This study examined the influences of coated folic acid (CFA) and coated riboflavin (CRF) on bull performance, nutrients digestion and ruminal fermentation. Forty-eight Angus bulls based on a randomised block and 2 × 2 factorial design were assigned to four treatments. The CFA of 0 or 6 mg of folic acid/kg DM was supplemented in diets with CRF 0 or 60 mg riboflavin (RF)/kg DM. Supplementation of CRF in diets with CFA had greater increase in daily weight gain and feed efficiency than in diets without CFA. Supplementation with CFA or CRF enhanced digestibility of DM, organic matter, crude protein, neutral-detergent fibre and non-fibre carbohydrate. Ruminal pH and ammonia N content decreased and total volatile fatty acids concentration and acetate to propionate ratio elevated for CFA or CRF addition. Supplement of CFA or CRF increased the activities of fibrolytic enzymes and the numbers of total bacteria, protozoa, fungi, dominant fibrolytic bacteria and Prevotella ruminicola. The activities of α-amylase, protease and pectinase and the numbers of Butyrivibrio fibrisolvens and Ruminobacter amylophilus were increased by CFA but were unaffected by CRF. Blood concentration of folate elevated and homocysteine decreased for CFA addition. The CRF supplementation elevated blood concentrations of folate and RF. These findings suggested that CFA or CRF inclusion had facilitating effects on performance and ruminal fermentation, and combined addition of CFA and CRF had greater increase in performance than CFA or CRF addition alone in bulls.

To investigate the influences of cobalt (Co) and folic acid (FA) on growth performance and rumen fermentation, Holstein male calves (n 40) were randomly assigned to four groups according to their body weights. Cobalt sulphate at 0 or 0·11 mg Co/kg DM and FA at 0 or 7·2 mg/kg DM were used in a 2 × 2 factorial design. Average daily gain was elevated with FA or Co supplementation, but the elevation was greater for supplementing Co in diets without FA than with FA. Supplementing FA or Co increased DM intake and total-tract nutrient digestibility. Rumen pH was unaltered with FA but reduced with Co supplementation. Concentration of rumen total volatile fatty acids was elevated with FA or Co inclusion. Acetate percentage and acetate to propionate ratio were elevated with FA inclusion. Supplementing Co decreased acetate percentage and increased propionate percentage. Activities of xylanase and α-amylase and populations of total bacteria, fungi, protozoa, Ruminococcus albus, Fibrobacter succinogenes and Prevotella ruminicola increased with FA or Co inclusion. Activities of carboxymethyl-cellulase and pectinase increased with FA inclusion and population of methanogens decreased with Co addition. Blood folates increased and homocysteine decreased with FA inclusion. Blood glucose and vitamin B12 increased with Co addition. The data suggested that supplementing 0·11 mg Co/kg DM in diets containing 0·09 mg Co/kg DM increased growth performance and nutrient digestibility but had no improvement on the effects of FA addition in calves.

To investigate the influences of dietary riboflavin (RF) addition on nutrient digestion and rumen fermentation, eight rumen cannulated Holstein bulls were randomly allocated into four treatments in a repeated 4 × 4 Latin square design. Daily addition level of RF for each bull in control, low RF, medium RF and high RF was 0, 300, 600 and 900 mg, respectively. Increasing the addition level of RF, DM intake was not affected, average daily gain tended to be increased linearly and feed conversion ratio decreased linearly. Total tract digestibilities of DM, organic matter, crude protein (CP) and neutral-detergent fibre (NDF) increased linearly. Rumen pH decreased quadratically, and total volatile fatty acids (VFA) increased quadratically. Acetate molar percentage and acetate:propionate ratio increased linearly, but propionate molar percentage and ammonia-N content decreased linearly. Rumen effective degradability of DM increased linearly, NDF increased quadratically but CP was unaltered. Activity of cellulase and populations of total bacteria, protozoa, fungi, dominant cellulolytic bacteria, Prevotella ruminicola and Ruminobacter amylophilus increased linearly. Linear increase was observed for urinary total purine derivatives excretion. The data suggested that dietary RF addition was essential for rumen microbial growth, and no further increase in performance and rumen total VFA concentration was observed when increasing RF level from 600 to 900 mg/d in dairy bulls.

To evaluate the impacts of guanidinoacetic acid (GAA) and coated folic acid (CFA) on growth performance, nutrient digestion and hepatic gene expression, fifty-two Angus bulls were assigned to four groups in a 2 × 2 factor experimental design. The CFA of 0 or 6 mg/kg dietary DM folic acid was supplemented in diets with GAA of 0 (GAA−) or 0·6 g/kg DM (GAA+), respectively. Average daily gain (ADG), feed efficiency and hepatic creatine concentration increased with GAA or CFA addition, and the increased magnitude of these parameters was greater for addition of CFA in GAA− diets than in GAA+ diets. Blood creatine concentration increased with GAA or CFA addition, and greater increase was observed when CFA was supplemented in GAA+ diets than in GAA− diets. DM intake was unchanged, but rumen total SCFA concentration and digestibilities of DM, crude protein, neutral-detergent fibre and acid-detergent fibre increased with the addition of GAA or CFA. Acetate:propionate ratio was unaffected by GAA, but increased for CFA addition. Increase in blood concentrations of albumin, total protein and insulin-like growth factor-1 (IGF-1) was observed for GAA or CFA addition. Blood folate concentration was decreased by GAA, but increased with CFA addition. Hepatic expressions of IGF-1, phosphoinositide 3-kinase, protein kinase B, mammalian target of rapamycin and ribosomal protein S6 kinase increased with GAA or CFA addition. Results indicated that the combined supplementation of GAA and CFA could not cause ADG increase more when compared with GAA or CFA addition alone.

Supplementing palmitic acid (C16 : 0) in combination with modifying the dietary n-6:n-3 fatty acid (FA) ratio may benefit energy metabolism and milk responses of dairy cows. Twelve Holstein cows (70 (sd 11) days in milk) were used in a replicated 4 × 4 Latin square and allocated to four low-fibre diets (18·5 % forage neutral-detergent fibre) supplemented with no FA (CON), or 2·4 % C16 : 0-enriched supplement (PAL), 2·4 % mixture (2:1) of C16 : 0 and n-6 FA (PW6), and mixture (2:1) of C16 : 0 and n-3 FA (PW3). The dietary ratio of n-6:n-3 was increased with PW6 (10:1) and decreased with PW3 (2·8:1), whereas PAL alone made no change in the ratio (about 7:1). Compared with CON, all FA-supplemented treatments increased milk yield. However, feed and energy intakes were higher in PAL than PW3 or PW6, resulting in greater feed efficiency for PW3 and PW6 than PAL. Dietary FA supplements decreased milk protein concentration but tended to increase protein yield. Compared with CON and FA mixtures, PAL increased milk fat content and tended to increase milk SFA and atherosclerotic index. The concentration of milk n-3 FA was similar between CON and PW3. Feeding PAL increased milk energy output and decreased energy partitioning towards body reserves (−4·2 %), while this measure was positive for other treatments. Blood TAG and NEFA concentrations, but not β-hydroxybutyrate, were increased by FA-supplemented treatments. Feeding C16 : 0 combined with either n-6 or n-3 FA enhanced feed efficiency, alleviated the negative impacts on body energy reserves, but lowering the dietary n-6:n-3 ratio improved the FA profile of milk.

Coated copper sulphate (CCS) could be used as a Cu supplement in cows. To investigate the influences of copper sulphate (CS) and CCS on milk performance, nutrient digestion and rumen fermentation, fifty Holstein dairy cows were arranged in a randomised block design to five groups: control, CS addition (7·5 mg Cu/kg DM from CS) or CCS addition (5, 7·5 and 10 mg Cu/kg DM from CCS, respectively). When comparing Cu source at equal inclusion rates (7·5 mg/kg DM), cows receiving CCS addition had higher yields of fat-corrected milk, milk fat and protein; digestibility of DM, organic matter (OM) and neutral-detergent fibre (NDF); ruminal total volatile fatty acid (VFA) concentration; activities of carboxymethyl cellulase, cellobiase, pectinase and α-amylase; populations of Ruminococcus albus, Ruminococcus flavefaciens and Fibrobacter succinogenes; and liver Cu content than cows receiving CS addition. Increasing CCS addition, DM intake was unchanged, yields of milk, milk fat and protein; feed efficiency; digestibility of DM, OM, NDF and acid-detergent fibre; ruminal total VFA concentration; acetate:propionate ratio; activity of cellulolytic enzyme; populations of total bacteria, protozoa and dominant cellulolytic bacteria; and concentrations of Cu in serum and liver increased linearly, but ruminal propionate percentage, ammonia-N concentration, α-amylase activity and populations of Prevotella ruminicola and Ruminobacter amylophilus decreased linearly. The results indicated that supplement of CS could be substituted with CCS and addition of CCS improved milk performance and nutrient digestion in dairy cows.

The present study investigated the association between fibre degradation and the concentration of dissolved molecular hydrogen (H2) in the rumen. Napier grass (NG) silage and corn stover (CS) silage were compared as forages with contrasting structures and degradation patterns. In the first experiment, CS silage had greater 48-h DM, neutral-detergent fibre (NDF) and acid-detergent fibre degradation, and total gas and methane (CH4) volumes, and lower 48-h H2 volume than NG silage in 48-h in vitro incubations. In the second experiment, twenty-four growing beef bulls were fed diets including 55 % (DM basis) NG or CS silages. Bulls fed the CS diet had greater DM intake (DMI), average daily gain, total-tract digestibility of OM and NDF, ruminal dissolved methane (dCH4) concentration and gene copies of protozoa, methanogens, Ruminococcus albus and R. flavefaciens, and had lower ruminal dH2 concentration, and molar proportions of valerate and isovalerate, in comparison with those fed the NG diet. There was a negative correlation between dH2 concentration and NDF digestibility in bulls fed the CS diet, and a lack of relationship between dH2 concentration and NDF digestibility with the NG diet. In summary, the fibre of CS silage was more easily degraded by rumen microorganisms than that of NG silage. Increased dCH4 concentration with the CS diet presumably led to the decreased ruminal dH2 concentration, which may be helpful for fibre degradation and growth of fibrolytic micro-organisms in the rumen.

The influence of phenol-rich dietary grapeseed extract on performance, energy and N balance and methane production was determined in sixteen lambs and thirteen goat kids (body weight 20·5 and 19·0 kg, 2 months of age, day 1 of study). Half of the animals received a concentrate containing grapeseed extract, and the others received concentrate without grapeseed extract (total extractable phenols analysed 27 v. 9 g/kg dietary DM; concentrate and hay 1:1). Diets were fed for 7 weeks with 1 week for determining intake, excretion and gaseous exchange in metabolism crates and respiration chambers. Overall, there was an adverse effect of the phenolic diet on apparent N digestibility and body N retention. Faecal N loss as proportion of N intake increased while urinary N loss declined. Relative to N intake, total N excretion was higher and body N retention lower in goat kids than lambs. Diets and animal species had no effect on methane emissions. The saliva of the goat kids had a higher binding capacity for condensed tannins (CT). Goat kids on the phenolic diet had higher CT concentrations in faeces and excreted more CT compared with the lambs (interaction species × diet P < 0·001). The lambs had overall higher (P < 0·001) urinary phenol concentrations than the goat kids (2·19 v. 1·48 g/l). The negative effect on body N retention and lack of effect on methane emissions make the use of the extract in the dosage applied not appealing. Species differences need to be considered in future studies.

We evaluated the differences between the supplementation of urea in rumen and/or abomasum on forage digestion, N metabolism and urea kinetics in cattle fed a low-quality tropical forage. Five Nellore heifers were fitted with rumen and abomasum fistulas and assigned to a Latin square design. The treatments were control, continuous infusion of urea in the abomasum (AC), continuous infusion of urea in the rumen, a pulse dose of urea in the rumen every 12 h (PR) and a combination of PR and AC. The control exhibited the lowest (P < 0·10) faecal and urinary N losses, which were, overall, increased by supplementation. The highest urinary N losses (P < 0·10) were observed when urea was either totally or partially supplied as a ruminal pulse dose. The rumen N balance was negative for the control and when urea was totally supplied in the abomasum. The greatest microbial N production (P < 0·10) was obtained when urea was partially or totally supplied in the abomasum. Urea supplementation increased (P < 0·10) the amount of urea recycled to the gastrointestinal tract and the amount of urea-N returned to the ornithine cycle. The greatest (P < 0·10) amounts of urea-N used for anabolism were observed when urea was totally and continuously infused in the abomasum. The continuous abomasal infusion also resulted in the highest (P < 0·10) assimilation of microbial N from recycling. The continuous releasing of urea throughout day either in the rumen or abomasum is able to improve N accretion in the animal body, despite mechanism responsible for that being different.

Although lignin has been negatively correlated with neutral-detergent fibre (NDF) digestibility (NDFD) in ruminants and used to predict potential extent of NDF digestion of forages, selection of an analysis, Klason lignin (KL) or acid-detergent lignin (ADL), to describe that the nutritionally relevant lignin has not been resolved. Dismissed as an artifact is the difference between KL and ADL (ΔL). A question is whether ΔL influences NDFD. We evaluated the relationships of ΔL, KL and ADL with NDFD in order to determine the nutritionally homogeneous or heterogeneous nature of KL. Data sets from two laboratories (DS1 and DS2) were used that included ADL, KL and in vitro NDFD at 48 h (NDFD48). DS1 contained seven C3 grasses, seventeen C4 maize forages and nineteen alfalfas, and DS2 had fifteen C3 grasses, eight C4 forages and six alfalfas. Mean ΔL was greater than ADL in C3 and C4 samples and less in alfalfas. Within forage type and laboratory, ΔL was not correlated with NDFD48 (r −0·34–0·49; all P > 0·17). ADL was more consistently correlated with NDFD48 (r −0·47–−0·95; P < 0·01–0·21) than with KL (r 0·03–−0·91; P < 0·01–0·94). ΔL as a proportion of KL was correlated with NDFD48 in C3 and C4 samples (r 0·44–0·76; P < 0·01–0·08). The differing behaviours of ΔL and ADL relative to NDFD48 indicate that KL is a nutritionally heterogeneous fraction, the behaviour of which may vary by forage type and ratios of ADL and ΔL present.

This study evaluated the effects of rumen-protected folic acid (RPFA) and betaine (BT) on growth performance, nutrient digestion and blood metabolites in bulls. Forty-eight Angus bulls were blocked by body weight and randomly assigned to four treatments in a 2 × 2 factorial design. BT of 0 or 0·6 g/kg DM was supplemented to diet without or with the addition of 6 mg/kg DM of folic acid from RPFA, respectively. Average daily gain increased by 25·2 and 6·29 % for addition of BT without RPFA and with RPFA, respectively. Digestibility and ruminal total volatile fatty acids of neutral-detergent fibre and acid-detergent fibre increased, feed conversion ratio and blood folate decreased with the addition of BT without RPFA, but these parameters were unchanged with BT addition in diet with RPFA. Digestibility of DM, organic matter and crude protein as well as acetate:propionate ratio increased with RPFA or BT addition. Ruminal ammonia-N decreased with RPFA addition. Activity of carboxymethyl cellulase, cellobiase, xylanase, pectinase and protease as well as population of total bacteria, protozoa, Fibrobacter succinogenes and Ruminobacter amylophilus increased with RPFA or BT addition. Laccase activity and total fungi, Ruminococcus flavefaciens and Prevotella ruminicola population increased with RPFA addition, whereas Ruminococcus albus population increased with BT addition. Blood glucose, total protein, albumin, growth hormone and insulin-like growth factor-1 increased with RPFA addition. Addition of RPFA or BT decreased blood homocysteine. The results indicated that addition of BT stimulated growth and nutrient digestion in bulls only when RPFA was not supplemented.

The inclusion of fibre-rich ingredients in diets is one possible strategy to enhance intestinal fermentation and positively impact gut ecology, barrier and immunity. Nowadays, inulin-type fructans are used as prebiotics in the feed of piglets to manipulate gut ecology for health purposes. Likewise, some by-products could be considered as sustainable and inexpensive ingredients to reduce gut disorders at weaning. In the present study, chicory root and pulp, citrus pulp, rye bran and soya hulls were tested in a three-step in vitro model of the piglet’s gastro-intestinal tract combining a pepsin-pancreatin hydrolysis (digestion), a dialysis step using cellulose membranes (absorption) and a colonic batch fermentation (fermentation). The fermentation kinetics, SCFA and microbiota profiles in the fermentation broth were assessed as indicators of prebiotic activity and compared with the ones of inulin. The immunomodulatory effects of fermentation supernatant (FS) were investigated in cultured intestinal porcine epithelial cells (IPEC-J2) by high-throughput quantitative PCR. Chicory root displayed a rapid and extensive fermentation and induced the second highest butyrate ratio after inulin. Citrus pulp demonstrated high acetate ratios and induced elevated Clostridium clusters IV and XIVa levels. Chicory root and pulp FS promoted the intestinal barrier integrity with up-regulated tight and adherens junction gene expressions in comparison with inulin FS. Chicory pulp FS exerted anti-inflammatory effects in cultured IPEC-J2. The novel approach combining an in vitro fermentation model with IPEC-J2 cells highlighted that both chicory root and pulp appear to be promising ingredients and should be considered to promote intestinal health at weaning.

Variations in feeding behaviour between animals result from individual variations in their metabolism as affected by diet composition. The study aimed to link the within-day dynamics of voluntary feed intake and those of blood metabolites and insulin in growing pigs having ad libitum access to feed and receiving diets differing in dietary fibre levels and aleurone supplementation. A total of forty pigs (body weight: 35 kg) had access to diets provided ad libitum, which differed by fibre content (13 or 18 % neutral-detergent fibre) and aleurone supplementation (0, 2 or 4 g/kg). Feeding behaviour was individually recorded for 1 week. The kinetic of plasma metabolites and insulin was followed for 1 h after a voluntary test meal. Dietary fibre level did not affect the daily feed intake but increased meal size and meal duration. Aleurone supplementation (4 g/kg) decreased the daily feed intake and number of meals. Dietary fibre level only decreased insulin concentration measured 15 min after meal beginning. Aleurone supplementation (4 g/kg) decreased glycaemia in the first hour after the meal and insulinaemia 15 min after the meal. Free access to feed led to high variability in pre-prandial metabolites and insulin concentrations, resulting in different test meal size irrespective of diet composition. Animals were then spread over different profiles combining feeding behaviour and fasted status to explain different profiles of regulation of feed intake. Plasma metabolites and insulin kinetics were affected by diet composition but also by animal characteristics. Individual variability should be considered when diet composition is used to modulate feeding behaviour.

The objective of this study was to determine whether a combination of crude glycerin (CG) and soyabean oil (SO) could be used to partially replace maize in the diet of Nellore steers while maintaining optimum feed utilisation. Eight castrated Nellore steers fitted with ruminal and duodenal cannulas were used in a double 4×4 Latin square design balanced for residual effects, in a factorial arrangement (A×B), when factor A corresponded to the provision of SO, and factor B to the provision of CG. Steers feed SO and CG showed similar DM intake, DM, organic matter and neutral-detergent fibre digestibility to that of steers fed diets without oil and without glycerine (P>0·05). Both diets with CG additions reduced the acetate:propionate ratio and increased the proportion of iso-butyrate, butyrate, iso-valerate and valerate (P<0·05). Steers fed diets containing SO had less total N excretion (P<0·001) and showed greater retained N expressed as % N intake (P=0·022). SO and CG diet generated a greater ruminal abundance of Prevotella, Succinivibrio, Ruminococcus, Syntrophococcus and Succiniclasticum. Archaea abundance (P=0·002) and total ciliate protozoa were less in steers fed diets containing SO (P=0·011). CG associated with lipids could be an energy source, which is a useful strategy for the partial replacement of maize in cattle diets, could result in reduced total N excretion and ruminal methanogens without affecting intake and digestibility.

The purpose of the present study was to compare digestibility of grass hay, faecal and plasma volatile fatty acid (VFA) concentrations, and faecal bacterial abundance in overweight and moderate-condition mares. Five overweight adult mixed-breed mares and five adult mixed-breed mares in moderate condition were housed individually and limit-fed orchard grass (Dactylis glomerata) hay at 20 g/kg body weight (as fed) daily for 14 d. Forage DM and fibre digestibility were determined using AOAC methods; digestible energy was measured using bomb calorimetry; plasma and faecal VFA concentrations were determined by use of GC and MS; faecal Firmicutes, Bacteroidetes, Fibrobacter succinogenes, Ruminococcus flavefaciens and total bacteria abundance was determined by quantitative real-time PCR using previously designed phylum-specific 16S ribosomal RNA gene primers. No differences in hay digestibility, faecal VFA concentrations or faecal bacterial abundance were detected between overweight and moderate-condition mares. Mean plasma acetate concentrations were higher (P = 0·03) in overweight (1·55 (range 1·43–1·65) mmol/l) v. moderate-condition (1·39 (range 1·22–1·47) mmol/l) mares. We conclude that the higher plasma acetate in overweight mares should be further investigated as a potential link between gut microbes and obesity in horses.

Three ruminally cannulated non-lactating dairy cows were used to investigate the effects of six different bag cloth types with pore size (μm): free surface area (%) ratios of 200: 45, 41: 33, 16: 5, 10: 2, 6: 5 and 1: 2 respectively on the disappearance of grass silage DM and neutral-detergent fibre (NDF), and on particle-associated carboxymethylcellulase (EC 3.2.1.4; CMCase) and xylanase (EC 3.2.1.8) activties extracted from feed residues. Another objective was to compare microbial activity inside the bags and in rumen ingesta. Rumen incubation periods were 3, 6, 12, 24, 48 and 96 h. DM and NDF disappearance and particle-associated enzyme activities were greatly reduced with the smaller pore size and/or open surface area. Re-analysing some of the data as a 2 x 2 factorial (pore size x free surface area) indicated that, generally, free surface area rather than pore size affected the disappearance of feed components and particle-associated enzyme activities. Enzyme activities were highly correlated with NDF disappearance at 6–48 h of incubation. Cumulative area under CMCase and xylanase activity curves explained 0·79 and 0·88 of the variation in NDF disappearance when different cloth type and 6–48 h incubation data were combined. Weighted mean enzyme activities inside the bags were less than 0·35 those in rumen ingesta. The highest activity values inside the bags (24 or 48 h) were less than 0·50 those found in rumen ingesta. The lower microbial activity inside the bags explains the slower rates of NDF digestion reported with in situ techniques than with rumen evacuation techniques. The general assumption of similar microbial activity inside the bags and in rumen ingesta is not justified by the present results, and caution must be taken in interpreting in situ results quantitatively for feed evaluation systems.

Fifty-four roughages of known voluntary dry-matter intakes (DMI; range 7·8−35·2 g/kg live weight per d) were examined in vitro in a gas production test. Samples (200 mg) of roughage and roughage neutral-detergent fibre (NDF) respectively were incubated in a mixed suspension of rumen contents for 96 h and the gas volumes recorded after 4,6,8,12,24,30,36,48,54,60 and 96 h. The kinetics of gas production were derived from the volume recordings described by the exponential equation Y=A+B(l—e-ct) where A is the intercept and ideally reflects the fermentation of the soluble and readily available fraction of the feed, B describes the fermentation of the insoluble (but with time fermentable) fraction and c the fractional rate at which B is fermented per h; A+B describes total fermentation. In vitro true dry matter (TD) and NDF degradabilities (NDF-D) after 24 h incubation were also determined. Of the variation in DMI, 75% was accounted for by the in vitro gas production parameters A, B and c in stepwise multiple regressions; 82% of the variation in DMI was explained by the parameters (ANDF+BNDF) and cNDF as obtained from the incubation of roughage NDF. The rate constants (c) were less important than parameters related to the extent of gas production, accounting for only 6·5 (whole roughage) and 4·1% (NDF) of the variation in DMI. There was no statistical advantage in the use of the exponential model describing extent and rate of fermentation over some of the simple gas volume measurements: 75% of the variation in DMI was accounted for by in vitro gas production of whole roughage after 8 h of incubation. On average gas production from NDF measured from 24–96 h accounted for 81% of the variation in DMI. A combination of gas volume measurements after a short period of incubation (4–8 h) with a concomitant determination of NDF-D after many hours (≥24 h) can render NDF preparations and long incubation times redundant. A method is suggested to obtain two results for DMI prediction in one single incubation. Of the variation in DMI 80% was accounted for by the incubation of 500 mg whole roughage when incubation was terminated after 24 h and the residual undegraded substrate quantified.

Digestible energy (DE) values were measured in a selection of feedstuffs for the tilapia (Oreochromis niloticus Linn.) and used to develop equations for predicting DE values of a wider range of feedstuffs from chemical analyses. Preliminary work examined the influences of substitution level in a reference diet and adaptation over time on DE values for soya-bean meal. Length of adaptation period significantly affected DE values (P < 0.01), but substitution level, over the range 200–600 g soya-bean meal/kg reference diet, did not. The DE values of sixteen feedstuffs, thirteen derived from plant sources and three animal by-products, were subsequently determined. DE values for plant-derived feedstuffs were found to be higher than those quoted in the literature for trout (Oncorhynchus mykiss) and catfish (Ictalurus punctatus), whereas DE values for animal-derived feedstuffs were lower than those for trout and pigs. It was concluded that energy values quoted in tables of feed composition for other species are inaccurate when used as proxy values for tilapia. Regression equations were therefore computed using data from the present study to provide a rapid means of predicting DE values of feedstuffs for tilapia. Equations using neutral-detergent fibre as an independent variable were found to predict DE values of plant-derived feedstuffs reliably. Where fibre values were not used as independent variables, available carbohydrate and crude protein (nitrogen × 6.25) were found to be useful predictors of DE values. These equations offer the possibility of reducing the need for time-consuming digestibility trials with tilapia when formulating least-cost production diets for this species.