Despite milk being an important product from sheep, there are very few reports of milk production from the complete lactation of dairy sheep. The Improved Awassi in Israel is kept under an intensive system of management with lambs being weaned soon after birth. Records from one such flock were analysed to investigate the suitability of various mathematical functions for describing milk yield from the complete lactations of dairy sheep. This included a consideration of whether the functions could cope with short lactations, a characteristic of dairy sheep, and a limited number of test-day records per lactation.

Four non-linear mathematical functions were investigated (Wood, Morant, Grossman and Pollott), two of which could also be fitted in a linear and a linear weighted form (Wood and Morant). These functions were fitted to weekly data from a ‘typical Awassi lactation curve’, represented by least squares means of daily milk yield from each week of a 40-week lactation derived from an analysis of 25605 test day records. Characteristics of the lactation were calculated from the functions, such as total milk yield, day and level of peak yield and persistency. These functions were also fitted to 1416 individual lactation records of up to 10 test-day records per lactation. The value of the functions was investigated using the residual mean square (RMS) of the fitted curve as an indicator of how well each function described the lactation. Forms of these functions with a reduced number of parameters were also investigated.

The non-linear functions always fitted the data with a lower RMS than their linear equivalent and the weighted form of the linear functions always had a lower RMS than the unweighted form. Of the linear functions, Morant fitted better than Wood. Of the non-linear functions Grossman, Morant and Pollott (additive and multiplicative) fitted the data equally as well but better than Wood. The various functions predicted characteristics of the lactation curve differently; the Wood functions tended to overestimate yield in early lactation and the Morant functions underestimated peak yield.

No function was better suited to short lactations than another. However the three-parameter function of Morant, Pollott multiplicative and Pollott additive were considered to be the most suitable for describing the complete lactation of dairy sheep.

A genetic study was carried out to: (1) conduct a genetic analysis of longevity of Swedish Yorkshire sows, (2) study the environmental and genetic factors that influence the presence and severity of osteochondrosis, and (3) investigate the relationship between breeding values for osteochondrosis and longevity of sows. The data for the longevity analyses were extracted from the Swedish litter-recording scheme data bank (Quality Genetics, former Scan Avel HB). After editing original data, records of 9814 Yorkshire sows with 7553 (77%) uncensored and 2261 (23%) censored born 1986 through 1997 were used in the analyses. Litter size at first and last farrowing, age at first farrowing, backfat thickness, daily gain and weight at completion of performance test (~170 days) were included as fixed effects in all analyses. The combination of herd-year effect was treated as fixed or random, time-independent or time-dependent in different analyses. Sire effect was considered as the source of genetic variation and thus a sire model was used. The analyses of osteochondrosis were based on information on 14 388 Landrace and 14 458 Yorkshire pigs from the Swedish pig progeny-testing scheme, recorded from 1987 through 1997. The birth herd and the combination of sex, testing station, year and month for start of test were included as fixed effects in the statistical model. Variance and covariance components for osteochondrosis recorded at elbow and knee joints were estimated in a bivariate animal model by the restricted maximum likelihood method within each breed. In the survival analyses (Yorkshire sows), the fixed effects of herd-year (when it was treated as fixed effect), litter size at first and at last farrowing, age at first farrowing, backfat and gain at completion of performance test were highly significant (P < 0·01). Herd-year combination was the major cause of variation for risk of culling, compared with other fixed effects. The risk of being culled at a certain time decreased as the litter size at first and at last farrowing, or backfat of the gilt at completion of performance test increased. With increasing age at first farrowing, the risk of being culled at a certain time increased. Heritability in the original scale for longevity ranged from 0·21 to 0·31. The results for osteochondrosis showed that the combined effect of sex-testing station-year and month of start of test was highly significant (P < 0·01). Estimates of heritabilities for osteochondrosis score were similar for both Landrace and Yorkshire breeds and was, on average, 0·21. The correlations between breeding values for longevity and osteochondrosis were low (on average 0·07, adjusted for genetic trends) but were significant (P < 0·01) and in a favourable direction: higher osteochondrosis load associated with higher risk of being culled.

A series of 24-h video studies on four commercial Danish pig herds investigated the behaviour of pregnant sows kept in dynamic groups (72 to 200 sows) with electronic sow feeding (ESF). The herds mainly differed with respect to provision of a layer of unchopped straw as bedding material, the frequency of introduction/removal of animals, space allowance in the lying area, group size and number of feeding stations, and starting times for the feeding cycle. All herds had one feeding cycle per 24 h. Six 24-h video recordings in the most settled period with respect to rank relationships (2 to 12 days after the last introduction) were collected from each herd and analysed according to activity and aggressive interactions. In two of the four herds recordings were collected in two separate periods with different starting times for the feeding cycle. In all herds peaks of activity in the morning and in the afternoon were evident. Furthermore, the diurnal variation was overlaid by an increase in activity at the start of the feeding cycle. The activity and aggression in the feeding area as well as the duration per sow of the period with high feeder occupation were lowest in herds where a layer of straw bedding was provided. Changing feeding start from day time to night time was accompanied by a reduced feeder occupation in the period following start of the feeding cycle. However, in herd 1 (characterized by no litter, high stocking density and high frequency of grouping) the level of aggression was reduced, whereas it was increased in herd 2 (characterized by straw in lying area, low stocking density, low frequency of regrouping). Number of regroupings and space allowance apparently had no obvious effects on the average frequency of aggression or the aggression per sow at risk in periods between introduction of new animals, but space allowance may have improved social function by weakening the association between activity and aggression. Due to the small number of herds included the present results were descriptive rather than conclusive. However, the study supports the suggestions that provision of unchopped straw as bedding material and starting the feeding cycle in the evening for overnight feeding may improve behaviour in dynamic group housing systems with ESF. However, the benefits of starting the feeding cycle in the evening may depend on low disturbance in daytime from other management procedures.

Possible correlated or pleiotropic effects on food intake, growth rate and carcass traits (meat quality and carcass composition traits) were investigated for the prolificacy related Booroola gene (FecB). This gene was introgressed in a Texel sheep population by systematic backcrossing with sires carrying the Booroola gene. In a 3-year experiment 273 spring-born male lambs were offered concentrates ad libitum, after weaning. Lamb carrier effects and dam carrier effects for the FecB gene were estimated, accounting for year, backcross generation, slaughter weight, weaning age, concentrate type, rearing code, age of the dam and litter size as fixed effects. Carriers of the FecB had a higher dressing percentage (+1·15%), longissimus muscle depth (+0·26 mm to +0·7 5mm) and cross-sectional area (+0·98 cm2). However the gene also had a positive effect on subcutaneous fatness (+0·12 points), KKCF (+30 g) and a substantial positive lamb carrier effect on carcass fat tissue content (+3·09%), which reflects a relative increase in fatness of 11·9%. The FecB gene had also very small, significant effects on some meat colour traits. The dam carrier effects of the Booroola gene were negatively associated with food efficiency and with longissimus muscle depth (–1·38 to –2·25 mm). The increased fatness of lambs could imply that for optimal use of FecB, non-carrier terminal sires should be used on heterozygous dams.

A total of 16 multiparous (range 2nd to 9th parity) Camborough sows were housed over parturition and during lactation in a group farrowing system. Two replicates of four sows were studied in the large pen (L) and two further replicates studied in the small pen (S) configuration, providing 13•4 m2 and 8•6 m2 per sow, respectively. A lactation diet providing 14 MJ/kg digestible energy (DE) and 180 g/kg crude protein was supplied ad libitum via a sow-operated feeder, adapted to facilitate the calculation of individual food intakes. The influence of environmental, physical, and production factors upon food intakes and feeding strategies was investigated. Daily food intakes of 7•69 (s.e. 0•31) kg and 7•72 (s.e. 0•35) kg were achieved during lactation by sows in the L and S pen configurations, respectively, accomplished by sows taking a series of small meals throughout the day. Sows in the S pen area made fewer visits to the feeder and spent less time per day feeding than sows in the L pen area. However, more food was consumed per visit by the S sows, resulting in similar daily food intakes in both treatment groups. There was a depression in lactation food intakes of sows housed at a mean daily ambient temperature of 17•5ºC compared with the other sow groups. The majority of sows consumed energy well in excess of their predicted requirements during lactation, whilst one sow in the L and two in the S pen areas had intakes within ±3•4 MJ DE of predictions. Only two of the L sows had food intakes that produced an energy deficit. In the L pen area, two distinct peaks of feeding activity occurred from 09:00 to 13:00 h and from 15:00 to 23:00 h that accounted for 0•77 of the total feeding time. In contrast, in the S pen configuration, one prolonged period of feeding occurred from 09:00 to 22:00 h accounting for 0•88 of total feeding time. The results show that S sows adapted their feeding patterns to maintain similar daily intakes to those of L sows, thus demonstrating that an ad libitum food supply enabled sows to obtain individual nutritional requirements in a communal farrowing system.

A dynamic model for simulation of growth in pigs, extended to describe thermoregulatory processes, was made stochastic to simulate groups of pigs with between-animal variation in mature body protein (Pα) and lipid mass (Lα), in the potential rate at which mature mass is attained (B⋆), and in the distribution of body protein and lipid over pools and depots. The resulting variation in body composition leads to variation in energy requirements for protein turn-over and thermoregulation, causing between-animal variation in maintenance requirements (MEmaint).

Simulated population means for Pα, Lα /Pα and B⋆ were varied in three steps each. Excluding unrealistic parameter combinations this led to 33 – 6 = 21 simulated genotypes. Simulated within-population coefficients of variation (CV) were 7, 15 and 3%. Random replicates of each genotype were simulated five times, in climatic conditions that were in turn severely cold, mildly cold (about 5 and 1ºC below lower critical temperature), thermoneutral, mildly hot and severely hot (about 1 and 5ºC above upper critical temperature), during the entire growth period of 23 to 100 kg live weight. Simulated food intake was ad libitum.

Simulated thermoneutral within-population standard deviations of body protein and lipid content were 0·21 to 0·46 kg and 0·78 to 2·14 kg at 100 kg body weight. On average, the corresponding values in cold and hot conditions were slightly higher.

MEmaint showed a protein-turn-over-related within-population CV of 1·5% at thermoneutrality. Thermoregulatory action contributed about 4% extra variance in cold and hot conditions but CV values were not affected. A genetic increase in the maximum protein deposition rate from 100 to 250 g/day would increase MEmaint as related to protein turn-over and thermoregulation by 11% at thermoneutrality, and by 6 to 11% in cold or hot conditions. Two relevant groups of genotypes could be distinguished based on the within-population regression coefficients of MEmaint on daily or cumulative protein deposition (bdailyPdep, bcumPdep). These ranged from 0·250 to 0·428 kJ/kg0·75 per day per g/day and from 2·77 to 5·45 kJ/kg0·75 per day per kg, respectively, in 12 ‘conventional’ genotypes at thermoneutrality. On average, bdailyPdep was increased by 48%, 20%, –11% and –36% in the other climatic conditions mentioned above, respectively. The corresponding increase of bcumPdep was 32%, 14%, 8% and 48%. Three fast-growing lean genotypes showed similar bdailyPdep and bcumPdep at thermoneutrality, but much more pronounced increases in cold and hot conditions: 137%, 49%, –12% and + 88% for bdailyPdep and 248%, 108%, 17% and 196% for bcumPdep.

It is concluded that differences in body composition traits between pig genotypes do not cause important between-genotype differences in thermoregulatory MEmaint, and that thermoregulatory processes contribute little body-composition-related variation to hot or cold MEmaint within most genotypes.

The inferences to be made from this with regard to experimental design are discussed. The verification of the above predictions will require a very elaborate and large-scale experiment.

This paper describes a two-part study evaluating the use of growth hormone (GH) response to infusions of somatostatin (SRIF) and GH-releasing factor (GRF) as an indicator of genetic merit in dairy cattle. The first part of the study involved the development of a standard test procedure. This was achieved in two sub-trials, the first testing three SRIF doses, 0·6, 3·0 and 15 µg/kg per h, given over an infusion length of 3 h. The second involved infusing 3·0 μg/kg per h SRIF for either 20 min, 1 h or 3 h. The GH response was measured in serum during the infusion and for 2 h following cessation, during which time a rebound response in GH occurred. As a consequence of these trials, a total of 40 British Friesian calves from two lines genetically divergent for yield traits were infused with 3·0 μg/kg per h for a period of 1 h and at the end of the infusion given 0·1 μg/kg GRF. Five hours later, GRF was administered without SRIF. The protocol was repeated on 2 days consecutively. Results suggested that calves of high merit had lower SRIF-induced baselines than their low-line contemporaries. There were no differences in the GH concentrations immediately following GRF administration, either with or without SRIF. The results from this trial, together with previously published work, indicated that calves selected for high milk fat plus milk protein production may be more sensitive to SRIF such that GH is regulated by lower concentrations of SRIF. However, the practical use of GH response to SRIF as an indicator of genetic merit for dairy cattle may be limited because of logistical difficulties in implementation.

Differences in the metabolic resource situation between non-reproductive male and female mice of a line selected for high litter size at birth (average of 22 born per litter) and a non-selected control line (average of 10 born per litter) were investigated in two replicates. Brody curves were fitted to individual data on body weight against age and linear regression lines were fitted to individual data on cumulative food intake against age. Mature body weight and mature daily food intake were higher in selected mice than in control mice and higher in males than in females. Selected males matured faster than selected females and control mice. In general, differences in growth and food intake curves between species or lines can mostly be explained by differences in mature size. Therefore, parameters were subsequently scaled by individual estimates of mature body weight. Differences that remain after scaling are a consequence of what have been called specific genetic factors. Scaled mature food intake was higher in selected mice than in control mice and higher in females than in males. Scaled maturation rate was higher in selected mice than in control mice and higher in selected males than in selected females. This shows that in the present study, specific genetic factors have been detected for both body weight and food intake, which suggests that selection for increased litter size has disproportionally changed the resource allocation pattern.

In the accompanying paper, specific genetic factors for body weight and food intake were identified in non-reproductive male and female mice of a line selected for high litter size at birth (average of 22 born per litter) and a non-selected control line (average of 10 born per litter). The existence of these factors are indicated by variation in efficiency parameters such as growth efficiency and maintenance requirements. Residual food intake (RFI) and Parks’ estimates of growth efficiency (AB) and maintenance requirements (MEm) were used to quantify these factors. In the growing period, females had a higher RFI (are less efficient) than males. At maturity, selected mice had higher RFI than control mice and selected females had higher RFI than selected males. AB was higher in selected-line mice than in control-line mice, and higher in males than in females. MEm was higher in selected-line mice than in control-line mice, and higher in females than in males. The results indicate the existence of specific genetic factors for both growth efficiency and maintenance requirements. Selected females may increase RFI in the adult state to anticipate the metabolically stressful periods of pregnancy and lactation, to support a genetically highly increased litter size.

Data on lamb survival from birth to weaning of Rambouillet and Finnsheep were analysed with sire-maternal grandsire linear (LM) and threshold (TM) models. Models for the Rambouillet included the effects of year, type of birth, age of dam and sex as fixed effects and sire, maternal grandsire and residual as random effects. For the Finnsheep, fixed effects were year-age of dam combination, type of birth-rearing and sex, and random effects were sire, maternal grandsire and residual. Restricted maximum likelihood (REML) and marginal maximum likelihood (MML) estimates of variance and covariance components were obtained under LM and TM, respectively. The performance of LM and TM was assessed in terms of goodness of fit and predictive ability. Within the Rambouillet breed, heritabilities of additive direct (0·06), additive maternal (0·04) and total effects (0·13) obtained with TM were 2, 1·3 and 1·9 times greater than those obtained with LM, respectively. For the Finnsheep data, estimated heritabilities of direct (0·17), maternal (0·26) and total (0·34) effects using TM were 1·9, 1·4 and 1·6 times greater than the estimates using LM, respectively. Estimated genetic correlation between direct and maternal effects was 0·14 for the Finnsheep with both the LM and TM. Higher genetic correlations between direct and maternal effects (0·44 and 0·62 with LM and TM, respectively) were obtained for the Rambouillet data. Goodness of fit and predictive ability of the models used with the Finnsheep data were better than for the models used with the Rambouillet data, but within breed, no sizable or significant differences were detected between LM and TM. Results indicate that maternal effects were important in lamb survival to weaning; especially in the more prolific Finnsheep breed.

Plasma leptin concentration shows pulsatility and diurnal rhythm in humans. However, there are few reports concerning the 24-h profile of circulating leptin levels in ruminants. Five crossbred ewes were housed in metabolism cages under a 1-h light-dark cycle. The ewes were offered alfalfa hay daily to meet their energy requirement. Blood samples were collected at 15-min intervals for 24 h. Plasma leptin concentrations were determined using a radioimmunoassay and the profile of plasma leptin levels was analysed by the PULSAR algorithmic program for detecting pulse. Plasma leptin concentration changed in a pulsatile fashion. The mean leptin concentration was 2·93 ng/ml. The mean pulse frequency was 4·8 pulses per day and the mean pulse amplitude was 0·67 ng/ml with an average pulse length of 1:13 h. Plasma leptin level was not affected by feeding or lighting cycle. These results indicate that plasma leptin level in sheep shows pulsatility but diurnal rhythm is not exhibited.

The components of phenotypic variance attributable to maternal lineage for production traits of the UK Holstein Friesian dairy population were estimated. First lactation production records of 55 230 cows calving between 1996-1998 in the UK Holstein Friesian population were used in the analysis. Maternal pedigree records were traced back to 1960 to establish maternal lineages. The tracing resulted in 36 320 cows being assigned to 11 786 cow families with more than one cow per maternal lineage. Using test day records it was possible to explore aspects of the lactation curve in terms of persistency and different periods of production. The traits analysed were 305-day milk yield and composition traits, the first three milk yield tests of lactation and two measures of persistency. A contemporary record design was used to minimize pair-wise additive direct genetic relationships between cows within a maternal lineage and to remove both the effect of heterogeneous variance over time and the complications of permanent environment effects. No significant component of variance attributable to maternal lineage was found for yield traits. When data were restricted to maternal lineages with five or more records, persistency, as a ratio of cumulative yield in the last third to that in the first third of a 300-day lactation, was estimated to have a 4·4% component due to maternal lineage variance significant at the 5% level. The study also investigated the preferential treatment of cow families. Some evidence of maternal lineage × herd interaction was found.

Restricted maximum likelihood analyses fitting an animal model were conducted to estimate genetic parameters with a pooled-data set of performance tests (growth traits and food intake) on 661 bulls and progeny tests (growth traits and carcass traits) on 535 steers. Traits studied included concentrate intake (CONC), roughage intake (ROU), TDN conversion (TCNV), TDN intake (TINT) of bulls; rib eye area (REA), marbling score (MARB), dressing proportion (DRES) and subcutaneous fat depth (SCF) of steers. Body weight at start (BWS), body weight at finish (BWF) and average daily gain (ADG) of all animals were measured. Estimated heritabilities were 0·18 (CONC), 0·71 (ROU), 0·11 (TCNV) and 0·36 (TINT); 0·02 (REA), 0·49 (MARB), 0·15 (DRES), 0·15 (SCF), and from 0·20 to 0·38 for growth traits. Genetic correlations of ROU were different from those of CONC, probably due to inconsistent restrictions on concentrate intake; those of TINT with the weights, ADG and SCF were high. MARB showed positive genetic correlations with growth traits and low correlations with TINT and SCF. High potentiality for improvement of marbling score was suggested.

There is some evidence to show that loose housing during gestation has a negative influence on the welfare of sows subsequently in farrowing crates. However, little is known about the effects of the gestation housing on the initial responses of gilts to the farrowing crate or of the effects on gilt welfare throughout lactation. The objective of this experiment was to evaluate three gestation housing systems (1) stalls (ST); (2) loose-bedded (LB) and (3) loose-unbedded (LU) pens for their effect on behaviour, heart rate and skin lesion scores of gilts in farrowing crates. During the 1st h in the farrowing crate, LB gilts were more active, while ST gilts were more vocal. Although heart rates during the 1st h did not differ significantly between treatments, LB and LU gilts had significantly higher mean heart rates during the first 5 min in the crate, compared with 35 min later. Posture changes during the first 24 h did not differ between treatments. However, there was a significant reduction in the number of posture changes made by ST gilts but not LB or LU gilts by day 8. Increases in the skin lesion score of gilts in all three treatments were observed after 24 h in the crate, further increases were observed post farrowing in both loose treatments. Although no significant differences in the skin lesion score were observed during lactation, LB gilts were weaned with lower lesion scores than LU or ST gilts. Gilts from both loose treatments experienced greater distress at first introduction to the farrowing crate. However, the change in environment also had an adverse affect on the welfare of ST gilts. Skin damage and continuing discomfort of the loose-housed gilts post farrowing suggests that they experienced more stress at parturition. Bedding during gestation had a beneficial effect on skin health in the farrowing crate that persisted until weaning.

This work was designed to investigate whether the profile of non-esterified fatty acids (NEFA) remains constant in response to underfeeding in beef cows, since in other species mobilization of fatty acids (FA) has been shown to be dependent on fatty acid structure. Twenty-one Charolais cows (multiparous, no. = 16 and primiparous, no. = 5) were assigned to one of two diets (control diet (CD; no. = 11) or restricted diet (RD; no. = 10) which, respectively supplied all or 0·7 of energy and protein requirements) after calving. Blood and composite milk samples were collected weekly before the morning meal over a period of 9 weeks. The cows given RD lost significantly more body weight (P < 0·001) and condition score (P < 0·05) than the cows given CD. Plasma NEFA and urea concentrations were also higher in RD than CD during the experiment (P < 0·05 and P < 0·001, respectively). The level of feeding did not affect plasma insulin and glucose concentrations. Milk FA were affected by feeding level. The proportions of short- and medium-chain FA were lower in RD compared with CD (C8: 0, C10: 0, C12: 0 and C14: 0, P < 0·05 and C14: 1 and C16: 0, P < 0·10) while some of the long-chain FA were higher (C17: 0 and C18: 1, P < 0·001 and C17: 1, P < 0·10). Feeding level affected the profile of FA in plasma NEFA. The proportions of C16: 0 and C18: 1 were higher (P < 0·05) in RD cows compared with CD cows while the proportions of C12: 0, C14: 1, C17: 0, C18: 0 and C18: 2 were lower (P < 0·05). There was a feeding level × time interaction in the proportions of different FA in the plasma NEFA and in the total plasma NEFA concentrations. The differences caused by feeding level gradually disappeared with time. The proportion of C18: 1 in plasma NEFA and the total plasma NEFA concentrations were quadratically positively correlated, while the correlation between C18: 2 and total plasma NEFA was negative. In conclusion, when post-partum beef cattle are underfed, plasma NEFA concentrations increase (indicating lipomobilization) and this increase is associated with an increase in the proportion of oleic acid (C18: 1) in plasma NEFA.

In The Netherlands a system has been developed for classification of live pigs based on lean meat proportion. Ultrasonic backfat thickness measurements were taken on 377 live pigs to assess the ability of the developed system for estimation of the lean meat proportion. The measurements were made at the sites mid point (half the distance from the occipital bone to the base of the tail), mid point –2·5 cm and mid point +2·5 cm, 5 cm off the dorsal mid line.

On the same day, these pigs were slaughtered and the lean meat proportions of the carcasses were estimated using the Hennessy Grading Probe (HGP). HGP measurements were taken between the third and fourth from last rib, 6 cm off the dorsal mid line on carcasses (3/4 LR). The day after slaughter, 88 left carcass sides were randomly chosen to be dissected according to a simplified European Union (EU) reference method.

From the fat thicknesses measured, the one at the site mid point on live pigs was the most accurate predictor for the EU lean meat proportion. The use of multiple site measurements, compared with a single site measurement, significantly reduced the residual standard deviation for the estimation of lean meat proportion.

The site mid point on live pigs differed in longitudinal as well as in dorsal-ventral direction from 3/4 LR on carcasses.

The results of this study suggest that multiple backfat thickness measurements on live pigs can be used for prediction of lean meat proportion with sufficient precision for practical use. Therefore, the developed system can serve as a classification system for live pigs.

Six multiparous Holstein-Friesian cows were offered a total mixed ration based on maize silage in a repeated measure design to evaluate the partition of gross energy (GE) during early to mid lactation. Four measurements were made at 6-week intervals with energy and nitrogen balances carried out in open-circuit respiration chambers over 6 days during lactation weeks 6, 12, 18 and 24. The intakes of total diet dry matter (DM) corrected for volatile losses (VCDM), organic matter (OM) and GE declined significantly (P < 0•01) as lactation progressed, although apparent digestibility of these fractions was not altered, resulting in a significant (P < 0•01) decline in digestible nutrient intake at each stage of lactation. Methane and urine energy losses were not significantly affected, resulting in significantly (P < 0·001) higher amounts of digestible energy (DE) partitioned to methane and urine as lactation progressed with associated significant reductions in metabolizable energy (ME) intake (MEI) (P < 0·01) and ME as a proportion of DE (P < 0·001) and GE (q) (P < 0·05). With advancing lactation there was a significant (P < 0·001) increase in the amount of ME partitioned to heat (HP/MEI), but no significant change in the amount partitioned to milk and tissue. Individual values for diet metabolizability (ME/GE) at actual (production) levels (qa) (mean 0·625 MJ/MJ) were corrected to an equivalent value at maintenance (qmc) (mean 0·666 MJ/MJ). The overall ME intakes (MJ/day) were: ad libitum, 246, corrected for level of feeding effect, 263, with a predicted ME requirement according to AFRC (1993) (MER93) of 242. Substitution of the calculated qmc into the predictive equations (AFRC, 1993) resulted in a mean maintenance requirement of 57·6 MJ/day (0·464 MJ/kg M0·75/day) whilst the mean value derived from the linear model describing the experimental data was 82·5 MJ/day (0·664 MJ/kg M0·75/day). The mean efficiencies of utilization of ME for milk production derived from AFRC (1993) and the linear regression model were 0·653 MJ/MJ and 0·625 MJ/MJ respectively.