Environmental issues, linked to the production of animals in lowland farming systems, are reviewed, using pigs, poultry and dairy cattle as examples. The principal livestock production factors influencing their environmental impact are identified as the balance between different farm animal types and the husbandry practices used for these species, the variable potential which exists for the recycling of wastes and the modification of inputs to systems, the extent to which animal production can be integrated into more holistic farming systems and the impact of livestock on 'wildlife' (plant and animal) biodiversity. The production of large quantities of nitrogenous waste, resulting from the importation of large quantities of nitrogen to intensive animal production units, is identified as the major environmental problem for lowland animal production. The development of-systems which allow these waste products to be re-used at sites of primary crop production is seen as a sustainable solution to this problem.

The extent to which grazing intensities of animal production systems in the uplands of the United Kingdom cause impacts on vegetation, soils, birds, mammals and invertebrates, and influence landscape value and water quality are reviewed. It is argued that these impacts need to be considered in an integrated manner in relation to their responses at the field and landscape scales. Evidence is presented which suggests that a range of grazing intensities is required to obtain significant benefits to the natural heritage. This suggests that new approaches are required to the mechanisms of delivering environmental benefits from grazing systems.

Countries, which import a significant amount of semen, embryos and animals, are faced with the problem of how properly to evaluate the animals in the national evaluation system when information on the foreign parents is generally missing. Additional problems arise when the foreign parents obtain an evaluation, usually on the basis of progeny, in the country of import with the result that there are two published evaluations for the animals in question. This paper presents a post-iterative method of incorporating foreign information into home country evaluations. The foreign information is initially converted to the same scale and base as in the home country using procedures recommended by the International Bull Evaluation Service. The method consists essentially of calculating a combined evaluation for animals with home and foreign information as a weighted average of yield deviations, parent averages and progeny contributions from the foreign and home countries. The combined evaluations are used to adjust the evaluations of progeny on the basis of formulae derived from the usual mixed model equations. The same principles were used to combine reliabilities from the respective home and foreign reliabilities. The results from the application of the method to the United Kingdom (UK) Holstein Friesian population are presented. There was re-ranking of both bulls and cows, especially foreign bulls with few UK daughters.

Two experiments were carried out to evaluate the effect of incremental levels of Gliricidia sepium (gliricidia) and Leucaena leucocephala (leucaena) on forage intake, diet apparent digestibility, microbial nitrogen supply and live-weight changes in cattle. In each experiment, 20 intact and five fistulated crossbred steers (Bos taurus × Bos indicus) were used. Napier grass basal diet was offered ad libitum alone or supplemented with 7·5, 15, 22·5, or 30 g dry matter (DM) per kg metabolic body weight (M075) gliricidia (experiment 1) or leucaena (experiment 2). In experiment 1, total DM intake increased but not significantly with gliricidia supplementation (5·2, 5·1, 5·2, 5·4, 5·7 (s.e.d. 0·21) kg/day) while intake of napier grass was depressed linearly (5·2, 4·7, 4·5, 4·3, 4·2 (s.e.d. 0·21) kg/day; P < 0·05). In experiment 2, supplementation with leucaena increased the total DM intake linearly without depressing the intake of napier grass (5·2, 5·8, 6·2, 6·6, 6·7 (s.e.d. 0·31) kg/day; P < 0001 and 5·2, 5·3, 5·3, 5·3, 5·0 (s.e.d. 0·21) kg/day respectively). Neither gliricidia nor leucaena supplementation affected the apparent digestibility of the diet or in sacco DM degradation characteristics of the forages. Rumen ammonia and live-weight gain were increased linearly (P < 0·05) by supplementation from 130 to 215mg/l (experiment 1) and 75 to 113mg/l (experiment 2), from 306 to 478 g/day (experiment 1) and from 538 to 850 g/day (experiment 2), respectively. However, since the responses were linear, further experiments are required to quantify the responses to higher levels of these legume supplements.

The main purpose of this work was to compare two breeds of improved rams (Suffolk and Merino Precoce) with the local Galego Bragangano breed for the production of crossbred slaughter lambs and to evaluate which cross was more adapted for meat production from the local breed in locations in the north-east of Portugal. The experiment was carried out over a 4-year period in three locations: (1) a farm with an intensive management; (2) an upland farm (400 to 600 m); and (3) a hill farm (above 800 m). Within each flock with 90 Galego Bragangano ewes, two rams of each of the sire breeds were used: Galego Bragangano, Suffolk and Merino Precoce. The lambs were slaughtered at 20 and 40kg, to obtain the carcass weight range of 8 to 14kg. The left sides of 151 carcasses were dissected into muscle, subcutaneous fat, intermuscular fat and bone. The lambs from location 1 had the highest carcass muscle proportion and the lowest carcass intermuscular fat proportion; their differences over locations 2 and 3 were 13 and 16 g/kgfor muscle proportion and 11 and 19 g/kgfor intermuscular fat proportion, respectively. The differences between breeds were relatively small and not significant. However the Suffolk crosses had less kidney, knob and channel fat than the other genotypes (5 and Uglkg less than Merino crosses and Bragangano, respectively). The Suffolk crosses tended to have less body fat.

Coopworth sheep from the high and low backfat selection lines and from one random-bred line developed at Invermay, New Zealand, were used to examine the interaction between the degree offatness of sheep and a period of maintenance feeding (from 20 to 35 weeks of age), in terms of total fat, carcass lean tissue, bone and viscera weights, relative to the empty body weight measured in computer-aided tomography scans. A similar study was done for weights of subcutaneous, intermuscular and internal (abdominal) fat partitions relative to total fat weight. During two a d libitum feeding periods (front 10 to 20 and from 35 to 43 weeks of age) changes in the same components relative to the ‘whole’ (empty body weight or total fat, depending on the case) were expressed in allometric terms.

Only the allometric growth coefficient for total fat weight was significantly affected by the period of restricted feeding, decreasing after this period. No line or sex effect was observed on any growth coefficient. As a result, there were no changes in differences between treatments over the ad libitum feeding periods, with fat animals and females having significantly higher proportions of fat, while females also had a significantly lower proportion of intermuscular fat.

During maintenance feeding, line and sex differences in dissectible component weights did not change significantly, except in terms of intermuscular fat, with the males showing an increasingly larger proportion than the females.

An investigation was conducted into the effect of buserelin treatment on day 12 post mating on ovarian function and plasma progesterone and oestradiol-17 ji concentration in ewes. Changes in ovarian structures were determined, at laparoscopy or after slaughter, 48 h after treatment in 38 control and 41 ewes given an injection of 4 jig buserelin on day 12 post mating (day 0 = day of oestrus). Ovulation rate was higher in buserelin treated (2·76) than in control (1·87) ewes (P < 0·05). Moreover, 41 of the 113 corpora lutea observed in the buserelin group were the result of ovulations that had occurred within 48 h post treatment, compared with two of 71 in the controls (P < 0·05). Blood samples were collected from 10 ewes in each treatment group for hormone analysis. Mean progesterone concentrations tended to be higher, in samples collected from 2 to 6h post treatment, in the buserelin group (4·31 (s.e. 1·32) p.g/1) compared with the control group (3·71 (s.e. 1·44) fig/l. Furthermore, overall mean oestradiol concentrations were lower, during days 13, 14 and 15 post mating, in buserelin treated (0·72 (s.e. 0·26) ngfl) than in control (0·94 (s.e. 0·50) ng/l) ewes (P < 0·05). The results of this study indicate that buserelin treatment on day 12 post mating changes ovarian function.

The effects of undernutrition and of the duration of undernutrition on changes in whole animal energy metabolism and on spontaneous physical activity were studied in adult, non-lactating, non-pregnant ewes. Animals were first given food at 338 kj metabolizable energy (ME) per day per kg live weight 0·75 (LW0·75), (M). Intake was then reduced by half (0·5M) during 7 weeks. Diet apparent digestibility decreased slightly with undernutrition. Rate of LW loss remained constant throughout the whole 0·5M period. Heat production (HP) declined by proportionately 0·18 within the 1st week at 0·5M and by another 0·05 in the following weeks. These latter changes were directly related to LW loss, since HP scaled to metabolic LW remained unchanged over the whole 0·5M period. Time spent standing decreased with undernutrition from proportionately 0·42 at M to 0·33 at 0·5M as a result of a reduction in the average duration of individual standing periods at all hours of the day except in a few hours preceeding feeding times. Energy cost of standing could not be dissociated from the energy cost of eating; these overall costs were not clearly modified by undernutrition but varied with time post prandially. Standing cost alone was estimated at 10 J/min per kg LW. Efficiency of ME utilization for maintenance (km) was calculated at 0·71, and ME requirements for maintenance (MEm) at 338 kj/day per kg LW0'75. No adaptation was noted with the duration of undernutrition. Accounting for the behavioural adaptation tended to increase km to 0·74 but MEm remained unchanged, suggesting that the behavioural changes observed with undernutrition were not sufficient to significantly modify whole animal energy metabolism.

Faecal egg counts and live weights were measured on approximately 200 predominantly twin-born Scottish Blackface lambs each year for 3 years, from 1 to 6 months of age. Measurements were made at 4-week intervals following anthelmintic treatment. Heritability estimates (with s.e.s) of log transformed faecal egg count at each age were 0·01, 0·00, 0·12 (0·10), 0·14 (0·12), 0·15 (0·07) and 0·22 (0·13), for ages 1 to 6 months respectively. Therefore, genetic variation exists for acquired but apparently not for innate resistance to infection. Maternal common environmental effects (with s.e.s) were 0·36 (0·11), 0·20 (0·05), 0·27 (0·09), 0·06 (0·08), 0·15 (0·09) and 0·16 (0·08), for ages 1 to 6 months respectively. Genetic correlations between faecal egg counts in lambs older than 3 months were not significantly less than 1·0, indicating that faecal egg counts at different ages are expressions of the same trait. Phenotypic correlations between faecal egg counts were generally positive but small. Measurement error contributed one-third of the observed variation for individual egg counts. The heritability of mean faecal egg count from 3 to 6 months was 0·33 (s.e. 0·15), indicating that selection decisions can be made more accurately using multiple egg counts per animal. Phenotypic correlations between faecal egg counts and live weight were generally negative but close to zero. However, genetic correlations between faecal egg counts and live weight in lambs older than 3 months were close to -1·0, indicating that resistance to gastrointestinal parasites may be an important genetic determinant of growth rate in this environment.

Genetic parameters were calculated for fibre traits measured on patch samples taken at 5 months of age on a crossbred population of cashmere goats, comprising goats of Scottish feral, Icelandic, Tasmanian, New Zealand and Siberian origin. Within-strain heritabilities, fitting genetic origin as a covariable, were: live weight, 0·71 (s.e. 0·08), fibre diameter, 0·63 (0·07), diameter standard deviation 0·43 (0·08), logipatch cashmere weight), 0·60 (0·06), log(estimated annual cashmere production), 0·51 (0·07), and fibre length, 0·49 (0·15). Including between strain information increased these values to 0·74, 0·68, 0·45, 0·73, 0·67 and 0·57, respectively. Maternal effects for all fibre traits were negligible. Expressions were derived to relate cashmere weight to fibre diameter and length, using functional relationships between these traits. Logarithmic regressions showed cashmere weight to be proportional to diameter2'7, indicating that selection to reducefibre diameter will have a disproportionate effect on cashmere weight. Analyses of subsets of the data confirmed this and showed that correlated responses to selection can be accurately predicted by considering thefunctionally relatedfibre traits on the log scale.

The behaviour of two groups offour pigs in concentrations of ammonia gas that are frequently recorded in piggeries (0, 10, 20 and 40 p.p.m.), was continuously observed for 14 days each in a choice test. An octagonally shaped (annular), eight compartment preference chamber was built to house the pigs. Each compartment supplied the pigs with ample food, water and bedding material but differed in the level of atmospheric contamination. Adjacent compartments were separated by plastic curtains, allowing the pigs free access to neighbouring compartments while reducing cross-contamination of the pollutant. The position of the contamination was changed weekly to eliminate positional preferences. The location of the pigs was scan sampled every 15 min and their behaviour at this time was instantaneously recorded. A significantly greater proportion of their time was spent in the unpolluted compartments (53·4%) than in the 10 p.p.m. (26·9%), 20 p.p.m. (7·1%) or 40 p.p.m. (5·1%) compartments (P < 0·001). This avoidance was maintained for the 14 days that each group inhabited the chamber (P > 0·05). The occupancy pattern was resumed following the rotation of the ammonia concentrations and/or following cleaning (P > 0·05). Higher concentrations of ammonia were visited less often (P = 0·005) and once there, the pigs stayed for a comparatively shorter time (P = 0·003) for approximately 35 min. As the aversion was not immediate, it is suggested that aversion was not due to the odour of ammonia initially experienced on entry. Instead the insidious aversion may be due to a sense of malaise that may develop while a pig is in a polluted atmosphere. The pigs chose to rest (P = 0·002), sit (P = 0·007), feed (P = 0·007) and forage (P = 0·013) more in the unpolluted compartments. Overall more feeding behaviour was observed in the fresh air and more food was consumed in these compartments of the chamber (P = 0·002).

It has been suggested that feeding motivation, and hence the incidence of certain abnormal behaviours, may be reduced by the gastro-intestinal distension which results from the ad libitum provision offoods with a high water-holding-capacity (WHO. Therefore, the aim of this experiment was to test the hypothesis that feeding motivation is inversely proportional to the WHC of meals which supply the same level of nutrients. Eight male pigs (initial body weight 54·0 (s.d. 5·76) kg) were offered a series of four foods which possessed increasing WHC due to a linearly increasing inclusion of dried citrus-pulp (2·18, 3·52, 5·60 and 6·92 g water per g dry food for LB, MLB, MHB and HB), but all with similar ratios of digestible energy (DE) to crude protein and minerals. The experimental treatments were allocated in two 4X4 Latin squares with blocks of 6 days separated by a 4-day change-over period. On each day the pigs were given food twice (09.00 h and 16.00 h) to provide the same level of scaled DE across the meals of each treatment. The resulting level offeeding motivation was measured in two successive tests at 30-min intervals before, and eight successive tests at 30-min intervals after, the ingestion of the first meal using a second-order operant conditioning schedule; a methodology which requires pigs to press a paddle to obtain a feeding-related stimuli (CS). It was found that the number ofCS earned pre-feeding was not affected by treatment (17·2, 17·2, 17·6 and 17·5 (s.e.d. 1·58) CS per test for LB, MLB, MHB and HB; P > 0·05), whereas, the number of CS earned post feeding was significantly affected by treatment (16·6, 15·5, 14·6 and 13·8 (s.e.d. 1·15) CS per test for LB, MLB, MHB and HB; P < 0·001). Within each treatment, the number ofCS earned did not alter significantly across any of the time-periods post feeding. As the number of CS earned has been shown to be a reliable indicator of feeding motivation, the results from this experiment support the hypothesis that feeding motivation is inversely proportional to the WHC of iso-energetic meals.

Polynomial models with random regression coefficients were used to describe cumulated food intake and gain as a function of number of days on test for gilts and castrated male pigs which were on test from 30 to 115 kg live weight. Growth rate and daily food intake were expressed as the derivative of the curves. The applied models allowed a separation of between and within animal variation. Confidence limits for average curves and prediction limits for individual curves were also obtained. A similar model was used to describe gain as a function of cumulated food intake. From this function food efficiency was obtained. The application of the results in stochastic simulation models is discussed.

Growth rate and daily food intake had a more curvilinear progress for castrated males than for gilts. It was estimated that 98% of the castrated males and 96% of the gilts had a lower growth rate at day 80 than at day 50; 74% of the castrated males and 48% of the gilts had a lower daily food intake at day 100 than at day 80. On average food efficiency of gilts was higher than food efficiency of castrated males and the difference increased through the test period.

An experiment was conducted to examine the effects of dietary lysine/digestible energy (DE) ratio (g/MJ) on growth performance and body composition of boars, gilts and castrated males from 25 to 90 kg live weight. Twelve pigs (four of each sex) were assigned to each dietary treatment consisting of lysine/DE ratios from 0·4 to 1·4, in 0·2 g/MJ increments. Food was provided at proportionately 0·90 ad libitum and at 90 kg all pigs were slaughtered and the body composition of two pigs per sex per treatment was determined. Responses to lysine/DE ratios were similar for all sexes up to the optimum level after which daily live-weight gain (DLWG) and nitrogen deposition rate (NDR) deteriorated in gilts and castrated males. This deterioration may have been due to energy used for deamination of excess protein not being availablefor growth processes. Lipid deposition rate (LDR) remained constant from the 0·4 to 0·8 g/MJ lysine/DE ratios and then decreased sharply to a lower plateau from the 1·0 to 1·4 g/MJ lysine/DE ratios suggesting that DE levels were adequate to meet NDR requirements up to the 1·0 g/MJ lysine/DE ratio when it became limiting. The optimum lysine/DE ratio for the genotype tested from 25 to 90 kg live weight was of the order of 0·95 to 1·0 g/MJ. The maximum NDR of the genotype tested appears to be of the order of 28 to 30 g/day (175 to 187 g/day protein deposition rate).

The time taken to lie down and stand up was determined for dry sows from different housing systems and related to physical and genetic parameters. In experiment 1, the times taken to lie down by 32 sows in two dry-sow housing conditions were measured. Sows housed long-term in stalls took longer to lie down than group-housed sows (20·42 v. 9·28 s, P < 0·001). Group-housed sows took longer to lie down in the open than to lie down against a wall (11·07 v. 7·48 s, P = 0·004). The length of time taken for stall-housed sows to lie down had strongest association with body length (P = 0·033, R2 = 0·718). The length of time taken for group-housed sows to lie down in the open had strongest association with the proportional weight of the extensor carpi radialis, (P = 0·001, R2 = 0·915). In experiment 2, the times taken for 30 sows in stalls to lie down and stand up were measured and genotype differences investigated. There were no differences between genotypes in total times taken to stand up or lie down, but total times taken to stand up quickly and lie down had strongest association with body length (P = 0·032, R2 = 0·185, and P < 0·001, R2 = 0·574 respectively). The results indicate that sows housed long-term in gestation stalls experience difficulty of movement when standing up quickly and lying down. Although the chronic effects of lack of exercise and the acute effects offloor type may contribute to this difficulty, the major factor is likely to be space restriction as the times taken to lie down and stand up quickly both increase as body length, and hence dynamic space requirement, increases. Lying down in an unrestricted environment is under muscular control and the degree of control depends on the proportion of muscle weight to total body weight. The factors affecting lying down and standing up should be considered when designing dry-sow and farrowing accommodation.

Carcass and meat quality traits were studied in 94 barrows of known HAL-1843 genotype (homozygote - - and heterozygote - +) from three crosses slaughtered at 90 kg and 120 kg live weight. Crosses A and B were produced from sires of which 0·5 of the genes were from a heavy-muscled × Large White line and 0·5 from pure Pietrain (A) or Duroc (B). Sows for both A and B were from a hyperprolific Large White × Landrace line and this latter line produced the pigs for cross C (control). The carcasses from cross A which were sired from boars having 0·75 of their genotype from heavy-muscled genetic lines had higher killing-out proportion, better conformation scores, greater loin depth and loin area, less carcass length and bone content, and greater lean proportion compared with cross C. The carcasses from cross B, with roughly 0·25 of Duroc genes had a significantly higher backfat thickness than either cross A or C. However their composition was not significantly different from C, since the higher backfat was counter-balanced by their better conformation. Carcass and lean weight distribution was more favourable in crosses A and B, which had significantly higher economic yields than cross C. The halothane genotype did not have a significant effect on fatness and carcass composition, although the heterozygote showed a significantly greater muscle depth. In addition, no significant effect of the halothane gene on carcass and lean weight distribution was observed. In relation to the light group of carcasses (72·8 (s.e. 4·4) kg), the heavy group (100·1 (s.e. 2·7) kg) had greater killing-out proportion, better conformation, longer carcasses, higher fatness and greater loin depth and area. These genetic lines showed a 9·6 g/kg increase in fat and a 6·1 g/kg decrease in lean for every 10 kg of increasing slaughter weight. With the slaughter weight increase the longissimus dorsi muscle became less pale due to a higher muscle pigment concentration, and the backfat had a higher proportion of oleic and lower proportion of linoleic and palmitic fatty acids. The halothane genotype had a significant effect on meat quality. The heterozygote produced much paler, softer and more exudative meat. However, the effect of cross and carcass weight was not significant for meat quality measurements determining pale, soft and exudative meat.

The timing for introducing ad libitum lactation feeding around farrowing was studied in an experiment comprising 20 blocks of three littermate sows and a total of 1085 piglets in 90 farrowings. Sows were given 2·2 kg daily of a diet containing 12·1 MJ metabolizable energy and 120 g digestible crude protein per kgfrom service until ad libitum feeding was introduced either 4 days before expected farrowing (treatment BEF), the farrowing day (AT), or 3 days after farrowing (AFT). BEF and AT sows had a higher total daily food consumption during the experimental period, 5 days before farrowing to weaning at 35 (±3) days, than AFT sows (7·2, 7·2 and 6·6 kg, respectively). BEF and AT sows tended to have deeper ultrasonic backfat (BF) than AFT sows. The mobilization of body reserves during lactation was modest as reflected in negligible losses of live weight (LW) and BF, low serum concentrations of non-esterified fatty acids and creatinine, together with increasing serum urea concentrations as well as the high proportion of sows served within 8 days after weaning (over 90% of the primiparous sows). Sow rectal temperature increased on the farrowing day but was also influenced by feeding level. No cases of agalactia post partum were diagnosed. Litter size, piglet LW, piglet mortality and creep food consumption did not differ between treatments, although there was a tendency for lower LW and increased mortality, particularly of piglets crushed by the sow, in litters of AFT sows. It can be concluded that there was no advantage in delaying the introduction of ad libitumfeeding until after farrowing.

Two experiments were made to investigate the effects of road transport on stress hormone responses in pigs. In experiment 1, seven 40-kg pigs, prepared with jugular catheters, were loaded onto a livestock lorry and transported over a 2-day period on routes characterized, by means of an accelerometer, as rough or smooth. Two 100-min journeys, one rough and one smooth, separated by a 100-min rest period, were conducted each day. The experimenters travelled with the animals and blood samples were taken for hormone analysis from each pig at 20-min intervals. On the 3rd day, samples were collected from the pigs when housed in their home pen (control). Plasma concentrations of cortisol increased after loading, remained higher for longer on rough compared with smooth journeys and were higher during both journeys on day 1 compared with day 2. Concentrations of beta-endorphin increased after loading on day 1 but neither beta-endorphin nor lysine vasopressin showed clear changes in secretion pattern during rough or smooth journeys. On day 3 (control), mean concentrations of all three hormones were significantly lower than on days 1 and 2, indicating that the responses observed were not due to a diurnal rhythm. In experiment 2, six 35-kg catheterized pigs were loaded on a lorry (09.30 h) that remained stationary while blood samples were taken at 30-min intervals during the next 8 h (control). Two days later, this procedure was repeated with the vehicle in motion for 8 h. Plasma concentrations of lysine vasopressin during driving increased between 2 and 4·5 h which coincided with behavioural observations indicating that the pigs were travel sick.

The objective of the study was to investigate the influence of amino acid-deficient diets and high dietary leucine levels offered for intervals of either 21 or 35 days pre-slaughter on the intramuscular fat content of pig muscle. Twenty-four hybrid gilts were offered individually ad libitum one of four diets that comprised combinations of amino acid levels (supplemented or deficient: 5·6 v. 4·0 g lysine per kg) and low or high leucine levels (10·3 v. 30·3 g leucine per kg). Live weights at the start of the study were 74·5 and 87·0 kg for pigs on the 35- and 21-day feeding regimens, respectively. There were no significant effects of dietary amino acid level, leucine level, or feeding interval on food intake, daily live-weight gain or food efficiency. Pigs given the amino acid-deficient diets had lower killing-out proportions (736 v. 747 (s.e. 3·4) g/kg; P < 0·05) and longissimus dorsi muscle areas (32·6 v. 35·7 (s.e. 0·71) cm2, P < 0·01) than those given amino acid-supplemented diets. Subjective score for longissimus dorsi colour indicated that pigs on the high leucine diet had darker muscle than those on the low leucine diet. Intramuscular fat contents of the longissimus dors i muscle at the Wth/llth rib and the 3rd/4th lumbar vertebra and of the semimembranosu s muscle were increased by 19, 18 and 18 g/kg, respectively (P < 0·05), for pigs given amino acid-deficient compared with those given amino acid-supplemented diets. Dietary leucine level had no significant effect on intramuscular fat level. Pigs given the diets for 35, compared with 21, days had higher fat levels in the semimembranosu s muscle (52 v. 30 (s.e. 5·2) g/kg; P < 0·01) but not in the other two muscle locations studied. This study suggests that relatively short-term feeding of amino acid-deficient diets can produce substantial increases in intramuscular fat levels.

Responses in gilt live weight, backfat depth and food intake during lactation and in pre-weaning piglet growth rate were examined after seven generations of divergent selection for daily food intake (DFI), lean food conversion (LFC) or lean growth rate (LGA) on ad-libitum feeding or lean growth on restricted feeding (LGS). There were 252 Large White gilts in the study. Selection for low DFI resulted in gilts with less backfat (25·7 v. 30·7 (s.e.d. 2·21) mm) at farrowing and a substantially lowerfood intake (129 v. 146 (s.e.d. 5) kg) during lactation, but similar reductions in live weight (42 (s.e.d. 6) kg) and backfat depth (8·4 (s.e.d. 1·7) mm) than with selection for high DFI. Therefore, the lower piglet growth (167 v. 295 (s.e.d. 11) g/day) with selection for low DFI compared with selection for high DFI was primarily due to lower food intake of the gilts, as energy for milk production from food was reduced. In contrast, selection for high LFC resulted in relatively smaller changes in live weight (37 v. 48 (s.e.d. 5) kg) and backfat depth (7·6 v. 8·9 (s.e.d. 1·3) mm) than selection for low LFC, which combined with a lower food intake (132 v. 148 (s.e.d. 4) kg) during lactation, resulted in lower piglet growth (181 v. 200 (s.e.d. 11) g/day). The higher food intake of high LGA gilts (137 v. 121 (s.e.d. 4) kg) compensated for the relatively lower reductions in live weight (41 v. 46 (s.e.d. 5) kg) and backfat depth (5·5 v. 6·7 (s.e.d. 1·3) mm) during lactation compared with the low LGA line, such that piglet growth was similar (195 v. 289 (s.e.d. 11) g/day) in the two selection lines. In the high and low LGS selection lines, piglet growth was similar (195 v. 186 (s.e.d. 11) g/day) as was gilt food intake (125 v. 227 (s.e.d. 5) kg) and the changes in live weight (39 v. 41 (s.e.d. 6) kg) and backfat depth (8·1 v. 7·7 (s.e.d. 2·2) mm) during lactation. An examination of the daily energy used in litter gain and the energy available from gilt food intake and mobilization of body lipid indicated that one equation to predict the amount of body lipid mobilized during lactation was not appropriate for different genotypes. Responses in gilt food intake and the changes in live weight and backfat during lactation were selection strategy dependent. However, in general, the selection strategies which reduced gilt voluntary food intake during lactation or resulted in lower live weight and backfat depth at farrowing were detrimental to piglet growth rate.