The influence of microbial activity on the pattern of gas production and the possibility of developing a protocol for standardising the microbial activity of ruminal fluid were investigated (chapter 2). Firstly, ruminal fluid was diluted 1:2, 1:5, 1:9 and 1:20 (v/v) with buffer solution and the <i>in vitro</i> gas production from cellulose, glucose, hay and ryegrass studied. Gas production at 4, 24 and 120h were significantly decreased as the ratio of buffer solution to ruminal fluid increased. For each substrate, incubation with ruminal fluid diluted 1:2 (v/v) gave a higher (P<<i>0.5</i>) asymptotic value 'B', rate 'C' of gas production and lower lag time than when incubated in the 1:5, 1:9 and 1:20 dilutions. Secondly, the effect of frequency of ovine ruminal sampling on microbial activity and substrate fermentation was investigated. The bacterial DM, bacterial absorbance and blank's gas volume were significantly affected by daily sampling of ruminal fluid. Daily sampling of rumen fluid did not significantly affect the gas production degradability parameters (B,C and Lag) until the bacterial DM fell below 0.09 g DM/10 cm³ strained rumen fluid. Bacterial DM was strongly related to the absorbance of ruminal fluid (R² = 0.99, <i>p</i><0.001), suggesting that standardisation of the level of microbial activity between days and laboratory can be achieved by estimating the bacterial DM from the absorbance of the strained ruminal fluid. The influence of donor diet on initial microbial concentration and gas production degradability was studied (chapter 3). Bacterial DM, bacterial absorbance and the blanks' gas volume increased significantly as the ratio of sheep pellet to hay increased. The gas production degradability parameters (B, C and Lag) were also affected by changing the donor diet ratio of sheep pellets and hay. NDF digestibility of cellulose and hay was not significantly affected by donor diet. Bacterial DM was strongly related to the absorbance of ruminal fluid and the blanks' gas volume (R² = 0.99, <i>p</i><0.001). These results suggest that changing the ratio of concentrate to hay reduced the initial bacterial concentration and affected the gas production degradability parameters.

The purpose of this work was to test the ability of sheep to select between foods of different energy density (ED; MJ ME/kg food) a diet which meets their physiological needs, in a series of five experiments. Experiment 1 used pregnant and contemporary non pregnant ewes to test the hypothesis that sheep can select a diet which meets their needs for energy and protein. Four foods were offered: foods A and B were of high ED and foods C and D were of low ED; foods A and C were of high crude protein (CP) concentration, foods B and D were of low CP concentration. The treatments included paired choices of foods A and B, and foods C and D, which were offered to the ewes. The results showed that when the ED of foods is high, pregnant ewes select a diet that reflects their enhanced protein requirements, in contrast with non pregnant contemporaries. When ED is low, priorities other than physiological state may motivate dietary choice. It was suggested that such a priority is the maintenance of a balanced rumen environment. Experiment 2 and 3 tested the effect of ED of foods offered as a choice on the diet selection of sheep and the relationship between the rumen environment and the diet selected from a pair of foods of different ED. The diet selected by the choice fed sheep in Experiment 2 did not consist solely of the high ED food (H), but contained a substantial quantity of the low ED food (L). It was hypothesised that the diet selected by sheep from a choice of low and high ED foods is affected by the rumen conditions, specifically rumen pH and osmolality, and that sheep will alter their pattern of dietary choice to maintain optimal rumen conditions.

This thesis was made to create a new version of the Edinburgh Model Pig computer program, to build a factorial computer model of nutrient utilisation by the breeding sow, to validate the Model Pig and the Model Sow, to expand the modelling function to economic aspects and to link with other packages relating to the feeding and management of pigs, and thereby to build a bio-economic model package of pig production systems. Chapter 1 describes the background of the use of models in pig science and production and indicates the purpose of the present work. Chapter 2 introduces the concepts of compensatory growth theory, heat production and the Gompertz growth function. The use of the Gompertz function and ash contents of the food to predict daily protein retention and water retention was also introduced into the model to test whether these concepts can be used in the new version of the Edinburgh Model Pig. The result shows that the trend of response is consistent with published experiments. Chapter 3 uses those concepts described in the Chapter 2 to create a version 2 of the Edinburgh Model Pig. This model has been extended to include the important weaning period (5 to 20 kg) by improving the prediction of voluntary food intake through the effect of ambient temperature, bulk of the diet, and stocking density. The model provides a graphic function to plot comparative daily biological data. The model indicates the limiting factors in terms of diet (protein, lipid or ash), environment (cold or hot), the situation of food intake control (bulk, allowance or refusal) and compensation of protein growth.

This thesis contains 86 publications on the general subject of the regulation of metabolism and performance in poultry by nutrients, mainly lipids and vitamins, and related biochemical and genetic interactions. There are 6 sections: 1) fat nutrition, 2) body fatness, 3) fatty liver and kidney syndrome, 4) vitamin nutrition and biochemistry, 5) bone biology, 6) sodium nutrition. Section 1 describes the effects of saturated fatty acids in impairing mineral absorption in young chickens, fat utilisation and requirements in turkeys and the role of unsaturated fatty acids in enhancing egg weight in laying hens. It is concluded in these last studies that egg weight is increased by a novel mechanism involving stimulation of oviduct protein synthesis. The problem of excessive fatness in broiler chickens is addressed in Section 2. The studies show how body fatness during growth is related to plasma very low density liprotein (VLDL) concentration and the use of this criterion in a 10-year divergent selection programme to breed lines of lean and fat broilers. Nutritional responses in the lines are compared and the bases of improved leanness and more efficient nutrient utilisation are related to differences in amino acid and lipid metabolism. The differences in immature body composition are shown to persist to maturity and the poorer reproductive performance of the fat line is shown to be associated with extremely high plasma oestrogen levels. It is concluded that genetically leaner broilers are metabolically and reproductively more efficient. The studies identify the practical procedures which have allowed the VLDL method to be applied in industry for the selection of more efficient broilers and turkeys.

Many pig breeding companies now test their young pigs on electronic feeders. This provides them with the means of obtaining accurate estimates of individual feed intake in a group housing system, which is similar to the production environment. Electronic feeders also provide new information on the pattern of how this feed was consumed, such as number of meals in a day or the size of each meal. The aims of this thesis were: to describe these feeding pattern traits in an objective manner; to determine any major environmental or social effects on them; to estimate genetic and phenotypic parameters for them; to predict their potential benefit for increasing the accuracy of selection for land growth and feed efficiency. The data used in these analyses were compiled between 1992 and 1995 from individual records of 1832 pigs from 70 sire families of a Large White derived sire line selected for lean tissue growth rate. Pigs were fed <I>ad libitum</I>, in single sex pens of 12 pigs (s.d. 0.87), using FIRE (feed intake recording equipment) system from Hunday Electronics Ltd. at the Cotswold Pig Development Company. Pigs were on test between 45 kg (s.d. 2.76) and 95 kg (s.d. 6.78). Daily feed intake (DFI kg), feed intake per visit (FIV kg), number of visits per day (NV), duration of each visit (TV mins.), time in the feeder day (TD mins.), feeding rate (FR kg/min) and number of non-feeding visits per day (NFV), were measured as means of test and also as means of bi-weekly periods of test (which were the means of weeks 2-3, weeks 4-5, weeks 6-7 and weeks 8-9). Feeding patterns were affected by the social and physical environment within a pen, particularly time spent in the feeder per visit and per day. DFI and performance test traits were unaffected which suggests that pigs were able to adapt their behaviour to compensate for different social structures within a pen. Genetic and phenotypic parameters were estimated by restricted maximum likelihood with a multivariate individual animal model. DFI had a heritability of 0.21 ranging from 0.18 to 0.26 over the four bi-weekly test periods. Correlations between part and whole test records of feeding patterns and DFI were high (r<SUB>g</SUB> = 0.75 to 0.99). DFI was highly correlated with performance test traits, but had low correlations with feeding pattern traits. The heritabilities of feeding pattern traits were low except for FIV and NV, but correlations between feeding pattern traits were high. FIV, NV and TV were moderately correlated with performance test traits.

The foraging strategies of cattle and goals were studied in different seasons in order to determine how animals respond to changes in the abundance of food resources in semi-arid rangelands in Zimbabwe. Three different approaches were used to test hypotheses concerning dietary shifts under different resource availability scenarios in the two species. In a field experiment, the foraging behaviour of cattle and goats responded to different cues in their environment: cattle were sensitive to changes in grass biomass while goats responded to browse. However, goats showed more flexibility in their response in changes in range condition by increasing the proportion of time spent on the more available resource. Similar results were found in an observational study of the foraging behaviour of goats and cattle in a communal area. Here the dry season strategy of goats consisted of increasing browsing and expanding their diet breath, while cattle depended more on stover in the fields as a forage supplement in the dry season.  key resources such as grass on contour bunds, riverine areas and fallow fields were also important, at the large scale. In both the field experiment and the communal area study; cattle and goats how low diet overlap indices, particularly for browse because cattle avoided species with thorns and selected broad-leaved species. In general, cattle had higher instantaneous intake rates on grass than browse, but matched intake rates on grass when they browsed on broad-leaved species. Goats were equally efficient on grass and browse, and had higher intake rates per metabolic body weight than cattle. Instantaneous intake rate under controlled conditions predicted that cattle would select broad-leaved browse species in preference to species with small leaves and thorns. This concurs with the results of the cattle in a field experiment and communal area study. The instantaneous intake rate of goats was only affected by a reduction in leaf density. Evaluation of range resources for sustainable animal production should take into account the animal species involved. Management interventions should be targeted at the dry season and key resource areas need to be protected so that they can be used strategically.

The use of silage additives can improve the reliability of the ensilage process. The fact that chemical additives are both highly corrosive to expensive farm machinery and hazardous to handle has prompted the development of alternative silage additives. Bacterial inoculants were developed to provide a source of homofermentative lactic acid bacteria. The objective of this study was to develop a bacterial silage inoculant that would improve the efficiency of ensilage of a crop. An inoculant termed SilCare II was developed. SilCare II, comprising a ratio of three <i>L.plantarum</i> L71 cells to every one <i>P.pentosaceus</i> P826, was always applied in the presence of molasses at the rate of 9 ml kg<SUP>-1</SUP> herbage. The freeze-drying of SilCare II in the presence of rehydrated skimmed milk was found to be the most successful protectant for preserving the cultures, both during the freeze-drying process and over a long storage period. Freeze-dried SilCare II preserved herbage more successfully than a commercially available inoculant, Lactomol, or an untreated control. Improvements in the efficiency of the ensilage process were achieved by increasing the application rate of the inoculant to the herbage. However, large scale ensilage experiments indicated that the inoculant had little effect on the fermentation of herbage in bunker silos. Animals trials using grey-faced wether lambs indicated that SilCare II did not significantly improve the nutritional quality of the silage.

This study contributes to the understanding of spatial aspects of plant-herbivore interactions within a grazed grass-shrub mosaic in heather moorland, an internationally important ecosystem dominating much of the Scottish uplands. A three-year field experiment was conducted to observe plant-herbivore interactions between Scottish Blackface sheep and heather-grass mosaics. Modelling tools were used for virtual experiments, thus complementing and extending the field data. The experiment showed the importance of a spatially explicit approach to understanding the interactions. The pattern of use of the vegetation mosaics by sheep was strongly heterogeneous, with spatially limited areas of intensive use intermixed with large areas of extensive use. Foraging and ruminating behaviour showed distinctively different patterns of impact, indicating that multiple processes determined herbivore of vegetation mosaics and their concomitant impacts on the dynamics of the vegetation. Application of a spatial interaction model, previously used in human geography, to the field data revealed that the amount of grass in an area was a good predictor of the local heather defoliation. Heather defoliation was highest near large grass patches and lowest away from small grass patches. The virtual experiment showed further that cognitive aspects of foraging behaviour could play an important role in determining the pattern of use by herbivores. Performance of foraging strategies was strongly affected by the heterogeneity of the vegetation, suggesting that herbivores could adapt their foraging strategies depending on the pattern of vegetation. This study provides new insights into the spatial aspects of plant-herbivore interactions in grass-shrub mosaics and offers a starting point for more detailed investigations. At the same time the results necessitate the increased use of spatially explicit approaches in the management of grazed ecosystems.

1. High temperature decreased food intake, growth rate, energy retention as protein and water retention but had no effect on energy retention as fat and little effect on total energy retention. 2. High temperature did not affect the proportion of apparent metabolisable energy retained as protein or the efficiency of energy retention from apparent metabolisable energy. However, it increased the proportion of apparent metabolisable energy retained as fat. The latter effect was attributable to the reduced energy requirement for thermoregulation. 3. Offering the birds a choice between a "high-protein" and a "high-energy" diet did not improve growth when the animals were exposed to high temperature (30°C). Moreover, choice-feeding led to decreased food intake and, in turn, growth rate in birds kept at moderate temperature (20°C) when compared with those given a complete compound diet. 4. Compared with birds kept at 20°C and fed <I>ad libitum, </I>there was a lower concentration of plasma triiodothyronine in birds kept at 20°C but pair-fed the same amount of food as birds eating <I>ad libitum </I>at 30°C. Triiodothyronine concentration was shown to be affected directly by food intake as well as by ambient temperature. 5. High temperature reduced the quantity of digesta in the gut. It was suggested that this may be due to a reduced ability of the gut to process food. 6. Dietary modifications such as finely grinding and wetting the food led to increased food consumption and growth rate at high temperature. 7. Increasing the sodium chloride content of a finely ground and wetted diet improved metabolisability. 8. Increasing the protein content of a finely ground and wetted diet increased body weight gain but caused decreases in total energy retention, protein retention efficiency and metabolisability. 9. It is usually assumed that reduced food intake at high temperature results firstly from a decrease in energy requirements for thermoregulation and secondly from the need to minimise heat stress by reducing the heat increment of feeding. However, the results of this project suggest that the domestic fowl reduces its food intake at high temperature not to reduce or limit heat production, but because the gut is limited in the amount of food it can process. If the gut is "assisted" by providing wetted or finely ground food, food intake, growth rate and heat production all increase.

The aims in this thesis were to investigate more comprehensive breeding goals in dairy cattle, to investigate the possible importance of genotype by environment interaction and to identify possible sources of genetic variation in feed utilisation. On the basis of a literature review, longevity was examined as component of the breeding goal, and four linear type traits (foot angle, udder depth, teat length and angularity) with the closest genetic association with longevity were used as index measurements. Economic values were derived for protein, fat and milk yield and for longevity using dynamic programming. The derived values were in genetic standard deviations relative to protein yield 1.0, 0.2, -0.2 and 0.8, respectively. Three completely additive indices were derived, assuming that the breeding goal was for: (i) yield only (YIN), (ii) longevity only (LIN) or (iii) yield and longevity, hence profit (PIN). Selection on PIN was expected to give a 5% higher annual rate of genetic progress in economic merit compared with selection on YIN, and PIN was robust to most assumptions made in the calculation. Genotype by environment interaction was investigated for a range of traits. Selection (S) and Control line (C) cows, housed and managed at the Langhill Dairy Cattle Research Centre, have been offered <I>ad libitum</I> complete mixed diets, with proportions (in total DM) of concentrates, silage, brewers' gains of either 20:5:75 (LC; 1.0 ton concentrate per annum) or 45:5:50 (HC; 2.5 ton concentrate), over a full lactation. No diet x genetic line interactions were observed for a number of traits, describing milk production, feed intake, efficiency and body tissue mobilisation. However, regression coefficients of milk yield and condition score on pedigree index for fat plus protein yield were significantly different between LC and HC. Phenotypic and genetic variances were generally larger on HC than on LC, but difficulties in separating the permanent environmental variance from the additive genetic variance might have obscured some of the comparisons.