This paper presents an approach for ethical evaluation of biotechnology, which is illustrated by cloning in dairy cattle breeding. The approach for ethical decision-making that was followed, the so-called network model, started with determining fundamental moral principles that together constitute a framework for the detection of ethical issues in a given situation. In terms of human ethics, four fundamental principles have been defined: beneficence, non-maleficence, justice and autonomy. In the context of animal biotechnology, the same four fundamental moral principles as in humans are considered, except that autonomy is replaced by respect for the integrity of the animal. Based on a Dutch advisory committee concerning animal biotechnology, two additional principles were considered: irreversibility and veriflability. For each individual case, moral decision-making means balancing of moral intuitions, moral principles and morally relevant facts. The final outcome on moral acceptability of an action, therefore, will differ among people. The latter is demonstrated using three types of people (the ruler, the steward and the partner/participant). For each type of person, the fundamental attitude toward animals and nature is defined and a possible decision on the application of cloning in dairy cattle breeding is given. The authors' opinion, however, is that the benefits of cloning are relatively small, especially in the long term, whereas the moral drawbacks of cloning are considerable. The authors conclude that other ways of furthering animal husbandry should have priority.

To investigate strategies for concentrate allocation, 155 multiparous Holstein-Friesian cows received grass silage ad libitum during lactation weeks 4 to 22 in two experiments. A standard concentrate (196 g crude protein (CP) per kg dry matter (DM)) was offered at the same average amount across the lactation period either at aflat rate or as two rates with a change of 6 kg in week 13. Treatments, described as kg concentrate DM per day in the first and second periods of lactation, were 3: 3, 6: 0, 3: 9 (experiment 1), 6: 6, 9: 3, 9: 9 and 12: 6 (experiments 1 and 2). In experiment 2, two additional treatments maintained CP supply from concentrate after a change from 9 and 12 kg DM by giving 3 and 6 kg high-protein concentrate (HP) DM per day (575 and 383 g CP per kg respectively).

Generally as concentrate intake increased silage intake fell, DM intake rose and yields of milk and all milk solids increased linearly. However, in experiment 1 from a base of 3 kg DM, as concentrate intake increased so did fat yield overall and protein yield in mid-lactation but at a declining rate. Fat concentration was depressed at the lowest and highest concentrate levels. Protein concentration rose as concentrate allowance increased particularly at low levels. Increasing concentrate in mid lactation improved protein but not fat concentration. Additional concentrate either spared weight loss or increased weight gain. Distributing more of a fixed allowance of concentrate to early lactation did not affect cow performance overall.

Silage intake, milk, protein and lactose yields and protein concentration increased and fat concentration was reduced by giving HP concentrates, effects being greater with 3 than with 6 kg DM per day. Reducing concentrate level while maintaining CP intake did not influence milk and solids yields. The ratios of responses to concentrate CP (3 kg DM per day HP v. 3 kg DM per day standard) compared with concentrate energy (9 kg DM per day standard v. 3 kg DM per day HP) were 2·1, 2·2, 7·4 and 1·9 for milk, fat, protein and lactose yields respectively. The results suggest feeding strategies for manipulating the production and quality of milk from grass silage diets.

Twenty-four Holstein Friesian cows in experiment 1 and 40 cows in experiment 2 were used over 12 weeks in continuous designs, to evaluate winter wheat as a forage for dairy cows. In experiment 1, whole-crop wheat was ensiled at 350 g dry matter (DM) per kg (EW) or harvested later and stored at 550 g DM per kg following treatment with 40 g urea per kg DM (40W). In addition to these treatments in experiment 2, whole-crop wheat of 550 g DM per kg treated with 20 g urea per kg DM (20W), and combine harvested wheat grain plus wheat straw (40: 60 ratio) treated with sodium hydroxide (SG) were also included. These whole-crop wheat forages were incorporated into the diet at a level of 400 g/kg of the forage DM in experiment 1, and 330 g/kg in experiment 2, with the remainder of the forage being grass silage. A control treatment of grass silage (GS) offered as the sole forage was also included. The forages were mixed in a mixer wagon and offered ad libitum. A fixed level of 8 kg/day of concentrates was offered in experiment 1 and 7 kg/day in experiment 2. Total DM intake was significantly greater for EW and 40W than for GS in experiment 1, but not in experiment 2. Milk yield was not significantly affected by diets (mean 28·8 kg/day in experiment 1 and 29·6 kg/day in experiment 2). Milk fat, protein and lactose contents and yields were also not significantly affected by diets. The additional total metabolizable energy (ME) intake of the whole-crop diets compared with GS was partitioned to live weight. There was no evidence of ME intakes being substantially greater than ME requirements as found in previous studies. A 5 × 5 Latin square digestibility experiment was carried out with Holstein-Friesian heifers offered the individual forages used in experiment 2. Intake was significantly greater for the urea-treated whole-crop wheat forages than for grass silage and for sodium hydroxide-treated grain and straw. The digestible organic matter in the dry matter (DOMD) of grass silage was 692 g/kg and the whole-crop wheat diets ranged from 626 g/kg for ensiled whole-crop wheat to 682 g/kg for the sodium hydroxide-treated grain plus straw (40: 60 ratio), with the DOMD of urea-treated whole-crop being intermediate. The results indicated that whole-crop wheat had higher intake characteristics than grass silage in spite of its lower digestibility.

One hundred and two multiparous Holstein/Friesian dairy coivs were used to examine the effect on food intake and milk production of replacing 330 g/kg dry matter (DM) of first-cut perennial ryegrass silage (GS) with either fermented (FW) or urea (40 g/kg DM) treated (UW) whole-crop wheat, brewers' grains (BG), fodder beet (FB) or maize silage (MS). A forage mixture containing GS and 750 g/kg DM of maize silage (MSH) was also offered. Dietary crude protein (CP) concentrations were equalized by offering all cows 6 kg DM per day of an appropriate concentrate supplement. DM intake and milk yield of cows offered GS (15·3, 23·0 kg/day) were increased (F < 0·01) by the inclusion of BG (16·3, 26·7 kg/day), FB (17·6, 25·9 kg/day), MS (16·6, 26·4 kg/day) and MSH (18·5, 27·6 kg/day). BG, FB and MS all produced similar increases in milk yield. Although the inclusion of FW and UW increased (F < 0·05) DM intake, milk yield was not significantly affected. The differences between the highest and lowest milk fat concentration, produced by FB (42·9 g/kg) and BG (40·1 g/kg) respectively, and the highest and lowest milk protein concentration produced by FB (32·1 g/kg) and GS (29·9 g/kg) respectively, were statistically significant (F < 0·001). Milk fat and protein yields of cows offered GS (0·97, 0·71 kg/day) were increased (F < 0·01) by BG (1·07, 0·82 kg/day), FB (1·11, 0·83 kg/day), MS (1·11, 0·83 kg/day) and MSH (1·12, 0·88 kg/day). Intermediate values were recorded for FW (1·01, 0·72 kg/day) or UW (1·00, 0·73 kg/day). Results show that farmers should consider replacing a major part of the grass silage component with an alternative forage or forage substitute.

Mature dry cows, 18-month-old heifers and 7-month-old female calves of the Charolais breed (eight animals of each age) were strip-grazed on moderate (15cm) or tall (52 and 32cm) cocksfoot swards over two 15-day experimental periods. The swards were leafy regrowths of 2 (short) and 8 (tall) weeks of age. Herbage intake was estimated by the alkane method, and grazing behaviour was visually recorded.

Bite weight was lower, grazing time and biting rate greater, and organic matter in take of grass (OMIG) higher on shorter swards, the latter being related to a higher in vitro digestibility and nitrogen concentration. Average OMIG per kg live weight (M) was 15·2, 19·1 and 17·0 (s.e.d. 0·49)g in cows, heifers and calves respectively. The average allometric coefficient between OMIG and M was 0·88 (similar between swards), and close to that obtained with hay after the grazing trial.

Faecal markers suggested a greater selectivity of calves compared with older cattle. Average bite weights (BW) were 310 (s.e. 40), 467 (s.e. 112) and 718 (s.e. 144)mg OM in calves, heifers and cows respectively. The allometric coefficient between BW and M was higher in short swards than in tall ones. Biting rates increased and grazing times tended to decrease with age of cattle.

The relationship between puberty and the cessation of luteal activity after nutritional restriction was evaluated in 15 Angus × Hereford heifers. Heifers attained puberty at a body weight of 297 (s.e. 6) kg and a body condition score (BCS) of 5·5 (s.e. 0·1) on a scale of 1 (emaciated) to 9 (obese). After 154 (s.e. 16) days of nutritional restriction, heifers became anoestrus at a weight of 273 (s.e. 8) kg and a BCS of 3·0 (s.e. 0·2). The Spearman's rank correlation coefficient between date of-puberty and date of cessation of luteal activity was (r = –0·49; P < 0·06). This indicates that the heifers that attained puberty first were the last to cease luteal activity during nutritional restriction.

Mature non-lactating cows, 18-month-old heifersand7-month-oldfemale calves of the Charolais breed (eight animals of each age) were strip-grazedon short (S, sward surface height: 7·5 cm), medium (M,10·2 cm) and tall (T, 21·1 cm) cocksfoot (Dactylis glomerata) swards at a daily herbage allowance of 40 g dry matter per kg live weight over three 15-day experimental periods. The swards were leafy regrowths of 1 (S), 2 (M)and 3·5 (T)weeks of age. Herbage intake was estimated by the n-alkane method, and grazing behaviour was recorded visually.

Intakes per kg live weight of calves and heifers were higher than those of cows on all swards(P < 0·01). Calves tended to maintain intake, whereas cows and heifers reduced their intake as sward height decreased. Bite weight decreased linearly with sward height, this decrease being greater for larger animals (P < 0·001). As sward height decreased, the scaling factor of bite weight to live weight decreased. Biting rates increased as sward height decreased (P < 0·001) and were lower in calves than in older cattle (P < 0·001). Calves had lower intake rates per kg live weight than cows in T swards, and the reverse was true in S swards (P < 0·001). Younger cattle always grazed longer than cows, but differences were reduced as sward height decreased (P < 0·01). Cows and heifers wereunable to increase their grazing time sufficiently to compensate for lower intake rates in S and M swards, resulting in reduced herbage intakes. Faecal markers and grazing behaviour suggested a greater selectivity by calves compared with older cattle.

The effects of fertilization (100 kg nitrogen plus 10kg phosphorus per ha per year) of Cymbopogon-Themeda veld on certain chemical components in vitro dry-matter digestibility (IVDMD), intake by either lactating cows or steers as well as daily gain of pre-weaned calves or steers, were determined in the summer rainfall area of the Republic of South Africa.

The nutritive value of fertilized veld herbage was, in terms of crude protein and acid-detergent fibre content, IVDMD and herbage intake, superior to that of unfertilized veld. Species, classified as palatable according to the literature, were more abundant on the fertilized veld than on the unfertilized veld. In spite of the fact that the stocking rate on the fertilized veld was double that of the unfertilized veld, the average daily gain (ADG) of steers on the former was higher than that of steers on the latter. There was no significant difference in the ADG of the pre-weaned calves, whose dams grazed either fertilized or unfertilized veld. It was calculated that the fertilization of veld for growing steers can (under the conditions that prevailed) be a financially viable proposition if the beef price per kg live weight amounts to at least 0·65% of the cost of fertilization per ha.

The study compared using chemical components (i.e. crude protein (CP), neutral- and acid-detergent fibre or lignin (NDF, ADF and ADL) g/kg dry matter (DM), the in vitro digestibility (Tilley and Terry, 1963), the in situ (nylon bag) DM degradation (g/100 g DM) and gas production (ml/200 mg DM) techniques to predict voluntary daily intake (g DM per kg M0.75) and in vivo apparent DM digestibility (DMD) of 10 graminaceous hays individually offered ad libitum to four Merino male sheep. Gas production or DM degradation were determined after 6, 12, 24, 48, 72 or 96 h incubation and their characteristics described using the equation p = a + b (1 – e-ct). Intake and in vivo DMD of the hays were variable and poorly related (r = 0·52; P > 0·05). The in situ DM degradation was significantly (Y < 0·05) related to in vivo apparent DMD at 48 to 96 h incubation (i = 0·76 to 0·75) and to intake at 24 to 96 h (r = 0·71 to 0·75) incubation. However, fibre components, the in vitro digestibility or gas production were either related to daily intake or in vivo apparent DMD, but not to both on the same occasion. Accurate prediction of intake (r = 0·90; P < 0·05) and in vivo apparent DMD (r = 0·88; P < 0·069) were achieved using NDF, ADF, ADL and CP in a multiple regression. Using the (a + b) and the rate (c) of in situ DM degradation, both in vivo apparent DMD (r = 0·77; P < 0·05) and intake (r = 0·83; P < 0·05) were predicted with accuracy. However, using the (a + b) and (c) of gas production, only intake was predicted accurately (r = 0·87; P < 0·01). The lower performance of the gas test was attributed to the small contribution to gas production and higher buffering capacity resulting from protein fermentation. When data of the graminaceous and other data from leguminous hays were combined, the most accurate prediction of both intake and apparent digestibility was by using characteristics of in situ DM degradation followed by those of gas production. The latter was more accurate than using chemical components or the in vitro digestibility. Addition of CP in the multiple regression improved the prediction of intake and in vivo apparent DMD from characteristics of gas production. It was concluded that despite the need to overcome the problem of protein fermentation in the gas test, accurate prediction of both intake and apparent digestibility can be achieved simply from the degradation characteristics of foods.

This study was aimed at evaluating (1) the dry matter (DM) and nitrogen (N) degradabilities (by sheep) of Sesbania sesban leaves, Acacia albida (Faidhabia albida), A. siberiana and A. tortiris frazfs and (2) the effect of feeding a mixture of oilseed cake and tanniferous browses on the degradability of teff (Eragrostis tef) straw and oilseed cake and on particulate and liquid passage rates. Thirteen rumen cannulated sheep in a randomized complete-block design experiment were given teff straw ad libitum supplemented (DM per day) with 155 g oilseed cake (50: 50 mixture of noug (Guizotia abyssinica) and groundnut cake; oilseed cake diet) or with 92 g oilseed cake plus either 118 g S. sesban (sesbania diet) or 179 g A. albida fruits (albida diet). Diets were formulated to supply 10·5 g supplementary N per day.

Sesbania had the highest DM washing loss, potential degradability and fastest DM degradation rate compared with the acacias. The N washing loss (g/kg) was highest for A. siberiana (500) followed by A. albida (431), A. tortilis (431) and sesbania (241). The potential N degradability (g/kg) was greater (P < 0·05) for sesbania (978) than for A. albida (848), A. tortilis (865) or A. siberiana (785). The degradation rates of oilseed cake-N and teff straw-DM, particulate and liquid passage rates, basal intake and DM digestibilities were similar among diets, but organic matter intake was highest for the albida diet. At the present levels of offer, supplementation with mixtures of oilseed cake and browses does not significantly reduce rumen degradation rates of oilseed cake-N or teff straw DM.

The hierarchic behaviour of 71 goats of the Verata breed from two independent flocks on two different farms (no.2 = 34 and no. 2 = 37) was studied during the milking-period of a year. Two statistical procedures: canonical correspondence analysis (CCA) and a log-linear model were used. In the CCA, the total frequencies of each goat in each of the 12 positions at the milking stalls during 210 days of lactation were computed as original variables. Canonical variables were the average values of dairy milk production, proportion of protein and proportion of fat. The first two CCA axes accounted for proportionately 0·72 of the variance. The first axis showed the effect of rank order (V1 = 0·45) and the second axis the variability between animals for the same group of hierarchy (V2 = 0·27). The first position (AC1, = 0·58; AC2 = 0·23) and the last position (AC1 = 0·11; AC2 = 0·26) of the milking stalls were in absolute contributions the most significant in the distribution of the four observed groups of hierarchy, as defined by their frequency of access to the first position. Two parameters were highly correlated with both axes: milk protein content (r1, = - 0·8; r2 = 0·9) and the day milk production (r1 = 0·6; r2 = 0·5). These results showed a direct relationship between rank order and milk production, whereas the variability in each group was best defined by the protein content. The log-linear model showed that goats of lactation numbers one, two and three have a similar behaviour with respect to their locations at the milking stalls, being distributed more randomly (â1, = 0·68 < â2 = 0·7 < â3 = 0·71) than goats of a lactation number four (â4 = 2·95). This effect could be a consequence of two different but related factors: either the animals develop a preference for a particular position over time, or that animals in their fourth lactation showed a more stable behaviour pattern which affected their body condition.

The effects of months and seasons of the year on the testis size, libido and plasma testosterone (T) concentrations were studied in 10 young sexually mature male goats over a 12-month period. Testis size was expressed in terms of scrotal circumference, testis diameter and paired testes width while the libido was measured as stimulation and reaction times,number of mounts and services, and libido index. These measurements were recorded fortnightly.

The body weight increased from 33·5 (s.e. 0·95) kg in October, to a maximum of 62·8 (s.e. 1·55) kg in July. All parameters of testis size decreased from October through January with a sharp increase noted in July and August. These parameters were significantly higher in the summer and autumn than in the winter and spring (P < 0·01). The libido in the goats increased from October to a maximum in December, declined gradually to reach the minimum in July, and then improved again in August. Thus, libido was lowest during the summer and highest in winter, the difference being significant (P < 0·01). Plasma T concentrations were high during October, with the lowest values noted in June. An increase in the T concentrations was observed in August, reaching the maximum value in September, the difference being significant (P < 0·01). The T concentrations were significantly higher during autumn than in all other seasons (P < 0·01).

Reproductive records from 1072 Large White sows (3589 litters) were used to examine the effect ofbackfat depth (Bd) and live weight (Lw) at selection on first litter and lifetime reproductive performance. The variables investigated included mating age, total piglets born, total born alive, piglet birth weight, number weaned, piglet weaning weight, weaning to remating period and number of litters produced. Using backfat depth, sows were categorized into three groups: L, 9 to 13 mm; M, 14 to 16 mm; and F, ≥17 mm. Bd had no significant effect (P > 0·05) on the measured traits for the sows grouped by fat at first parity. When lifetime reproductive data were analysed, the L group had fewer litters (P < 0·05), and weaned fewer pigs (P < 0·01) than either the M or F group. The birth weight for piglets from L sows were significantly higher (P < 0·05) than for piglets from the other groups. The data suggest that the reproductive performance of L sows is not as good as that of M or F sows.

There have been suggestions that positive feedback from the meal may stimulate the performance of undesirable behaviour, such as stereotypies, in the post-feeding period. The aim of this work was to measure the short-term effects of a meal on the post-feeding behaviour of the pig.

Pigs (70 kg live weight) were offered food ad libitum for 7 days and then given a restricted food allowance for 23 days (experiment 1) or 14 days (experiment 2). In experiment 1, the food allowance was given in one, two or three meals per day. In experiment 2, the food allowance was given in two meals per day and comprised ingredients which supplied glucose at different rates (starch, glucose/starch, starch/guar gum). The pigs' posture and behaviour were recorded in three 1-h periods during the day on 2 days consecutively during the ad libitum feeding period (time period 1) and on pairs of days on three (periods 2, 3 and 4: experiment 1) or two (periods 2 and 3: experiment 2) subsequent occasions during food restriction. The meal frequency or estimated rate of glucose supply had no significant effects on posture or behaviour. In both experiments activity increased with the duration of food restriction. In periods 1, 2, 3 and 4 of experiment 1 the proportions of observations spent were: standing 0·244, 0·260, 0·706 and 0·728 (s.e.d. 0·0346); nosing 0·073, 0·094, 0·229 and 0·253 (s.e.d. 0·0169); rooting 0·032, 0·037, 0·108 and 0·133 (s.e.d. 0·0133). In periods 1, 2 and 3 of experiment 2 the proportions were: standing 0·247, 0·302 and 0·446 (s.e.d. 0·0240); nosing 0·045, 0·075 and 0·138 (s.e.d. 0·0113); chewing 0·101, 0·126, 0·155 (s.e.d. 0·0121). In experiment 1 feeding rate increased with duration of restriction (P < 0·001), but was unaffected by treatment. It is concluded that post-feeding behaviour was influenced more by long-term changes in state (i.e. prolonged food restriction) than by short-term effects of the meal.

Computerized food intake recording systems of various designs are used by research, centres and breeding companies to monitor the individual food intake of pigs kept in groups. In the present experiment, three feeder designs are compared in order to estimate the effect on performance and feeding behaviour. Ninety entire male pigs (34 (s.e. 0·6) kg) were allocated in three replicates to pens of 10 pigs; each pen containing one of three different feeder entrance designs: low (head-guard), medium (full-length standard race), and high (enclosed pneumatic race) protection against disturbance of the feeding pig. No significant differences were found between treatments in daily food intake, daily live-weight gain, food conversion ratio, number of visits per day and daily feeder occupation. Pigs with access to an enclosed race had longer visits than pigs on the two other treatments, and they also ate more per visit (4·8, 4·8 and 6·2 (s.e.d. = 0·37) min per visit; 172, 157, and 202 (s.e.d. = 11·1) g per visit; means of low, medium and high protection, respectively). The enclosed race was not only the most protective, it was also the most difficult to enter, and this decreased accessibility of the food may have given rise to the changed feeding pattern displayed by pigs using this type of race. Access to a feeder with low protection resulted in a faster rate of eating (36·9, 33·2, and 32·8 (s.e.d. = 0·82) g/min; means of low, medium and high protection, respectively) indicative of a more forced feeding behaviour.

Six sport horses were given 1·26 times the measured maintenance energy requirement (MEm) from each of the four following diets: H1, meadow hay in the long form (organic matter digestibility OMD = 0·541); HMI, 700g/kg the same hay and 300 g/kg pelleted maize; HSBPI, 600g/kg hay and 400g/kg pelleted dehydrated sugar-beet pulp; SCFI, 500g/kg wheat straw and 500g/kg pelleted compound food (experiment 1). In experiment 2, eight sport horses were equipped with a portable device for recording feeding behaviour and fed at 1·31 MEm diet HI (meadow hay in the long form: OMD = 0·574).Circadian energy expenditure (EE) of horses was determined by indirect calorimetry using two large open-circuit respiration chambers. Horses were continuously standing. Increase in metabolic rate (IMR) during eating was calculated from the difference between the mean EE obtained during each eatingperiod and the corresponding resting EE. The mean daily ingestion rate of hay H2 amounted to 148 (s.d. 27)mg dry matter per kg metabolic body weight per min. IMR during the two main meals averaged 0·388 (s.d. 0·059) and was not significantly different between diets H1, H2, HM1 and SCF1. Expressed per kg dry matter intake, energy cost of eating (ECE) was similar for diets H2, H1 and SCF1 but significantly lower for HSBP1 and HM1 (P<0·05). ECE of simple foods was calculated from those of the diets and of hay: proportionately 0·010, 0·042, 0·102 and 0·285 metabolizable energy intake for pelleted maize, pelleted SBP, long hay and wheat straw, respectively.

Quantitive collection of night-time urine every 2nd week was used to elucidate the temporal relationship between changes in melatonin production, weight gain, and priming of the winter pelage under natural light conditions and under a long photoperiod (18L: 6D) imposed at different times in four groups of female mink during autumn. In mink maintained outdoors under the natural photoperiod (group 1), melatonin excretion was 1·80 (s.d. 0·80) ng per night (no. = 3) at the beginning of September, thereafter it began to decrease slowly until November when it was only 0·62 (s.d. 0·32) ng per night. These mink had a normal autumn moult during the first half of October and the winter pelage was mature at the end of November. When mink maintained outdoors were transferred to 18L: 6D on 7 September (group 2), excretion of melatonin decreased promptly, autumn moult was disrupted and winter fur priming delayed. Long photoperiod after mid October (group 3), did not interfere with pelage maturation despite the decrease in melatonin production. Mink implanted with melatonin on 14 July (group 4) showed very high melatonin excretion in early September. In these animals the winter pelage growth was unaffected by the long-day conditions; the pelage matured in mid October. Body weight increased in mink of all groups during autumn. This increase levelled off or a weight loss (group 2) was seen along with the final hair maturation. Activation of hair follicles occurred during a 4 to 6 week period. According to these results, melatonin is the photoperiodic signal to autumnal weight increase and autumn moult but seems not to be necessary for later pelage growth and maturation.

On the basis of their respective metabolizable energy contents, supplemental marine oil was partially replaced with a maize starch/dextrin mixture in diets containing approximately 500 g crude protein per kg food thus lowering the total dietary lipid concentration from 141·6 to 100·9 g/kg. The basal diets were then supplemented with thiamin hydrochloride at 0·5, 5·0 and 10·0 mg/kg food to produce six test diets in all. After 182 days of feeding, the fish given the diets containing the higher lipid content with a thiamin supplement of 10·0 mg/kg had shown significantly better growth and food utilization (food conversion efficiency and apparent net protein utilization) than the fish presented with the remaining five diets. Thiamin supplements of 5 mg/kg or less resulted in poorer performance amongst the fish given the high lipid diets. The fish given all three of the low oil diets performed poorly by comparison with those given the high lipid, high thiamin diet and performance was comparable with that of the fish given the high lipid diets containing thiamin at 5 mg/kg or less. Proximate carcass composition in terms of moisture, protein and lipid was unaffected in response to diet. Glycogen accumulation in the liver was significantly elevated amongst the fish given the low lipid diet where the thiamin supplement was 5·0 mg/kg or more. The haematology of the fish was not significantly altered in response to diet although elevated haematocrit was associated with decreasing thiamin supplementation. It would appear that for the seabream, lipid and carbohydrate are not exchangeable on the basis of their metabolizable energy content. When given high lipid diets the thiamin requirement of the seabream is in excess of 5·0 mg/kg and increasing the thiamin supplement to 10 mg/kg in high carbohydrate diets made no impact on performance.