogen was estimated by addition of starch to the substrate. The response in gas production was linear in the range 0–200 mg starch, when starch was added to 0–200mg feedingstuff dry matter and 30 ml rumen fluid-medium mixture.

3. Linear regression between NH3-N concentration (y, mg) and gas production (x, ml) yielded an intercept (b0) representing that amountof NH3-N which would be released when no fermentable carbohydrates were available and consequently no bacterial protein synthesis took place.

4. The difference between this intercept b0 and NH3-N content in the blank (rumen fluid without substrate added) indicated the amount of NH3 liberated from protein and other N-containing compounds of the feedingstuff incubated. In vitro-degradable N (IVDN) was calculated as a proportion of total N by the equation:

1. To evaluate the role of gestational protein-energy malnutrition on fetal hard-tissue growth and metabolism, we measured several variables in the growth centres of mandibles and long bones of newborn rats.

2. Control pups and pups of malnourished dams had approximately the same extent of reduction in body-weight, mandibular weight and long-bone weight.

3. The malnourished group had more cells in the mandible although cell size was the same as that of controls.

4. In contrast, in the long bones, the malnourished group had fewer cells than did controls whereas cell size was unchanged.

5. Calcium content was the same in long bones of both groups, but was less in the mandibles of pups from malnourished dams. Ca metabolism as measured by 45Ca uptake was unchanged in the long bones, but was increased in the mandibles of the malnourished group shortly after birth. Calcification patterns at birth in these bones correlated well with alkaline phosphatase (EC 3.1.3.1) activity.

6. These findings indicate that the mandibles and long bones of offspring are affected differently by protein–energy malnutrition during the mother's gestation. Prenatal nutritional stress resulted in a disturbance of the pituitary–adrenal system. Increased adrenal corticosterone could possibly be related to the different observed changes in bone metabolism.

1. Kale (Brassica oleracea) and ryegrass (Lolium perenne)–clover (Trifolium repens) pasture grown under similar soil conditions were grazed in the vegetative state by growing lambs of 23·6 kg initial live weight for 24 weeks. Forty-eight lambs grazed each forage. The kale and pasture contained respectively 4 and 14 mg copper/kg dry matter (DM), 7·2 and 3·1 g total sulphur/kg DM and 0·4 and 1·1 mg molybdenum/kg DM.

2. Subcutaneous injections of Cu (12 mg) were given to half the animals grazing each forage during weeks 1, 6, 12 and 18.

3. All ninety-six animals were slaughtered at the end of the experiment and an additional group of twelve animals was slaughtered when the experiment commenced. Liver Cu was determined on all slaughtered animals and heart muscle cytochrome oxidase (EC 1.9.3.1) activity on those slaughtered at week 24. Blood samples removed at 6-week intervals were assayed for activity of superoxide dismutase (EC 1.15.1.1; SOD) and serum Cu concentration determined. Wool growth, live-weight gain and cytochrome oxidase activity of biopsied hind-limb muscle were also measured at 6-week intervals.

4. Control animals grazing pasture showed an accumulation of total liver Cu during the experiment. Animals grazing this diet and given Cu injections showed an additional accumulation of liver Cu equivalent to the supplementary Cu administered, but Cu supplementationdid not affect the activity of any of the Cu-containing enzymes measured and did not affect live-weight gain or wool growth.

5. Control animals grazing kale showed a depletion of total liver Cu and reductions in serum Cu concentrations during weeks 18 and 24. However, cytochrome oxidase activity inheart and hind-limb muscle was similar to that of pasture-fed lambs. Blood SOD activity was reduced by kale feeding, but a laboratory study showed this enzyme was not inactivatedby dimethyl disulphide produced from rumen fermentation of S-methyl-L-cysteine sulphoxide(SMCO) which occurs in kale.

6. Cu supplementation of kale-fed lambs had no effect on hind-limb cytochrome oxidaseactivity, but considerably increased total cytochrome oxidase activity in the heart and minimized the reduction in blood SOD activity. Both are considered to represent defence mechanisms against the haemolytic anaemia caused by SMCO. However, Cu supplementation did not affect live-weight gain or wool growth.

7. The proportion of injected Cu not recovered in the liver at slaughter was greater for lambs grazing kale than ryegrass–clover pasture and the same result was calculated from Barry et al. (1981) for growing cattle. It was therefore concluded that Cu requirements must be greater for growing sheep and cattle grazing kale than ryegrass-clover pasture.

1. Changes in concentrations of metabolites in blood and rumen fluid were measured in cows when fat and when thin and when given hay and concentrates at a restricted level or ad lib.

2. When fed ad lib., cows consumed 24% more food when thin than when fat.

3. Concentrations in rumen fluid of acetate and propionate after feeding at both levels did not differ significantly between fat and thin cows.

4. Concentrations in blood of the lipogenic substrates acetate and 3-hydroxybutyrate did not differ between fat and thin cows after feeding, but glucose concentration fell more rapidly in thin cows especially when fed ad lib. Propionate rose more in thin cows than in fat cows when fed ad lib. Non-esterified fatty acids were higher in fat cows before feeding, but fell to similar levels to those in thin cows after feeding.

5. The results are interpreted as indicating a more rapid rate of fatty acid synthesis in adipose tissue of thin cows. This in turn tends to lower blood levels of fatty acid precursors and so enhance the rate of their absorption from the rumen. This in turn permits higher intake of food before short-term chemical regulators of intake start to operate.

1. The absorption of methionine and methionine sulphoxide was compared in inverted rings from three sections of rat intestine.

2. In all sections, absorption of unoxidized methionine was 20–40% higher than that of methionine at low physiological substrate concentrations. At higher substrate levels, the difference was not so pronounced.

3. Inhibition studies indicated that absorption of methionine and of methionine sulphoxide took place by different absorption systems.

4. Glutathione increased the absorption of methionine and, to a greater extent, the absorption of methionine sulphoxide.

1. Chicks were given biotin-deficient diets containing either suboptimal (low) or supraoptimal (high) concentrations of protein from 1-d-old until they were used during their fourth week of life. The low-protein diet predisposed chicks to develop fatty liver and kidney syndrome and the high-protein diet to develop classical biotin deficiency signs. Two other groups, as controls, received biotin-supplemented rations.

2. Low dietary protein increased lipogenesis by isolated hepatocytes but had little effect on gluconeogenesis compared to high dietary protein.

3. Low dietary protein decreased activities of hepatic isocitrate dehydrogenase (EC 1.1.1.42), fructose-1,6-bisphosphatase (EC 3.1.3.11) and glucose-6-phosphatase (EC 3.1.3.9; GP) and increased activities of fatty acid synthase (FAS), citrate cleavage enzyme (EC 4.1.3.8; CCE) and malate dehydrogenase (decarboxylating) (EC 1.1.1.39).

4. When biotin deficiency was superimposed, the rate of lipogenesis by isolated hepatocytes (from fed birds) was decreased. Gluconeogenesis from lactate and glycerol was also depressed.

5. Activity of GP was further decreased by biotin deficiency on the low-protein regimen and FAS and CCE were further increased. PK activity was increased by biotin deficiency.

1. By using continuous infusions of 3H- and 14C-labelled substrates, three-pool models, incorporating rumen propionate, plasma glucose and blood carbon dioxide were constructed to determine the contribution of propionate to glucose in non-pregnant, pregnant (mid and late) and lactating hill ewes.

2. Although the intakes of non-pregnant and pregnant ewes were the same (1200 g driedgrass/d) and resulted in similar levels of propionate production (33 g C/d), glucose production rate (GPR) increased from 44 g C/d in the non-pregnant ewes to 62 g C/d in the ewes carrying twins in late pregnancy. In lactating ewes given 2500 g dried grass/d, propionate production increased to 56 g C/d and GPR increased to 93 and 104 g C/d in ewes suckling single and twin lambs respectively.

3. There was an increase in the percentage of the propionate resource which was diverted to glucose, from 37% in the non-pregnant ewes and ewes in mid-pregnancy, to 55% in late pregnancy and 60% in lactation. In spite of this apparent metabolic adaptation to the additional requirements for glucose, approximately 55% of the glucose-C was supplied by metabolites other than propionate and CO2.

4. From the determination of plasma glycerol concentrations it was estimated that themaximum possible contribution of glycerol-C to glucose was only 8–12 g C/d. Theremaining 40% of the glucose-C could not be accounted for and could have been derived from non-essential amino acids (NEAA).

5. In the non-pregnant and pregnant ewes only 62% of the GPR was oxidized to CO2. In the lactating ewes only 49 and 30% of the GPR was oxidized to CO2 in the ewes suckling single and twin lambs respectively.

6. In the majority of cases there was a marked similarity between the amounts of glucose-C apparently derived from NEAA and the amount of glucose-C which was not oxidized to CO2.

1. Various methods of estimating intestinal mass and cell proliferation were used to compare the effects of a pelleted laboratory diet (diet A), an elemental diet (Flexical; diet B), Flexical plus kaolin (diet C) and Flexical plus cellulose (diet D) on the mucosa and other tissues of the mouse small intestine and colon.

2. The weight of the distal third of the small intestine was significantly decreased in mice given diets B, C and D (P < 0·001). The length of the colon was significantly decreased in the mice given diets B (P < 0·001), C (P < 0·01) and D (P < 0·05); however, the weight of the colon was only decreased in the mice given diet B or diet C (P < 0·001).

Similar changes were noted in the weights of the intestinal muscle and serosa layers; however, no such changes were noted in the weight or DNA content of the mucosa.

3. No changes were observed in the crypt cell production rate in the small intestine, apart from an increase in the mid-region (P < 0·01) of mice given diet C, which was associated with a similar increase in the mucosal content of DNA.

Crypt-cell production was significantly decreased (P < 0·01, 0·001) in the two sites of the colon studied in the mice given diets B or C, but the crypt-cell production rates were not significantly different from control levels at either site in the mice given diet D.

4. The length of the colonic crypts was significantly decreased (P < 0·05, 0·02) in mice given diets B and C, but not in those given diet D; total cell number showed a similar change (P < 0·01, 0·05).

The correlation coefficient between the length of the colonic crypts and absolute cell number was 0·513.

5. The conclusion of the present study was that dietary fibre (cellulose), but not inert bulk (kaolin), prevents mucosal atrophy of the colon of mice given a fibre-free diet, and that many of the observed changes in intestinal weight are due to reduction in the bulk of the muscle layer.

1. Groups of kids were reared from birth to 5 d on goat's milk. On the 6th day five of the kids received by bottle a morning feed of goat's milk with [3H]folic acid added to saturate the folate-binding proteins (FBP) (Expt 1); three kids received raw goat's milk containing only the endogenous folate and hence a large surplus folate-binding capacity (FBC) (Expt 2). The contents of the stomach, duodenum, jejunum and ileum were recovered by washing out 1·5 h after feeding (Expt 1) or at 0·5, 1 and 3·5 h after feeding (Expt 2).

2. Recovery of [3H]folic acid 1·5 h after feeding (Expt 1) in all segments was 58·4%, mainly in a soluble form, most of this being in the stomach (37·0%) and ileum (14·3%). No surplus FBC was found in any gut segment. Sephadex G-75 chromatography of the soluble fractions of the contents of the various gut segments showed that [3H]folic acid remained bound to FBP throughout the stomach and small intestine. The bound [3H]folic acid exhibited a molecular weight of 81000 in stomach contents, similar to that in the milk feed, presumably representing an aggregated form of the FBP, whereas in the intestinal contents its molecular weight was 39000 indicating dissociation to monomer due to dilution in the recovery process.

3. In Expt 2, the total recovery of free FBP in all four gut segments was 67, 54 and 23% respectively at 0·5, 1 and 3·5 h after the milk feed, and the distribution of FBP along the gut at 1 h was similar to that of [3H]folic acid-labelled FBP at 1·5 h in Expt 1. In mature goat's milk the endogenous 5-methyltetrahydrofolate was shown to be associated with species of molecular weight 80000 and 38000.

4. The results indicate that goat's-milk FBP is relatively resistant to digestionby gastric and intestinal enzymes in vivo in the kid and survives along the length of thesmall intestine.

5. The implications of the findings are discussed in relation to the possible influence of FBP on uptake of folate by mucosal cells and their relevance to neonatal folate nutrition.

1. Two dairy cows were maintained by intragastric infusion of volatile fatty acids and casein. Except when fasting, the casein-nitrogen was held constant, while total gross energy supply was varied from zero during fasting to 650 kJ/kg body-weight (W)0·75.

2. One cow was estimated to attain zero N balance at an energy intake of 255 kJ/kg W0·75 and the other at 307 kJ/kg W0·75, which was calculated to be substantially below the estimated energy required for zero energy balance.

3. When the cows were later given an N-free infusion for a period preceding the trial, N balance occurred at 98 kJ/kg W0·75 for one cow and 115 kJ/kg W0·75 for the other.

4. Four steers were similarly nourished by intragastric infusion and the energy nutrient increased from 0 at fasting to 450 kJ/kg W0·75. The protein was held constant at 1 g N/kg W0·75 except at fasting. The energy level at which N balance occurred was 154 (SE 38) kJ/kg W0·75 or approximately equal to the energy content of the protein. The practical implications of these findings are discussed.

1. Male weanling rats were maintained on diets either deficient or adequate in zinc for a period of 4 weeks. The rats on the deficient diet showed a reduction in food intakesand growth.

2. After 4 weeks both soleus muscles and the lateral portion of the diaphragm were studied histochemically to examine the relative frequencies of the fibre types.

3. The soleus muscles of the deficient animals showed a significant change in the proportion of slow and fast fibres. The diaphragm muscles of the deficient animals had a significant increase in the proportion of fast-twitch oxidative glycolytic fibres and a significant decrease in fast-twitch glycolytic fibres compared with the controls.

4. Stainable lipid increased in the diaphragm muscle of the deficient animals with respect to their pair-fed controls.

1. Everted sacs of rat duodenum and ileum were used to study the effect of anions and organic ligands on the absorption of zinc. The uptake per unit weight of tissue was greater in duodenum than ileum, and it was influenced by the Zn concentration and pH of the incubation medium.

2. The Zn uptake from inorganic salts in simple buffered medium varied in the order zinc sulphate > zinc chloride > zinc phosphate. Zinc acetate was more effective and zinc citrate less effective than ZnCl2. Addition of aspartic acid or histidine to ZnCl2 increased the uptake but galactose or lactose decreased it. 2-Picolinic acid greatly increased the Zn uptake but 4-picolinic acid reduced it.

3. When incubated with intestinal sacs after incorporation into a synthetic rat diet, only ZnSO4 and 2-picolinic acid increased Zn uptake compared with ZnCl2, but zinc citrate and 4-picolinic acid still tended to decrease it.

4. Metabolic balance studies showed no significant differences in the faecal excretion, total excretion or retention of Zn between rats receiving diets containing different forms of Zn. ZnSO4, zinc citrate and particularly 2-picolinic acid increased the urinary excretion of Zn.

5. The significance of these results is discussed in relation to the suitability of methods for investigating Zn absorption and the importance of Zn-binding ligands.

1. The effect of xylitol feeding on the iron content of rat tissues was studied.

2. Adult male rats were fed on the basal diet containing (g/kg) 200 glucose, or 50 or 200 xylitol, or the same diet containing no added carbohydrates for 8 weeks. Each feeding group comprised nine animals.

3. Xylitol at 200 g/kg diet retarded the growth rate of the rats, whereas 200 g glucose/kg increased the weight gains compared with animals given no added carbohydrates.

4. Xylitol at 50 g/kg did not affect the tissue Fe concentrations, but 200 g xylitol/kg increased the Fe content of the livers, duodenum wall, spleen, bone marrow and serum.

5. Cadmium and lead contents of the livers were similar in all groups.

6. Xylitol–Fe complexes are suggested to be responsible for the increased Fe absorption during xylitol feeding.

1. The effect of cobalamin inactivation by the anaesthetic gas nitrous oxide on the concentration of S-adenosylmethionine (Ado Met) in brain and liver of fruit bats (Rousettus aegyptiacus) was examined.

2. Test animals exposed to N2O–oxygen (50:50, v/v) developed ataxia and paralysis leading to death after an average of 9·8 weeks (n 6). Animals receiving pteroylmonoglutamic acid supplements in the diet became ataxic earlier (mean 8·8 weeks) while those receiving methionine supplements survived for significantly longer periods (12·5 weeks, P < 0·01).

3. Plasma cobalamin levels indicated severe depletion of cobalamin stores in N2O-exposed animals.

4. The mean concentration of Ado Met in the brain of N2O-treated bats was nearly 50% higher than that of untreated controls. Ado Met levels in treated bats receiving pteroylmonoglutamic acid or methionine supplements were respectively 18 and 25% higher than in controls. In contrast, the concentration of Ado Met in the liver of all the N2O-treated groups was slightly lower than in controls.

5. These results suggest that the N2O-induced neuropathy in the fruit bat is not related to a depletion of Ado Met in the nervous system.

1. The effects of feeding either a high-protein (HP) diet or a low-protein (LP) diet between 1·8 and 15 kg live weight (LW) and a low-energy (LE) or a high-energy (HE) intake but at the same protein intake subsequent to 15 kg LW on the performance and body composition of pigs growing to 75 kg LW were investigated.

2. During the LW period 1·8–15 kg, pigs given the LP diet exhibited poorer growth performance (P < 0·01) and at 15 kg contained more fat (P < 0·01) in their empty bodies than pigs given the HP diet.

3. On the LE treatment subsequent to 15 kg LW, pigs previously given the LP diet deposited protein at a faster rate and exhibited more rapid and efficient growth to 60 kg LW than those given the HP diet before 15 kg. However, on the HE treatment, pigs previously given the LP diet deposited protein at a slower rate and exhibited poorer growth performance (P < 0·05) between 15 and 45 kg LW but grew at a faster rate between 45 and 60 kg LW than pigs previously given the HP diet.

4. On the LE treatment subsequent to 15 kg LW the differences in body composition between the two protein groups were no longer significant at 45 kg. However, on the HE treatment, pigs previously given the LP diet remained fatter (P < 0·05) to 60 kg LW than those previously given the HP diet.

5. The results suggested that restricting protein intake between 1.8 and 15 kg LW reduced, temporarily, the upper limit of protein retention and growth performance during subsequent development. This finding is discussed in relation to the effects of protein nutrition in early life on the hyperplasic development of muscle tissue.

1. The effects of feeding either high-protein (HP) or low-protein (LP) diets between 1·8 and 15 kg live weight (LW) and a low-energy (LE) or high-energy (HE) intake subsequently on the cellularity of muscle and adipose tissue in pigs growing to 75 kg LW were investigated.

2. The effects of the nutritional treatments on muscle tissue were assessed from the weight and DNA content of the m. adductor. For adipose tissue the total DNA content and fat cell size of the subcutaneous adipose tissue contained in the left shoulder joint were determined.

3. Feeding the LP diets in early life reduced the weight and DNA content of the m. adductor (P < 0·01) and increased fat cell size (P < 0·01) at 15 kg LW.

4. Subsequent to 15 kg there was an almost linear increase in muscle DNA with increasing LW, and the difference between pigs from the initial protein treatments progressively diminished and was no longer apparent at 60 kg LW.

5. At 30 kg LW, pigs given the LP diets before 15 kg LW contained less DNA in the subcutaneous adipose tissue from the shoulder joint (P < 0·01) and had larger fat cells (P < 0·05) than pigs given the HP diets initially. However, adipose DNA and fat cell size increased with increasing LW and the differences resulting from the initial protein treatments progressively diminished. On the LE and HE treatments subsequent to 15 kg these differences were no longer evident at 45 and 60 kg respectively.

6. Pigs given the HE intake subsequent to 15 kg, contained less DNA in muscle tissue (P < 0·05) at 60 and 75 kg LW and had larger fat cells (P < 0·05) at 45, 60 and 75 kg LW, than pigs on the LE treatment.