The broad objective of the study was to determine feeding behaviour and gut health of pigs fed different fibres at varying inclusion levels. Maize cob, maize stover, sunflower hulls, veld grass, sawdust, lucerne and dried citrus pulp were used in growing and finishing pig diets to provide a wide range of physicochemical properties. Time spent eating, drinking, lying down, sitting/standing and other activities was observed using video cameras. Blood samples were collected at the end of the trial for both growing and finishing pigs for analysis of glycated haemoglobin, albumin, globulin, total protein, creatine kinase, urea and uric acid. Intestinal segments were collected at the end of trial for growing pigs to determine mucosal architecture of the intestines. Digestible energy (DE), bulk density (BD), acid detergent fibre (ADF) and water holding capacity (WHC) were the most important variables predicting time spent on different behavioural activities in growing pigs (P < 0.001). Water holding capacity, neutral detergent fibre (NDF) and ADF were the most important variables involved in predicting time spent on different behavioural activities in finishing pigs (P < 0.001). Physicochemical properties influenced time spent on different behavioural activities and they provided relationships with time spent on different behavioural activities in both growing and finishing pigs. Glycated haemoglobin, albumin, globulin, total protein and uric acid were factors that influenced time spent eating in growing pigs (P < 0.05). In finishing pigs, only serum total protein was selected as the best predictor variable influencing time spent eating (P < 0.05). The blood metabolites were correlated with time spent eating and drinking. They provided threshold values with time spent eating and drinking by pig. Hence, they can also be used as potential biomarkers that modulate neuronal pathways which reduce time spent eating and drinking. In this study, bulk density (BD) and NDF were the best predictor variables influencing villi height (VH) and apparent villi surface area (AVSA) in pigs (P < 0.05). Grouped pigs fed fibrous diet spent more time eating, lying down, standing, walking and fighting (P < 0.05). Skin lesions appeared the most on neck and shoulder region followed by chest, stomach and hind leg region, and finally head region (P < 0.05). Fibrous diet did not reduce aggressive behaviours, rather aggressive behaviours emanated out of frustration when queuing at the feeder. It was concluded that physicochemical properties of fibrous diets and nutritionally-related blood metabolites influence feeding behaviour. Mucosal architecture was also influenced by physicochemical properties of the fibrous diets. / Ph.D. University of KwaZulu-Natal, Pietermaritzburg 2013.

The broad objective of the study was to determine the effects of restricted water intake on the growth, blood chemistry and meat quality characteristics of indigenous chickens in semi arid environments. A total of 15 flocks in communal villages and 12 flocks in resettlement schemes of Msinga District in South Africa were monitored for 30 months to determine the effects of production system and season on flock size, dynamics and constraints faced by indigenous chicken producers. As a follow up, 281 and 233 chicks hatched in November 2011 from 18 and 9 households in communal villages and resettlement schemes, respectively were monitored using a structured checklist to determine survival and causes of mortality from hatching up to 12 weeks of age. Kaplan-Meier survival distributions and the odds ratios for effects of potential risk factors were determined using survival analysis and logistic regression models, respectively. Flock composition and structure were significantly affected by production system and month. Farmers in resettlement schemes had approximately one cock and three hens more (P < 0.05) than those in communal villages. The number of growers and chicks were similar (P > 0.05) between the two production systems but chicks were the predominant age group (38 %) of each flock. The cock to hen ratio in households in communal villages was 1:3.5, while that in resettlement schemes was 1:3.7, suggesting that inbreeding might have been reducing flock productivity. Flock sizes in communal villages peaked in March (45.1 ± 3.02) during the post rainy season and declined steadily to a low of 34.7 ± 3.63 in September during the hot-dry season. The largest flock sizes in resettlement schemes were observed in January (52.4 ± 5.09) during the hot-wet season and the smallest in August (36.1 ± 5.98) during the cold- dry season. Households in communal villages had more chicks (P < 0.05) than cocks, hens and growers throughout the year except in June when the number of growers was equal to that of chicks. In resettlement schemes, the number of chicks was lower (P < 0.05) than the number of hens and growers, except in June and July when it was equal to that of growers. Total entries were not affected (P > 0.05) by production system. Hatched chicks were the major mode of entry, accounting for more than 97 % of entries into flocks. The contribution of purchases, gifts and exchanges was negligible. Mortality was the main cause of exits from flocks, accounting for 70 and 63 % of total exits among households in communal villages and resettlement schemes, respectively. The major causes of mortality were aerial predators, sub-optimalnutrition and inclement weather conditions, which were similar between production systems. The number of birds that exited flocks through slaughter for household consumption was higher (P < 0.05) among households in resettlement (34% of total exits) than communal villages (21 % of total exits). The proportion of chickens exiting flocks through sales was higher in communal (9 %) than resettlement (4.3 %) households. Chick survival from hatching to 12 weeks was higher (P < 0.05) in communal villages (55 ± 3.14 %) than in resettlement schemes (41 ± 4.19 %). Mean chick survival time did not differ (P > 0.05) between communal (56 ± 3.30 days) and resettlement (49 ± 3.23 days) flocks. Provision of water ad libitum and treatment of sick birds were important covariates in prolonging the survival time of chicks. The effects of restricted water intake on growth performance, blood chemistry, physicochemical properties, and sensory characteristics of meat from Naked Neck (NNK) and Ovambo (OVB) chickens were also assessed. In the experiment, 54 pullets of each strain with an average weight of 641± 10g per bird were randomly assigned to three water restriction levels for 60 days in a completely randomized design. The treatments were ad libitum, 70% of ad libitum and 40% of ad libitum water intake. Each treatment group was replicated three times. The pectoralis (breast) muscle was sampled for meat quality, fatty acid composition and sensory quality analyses. Ovambo chickens had superior body-weight at 16 weeks of age, average daily gains (ADG) and average daily water intake (ADWI) than NNK chickens. Body weight of birds at 16 weeks of age, ADG, average daily feed intake (ADFI), ADWI and water to feed ratio (WFR) declined progressively (P < 0.05) with increasing severity of water restriction, while food conversion ratio (FCR) values increased (P < 0.05) as the severity of water restriction increased. Naked Neck chickens had better FCR at the 40 % of ad libitum water intake level than OVB chickens. The dressing percentage per bird was higher (P<0.05) in water-restricted birds than those on ad libitum water consumption, irrespective of strain. Heart weight was significantly lower in birds on 40% of ad libitum water intake than those on ad libitum and 70% of ad libitum water intake, respectively. Packed cell volume was higher (P < 0.05) in NNK than OVB chickens offered waterad libitum, but similar in birds offered 70 and 40 % of ad libitum. There were no differences in erythrocyte count (RBC) and mean corpuscular volume (MCV) values between strains, but MCV was higher in birds on 40 than 70 % of ad libitum water intake, irrespective of strain. Naked neck chickens had higher (P < 0.05) white blood cell count (WBC) values than OVB chickens at 40 % restriction level, but lower WBC than OVB at 70 % water restriction level. Uric acid, creatinine, triacylglycerides, total cholesterol, low density lipid cholesterol, total protein and globulin increased (P < 0.05) with each increment in water restriction, but the increase in creatinine and total cholesterol was more pronounced in OVB than NNK chickens. The opposite was observed for uric acid. Alanine transaminase, alkaline phosphatase and aspartate transaminase activities were not influenced by strain and water restriction. It was concluded that the two strains could withstand up to 40 % of ad libitum water restriction, but NNK chickens tolerated water stress better than OVB chickens. Water intake levels of 40% of ad libitum produced meat with significantly lower (P < 0.05) cooking loss, and higher (P < 0.05) redness (a*) values in NNK chickens compared with OVB chickens. Water intake level had no effect (P > 0.05) on lightness (L*) and yellowness (b*) values, shear force, moisture and protein contents in both strains. The fat content of NNK meat was 41 % lower (P < 0.05) than that of OVB meat at 70 % of ad libitum, but 31 % higher at 40 % of ad libitum water intake. The ash content was significantly elevated (P < 0.05) in birds on 70 % of ad libitum compared to those on ad libitum and 40 % of ad libitum water intake, which had similar (P > 0.05) ash contents. Birds on 40 % of ad libitum water intake had significantly higher (P < 0.05) proportions of octadecanoic acid (C18:0), cis, cis-9,12-octadecadienoic acid (C18:2 ῳ-6), cis-8,11,14,17-eicosatetraenoic acid (C20:4 ῳ-6), cis-7,10,13,16-docosatetraenoic acid (C22:4 ῳ-6), cis-4,7,10,13,16,19-docosahexaenoic acid (C22:6 ῳ-3), total polyunsaturated fatty acid (PUFA), total omega-3 PUFA and total omega-6 PUFA proportions, but lower (P < 0.05) cis-7-hexadecenoic (C16:1c7), cis-9-octadecenoic (C18:1c9), cis-11-octadecenoic acid (C18:1c11), cis-13-docosenoic acid (C22:1c13), total monounsaturated fatty acids than those on the 70% of ad libitum and ad libitum water intake, respectively. The proportion of trans-9-octadenoic acid (C18:1t9) was higher (P < 0.05) in NNK chickens on 40 % of ad libitum water intake than OVB chickens. It was concluded that water restriction at 40 % of ad libitum water intake resulted in favourable cooking loss values and meat redness (a*) values, omega-3 and 6 PUFA proportions and a high ῳ-6/ῳ-3 ratio. The high fat content of NNK chickens at 40 % of ad libitum water intake compared to OVB chickens suggests a superior adaptation to hydric stress. Naked Neck breast meat had higher initial impression of juiciness scores than that from OVB chickens, but only in birds on ad libitum and 70 % of ad libitum water intake. Sensory scores for first bite, connective tissue and tenderness decreased with increasing severity of water restriction (P < 0.05). Aroma, flavour and atypical flavour were not affected by strain or water restriction level (P > 0.05). There were significant strain differences for sustained impression of juiciness and tenderness, with the highest scores occurring in NNK chickens (P < 0.05). Aroma had a significant influence on the flavour of breast meat (P < 0.05). Fat content was significantly correlated with initial impression of juiciness, first bite and sustained impression of juiciness of breast meat. It was concluded that water restriction up to 40 % of ad libitum had a significant and adverse impact on juiciness and first bite scores of meat. / Ph.D. University of KwaZulu-Natal, Pietermaritzburg 2013.

A study was conducted to test the effects that different high fibre sources and their varying inclusion levels had on the characteristics of the excreta (faeces, urine and the slurry), and on odour from the slurry of growing pigs. Faeces and urine characteristics were tested from 52 pigs fed rations containing grass hay (GH), lucerne hay (LH), maize cobs (MC), maize stover (MS) and sunflower husk (SH) diets at inclusion levels up to 400 g/kg as fed basis. Faecal output, faecal consistency and nitrogen were influenced by fibre type (P <0.01) and inclusion level (P <0.01). Nitrogen content in faeces and urine was also affected by dietary fibre inclusion. Increasing fibre inclusion levelled to a reduction in urinary nitrogen content, indicating nitrogen repartitioning from urine to faeces, thereby minimizing nitrogen volatilization. The slurry from pigs fed on LH, MC and SH at levels up to 160 g/kg was tested for chemical composition and odour offensiveness. The slurry was incubated for 16 days. The pH and nitrogen content varied among fibre types and incubation period (P <0.05). Isobutyrate and butyrate concentrations varied with fibre type and the incubation period tested (P <0.01). Using panellists, the SH containing rations resulted in low odour offensiveness score. Maize cob-containing diets resulted in the largest odour scores, with (mean rank of 2.2 and 4.3 for SH and MC, respectively). To reduce odour offensiveness from piggeries, sunflower husk was recommended as an alternative feed ingredient for growing pigs. / Thesis (M.Sc.Agric.)-University of KwaZulu-Natal, Pietermaritzburg, 2012.

The broad objective of the current study was to determine the response of pigs fed on incremental levels of Acacia tortilis leaf meal-based diet. Eight trees of each of the following five dominant leguminous leaf meals; A. tortilis, A. robusta, A. nilotica, A. nigrescens and A. xanthophloea, were individually hand harvested from the same grazing camp at Makhathini Research Station, Jozini, South Africa. The leaf meals were harvested between April and May 2013 after the rainy season at advance stage of maturity. Following the nutritive evaluation of Acacia species, A. tortilis was selected for the feeding experiment. Thirty finishing male F1 hybrid (Landrace × Large White) pigs with an initial weight of 60.6 (s.d. = 0.94) kg were randomly allotted to six dietary treatments containing 0, 50, 100, 150, 200, 250 g/kg DM inclusion levels of A. tortilis leaf meal. Each treatment diet was offered ad libitum to five pigs in individual pens for 21 days. Average daily feed intake (ADFI), body weight, average daily gain (ADG) and gain: feed ratio was measured every week. Blood was collected at the end of the experimental period for the determination of nutritionally-related blood metabolites and activity of Aspartate aminotransferase (AST), alanine amino transferase (ALT) and alkaline phosphatase (ALP). Following a feed withdrawal period of 12 hours, pigs with a mean body weight of 80 ± 15 kg were slaughtered, eviscerated for the collection of livers, kidneys, lungs and heart from each pig and weighed individually using a digital scale. Acacia tortilis and A. xanthophloea leaf meals had the highest CP and fat content (P < 0.01) among all the Acacia species. The NDF and ADF varied significantly across Acacia species (P < 0.05). Acacia robusta had significantly the highest levels of non-structural carbohydrates, whilst A. tortilis had the lowest. The concentration of proanthocynidins varied across the species. Acacia tortilis had significantly high levels (23 g/kg DM) of phosphorus compared to the other species. Acacia nilotica had the highest (P < 0.001) concentration of iron, but it had the lowest (P < 0.01) level of manganese. There was a quadratic decrease in both ADFI and ADG (P < 0.001) with as A. tortilis leaf meal increased. The gain: feed ratio was linearly reduced (P < 0.001) with incremental levels of A. tortilis leaf meal in the diets. Serum concentrations of Fe, AST and ALP increased quadratically (P < 0.01) as A. tortilis leaf meal increased. There was a linear increase in ALT activity with increased leaf meal inclusion. Hepatosomatic index (HSI: liver weight/ body weight x 100), scaled kidney weight and scaled heart weight increased linearly (P < 0.001) as A. tortilis leaf meal increased. There was, however, a quadratic increase in the relative weight of lungs (P < 0.001) as leaf meal increased. In conclusion, the piecewise regression (brocken-stick) NLIN procedure (SAS, 2008; SASA institute Inc.) showed that A. tortilis leaf meal can be included in finisher pig diet up to 150 g/kg DM of feed without negatively affecting growth performance, nutritionally-related blood metabolites, liver enzymes and internal organs of pigs. / M.Sc.Agric. University of KwaZulu-Natal, Pietermaritzburg 2013.

Escherichia coli diarrhoea is the most important source of mortality in piglets. The most frequently isolated strain in enterotoxigenic E. coli diarrhoea is F4ab/ac. Recent studies in South Africa reported non-fimbrial strains such as PAA and EAST-1 to be prevalent. The objective of the study was to determine whether there are breed differences among pigs with respect to E. coli adhesion phenotypes and correlate them to polymorphisms at selected candidate genes in the South African population. A total of 225 pigs aged 3-12 weeks of the imported (Large White, Landrace and Duroc), local and crossbreds, were sampled from the Eastern Cape and Limpopo provinces of South Africa and genotyped for PCR-RFLP polymorphisms at four candidate genes associated with E. coli F4ab/ac resistance/susceptibility. These genes were Mucin 4 (MUC4), Mucin 13, (MUC13), Mucin 20 (MUC20) and Transferrin Receptor (TFRC). The TFRC and MUC13 genes were less polymorphic, the C allele was close to fixation and the homozygous CC genotype was the most frequent in all three pig populations. There was a significant difference (P <0.05) in allelic and genotypic distribution amongst breeds for the TFRC locus. The g.8227G>C polymorphism in MUC4 segregated in all three breeds and the marker was moderately polymorphic. There was a significant difference (P <0.05) in genotypic distribution amongst breeds for MUC4.The g.191C>T polymorphism in MUC20 segregated in the local and crossbred pigs and was close to fixation in the imported pigs. There was a significant difference (P <0.05) in allelic and genotypic distribution amongst breeds for MUC20, which was moderately polymorphic. There was a reduction in heterozygosity in both the TFRC and MUC13 loci, although MUC4 and MUC20 genes had higher heterozygosity levels. The MUC4 gene had a negative FIS value, indicating outbreeding at this locus. The MUC20, MUC13 and TFRC genes had a positive FIS value, indicating inbreeding at these loci. Overall, the studied population was outbred. Imported pigs in TFRC and MUC20 deviated from Hardy-Weinberg equilibrium (HWE). All breeds were in HWE at the MUC4 and MUC13 genes. There was no linkage disequilibrium observed amongst the analysed loci. iv A total of 109 piglets of three breeds (Large White, indigenous and crossbred) aged 3-5 weeks, were investigated for the susceptibility to E. coli F4, PAA strains and EAST-1 toxin. Adhesion tests were conducted on pig intestinal cells, which were viewed under a phase contrast microscope. Three phenotypes were identified as, adhesive, weakly adhesive and non-adhesive. There was a significant association (P <0.05) between breed and level of adherence of the F4 and PAA strains. Highest frequencies of adhesion phenotypes were observed in the indigenous pigs for both F4 and PAA E. coli strains. Large White pigs had the lowest frequency of non-adhesion in F4 and PAA E. coli strains. The F4 strain had a higher (P <0.05) level of adherence compared to PAA and EAST-1 in Large White pigs. Age of pigs had a significant effect on the level of E. coli adherence in indigenous and crossbred pigs (P <0.05). Adhesion of F4 and EAST-1 was higher in weaned indigenous and crossbred pigs, respectively, than in suckling piglets. There was no significant difference between F4 adhesion and the genotypes at all four candidate genes genotypes. The study showed that both imported and local pig populations carry receptors and are susceptible to F4, PAA and EAST-1 E. coli infections. Indigenous pigs were less susceptible than Large White to E. coli infection. Although polymorphic and segregating in the populations, the MUC4 g.8227G>C and MUC20 g.191C>T mutations were not associated with the adhesion phenotypes and cannot be used in the selection of susceptible animals. / M.Sc.Agric. University of KwaZulu-Natal, Pietermaritzburg 2013.

African horse sickness (AHS) is a non-contagious viral disease transmitted by an arthropod vector and is endemic to sub-Saharan Africa. The disease affects all equine species, but is more severe in horses and other equid species not native to Africa. Vaccination is the only demonstrated means of its prevention. The horse-owning public provides much anecdotal evidence of prophylactic strategies, such as repellents, stabling, alternate hosts, traps, paraffin, blankets, smoke or fans. The present study investigated the relationship of these strategies to the incidence of AHS, and evaluated alternate hosts, wind speed and repellents on the activity of males and females of the different Culicoides species.. Cypermethrin and citronella-containing repellents repelled the most female midges. Sheep and cattle offer an alternate blood meal to gravid and nulliparous female midges. Fans are very effective in keeping midges away from horses. Methods are summarised for the horse owner to implement in addition to vaccination to prevent AHS. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2008.

African horse sickness virus causes a non-contagious, infectious disease of equids. It is epizootic to sub-Saharan Africa and parts of the Middle East. The epizootics caused by the virus have caused widespread devastation amongst equids worldwide. Fortunately no epizootic has lasted more than 5 years outside of sub- Saharan Africa. It is vectored by species of Culicoides midges (Diptera: Ceratopogonidae) and most importantly by the two Avarita species of C. imicola Keiffer and C. bolitinos Meiswinkel. The literature pertaining to the study and research of the virus, the disease and the vectors is reviewed. Models allowing prediction of future possible outbreaks as well as details of control strategies and findings of researchers are presented and discussed. The virus needs a long term reservoir host in which to overwinter and various theories are discussed. Control measures in South Africa are suggested so that outbreaks of the disease can be reduced. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermarizburg, 2008.

No abstract available. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2006.

This research reported here was carried out to examine alternatives to antibiotic growth promoters as a result of their being banned in the animal feed industry. Four experiments were conducted to evaluate the efficacy of non-medicated feed additives as replacements for antibiotic growth promoters in broiler feeds. The additives used were enzymes (a new thermo-tolerant powder enzyme called TXAP, phytase, lipase and a new phytase enzyme derived from E. coli called phyzyme XP), organic acid (Acid Pak), prebiotic (Bio-Mos®) and probiotic (All-Lac XCL). Mashed maize-soya based feeds were used in all the experiments, which were conducted in litter-floor pens. The first experiment was a dose-response trial. Broilers in eight replicate pens of 50 males and 50 females were fed unsupplemented feeds and five additional feeds containing increasing levels of TXAP, from 0.5 to 2.5 g/kg to 42 d. The second experiment used enzyme TXAP with two different enzymes (phytase and lipase), individually or in combination. Six replicate pens of 50 males and 50 females were fed either unsupplemented feeds or one of six additional feeds treated with TXAP, lipase, phytase , a combination of TXAP and lipase, a combination of TXAP and phytase or a combination of all the three enzymes . This trial continued for 42 d. In the third experiment three types of TXAP (Lot 1, 2 and 3) were used, with fixed levels of xylanase and amylase but varying levels of protease activities (4000, 2000 and 1000 U/kg for Lot 1, 2 and 3, respectively) in combination with phyzyme XP for 35 d. The fourth experiment used mannan-oligosaccharide (Bio-Mos®), organic acid (Acid pak 2x), probiotic (All Lac XCL 5x), individually or in combination and an antibiotic growth promoter (Zinc bacitracin) for 42 d. The chickens in this experiment were challenged with Clostridium perfringens (CP) at 21, 22 and 23 d to determine the efficacy of these additives for replacing antibiotics in hindering the effects of CP on the villus surface area. The dose-response trial did not show any significant improvement in broiler performance with any level of inclusion of enzyme TXAP. The results from this study showed some beneficial effects with the use of enzyme TXAP when fed alone and at a young age. Its use when combined with other enzymes and at later stages of growth needs further investigation. Feed additives in experiment 4 prevented the negative effects of CP as the treated chickens did not have lesions on their villus surfaces. The conditions under which these trials were conducted appeared to be such that little benefit was derived from the use of any of the feed additives used. It is possible that under less-hygienic conditions such as those in commercial operations greater benefits from these additives may be realised. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2006.

Two experiments were conducted for this thesis, to determine whether an animal should be fed to its genetic potential in spite of this not being achievable due to an on-farm constraint. The first experiment was designed to compare the response of pigs housed either individually or in groups to a range of feeds limiting in lysine between 40 and 85 kg live weight. Two hundred and eighty-eight entire male Large White x Landrace pigs were used. The experiment was divided into two growth periods, i.e. from 40 to 60 kg and from 60 to 85 kg. In each period, pigs were subjected to feed containing one of four dietary lysine concentrations. In Period 1, the lysine concentrations were 11.03 (L1); 9.54 (L2); 8.00 (L3) and 6.51 (L4) g/kg, while in Period 2 these were 7.82 (T1); 6.71 (T2); 5.55 (T3) and 4.40 (T4) g/kg. Pigs fed an L1, L2, L3 or L4 diet in Period 1 were fed a T1, T2, T3 and T4 diet in Period 2, respectively. Three buildings provided the following group sizes and floor-space allowances: House 1 contained eight pigs per pen at 1.94 m2/pig; House 2 contained four or eight pigs per pen at 1.72 or 0.86 m2/pig; and House 3 contained one pig per pen at 1.72 m2/pig. The individually-housed pigs were divided into three feeding levels, i.e. ad libitum, or pair-fed so that feed intakes would match those of ad libitum-fed pigs housed in groups of either 4 (restricted-4) or 8 (restricted-8) pigs per pen in House 2. For all group sizes, feed intake increased linearly as the dietary lysine content increased. However, this increase was significantly lower for 8, when compared with 1 and 4 pigs per pen. The linear increase in feed conversion efficiency with dietary lysine content was similar for all group sizes. However, at any dietary lysine concentration, pigs housed in groups of 8 had significantly higher efficiencies than the pigs housed individually or in groups of 4. Average daily gain increased linearly as lysine intake increased, this increase being the same for all group sizes. However, pigs in smaller groups grew significantly faster than those in larger group sizes for any lysine intake. Protein and lysine retention were unaffected by group size, increasing linearly as lysine intake increased. The efficiency of lysine utilisation (0.45) was not impaired by group size. The pair-fed pigs housed individually (restricted-4 and -8) consumed significantly less feed than the individually-housed pigs fed ad libitum, and this was reflected in their average daily gains, which increased linearly as lysine intake increased, but with the restricted-8 growing significantly slower than the ad libitum or restricted-4 pigs. In all three treatments feed conversion efficiency increased linearly with dietary lysine content, although the restricted-4 and -8 had significantly higher efficiencies than the ad libitum-fed pigs at any dietary lysine content. Protein and lysine retentions were unaffected by feeding level and increased significantly with lysine intake. However, at any lysine intake the restricted-8 pigs had a significantly lower efficiency of lysine utilisation than the ad libitum or restricted-4 pigs. The pigs with floor-space allowances of 0.86 and 1.94 m2/pig consumed significantly less and grew slower than the pigs with floor-space allowances of 1.72 m2/pig at any dietary lysine content. Feed conversion efficiency was unaffected by floor-space allowance and increased significantly with dietary lysine content. Similarly, protein and lysine retentions were unaffected by floor-space allowance and increased linearly as lysine intake increased. The efficiency of lysine utilisation (0.45) remained unaffected by floorspace allowance. It was concluded that when animals are socially stressed, feeding according to the requirement for maximum protein growth produces the best biological performance and carcass composition, with the corollary that, if profitability and biological efficiency is to be maximised, pigs housed in stressful conditions, or those whose future performance is predicted to be below potential because of external stressors, should not be given feed of an inferior quality. The second experiment was designed to determine the extent to which grouping or floorspace allowance would alter the nutrient content of feed chosen by pigs given a choice of two feeds differing in protein: energy ratio between 40 to 85 kg live weight. Three hundred and eighteen entire male Large White x Landrace pigs were used. Two buildings provided the following group sizes and floor-space allowances: House 1 contained nine and eighteen pigs per pen at 1.72 or 0.86 m2/pig; House 2 contained four, nine and fourteen pigs per pen at 1.72; 0.86 or 0.49 m2/pig. Animals were given simultaneous ad libitum access to a high (236 g protein/kg as fed) and a low crude protein feed (115 g protein/kg as fed) in two hardened plastic self-feeder bins placed side-by-side. A training period of six days was used prior to the start of the trial, during which the two feeds were alternated daily. The reduction in the proportion of high protein feed chosen over time was significantly higher for the groups of four and eight, in comparison to the groups of nine and eighteen, contrasting with the steady increase for the groups of fourteen pigs. Similarly, the significant increase for pigs with floor-space allowances of 0.49 m2/pig differed from the significant decrease for pigs with floor-space allowances of 0.86 and 1.72m2/pig. Pigs housed in larger group sizes and smaller floor-space allowances consumed significantly less and grew slower than pigs housed in smaller group sizes and larger floorspace allowances. However, the feed conversion efficiency remained unaffected by group size and floor-space allowance. The non-significant effect on protein retention with increasing group size contrasted with the significant increase associated with increasing floor-space allowance. The results of the two studies were compared to determine whether pigs chose differently depending on the degree of stress and the implication of this choice. Average daily gain was significantly reduced as the group size increased for pigs fed a fixed lysine content and choice-fed. However, this reduction was less severe with choice-feeding than when feeding a fixed lysine content. Increasing the group size significantly reduced the feed intake in pigs fed a fixed lysine content only. The efficiency of protein utilisation remained unaffected as the group size increased for the pigs fed a fixed lysine content. However, at any group size pigs fed lower lysine contents had higher efficiencies than pigs fed higher lysine contents. On the contrary, increasing the group size significantly increased the efficiency of protein utilisation in choice-fed pigs. The average daily gain and feed intake was significantly improved as the floor-space allowance increased but was similar for pigs fed a fixed lysine content and choice-fed. Although the efficiency of protein utilisation remained unaffected by increasing the floor-space allowance for the pigs fed a fixed lysine content and pair-fed, at any floor-space allowance pigs fed higher lysine contents had higher efficiencies than pigs fed lower lysine contents. The results indicate that providing socially stressed pigs a choice between an appropriate pair of feeds differing in protein: energy ratio, does not overcome the reduction in potential growth, but does result in performance similar to that of pigs fed a fixed lysine content. It was concluded that the social stress of grouping or floor-space allowance has no influence on the ability of the animal to select an appropriate dietary combination allowing the expression of potential growth within the constraint(s) of the production system. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2005.