In early lactation, most of the dairy cows are in negative energy balance; the extent and duration depend in part on the feeding strategy. Previous studies showed an increased lactation milk yield by use of a live weight (LW) adjusted feeding strategy with a high energy diet before and a reduced energy diet after LW nadir compared with a standard diet throughout lactation. The objective of the present study was to examine how such an individualized feeding strategy affects plasma indicators of energy status. It was hypothesized that an energy-enriched diet until LW nadir will reduce the severity of the negative energy balance, and that the reduction in diet energy concentration from LW nadir will extend the negative energy balance period further. Sixty-two Holstein cows (30% first parity) were managed for 16 months extended lactation and randomly allocated to one of two feeding strategies at calving. Two partially mixed rations were used, one with a high energy density (HD) and a 50 : 50 forage : concentrate ratio, and one with a lower energy density (LD, control diet) and a 60 : 40 forage : concentrate ratio. Half of the cows were offered the HD diet until they reached at least 42 days in milk and a LW gain⩾0 kg/day based on a 5-days LW average, and were then shifted to the LD diet (strategy HD-LD). The other half of the cows were offered the LD diet throughout lactation (control strategy LD-LD). Weekly blood samples were drawn for analysis of plasma metabolites and hormones. Before the shift in diet, the HD-LD cows had higher glucose and lower beta-hydroxybutyrate and non-esterified fatty acids (NEFA) concentrations than the LD-LD cows. After the shift until 36 weeks after calving, plasma NEFA was higher in HD-LD than LD-LD cows. Insulin and insulin-like growth factor-1 were not affected by the feeding strategy. To conclude, in early lactation, the energy-enriched diet reduced the negative energy balance. Plasma NEFA was higher in HD-LD than LD-LD cows from diet shift until 36 weeks after calving, indicating a carry-over effect of the early lactation HD diet to late lactation metabolism.

The increase of sheep meat competitiveness in international markets can be attributed to the rise of the quantity and the improvement of the quality of the edible portion of sheep carcasses. Usually, carcass yield is established after the slaughter of the animals. Yet, when carcass yield is determined in vivo, it can be both a costly and subjective method. This study proposes models for predicting the physical characteristics of lamb carcass using bioimpedance analysis (BIA) in live animals. Thirty-one Texel × Ile de France crossbreed ram lambs were slaughtered at 20, 26, 32 or 38 kg of BW. Before the slaughter, values of resistance (Rs) and reactance (Xc) were collected using a single-frequency BIA equipment (Model RJL Quantum II Bioelectrical Body Composition Analyzer). Then, BIA main variables such as body bioelectrical volume (V), phase angle (PA), resistive density (RsD) and reactive density (XcD) were calculated. After slaughter, cold carcass weight (CCW), cold carcass yield (CCY), subcutaneous fat thickness (SFT), soft tissue weight (STW) and soft tissue yield (STY) were also measured. Multiple regression analyses were carried out using the physical characteristics as dependent variables and the bioimpedance values as independent variables. Predictive performance of the models was assessed using leave-one-out cross-validation. The prediction model of CCW was obtained using the V, PA and RsD (R2 = 0.97), STW through the V, RsD and XcD (R2 = 0.97), CCY by Rs, Z and XcD (R2 = 0.69), STY by V and XcD (R2 = 0.67), and SFT only for XcD (R2 = 0.84). The results indicated that BIA has the potential to predict carcass characteristics of lambs at different body masses.

1. Wessex Saddleback sows were fed 4 lb. or 2 lb. of a balanced meal per head per day during the first 13 weeks after mating with a Large White boar. All sows received 6 lb. of meal for the last 3 weeks of pregnancy, and approximately 14 lb. (4 lb. plus 1 lb. for each piglet) during lactation. From mating to weaning half the sows on each feeding treatment had access to small soil pens, and the other half to range grazing of lucerne or grass/clover swards.

2. During the first 3 months of gestation sows on soil receiving 2 lb. of meal gained 20 lb. while those on grazing gained 40 lb., equal to the gain of sows fed 4 lb. of meal.

3. Approximately 10·5 piglets were born per sow under all treatments. Piglet mortality was highest with sows which had been fed 4 lb. of meal during pregnancy and were confined in small soil pens, because this treatment produced the greatest number of over-fat sows. Mean piglet weight at 3 weeks (≏ 12 lb.) and 8 weeks (≏ 39 lb.) and total creep-fed meal consumption per piglet (≏ 30 lb.) were similar for all treatments.

4. In a preliminary trial 3½ lb. of dried grass plus ½ lb. of meal per head per day was fed to gilts during the second and third months of pregnancy without adversely affecting farrowing results.

5. From these experiments and the published work it is suggested:—

(a) Pasture alone would probably have been sufficient during the second and third months of pregnancy of the sows used in these trials.

(b) Feeding standards for pregnant sows of 250-350 lb. live-weight may be too generous, and produce ‘fat not fit’ sows at farrowing.

(c) In practice the reduction of meal feeding and the greater dependence on grazing should be applied individually to each sow according to her condition and response.

Five Karan Swiss cows were allowed to calve spontaneously and served as controls (Group A), while another 10 were induced to calve about 1 month prior to expected term using either 25 mg dexamethasone alone (Group B) or 25 mg dexamethasone in combination with 25 mg stilboestrol dipropionate (Group C). Blood samples were collected prior to, and for 50 days following, parturition for subsequent hormone analysis, and observations were made on the occurrence of retained placentae.

Parturition was successfully induced in all treated cows within 72 h of drug administration. Placental membranes were retained for 24 h in all treated animals and were manually removed in the farm veterinary clinic. No incidence of placental retention occurred in control cows.

Mean plasma oestradiol-17β in control cows showed a gradual rise from days 30 to 5 pre partum, followed by a sharp increase to peak value 1 day before calving. The rise in mean oestradiol-17β levels in the 2 days before parturition in controls was significantly greater than that recorded for the treated animals. A precipitous drop in plasma oestradiol-17β level occurred following parturition in all groups, with the levels staying low for the rest of the sampling period. Drug administration to animals in groups B and C resulted in a rapid and significant drop in plasma progesterone concentrations to a level similar to that for control animals at parturition. Subsequently, the plasma progesterone remained low in all treatment groups. Mean plasma cortisol levels showed a significant but transient increase at calving in control cows, but were suppressed for at least 48 h following drug administration in treated cows before exhibiting a rebound phenomenon to reach high values after calving.

It is concluded that for normal placental expulsion, oestradiol-17β levels are more critical than progesterone, and that the rise in oestradiol-17β and fall in progesterone levels pre partum must be gradual.

The importance of family sizes in adult multiple ovulation and embryo transfer (MOET) nucleus schemes with discrete generations of single trait selection was examined using Monte Carlo simulation. Two areas were investigated. Firstly, the number of sons and daughters per dam was varied in schemes using hierarchical mating designs. With four or eight sires and 32 dams selected, increasing the number of sons per dam from one up to four achieved 1 to 8% higher rates of response but at the expense of increased variation in response and 10 to 56% higher rates of inbreeding. With four or eight sires and 16, 32 or 64 dams selected, the number of daughters was set to two, four or eight (with one son per dam in each case). For schemes transferring equal numbers of embryos, responses were lower with two daughters per dam but were fairly similar with four or eight daughters per dam while inbreeding rates increased as fewer sires and dams were selected. Secondly, the effects of variation in family sizes due to biological factors and chance were investigated with eight sires and 32 dams selected and with hierarchical or factorial (two or four sires per dam) mating designs. When all selected cows yielded embryos, changes in family sizes due to differences in sex ratios, in survival rates of embryos to selection and to variation in the number of embryos per donor reduced response by 1 to 4%. However, when 20% or 33% of the superovulated females yielded no embryos, thus requiring the use of genetically inferior replacements, response was reduced by a further 9 to 13%

The water intake of individually penned ewes has been determined during the last 21 days of pregnancy and the first 21 days of lactation. Pregnant ewes carrying one lamb had lower water intakes than ewes carrying twins. Water intake per unit dry-matter intake increased as dry-matter intake declined.

Lactating ewes suckling twin lambs consumed more water than ewes suckling singles although both groups had similar dry-matter intakes.

Enterotoxigenic Escherichia coli (ETEC) is a major cause of diarrhoea in neonatal and postweaning pigs. F41 is one of ETEC fimbriae that adhere to the small intestinal epithelium and lead to development of diarrhoea. The genetic architecture of susceptibility to ETEC F41 remains elusive in pigs. In this study, we determined the in vitro adhesion phenotypes of ETEC F41 in a total of 835 F2 animals from a White Duroc × Erhualian intercross, and performed a genome scan using both F2 and half-sib analyses with 183 microsatellite markers to detect quantitative trait loci (QTL) for porcine susceptibility to ETEC F41. The two analyses consistently revealed a 1% genome-wide significant QTL on pig chromosome 4. Moreover, we determined F41 adhesion phenotypes in 14 purebred Erhualian and 14 White Duroc pigs. The results showed that both the founder breeds are segregating for the F41 adhesion phenotype, while less percentage of Erhualian pigs were adhesive to ETEC F41 compared to White Duroc pigs.

An experiment has been carried out at Bahia Blanca to estimate the trend in ovulation rate throughout the year shown by Corriedale ewes of similar live weight in the south-west district of the Argentinian pampa.

Five groups of 16 animals were each subjected to two 20-day observation periods and the 10 observation periods in the experiment were evenly spaced within a year.

The proportion of ewes showing oestrus and ovulation (OO), the frequency of twin ovulations (TO) and, hence, ovulation rate (OR) were estimated for each period by using vasectomized rams and performing laparatomies.

A significant and periodic relationship was found between each trait and the number of days elapsed from the longest day (×). In turn, changes in the values of ×, arising from the progress of the experiment, were to some extent associated with changes in nutrition and surgical interference. However, it was concluded that these factors could not have affected the similar and remarkable seasonality shown by each trait in this study, and that this seasonality was mainly associated with differences in the natural light regime to which the ewes were exposed before and during their corresponding observation periods.

The predicted maximum values for each parameter were reached 1 month before the shortest day. Prior to and after this date both OO and TO declined, so that the predicted OR values became zero for a period of approximately 2 months before the longest day

An experiment was undertaken to compare the carcass characteristics and meat quality of lambs from crossbred ewes produced by crossing Texel sires with Rouge de l’Ouest dams and Rouge sires with Texel dams, relative to lambs from purebred Texel and Rouge ewes. The ewes were crossed with Texel or Rouge sires and the relative performance of the sires assessed in terms of lamb carcass quality. Ewe and ram genotype had a significant effect on conformation classification. Lambs from Texel ewes had a greater conformation classification (4·0) than lambs from Rouge ewes (3·4, s.e. 0·09) (P <0·001) and Texel-sired lambs had a greater conformation classification (4·0) than Rouge-sired lambs (3·6, s.e. 0·06) (P <0·001). Fat depth was significantly influenced by ewe genotype with lambs from Texel × Rouge ewes having greater fat depths compared with lambs from Texel ewes (longissimus dorsi 2·3 and 1·9 (s.e.0·12) mm respectively) (P <0·05). Chemical composition of the carcass and meat quality measurements were unaffected by ewe or ram genotype. Significant heterosis effects on lamb live-weight gain (21 g/day; 15%) (P <0·01) from birth to slaughter, age at slaughter (–37 days; –12%) (P <0·01), weight of perinephric and retroperitoneal fat (76 g; 38%) (P <0·01) and fat depth over the longissimus dorsi (0·3 mm; 16%) and gluteus medius (0·5 mm; 17%; P <0·05) were obtained. The results from this study demonstrate that significant heterosis effects, particularly in terms of lamb growth rate, can be achieved by crossing Rouge and Texel breeds. In addition lambs from hybrid ewes have similar conformation characteristics to the Texel breed which has been intensively selected for carcass characteristics.

Dividing the meat carcass into its three main component tissues, bone, muscle and fat, it is known that they develop in a well-defined order, bone being the earliest maturing and fat the latest (Hammond, 1932). Of the three, muscle is the most important from the viewpoint of human nutrition and the factors influencing its growth obviously demand serious consideration. The literature on qualitative growth contains numerous references to microscopical meat studies, but, due to the diversity of conditions under which the observations were made, the results obtained cannot easily be co-ordinated. An investigation was therefore planned in which most of the major factors influencing growth and development of muscle in general, and the muscle fibre in particular, could be studied.

A low finishing weight and poor carcass characteristics are major causes of lower incomes in extensive sheep flocks; however, the use of terminal sire crossbreeding would improve lamb performance and carcass traits under these conditions. The aim of this study was to evaluate sire breed effects on the performance of lambs born to Corriedale ewes in extensive sheep systems in Western Patagonia. A total of 10 Corriedale, 10 Dorset, nine Suffolk and seven Texel sires, 16 of which were under a genetic recorded scheme and 20 selected from flocks not participating in genetic improvement programmes, were used across six commercial farms for 2 successive years. Data were collected from 685 lambs of the four resulting genotypes. Overall, Corriedale lambs were 0.47 kg lighter at birth than crossbred lambs (P<0.001). Suffolk and Texel sired lambs required more assistance (P<0.01) at birth than Corriedale or Dorset sired lambs, with Suffolk sired lambs requiring the most assistance (8%). Ewes sired with Suffolk rams had larger (P<0.05) litters than ewes sired with Texel or Corriedale rams. Lamb live weight gain from birth to weaning was higher (P<0.001) in crossbred lambs compared with Corriedale lambs, therefore, crossbred lambs averaged 2.9 kg heavier BW (P<0.001) than Corriedale lambs. A significant sire breed x sire source interaction was detected for lamb live weight gain (P<0.05) and lamb live weight at weaning (P<0.01), showing that the heaviest lambs were from recorded sires, except for Suffolk crossbred lambs. Mortality rate to weaning was increased (P<0.05) in Suffolk cross lambs (31%), with Corriedale lambs showing the lowest (17%) mortality. Terminal sire breeds increased (P<0.001) cold carcass weight, with 13.8, 16.0, 15.2 and 14.9 kg for the Corriedale, Dorset, Suffolk and Texel sired lambs, respectively. Carcass length, kidney knob and channel fat, fat grade, grade rule and fat depth measurements were not affected by sire breed (P>0.05). Carcass conformation was higher in Texel sired lambs compared with Corriedale lambs (P<0.05), with Dorset and Suffolk sired lambs being intermediate. Crossbred lambs showed a greater (P<0.001) eye muscle than Corriedale. Commercial cuts were affected by sire breed, as a result of the Corriedale lambs being smaller and having lighter carcass than crossbred lambs. Significant improvement in lamb weights at weaning and carcass traits could be expected when using a terminal sire on Corriedale ewes in Western Patagonia. However, no advantages were detected with the use of recorded sires under these production systems.

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.

Length and diameter of teats were measured on first-lactation cows at Danish progeny testing stations. The cows were also classified into four groups according to shape of teat and teat tip. All traits were corrected for yield level and interval after calving and correlated with milking characteristics. The conclusions drawn were:

1. Teat length influences both average rate of milk flow and yield on hand stripping. The most satisfactory milking is obtained in cows with short teats. In Red Danish, Black and White Danish and Jersey cows, respectively, desirable length of fore teat is ≤ 6, ≤ 6, and ≤ 5 cm, and of rear teat about 1 cm shorter.

2. Teat diameter is of no significant importance to average rate of milk flow, but yield on hand stripping is most satisfactory when the diameter is ≤ 2·4 cm.

3. Teat shape must be very extreme before an adverse influence is recorded on average rate of milk flow and yield on hand stripping.

4. The milking characteristics of cows are correlated with shape of the teat tip. The highest rate of milk flow and the lowest yield on hand stripping is found in cows with flat and concave teat tips.

5. A relationship exists between teat length and the shape of the teat tip, and partial correlation coefficients show that of the teat traits investigated the most important factor affecting average rate of milk flow is the shape of the teat tip whilst teat length—similar to the effect of teat diameter—is of greater importance to the yield on hand stripping. Investigations into the possible consequences for udder health must, however, be conducted before selection for a special shape of the teat tips is introduced.

The fibrous roughages, wheat straw (WS) and cotton stalks (CS), were left untreated or treated with 30 g sodium hydroxide per kg, and residual alkali was either neutralized with sulphuric acid or left unneutralized. All roughages were pelleted and used either in complete feeds or separately. The effects of these treatments on the nutritive value of the roughages were examined in digestibility trials in vitro and in vivo, and in a 7·month feeding trial of a factorial arrangement of 2 × 3 × 2, in which the roughages constituted 350 g/kg of the diets. Approximately 150 g/kg of the neutral-detergent fibre (cell wall) constituents of both roughages was solubilized by sodium hydroxide treatment, while the changes in acid-detergent fibre components (cellulose and lignin) were much smaller, although evident.

In vitro-dry matter digestibility of wheat straw was 22 % higher than that of cotton stalks, and its increase due to NaOH treatment was 50% and 30% for wheat straw and cotton stalks, respectively. In vivo digestibility of organic matter of wheat straw was increased from 0·47 to 0·59, and that of cotton stalks from 032 to 037. The increase was evident in all components except protein, in which digestibility was reduced. The increase in the metabolizable energy values of wheat straw from 6·86 to 7·45MJ/kg dry matter and that of cotton stalks from 4·77 to 5·10MJ/kg dry matter following sodium hydroxide treatment was not reflected in daily live-weight or carcass-weight gain in the feeding trial. The animals offered NaOH-treated roughages had significantly more fat trim, and more fat in the large depots. Including roughages in complete feeds revealed no advantage in metabolizable energy conversion into live weight.

Dry-matter (DM) intake and water consumption were recorded in Borana cattle (Bos indicus) which were subjected to infrequent watering regimes, prolonged walking and night-enclosure. DM intake and water intake were depressed in cows watered every 2nd day (0·04 and 0·16 respectively) and in cows watered every 3rd day (0·1 and 0·3 respectively) as proportions of the food and water intake of daily watered cows. DM intake was further reduced by proportionately 0·05 as a result of a 40-km walk every 3rd day throughout the dry season and by 0·05 as a result of night-enclosure. Digestibility of food was unaffected by treatments. Water intake increased in some months as a result of walking but not as a result of night-enclosure. It is thought that DM intake was depressed with decreasing frequency of watering owing to the physiological effects of dehydration resulting in inappetence. Water consumption fell owing to the constraint imposed by rumen volume at drinking. Walking and night-enclosure independently reduced DM intake as a result of the greatly decreased time available for grazing.

Six boars weighing about 65 kg were penned individually in preliminary observations to investigate their response to increasing levels of faeces in a concentrate diet fed as a gruel. During an introductory 5-day period pigs were restricted to 1·5 kg concentrates plus 100 g fresh faeces daily. The daily concentrate allowance was then increased to 2·2 kg and faeces increased by 50 g/day providing the ration was consumed completely. Two animals showed diarrhoea at intakes of 1 kg faeces per day while the remaining four animals showed no signs of ill health at the maximum allowance of 1·4 kg faeces per day.

In the main experiment littermate pairs of the same sex (16 boars plus 16 gilts in total) were allocated to control (concentrates only) or treated (concentrates plus fresh faeces) groups. Animals were housed in metabolism crates from about 55 kg live weight and the diet allowances were increased gradually up to 2·5 kg concentrates with or without 500 g fresh faeces per day. This was achieved on day 17 and continued to slaughter some 40 days later when pigs weighed around 85 kg live weight. Levels of skatole and indole were significantly higher in the caecal contents of treated pigs at slaughter but similar in faeces compared with controls. There was no effect of dietary treatment on skatole or indole levels in subcutaneous fat.

It seems unlikely that natural coprophagy would contribute significantly to skatole and indole levels in carcass tissues.

The effects of feeding total mixed ration (TMR) or pasture forage from a perennial sward under a management intensive grazing (MIG) regimen on grain intake and enteric methane (EM) emission were measured using chambers. Chamber measurement of EM was compared with that of SF6 employed both within chamber and when cows grazed in the field. The impacts of the diet on farm gate greenhouse gas (GHG) emission were also postulated using the results of existing life cycle assessments. Emission of EM was measured in gas collection chambers in Spring and Fall. In Spring, pasture forage fiber quality was higher than that of the silage used in the TMR (47.5% v. 56.3% NDF; 24.3% v. 37.9% ADF). Higher forage quality from MIG subsequently resulted in 25% less grain use relative to TMR (0.24 v. 0.32 kg dry matter/kg milk) for MIG compared with TMR. The Fall forage fiber quality was still better, but the higher quality of MIG pasture was not as pronounced as that in Spring. Neither yield of fat-corrected milk (FCM) which averaged 28.3 kg/day, nor EM emission which averaged 18.9 g/kg dry matter intake (DMI) were significantly affected by diet in Spring. However, in the Fall, FCM from MIG (21.3 kg/day) was significantly lower than that from TMR (23.4 kg/day). Despite the differences in FCM yield, in terms of EM emission that averaged 21.9 g/kg DMI was not significantly different between the diets. In this study, grain requirement, but not EM, was a distinguishing feature of pasture and confinement systems. Considering the increased predicted GHG emissions arising from the production and use of grain needed to boost milk yield in confinement systems, EM intensity alone is a poor predictor of the potential impact of a dairy system on climate forcing.

Growth performance and carcass quality (linear measurements) were studied in pigs grown from 25 to 90 kg on barley stored and prepared in different ways. Half the barley was dried to about 14% moisture content for storage and half was not dried but treated with 0·8% propionic acid. Each was given after hammer-milling (3·2 mm screen) or rolling. A total of 144 pigs was used at four centres.

The performance and carcass quality of pigs given acid-treated barley were as good as those given dried barley. While feed utilization, expressed on the basis of air-dry feed, was poorer for pigs fed on acid-treated barley this difference disappeared when account was taken of the different moisture contents of the barley samples. Pigs fed on diets based on milled barley grew more quickly, had better feed utilization but greater fat measurements than pigs fed on rolled barley. There was a significant interaction between storage method and feed preparation (milling or rolling). Pigs given the acid-treated barley grew more quickly when it was in the milled form than when it was rolled, whereas for the dried barley there was a small, non-significant difference in favour of the rolled form.

Traditionally silage additives have increased silage intake and the performance of beef and dairy cattle through improvements in silage fermentation. In a series of studies undertaken at this Institute with feaef and dairy cattle an inoculant based on a single strain of lactobacillus planterum (Ecosyl) produced leBirable benefits in terms of silage intake and animal performance without any improvement in silage fermentation. Most of these benefits occurred in studies where the untreated silage was well preserved. Uherefore the present studies were designed to evaluate the same inoculant under difficult-to-ensile conditions.

To determine whether lactating cows select a nutritionally appropriate diet when given grass silage ad libitum and a choice of concentrates high or low in digestible undegradable protein (DUP), 24 mid-lactation Holstein-Friesian cows were given, in different 3-week periods, a concentrate high in DUP (HP), one low in DUP (LP), or a choice between the two (CHOICE), all at an allowance of 5·4 kg dry matter per day. Milk yield was significantly lower, and silage intake slightly but significantly higher, on LP than on either HP or CHOICE. The mean proportion of HP taken in the CHOICE period was 0·47 which was greater than the proportion required to satisfy the cows’ calculated requirements for effective rumen degradable protein, DUP or metabolizable protein but not significantly different from the ratio of 0·50 expected if no choice was made. Individual cows had a strong tendency initially to continue to eat from the feeder from which they had obtained concentrates in the period before CHOICE but this changed to eating significant amounts of both foods within about a week. The fact that the relationship between the proportion of HP in the concentrate intake and milk protein output was significant and positive, and became more so as the CHOICE period progressed, supports the hypothesis that protein demand was directing diet selection but must be viewed with caution as it was heavily influenced by the results of one animal: the highest-yielding cow which also chose to eat the greatest proportion of HP.