1. Eight healthy male volunteers (aged 22–39 years) supplemented their normal daily diet with 15 g encapsulated fish oil (MaxEPA) for a 6 week period. Fasting blood samples were taken before, at the completion of and 3 months after the period of supplementation.

2. Evaluation of nutrient intakes showed that the intake of n-3 polyunsaturated fatty acids rose significantly (P < 0·01) during supplementation. This was reflected in changes in the fatty acid composition of platelet phosphatidyl choline (PC) and phosphatidyl ethanolamine (PE) without any changes in phosphatidyl serine, phosphatidyl inositol or sphingomyelin.

3. In both PC and PE there were significant (P < 0·05) increases in the levels of 18:1 n-9 and 20:5n-3 fatty acids and a significant (P < 0·05) decrease in 20:4n-6 during supplementation. 16:0 rose significantly in PC (P < 0·05) while in PE, 18:0 fell and both 22:5n-3 and 22:6n-3 rose significantly (P < 0·05).

4. There were no significant effects of fish-oil supplementation on serum lipids, platelet cholesterol: phospholipid, collagen-induced platelet aggregation or collagen-induced platelet thromboxane B2 production. However, there was a significant correlation (P < 0·001; r+0·63) between total phospholipid arachidonic acid and platelet thromboxane production.

5. The fluorescent probe 1,6-diphenyl-1,3,5-hexatriene was used to determine whether fish-oil supplementation altered fluorescence polarization of isolated platelet plasma membrane and, by inference, platelet plasma membrane fluidity. No significant effect of fish-oil supplementation on fluorescence polarization was seen.

1. Peripheral blood mononuclear cells from eight healthy volunteers were cultured, with or without concanavalin A (Con A), in a medium containing (ml/1) 100 normal autologous serum, 100 experimental autologous serum or 100 heterologous (fetal calf) serum.

2. The control and experimental autologous sera were obtained from the volunteers, before and after 15 d supplementation with 15 g fish oil (MaxEPA)/d to provide 1.5 g eicosapentaenoic acid (EPA; 20:5n-3)/d. The sera were frozen at -20°. The level of EPA increased from trace quantities in the control autologous serum to 14.3% w/w free fatty acids and between 6.9 and 8.1% w/w lipoprotein phospholipids in the experimental autologous serum. The heterologous fetal calf serum was enriched with EPA, complexed with bovine serum albumin, to provide a final concentration of 15μg/ml. All culture medium contained 10 ml fresh autologous serum/l and cells were obtained from the volunteers for the culture studies about 60 d after the end of EPA supplementation.

3. Portions of cells were removed from culture at 36, 48 and 72 h for phospholipid fatty acid analysis.

4. The level of EPA in phospholipids of cells cultured with exogenous EPA in fetal calf serum was increased significantly (P < 0.05) at all sampling times, both with and without Con A. By 48 h the levels had peaked at 15.8 (se 2.7) and 18.4 (se 4.5) % w/w respectively.

5. Resting cells, i.e. with no Con A present, failed to accumulate EPA when cultured in the experimental autologous serum containing 8.6% w/w total lipids as endogenous EPA. When cells were stimulated to divide by the mitogen Con A, there was a gradual increase in phospholipid EPA levels of mononuclear cells cultured with experimental autologous serum containing endogenous EPA. The maximum level obtained was at 72 h (7.2 (se 2.7)).

6. In a separate experiment, the effect of Con A at increasing doses (0, 1,5, 10 and 20 μg, final concentration) on lymphocyte blastogenesis was determined by estimating the incorporation of 3[H]thymidine. No significant effects due to serum in the culture medium were observed.

1. Anthropometric indices are presented for 402 healthy Chinese elderly subjects leading an active life in the community in Hong Kong.

2. Women had higher body-mass index (weight/height2) and body fat, while fat-free mass, arm-muscle circumference and corrected arm-muscle area were higher in men. Body-mass index, fat-free mass, arm-muscle circumference and corrected arm-muscle area did not decline with age. Total body fat was lower in women aged 75 years and above compared with those aged 60–64 years.

3. All values were lower than those for elderly Caucasians. A different criteria for severe wasting malnutrition among elderly Chinese should be established.

1. Five healthy volunteers whose usual fat and energy intakes were moderately high (fat intake 155 (SE 11) g/d; energy intake 13683 (SE 909) kJ/d) were given on two separate occasions (a) 96 g fat and (b) 96 g fat and intravenous (IV) glucose (250 g glucose/1; 100 ml followed by a 2 ml/min infusion for 180 min).

2. Subjects continued on a low-fat diet for 35 d (fat intake 25 (SE 4) g/d; energy intake 6976 (SE 539) kJ/d) and the tests repeated.

3. The gastric inhibitory polypeptide (GIP) response to oral fat was significantly attenuated by IV glucose whilst subjects were consuming their normal diets and the GIP response to fat alone was significantly diminished during the low-fat diet. Post-prandial plasma triglycerides, light scattering indices (LSI; an index of post-prandial chylomicronaemia) and paracetamol levels paralleled the integrated GIP responses on both normal and low-fat diets.

4. The study of oral fat with or without glucose was repeated on seven further volunteers consuming their usual diet, substituting 10 MBq 99Tcm-labelled tin colloid for the paracetamol to investigate the rate of gastric emptying by radionuclide imaging.

5. Plasma GIP, insulin, triglyceride and LSI levels were similar to those found in the first study. IV glucose almost doubled the gastric emptying time of the oral fat load (half emptying time (t½) 148 (SE 11) min after fat alone and 224 (SE 18) min after fat and IV glucose). Post-prandial plasma motilin levels were significantly depressed by IV glucose.

6. We conclude that (a) the GIP response to oral fat is attenuated both by IV glucose and by a low-fat diet, (b) the delay in gastric emptying induced by IV glucose may be partly responsible for the diminished GIP response to oral fat when IV glucose is infused, (c) it is possible that some of the changes observed with IV glucose are mediated via changes in motilin.

1. The thermogenic, cardiovascular and metabolic responses to the ingestion of a 30 kJ/kg body-weight test meal were studied in six normal weight, female subjects before and after a 7 d period of underfeeding at 60 kJ/ kg ideal body-weight per d.

2. With underfeeding there were decreases in body-weight, plasma insulin and 3,5,3'-triiodothyronine concentrations, resting metabolic rate and respiratory exchange ratio, with increased blood ketone levels. Baseline ‘arterialized’ venous plasma noradrenaline and adrenaline concentrations were not affected by underfeeding.

3. Ingestion of the test meal caused similar increases in heart rate and calf blood flow and changes in blood pressure in the fed and underfed states. There was a greater glycaemic response to the test meal in the underfed state compared with the fed state although the rise in plasma insulin concentration was similar and ketogenesis was suppressed. The increases in metabolic rate and plasma noradrenaline concentrations following the test meal were similar in the fed and underfed states.

4. Although the period of underfeeding in the present study led to considerable metabolic adaptation, and some alteration in physiological responses to ingestion of a test meal, there was no evidence that there were associated changes in sympathetic nervous system activation.

1. The effect of unsaturated long-chain non-esterified fatty acids (NEFA) on the human leucocyte sodium-pump was studied in vivo.

2. Plasma NEFA level was raised acutely from 0·28 (sd 0·10) to 2·54 (sd 0·48) mmol/l by infusion of ‘Intralipid 20%’ (trademark) at 90 ml/h with heparin, and the human leucocyte 22Na efflux rate constants were studied in eight normal weight males.

3. After 3 h, there was a significant lowering of the total (from 3·97 (sd 0·92) to 3·10 (sd 0·71)/h; P < 0·01) and ouabain-sensitive 22Na efflux rate constants (from 2·89 (sd 0·55) to 2·37 (sd 0·62)/h; P < 0·02). Ouabain-insensitive efflux rate constants showed a tendency to fall (from 1·08 (sd 0·51) to 0·73 (sd 0·23)/h). Leucocyte potassium content remained unchanged, but sodium content rose from 31 (sd 12) to 38 (sd 18) mmol/kg dry weight (P < 0·05). Total, ouabain-insensitive and ouabain-sensitive efflux rates did not change significantly during the Intralipid-heparin infusion.

4. Plasma insulin levels rose gradually throughout the 3 h infusion period.

5. In conclusion, NEFA, when raised to pathological levels, can inhibit the leucocyte Na-pump in vivo even in the presence of physiological levels of serum albumin, and may increase insulin secretion.

1. The effect of dietary methionine on the utilization of selenium from dietary selenomethionine ([Se]Met) for tissue Se deposition and for glutathione peroxidase (EC 1.11.1.9; GSH-Px) synthesis was studied in male weanling rats.

2. When rats were given 0.5 mg Se as [Se]Met/kg diet supplemented with 0, 4 or 9 g methionine/kg, Se in plasma, erythrocytes, liver and muscle increased significantly over the 20 d period for all methionine-treatment groups. The increases in erythrocyte and muscle Se, however, were significantly higher in rats fed on the methionine-deficient diet compared with the methionine-supplemented diets.

3. In contrast to the increases in tissue Se, GSH-Px activity in liver, plasma and muscle decreased in methionine-deficient rats given 0.5 mg Se as [Se]Met/kg whereas GSH-Px activity was maintained or increased in rats supplemented with methionine.

4. The percentage of tissue Se associated with GSH-Px was calculated from the measured Se concentration and GSH-Px activity. A significantly lower percentage of Se was associated with GSH-Px in methionine-deficient rats compared with methionine-supplemented rats.

5. These results show that Se from dietary [Se]Met is preferentially incorporated into body proteins rather than used for GSH-Px synthesis when methionine is limiting in the diet.

6. These results further suggest that [Se]Met might not be the optimum Se compound to use for Se supplementation because metabolism of dietary [Se]Met to a biochemically active form, such as GSH-Px, was impaired when [Se]Met was provided in diets low in methionine.

1. To study the efficiency of rehabilitation after different periods of protein–energy malnutrition, we used as a model preweaning malnourished rats. After weaning, male Wistar rats were fed on a protein-deficient diet (50 g casein/kg) ad lib. for the whole study (DR group) or rehabilitated with normal diet (180 g casein/kg; RR group) from weaning, week 0, or weeks 1, 3, 5, 8 and 16 thereafter.

2. Twelve animals from the DR group were killed at the beginning of each rehabilitation period. The twelve rehabilitated rats were killed after 2 weeks. Body-weight and epididymal adipose tissue weight, blood glucose, plasma immunoreactive insulin (IRI) and immunoreactive glucagon (IRG), and pancreatic contents of IRI and IRG were determined.

3. Food intake of RR rats rose significantly except during the last period where body-weight increased less than that during the previous periods. Fat-pad weights increased in the same manner in DR and RR groups.

4. Blood glucose fell and plasma IRG rose significantly without any change in plasma IRI after each rehabilitation period, except during the last period where blood glucose concentrations became stable. Pancreatic IRG and IRI showed the same type of response to those of the plasma.

5. All short-term rehabilitation periods were similarly efficient at producing catch-up growth. High insulin sensitivity of target cells was responsible for good recovery except after long-term malnutrition.

1. Protein fractional synthesis rate (FSR) was measured in some major tissues and in the whole body of six 1-week-old sucking lambs by a large injection of L-[3H]valine.

2. Upper estimates of tissue protein FSR (%/d), assuming that the tissue-homogenate free-valine specific radioactivity defined that of valyl tRNA, were 115.0 in liver, 24.1 in skin, 22.9 in the white M. tensor fasciae latae, 21.6 in the red M. diaphragma and 19–6 in the remainder (exsanguinated whole body without liver and gastrointestinal tract) of lambs.

3. Absolute synthesis rates (ASR) of tissue protein were 17, 19 and 42 g/d in the liver, skin and skeletal muscle respectively, and 112 g/d in the remainder. The ASR of whole-body protein, derived from the tissue values, was 146 g/d, i.e. 33 g/d per kg body-weight. The calculated whole-body protein FSR was 23.9 %/d.

4. The relative percentage contribution of liver, skin and skeletal muscle to whole-body protein synthesis was 11.7, 13.1, and 29.0.

5. We concluded that tissue protein FSR in lambs were in exactly the same decreasing order, from visceral tissues to skeletal muscles, as observed in rats. The ovine FSR estimates and the partitioning of protein synthesis between tissues were in the same range as values recently obtained by flooding-dose experiments in immature rats, piglets, and even in chicks. These findings suggest that inter-species differences are rather limited.

1. Male Sprague-Dawley rats weighing 146·5 (se 4·3)g were fed on a semi-synthetic diet containing 0, 25 or 150 mg cimaterol/kg for 12 d. Net changes in weight and composition of carcass, liver, heart, gastrointestinal tract, gastrocnemius plus plantaris muscles, skin and remainder were estimated by comparative slaughter.

2. Cimaterol increased protein gains in gastrocnemius plus plantaris muscles from 0·09 g in controls to 0·14 and 0·12 g in 25 and 150 mg cimaterol/kg groups respectively. Carcass protein gains increased from 6·27 g in controls to 8·00 and 7·05 g in 25 and 150 mg cimaterol/kg groups respectively.

3. Rats treated with cimaterol either gained less fat or actually lost fat from all tissues studied, whilst control rats gained fat. These changes were reflected in lower energy retention in cimaterol-fed rats.

4. Energy intake was not affected by treatment. Cimaterol increased heat production from 776 kJ/kg body-weight0·75 in controls to 863 kJ/kg body-weight0·75 in both treated groups. Gross efficiency was reduced from 17·4% in controls to 8·0 and 7·7% in rats fed on 25 and 150 mg cimaterol/kg diets respectively.

5. These results indicate that cimaterol increases protein gain at the expense of fat in rats. In addition, subcutaneous adipose tissue appears to be more sensitive than abdominal fat, whilst protein gains are particularly enhanced in skeletal muscle relative to other body tissues.

1. Concentrations of amino-nitrogen, glucose, reducing sugars and lactic acid in blood obtained from arterial and portal permanent catheters were measured together with the portal hepatic blood flow-rate during a post-prandial period of 8 h in twenty unanaesthetized pigs (initial mean body-weight 52·3 (SEM 0·9) kg) receiving experimental meals (200–1000 g) at 3–4 d intervals from 6–8 to 20 d after surgical implantation of the catheters and electromagnetic flow probe. The semi-synthetic starch-based diets contained variable amounts of fish meal giving crude protein (nitrogen x 6·25; CP) concentrations (g CP/kg) of 80 (seven meals), 120 (twenty-two meals), 160 (six meals) and 240 (nine meals).

2. After the meal the concentration of amino-N increased with increasing levels of protein intake and increased more in the portal than in the arterial blood. There were significant relations between amounts of amino-N appearing in the portal blood at various time-intervals after the meal and the level of protein intake. Values for the ratio, amount absorbed within 8 h: amount ingested (absorption coefficient: 0·633 for a mean intake of 13·4 g N) decreased with increasing level of protein intake.

3. There was a rise in glycaemia after the meal, increasing with the amount of carbohydrate eaten, and this was more marked in the portal than in the arterial blood. There were also significant relations between amounts of glucose absorbed and amounts of starch ingested. However, the appearance of glucose in the portal blood was less marked than that of amino-N since the absorption coefficient within 8 h was lower (0·504 for a mean intake of 291 g reducing sugars). This was most probably due to a larger uptake of glucose by the intestinal cell wall.

4. Amounts of lactic acid appearing in the portal vein during the post-prandial period did not depend on amounts ingested; they ranged from 3 to 1·6 g/h from the 1st to the 8th hour after the meal.

1. Concentrations of amino acids (AA) in blood obtained from arterial and portal permanent catheters were measured together with the portal hepatic blood flow-rate during a post-prandial period of 8 h, in nine conscious pigs (initial mean body-weight 49.3 (SEM 1.8) kg) receiving experimental meals (500, 600 or 1000 g) at 3–4 d intervals from 6–8 to 20 d after the surgical implantation of the catheters and electromagnetic flow probe. The semi-synthetic starch-based diets contained variable amounts of fish meal giving crude protein (nitrogen x 6.25; CP) concentrations (g/kg) of 80 (seven meals), 120 (seven meals), 160 (five meals) and 240 (three meals).

2. During the post-prandial period, variations in the individual blood AA concentrations were parallel to those of total amino-N (Rérat et al. 1988) to a greater or lesser extent according to the AA considered. Portal concentrations, which always exceeded arterial ones (except for glutamic acid and glutamine), increased quickly and, after a peak, returned slowly to reach initial values (small intake) or above (large intake) after 8 h.

3. Relations between amounts of each AA appearing during 8 h after the meal and amounts ingested were characterized by a highly significant linear regression (with the exceptions of glutamic acid and cystine). There were also close and significant relations between amounts of AA absorbed during the first 2 and 4 h after the meal and the amounts ingested.

4. For a mean intake of 90 g CP, aromatic AA showed the highest hourly absorption coefficients (about 010 /h), and sulphur-AA (0.07/h), lysine (007/h) and arginine (0.056/h) the lowest ones. Alanine was synthesized (amounts absorbed within 8 h exceeding those ingested) at the expense of glutamic acid (absorption coefficient 001/h).

5. For a given period of time, the AA abscrption coefficients decreased with increasing intake, but not in the same proportions for all AA, resulting in an enrichment (lysine, arginine, serine, proline) or depletion (branched- chain AA, histidine) of the absorbed mixture.

6. Some substances of the urea cycle were synthesized in rather large amounts in the gut wall (for a mean level of intake of 90 g CP: citrulline 2.41 g/8 h, ornithine 1.09 g/8 h). Blood glutamine was taken up by the gut wall in larger amounts (4.28 g/8 h).

1. Six non-anaesthetized pigs (mean body-weight 57.0 kg) were used to study the intestinal absorption of amino acids (AA) from either an enzymic hydrolysate of milk (PEP) containing a large percentage of small peptides (about 50% with less than five AA residues) and very few free AA (8%), or from a mixture of free AA with an identical pattern (AAL) infused intraduodenally in one of two amounts (55 or 110 g). Concomitant insulin and glucagon production rates were estimated.

2. Each pig was previously fitted, under anaesthesia, with an electromagnetic flow probe. around the portal vein, with permanent catheters in the portal vein, the carotid artery and the duodenum. Each infusion was performed after an 18 h fasting period and each pig received each infusion. The observation period lasted for 5 h.

3. The absorption of AA was greater, more rapid and more homogeneous after PEP infusion than after AAL infusion, independent of the amount infused.

4. For the majority of AA considered individually, the absorption coefficient was higher after infusion of PEP than after that of AAL. The exceptions were methionine with a higher absorption coefficient after AAL infusion, and isoleucine, aspartic acid + asparagine and glutamic acid + glutamine with identical coefficients for both infusions.

5. Some AA, such as asparagine, ornithine, citrulline and taurine, while absent in the infusates, appeared in the portal vein in appreciable amounts after the infusion of both solutions. While a small proportion of taurine may arise from recycling of taurine-containing bile salts, it seems that the gut wall is able to synthesize all four AA.

6. Insulin production did not differ according to the nature or amount of solutions infused. Glucagon production was greater after PEP infusion.

1. Pure swards of perennial ryegrass (Loliwn perenne cv. Melle) as a primary growth (May), trimmed primary growth (early June) and regrowth (late June), and white clover (Trifolium repens cv. Blanca) as a mature primary growth (July) and vegetative regrowth (August), were grazed by twelve Friesian steers (mean body-weights throughout experiment 152–231 kg) at daily allocation rates of forage which provided dry matter (DM) intakes ranging from 28 to 36 g/kg body-weight).

2. Total nitrogen contents of the three ryegrasses declined with season (37–20 g/kg DM), but in vitro organic matter (OM) digestibilities were relatively constant (0.76–0.79). The clovers had higher N (average 366 g/kg DM) and lower fibre and water-soluble carbohydrate contents than the grasses, and in vitro OM digestibilities of 0.70 and 076 respectively.

3. Duodenal digesta samples were obtained, using a portable sampling apparatus from the animals grazing the pasture, and estimates of the flow of nutrients into the small intestine were derived using two indigestible markers which were continuously infused into the rumen using a portable infusion apparatus. Forage intakes were calculated from estimates of faecal output of indigestible OM and the predicted in vivo OM digestibilities of the forages consumed. Coefficients of variation for OM flow to the small intestine and OM intake were 11.8 and 10.9 % respectively.

4. The apparent digestion of OM in the rumen ranged from 722 to 741 g/kg digestible OM intake and from 681 to 711 g/kg digestible OM intake for the grass and clover diets respectively. Substantial losses of ingested N before the small intestine were measured on all diets except the regrowth ryegrass. Losses were 0.30 and 040 g/g N intake on the primary growth ryegrass and the regrowth clover respectively; N contents were 37 and 39 g/kg DM respectively.

5. Comparison of the values obtained at pasture with that obtained when similar diets were offered to housed cattle (Beever et al. 1985) indicated that combined relations relating duodenal OM and non-ammonia-N (NAN) flows to OM and N intakes respectively could be established for each diet. For two diets (primary-growth grass and regrowth clover) the relations were curvilinear; for the remaining diets, the derived relations were linear.

6. The effects of forage species, stage of maturity and variations in the stem: leaf ratio in the grasses, on nutrient supply in grazing animals are discussed, and a combined relation for all grazed and harvested forages examined in this and in a previous study reported by Beever et al. (1985) is established. The relation Y = 1.430 - 0.0169X (r072, residual sd 0.140). where Y is NAN flow/N intake (g/g) and X is N content in the crop (gfkg OM), indicates that net losses of N across the rumen will occur on diets which contain more than 25.5 g N/kg OM.

7. Possible limitations in the techniques available to measure nutrient digestion and supply in grazing ruminants are discussed.

1. Two experimental approaches were employed to assess the relation between food consumption rate and maintenance requirements in male weanling rats. The first approach involved restricting food intake in rats previously given free access to food from weaning to 59 d of age. The second approach involved restriction of food intake to various levels after weaning. Maintenance requirements (g foodid per g body-weight (W)) were estimated by dividing the rate of food consumption by the resulting equilibrium W (EBW) for each animal. In addition, food consumption was partitioned into growth-independent (maintenance) and growth-dependent (gain) components by alternately setting W and specific growth rate (W') to zero in an equation relating food intake rate to W and W. Coupling coefficients representing maintenance consumption (g food/d per g W) and gain consumption (g food/g gain) were estimated for each animal by least squares.

2. Both techniques for estimating maintenance consumption provided similar estimates within and across experiments, and regardless of when food restriction was imposed or its severity, consumption for maintenance was about 5% W/d.

3. The EBW to which animals in each treatment group aspired was directly proportional to that group's food intake rate.

4. Coventional measures of growth efficiency were also related to food intake; efficiency decreased with decreasing food intake. Partitioning food consumption into maintenance and gain components revealed that as the rate of food intake decreased, the proportion of total intake consumed for maintenance increased. The results suggest that growth efficiency declines during food intake restriction because proportionately more of total intake is used for maintenance, leaving less available for gain.

1. [3H]leucine infused directly into the rumen of sheep labelled microbial protein and, when compared with the specific activity of the rumen free-leucine pool, indicated that 50% of the bacterial protein leucine originated from the rumen free-leucine pool.

2. The lower limit for bacterial protein turnover in the rumen was 0.37/d when calculated as the difference between the specific rate of disappearances of labelled bacteria from the rumen and the liquid-phase dilution rate.

3. Intravenously infused leucine also labelled the rumen bacteria. The build-up of specific activity in the rumen bacteria was sigmoidal and did not resemble that of the salivary protein which suggested that the rumen epithelium was a major endogenous protein input to the rumen. Additionally, bacteria isolated from the rumen epithelium had high radioactivity indicating that they were ingesting the rumen epithelial cells.

1. Sweet-potato (Ipomoea batatas L. Lam.) strips (SPS) mixed with maize powder (CP) in proportions 10:0, 9:1, 8:2, and 7:3 were ensiled for 1, 2 or 3 months.

2. Trypsin inhibitor activity (TIA) decreased during ensilage in samples of all treatments while the SPS–CP mixture (7:3, w/w) ensiled for 3 months contained the lowest TIA.

3. SPS–CP (8:2, w/w) dried or ensiled for 2 months, or ensiled for 2 months and dried, were each mixed with twice the amount of control diet (1:2, w/w) to make three diets. These three diets together with the control diet were used for a feeding experiment with rats to evaluate the nutritive value.

4. General composition analysis (including metabolizable energy), fatty acid composition and amino acid analysis (including percentage of essential amino acids) of the samples did not change during ensilage to an extent which could explain the improved performance of rats fed on ensiled diets.

5. Rats fed on diets containing dried SPS–CP (8:2, w/w) showed significantly lower (P < 0.05) body-weight gain than rats fed on the control diet or ensiled SPS diets, at the end of the 8th week. They also showed enlargement of the pancreas. The adverse effect of SPS was associated with TIA which seemed to be prevented to some extent by ensilage.

6. The possibility that the starch of SPS may also contribute to the adverse effect cannot be excluded at present.

1. Polysaccharide depolymerases and glycoside hydrolases involved in the breakdown of plant structural polysaccharides (hemicellulose and pectins) were monitored in three fractions of the liquid phase of horse caecum digesta: acellular fluid (AF), bacteria (B) and protozoa plus bacteria (PB).

2. Both bacteria and protozoa were found to be involved in the decomposition of pectic substances, with two enzymic activities: depolymerase (polygalacturonase, EC 3.2.1.15; and pectin lyase, EC 4.2.2.10) and esterase (pectinesterase, EC 3.1.1.11). The activity of the PB fraction was higher than that of B.

3. With hemicellulosic substrates, all three fractions showed a significant xylan endo-1,3-β-xylosidase (EC 3.2.1.32) activity. Mannan was hardly broken down.

4. Galactomannan and arabinogalactan were broken down more extensively by the PB fraction than by the B fraction. Glycosidase activities (xylan 1,4-β-xylosidase, EC 3.2.1.37 and α-L-arabinofuranosidase, EC 3.2.1.55) were also observed.