Thesis (MScAgric)--University of Stellenbosch, 2002. / ENGLISH ABSTRACT: Starch is the main component of cereal grains and is usually the primary energy source for pigs and poultry. Feed manufacturing can adopt several measures, including physical treatments such as milling or pelleting and other techniques, such as enzyme treatment, to disrupt cell structure. Grinding and pelleting are the most common food processing methods used for pigs. However, pelleting of complete balanced feeds is no longer such an economical proposition due to rising energy and equipment costs; therefore this cost has to be outweighed by an increased production efficiency. It has been known for many years that grinding is an essential prerequisite for the satisfactory blending of the ingredients of a multi-component food. Feed cost represents the major item in the cost of animal production. Without doubt, research and production efforts will continue to refine feed processing techniques to reduce the cost of feed and to increase the value of feed for a target animal. The possibilities for improvements in feed are endless; however the cost of each innovation must be carefully weighed against demonstrated improvements in animal performance. Experiment 1: The effect of feed processing techniques on weanling pig growth-performance The effects of processing of the carbohydrate source and the feed on growth performance of commercial Landrace x Large White piglets (n=480) weaned at 28 ± 2 d were investigated. Two processing combinations of the carbohydrate source were used with 3 processing conditions of the diet in a 2 x 3 factorial design. The pigs were blocked by weight (7.196 ± 2.03 kg BW) and then allotted randomly to 1 of 6 dietary treatments. Ten pens of 8 piglets each were fed with each dietary treatment. The two main processing conditions of the carbohydrate source were raw or extruded maize and the 3 processing conditions of the diet was meal or pelleted or extruded. No carbohydrate processing x diet processing interactions were observed (P > 0.05) for ADG, ADFI or FeR. In this experiment, extrusion of the maize led to an significant decrease in FeR efficiency (P < 0.05) (1.57 vs. 1.42) when compared to a raw maize diet. Pelleting a diet had no significant effect (P > 0.05) on ADG but tended to decrease ADFI (P < 0.07) and significantly improve FeR efficiency (1.49 vs. 1.66) when compared to a meal diet. Extruding the whole diet did not have any significant (P > 0.05) effect on ADG but tended to decrease ADFI (P < 0.07) and gave an significant improvement in FeR when compared to a meal diet (1.34 vs. 1.66). This processing technique also gave a significant improvement (P < 0.01) in FeR when compared to a pelleted diet (1.34 vs. 1.49). Experiment 2: The effect of pig feed processing conditions on pig metabolism parameters The effects of processing of the carbohydrate source and the diet on certain metabolism and production parameters of commercial Landrace x Large White pigs (n=24) were investigated. Two processing combinations of the carbohydrate source were used with 3 processing combinations in a 2 x 3 factorial design. Six diets were formulated on an iso-nutrient basis (14.48 MJ/kg metabolizable energy (ME), 23.01 crude protein (CP), 1.092% lysine, 0.742% methionine and cystine and 0.271% tryptophan on a DM basis). The pigs were blocked by weight (26.02 ± 0.25 kg BW) and then allotted randomly to I of 6 dietary treatments. The carbohydrate source was raw or extruded maize and the diets was meal or pelleted or extruded. No carbohydrate processing x diet feed form interactions were observed (P > 0.05) for dry matter intake (DMI), dry matter digestible energy (DE), Nitrogen (N) or dry matter intake (DMI). In a metabolism and nitrogen (N) balance study, apparent N digestibility, digestible energy and metabolizable energy contents were found not to be significantly (P > 0.05) influenced by carbohydrate or diet processing. / AFRIKAANSE OPSOMMING: Stysel is die hoof komponent in grane en is gewoonlik die primêre verskaffer van energie vir varke en pluimvee. Voermeulens kan verskeie metodes implementeer, insluitend fisiese behandeling soos bv. maal, verpilling en ensiem behandelings, om sel struktuur te verander. Maal en verpilling is die mees algemene prosessering metodes wat vir varkvoer gebruik word. Maar, verpilling van totaal geballanseerde voere is nie meer so ekonomies geregverdig nie as gevolg van stygende energie en masjinerie koste. Daarom moet die koste van voervervaardiging oorbrug word deur 'n verhoogde produksiedoeltreffendheid. Dit is al lankal bekend dat maal 'n voorvereiste is vir effektiewe vermenging van grondstowwe van 'n multi-komponent voer. Voerkoste verteenwoordig die hoof item van die koste van intesiewe diereproduksie. Voedingskudiges sal sonder twyfel voortgaan om voerprosessering te verfyn om so die koste van die rantsoen te verlaag en om die waarde van die rantsoen te verhoog vir die dier. Die moontlikhede is veelvuldig, maar die koste van elke ontwikkeling moet opgeweeg word teen verhoogde diereproduksie. Eksperiment 1: Die effek van voerprosesserings-tegnieke op speenvark groei en produksie 'n Proef is uitgevoer om die prosesseringseffek van 'n koolhidraat bron en voer op die groei van Landras x Groot Wit speenvarke (n=480) wat gespeen is op 28 ± 2 d te bepaal. Twee prosesserings-kombinasies van die koolhidraat bron en drie prosesserings-kondisies van die dieet is in 'n 2 x 3 faktoriaal ontwerp uitgevoer. Die varkies is geblok volgens massa (oorspronklik 7.196 kg ± 2 kg) en toe ewekansig in I van 6 dieët behandelings ingedeel. Die proefdiëte is vir lO hokke varkies gevoer, elke hok het 8 varkies in gehad. Die twee hoof prosesserings-kondisies van die koolhidraat bron was rou of gaar mielies en die drie prosesserings-kondisies van die dieët was meelof gaar of geëkstrueer. Geen koolhidraat prosesserings x dieët prosesserings interaksie van die dieët was opgemerk (P > 0.05) vir gemiddelde daaglikse toename (GOT), gemiddelde daaglikse voer inname (GDVI) of voer omset doeltreffendheid (VOD) nie. In die eksperiment was die VOD van die geëkstrueerde mielie dieët, statisties betekenisvol laer (P < 0.05) (1.57 vs.1.42) as die rou mielie dieët. Verpilling van die dieët het geen statisties betekenisvolle effek (P > 0.05) op GDV! gehad nie, maar die VOD was statisties betekenisvol (P < 0.05) beter (1.49 vs. 1.66) wanneer dit met die meel dieët vergelyk word. Ekstrusie van die dieët het geen statisties betekenisvolle effek op GOT en GOVI gehad nie, maar die VOD was statisties betekenisvol (P < 0.05) beter wanneer dit met die meel (1.34 vs. 1.66) en verpilde dieët (1.34 vs 1.49) vergelyk word. Eksperiment 2: Die effek van voer prosessering op vark metabolisme parameters 'n Eksperiment is uitgevoer op Landras x Groot Wit bere (n=24) om die effek van prosessering van 'n koolhidraat bron en voer te bepaal. Twee prosesserings-kombinasies van die koolhidraat bron en drie prosesserings-kondisies van die dieët is in 'n 2 x 3 faktoriaal ontwerp uitgevoer. Ses diëte is geformuleer op 'n iso-nutriënt basis (14.48 MJ/kg metaboliseerbare energie (ME), 23.01 ru-proteïen (RP), 1.092% lisien, 0.742% metionien and sistien en 0.271 % tryptofaan op 'n droeë materiaal (DM) basis). Die varke is geblok volgens massa en ewekansig aan 1 van 6 diëte toegeken. Die koolhidraat bron was gaar of rou mielies en die diëte was meel, verpil of geëkstrueer. Geen koolhidraat-prosessering x dieet interaksie is opgemerk (P > 0.05) nie. In 'n metabolisme en stikstof (N) balans-studie is daar gevind dat DM verteerbaarheid, skynbare N verteerbaarheid, verteerbare energie (GE) en ME inhoud nie beduidend deur die behandelings beïnvloed word nie.

A series of experiments was conducted to evaluate the value of jojoba meals in chick, rabbit and laying hen diets. Treated meals used in these studies had been subjected to solvent, chemical and microbial treatments for the removal of simmondsin toxicants. Feeding 5% untreated jojoba meal diets to either chicks, rabbits or laying hens resulted in a general poor performance as evidenced by reduced feed intake, poor growth and poor feed conversion; in addition, low egg production and egg output were obtained with laying hens. Incorporation of the untreated meal (containing 4.7% simmondsin compounds) at 10% of the diet caused 55.6% mortality in chicks. A pair-feeding study with chicks revealed that reduced feed intake rather than toxic effects was responsible for decreased growth and mortality. A similar study with rabbits showed that both high levels of simmondsin toxicants and palatability were responsible for poor growth. Of the seven meals tested with chicks, only the Lactobacillus acidophilus #1911 treated meal supported adequate growth when added to the diet at a level of 10%. However, feed conversion was poor at this level of suplementation. Lowering the simmondsin toxicants to .34% of the meal allowed adequate growth in weanling rabbits when incorporated in the diet at levels up to 15%. Slightly lower amino acid digestibilities were obtained for diets containing this meal. Studies of amino acid availabilities with the jojoba meals indicated a general low availability for the essential amino acids (33.7% for lysine and 43.4% for methionine). Starch digestibility was also low (22%) and true metabolizable energy varied from 1.71 to 1.90 kcal/g in laying hens. Jojoba meal also contains a trypsin inhibitor (TI) with activities reaching 5.48 TIU/g which was inactivated in 30 min at 15% moisture and 100 C. Tannic acid levels in the jojoba meals varied from 1.50 to 2.57%. Poor amino acid availabilities were still obtained after TI inactivation. The use of jojoba meal in the feeding of poultry or rabbits will require the removal of the simmondsin toxicants, denaturation of the TI as well as consideration of the poor amino acid availability and low energy content.

Rangeland resources play a significant role in household production and sustainability of livelihoods among pastoral communities in Kenya. Although wildlife is one of the rangeland resources, it is viewed by pastoralists as a competitor with livestock for grazing resources rather than an economic resource. It is assumed that competition between wild herbivores and cattle may have an impact on the forage biomass in rangelands as well as on livestock production. It is from this view point of competition between wildlife and livestock for forage resources, that this study assessed effects of forage utilization by wildlife on cattle diet, plant community composition, forage biomass and level of forage utilization in semi-arid lands in Kenya. The following hypotheses were tested: there is a decline in proportion of dominant grasses due to wildlife grazing; there is a decline in forage biomass due to grazing by wildlife and there are changes in the diet of cattle (Bos indicus Lichtenstein) due to grazing by wild herbivores. Grazing experiments were conducted at the Kenya Long-term Exclosure Experiment (KLEE) on Mpala Ranch, Laikipia District, Kenya. A number of techniques were used during data collection: line transects and 1m2 quadrats to assess plant species composition, richness and diversity. Forage biomass and increment in forage weight in grazed and ungrazed exclosures were determined by use of a disc pasture meter, whereas plant species consumed by cattle and wild herbivores were assessed by observation during feeding. The dietary forage composition of herbivores was done by micro-histological analysis of faecal samples of cattle, zebra, oryx, hartebeest and Grant’s gazelle. The results indicate that there was a high (>21 %) proportion of the tall coarse grasses (Pennisetum stramineum and P. mezianum) in the exclosures grazed by cattle with wildlife in wet and dry seasons, whereas 21 % in the exclosures grazed by wild herbivores. The proportions of Themeda triandra in the exclosures grazed by cattle with wildlife in the dry season was 18 %, whereas it was more than 25 % in the exclosures grazed by cattle with wildlife in the wet season. The results also indicate that there were very highly significant (p<0.0032 and p<0.0015) differences in percentage composition of dominant and less dominant grasses between the grazed and ungrazed exclosures during the dry seasons, whereas a significant (p<0.05) difference and a highly significant (p<0.01) difference in percentage composition between the grazed and ungrazed exclosures during the wet seasons. 5 – 6 % of the total herbaceous forage biomass (0.7 % dry matter intake) was consumed in the exclosures grazed by wildlife, whereas 13 – 17 % (2.8 % dry matter intake) was consumed in the exclosures grazed by cattle. There was a large decrease of forage biomass in the pasture grazed by cattle. However, there was no significant (p<0.133) difference in forage biomass in exclosures grazed by large wildlife or grazed by elephants (mega-wildlife). There was less than 12 % utilization of dominant grass species in the exclosures grazed by wildlife, whereas over 40 % utilization of dominant grass species in the exclosures grazed by cattle. The results indicate that there is no evidence that grazing by wild herbivores decreases forage biomass in the pasture. Wildlife, therefore, should not be hunted out on communal grazing lands because it has no significant impact on the available forage biomass for livestocks. Nonetheless, stocking rates of livestock should be consistent with forage production so that wildlife conservation is integrated in pastoral production systems.

Digitized by Kansas Correctional Industries

Bibliography: leaves 175-196. The influence of starch in grain-based diets on fat development in cattle and sheep is investigated.

The main objective of this research was to assess the nutritional status of sheep and goats grazing two rangeland types in Baluchistan, Pakistan. These types were Artemisia maritima/Haloxylon griffithii association represented by the Zarchi Field Station and Cymbopogon shoenanthus, found at the Tomagh Field Station. Seasonal above ground biomass production and variability in the quality of major plant species was determined at four phenological growth stages for two years. At the same time a relative palatability index of major plant species and preference of sheep and goats was also determined. The Analysis of Variance procedures (ANOVA) of Statistical Analysis System (SAS, 1985) were used to evaluate various null hypotheses. Forage quality decreased with the advancement of phenological growth stages in all plants studied. On both study sites (Tomagh and Zarchi), spring forage samples were of higher quality than those of other seasons. Advance in season brought significant changes in the nutrient content of plant species. Grasses were lower in crude protein concentration and higher in neutral detergent fiber (NDF) content than forbs and shrubs. Plant species were not statistically different in phosphorus concentration. From March through October, sheep and goat diets varied in their botanical composition on both study sites. Grasses remained a major component of animal diets at Tomagh throughout the grazing seasons. However, sheep and goats consumed a higher percentage of shrubs with the passage of time while grazing Zarchi area. Across all the grazing seasons, the diets of both animal species were deficient in protein and phosphorus. The lignin ratio technique was evaluated for its use in determining daily dry matter and apparent nutrient digestibility consumed by grazing animals. Highly variable digestion co-efficients of lignin (both positive and negative) were obtained by using sheep and goat rumen liquors. As much as 51.0% of forage lignin in samples was digested. Negative lignin digestion co-efficients were also obtained from forage samples. These ranged from -1.0% to - 103.0%. Lignin biodegradation and/or complexing during in vitro digestion invalidated its use as an internal marker in digestion studies on these rangelands. / Graduation date: 1991

Graduation date: 1964

Many studies have been conducted on the effect of size-grading in other fish species. However, there is a paucity of scientific information on the effects of size variation on cannibalism of juvenile dusky kob. Thus, a study focusing on the effect of size variation on juvenile dusky kob aggressive and feeding (browsing) behaviours was conducted. Three separate groups of hatchery-reared juvenile dusky kob of were obtained from Oceanwise (Pty) Ltd for use in a series of three experimental trials. In all trials, juvenile fish of the same age were size-graded and the COV-value was used to determine the size variation. The focal fish (largest fish) was exposed to groups of fish with different size variation for 30-min. behavioural observations before and after feeding in randomised trials. The first experimental trial (Chapter 2) quantified the effect of increasing size variation and observation time on the aggressive and browsing behaviours of juvenile dusky kob. On average, juvenile dusky kob weighed 3.60 ± 0.68 g fish-1 and measured 5.8 ± 0.41 mm. Each focal fish was observed (a) before feeding in the morning, (b) 2 h after feeding, (c) 6 h after feeding and (d) 12 h after feeding. Fish increased browsing behaviours (averaging 6.60 ± 0.56) and decreased intimidating aggressive behaviours (18.60 ± 1.39) 12 h after feeding. Other aggressive behaviours occurred but did not differ between observation times. Aggressive and browsing behaviours positively correlated with size variation variables, predominantly, 12 h after feeding. An average frequency of 19 chases were observed positively correlated with size variation, followed by average frequencies of 17 body bites and browses, and 11 tail bites per 30 min. Some behaviours including average frequencies of 0.2 chases, 4 tail bites, 2.4 intimidating and 0.3 browsing behaviours negatively correlated with the size variation, generally closer to the last meal. These preliminary observations thus showed that fish have become hungry approximately 6-12 h after feeding and substituted certain behaviours for others as time after feeding passed and as size variation increased. The second experimental trial (Chapter 3) consisted of the observations further testing the relationship between aggressive behaviours and size variation of juveniles of dusky kob averaging 0.43 ± 0.27 g fish-1. The focal fish was exposed to groups of fish of four size variation (COV) treatments for observations before and 12 h after feeding. Aggressive behaviours positively correlated with size variation both before and 12 h after feeding. An average frequency of 437 body bites positively correlated more often with COV, followed by average frequencies of 365 intimidating behaviours and 199 tail bites per 30 min., respectively, before and 12 h after feeding. The least often exhibited aggressive behaviours averaged 26 chases while positively correlating with size variation on the times specified. An average frequency of 311 intimidating behaviours before and after feeding including average of 28 tail bites after feeding negatively correlated with size variation. This may relate to shift of behaviours depending on the needs and capacity of the fish. Apparently, aggressive fish can change its behaviour as a function of COV-values rather than the mean size of the other fish. The third trial (Chapter 4) investigated the effect of the aggressor’s (focal fish) size in relation its aggressive and browsing behaviours to other fish (non-focal fish). Juvenile fish used for this trial, on average, weighed 30 ± 7.63 g fish-1. Aggressive and browsing behaviours were observed in four treatments of a) high COV and mean weight below, b) low COV and mean weight less, c) high COV and mean weight equivalent to and d) low COV and mean weight higher than that of the focal fish. Increased frequencies of aggressive and browsing behaviours per 30 min. occurred in treatment A, sharing similar frequencies in treatment C, compared to the other treatments (B and D) which shared certain frequencies. The intimidating behaviours predominated, followed by browsing, body bites, chases and tail bites, respectively. The results of the overall study suggest that the time passed after feeding and increasing size variation and differences facilitated aggressive and browsing behaviours in juvenile dusky kob. Dusky kob showed increasing aggressive behaviours as early as in the first two weeks after hatching, averaging 0.43 ± 0.27 g fish-1 with the frequency correlating with size differences. Consistent size-grading technique in the same-age fish should be used to manage size variation associated with aggressive behaviours. The period of about 4-6 h after feeding may explain the noticeable increased aggressive and browsing behaviours. Thus, fish should be fed immediately before or after evacuation of their guts to maintain less-aggressive behaviours of juvenile dusky kob. Fish generally increased aggressive and browsing acts before and long time after feeding than closer to after feeding. The study has provided the fundamental scientific groundwork for fish farmers and future researchers can further explore size variation, time after feeding and gut evacuation rate as critical components of aggressive behaviours. The scientific knowledge of aggressive and cannibalistic behaviours has essential application in farming management to achieve improved survival and growth rates in juvenile fish.

The aim of the first study was to determine if barley could replace maize as an energy source in concentrates fed to dairy cows grazing on kikuyu/ryegrass pasture without affecting the milk production, milk composition, or cause metabolic disorders. Sixty Jersey cows, in early to mid lactation were randomly allocated to one of five treatments (n = 12) based on feeding concentrates with different ratios of maize to barley, ranging from 100:0, 75:25, 50:50, 25:75 and 0:100, respectively. Concentrates contained 12 MJ ME kg -1 and 130g CP kg-1 DM and cows were fed 3 kg (as-is) concentrate after each milking for a period of 42 days (14-day adaptation and 28-day measurement). Cows strip-grazed the irrigated kikuyu/ryegrass pastures (15.7 ± 1.8 percent DM; 20.2 ± 4.3 percent CP; 44.7 ± 3.5 percent NDF). Milk weights were recorded daily and milk samples collected weekly and analyzed for milk fat and protein content. Body condition score and live weight were recorded at the start and end of the experimental period. Data of all the studies were subjected to a one-way ANOVA. Daily milk yield, FCM, MUN, milk fat yield, milk fat percent, protein yield, protein percent, live weight change, or body condition score change were not affected by treatment and values were 15.8 kg, 17.2 kg, 14.9 mg dl-1, 0.72 kg, 4.56 percent, 0.59 kg, 3.77 percent, 6.67 kg, and 0.15 BCS; 15.6 kg, 17.4 kg, 15.2 mg dl-1, 0.73 kg, 4.3 percent, 0.57 kg, 3.71 percent, 1.33 kg and 0.04 BCS; 17.2 kg, 17.9 kg, 15.2 mg dl-1, 0.74 kg, 4.36 percent, 0.63 kg, 3.71 percent, 0.33 kg and 0.08 BCS; 15.6 kg, 16.4 kg, 15.5 mg dl-1, 0.67 kg, 4.33 percent, 0.60 kg, 3.83 percent, -1.46 kg and 0.11 BCS; and 15.0 kg, 16.0 kg, 15.5 mg dl-1, 0.67 kg, 4.57 percent, 0.57 kg, 3.85 percent, 8.86 kg, and 0.05 BCS, respectively for the cows fed 100:0, 75:25, 50:50, 25:75 and 0:100 maize to barley ratio concentrate. According to these results, barley can replace maize without significantly affecting the milk production or milk composition. None of the cows presented any visible symptoms of acidosis. As it was clear from the results in the first study that maize could replace barley the aim of study 2A was therefore focused on determining the effect of feeding different levels of such a barley-based (2.4, 4.8 or 7.2 kg cow-1day-1) concentrate, on milk production, milk composition and live weight change of Jersey cows on kikuyu/ryegrass pasture (23.1 ± 2.95 percent DM, 11.1 ± 0.11 percent CP, 60.8 ± 0.58 percent NDF). Forty-five Jersey cows (early- to mid lactation), were randomly allocated to one of three treatments (n = 15) involving different levels of concentrate (12 MJ ME and 130g CP kg-1 DM) feeding for a period of 42 days (14-day adaptation and 28-day measurement). Milk weights were also recorded daily, and milk samples collected weekly, and analyzed for milk fat and protein. Body condition score and live weight were recorded at the start and end of the experimental period. The results of this study indicated that increasing the concentrate level from 2.4 to 4.8 and 7.2kg cow-1day-1 did not increase the milk yield (14.0 kg, 15.2 kg, 14.4 kg; P = 0.19). The FCM production increased from 15.8 to 17.5kg (P = 0.04) as the concentrate level increased from 2.4 to 4.8kg cow-1day-1. Increasing the concentrate from 4.8 to 7.2kg cow-1day-1 did not result in a significant increase in FCM. The milk protein percent increased significantly from 3.4 - 3.6 percent when the concentrate feeding level was increased from 2.4 to 7.2kg cow-1day-1. The MUN levels were 17.09 mg dl-1, 16.03 5 mg dl-1, and 16.36 mg dl-1 for the 2.4, 4.8 and 7.2kg cow-1day-1 concentrate levels, respectively. This is well within the recommended MUN levels (12 – 18 mg dl-1) indicating that sufficient protein was fed to cows. Increasing the concentrate level from 4.8 to 7.2 kg cow-1day-1 did not increase production, probably due to a higher pasture substitution rate. Supplementing large quantities of rapidly fermentable grains, such as barley, can suppress rumen pH and may have a negative effect on the rate and extent of fibre digestion in the rumen. Therefore the aim of study 2B was not only to determine the effect of feeding different levels of a barley-based concentrate, on milk production, milk composition and live weight change, but was also to determine the effect of a low (4.8 kg cow-1day-1) versus a high (7.2 kg cow-1day-1) level of barley-based concentrate supplementation on ruminal DM and NDF degradability of Westerworld ryegrass sampled from the pastures that these cows were grazing on. Sixty Jersey cows (early- to mid lactation), were randomly allocated to one of three treatments (n = 20) involving different levels of concentrate feeding. Concentrate (12 MJ ME, 130g CP kg-1 DM) was fed at 2.4, 4.8 or 7.2 kg cow-1day-1 for a period of 74 days (14-day adaptation and 60-day measurement). These cows stripgrazed irrigated kikuyu/ryegrass pastures (14.7 ± 4.37 percent DM, 25.1 ± 1.53 percent CP, and 44.4 ± 2.58 percent NDF) at a daily pasture allocation of 10 kg DM cow-1. Milk weights were recorded daily and milk samples collected weekly and analyzed for milk fat and protein. Body condition score and live weight were recorded at the start and end of the experimental period. Twelve Jersey cows, fitted with ruminal cannulae, were randomly allocated to two of the three treatments in the production study and received either 2.4 or 7.2 kg cow-1 day-1, of the same barley-based concentrate, in a two-period crossover design. These cows strip-grazed the same irrigated kikuyu/ryegrass pastures as the sixty cows in the production study. Each period consisted of 21 days for adaptation and seven days for data collection. Rumen liquor samples were collected every 4 hours within a 24-hour cycle and repeated once. Rumen pH was measured immediately, recorded, and the supernatant fluid preserved and frozen, pending VFA analysis. The in situ nylon bag technique was used to determine DM and NDF degradation and dried samples of Westerworld ryegrass pasture were incubated for 0, 4, 8, 12, 20, 30, 48, 72 and 96 hours. The data were fitted in the non-linear model p = a + b (1-exp-ct) (Ørskov & McDonald, 1979). Daily milk production, fat corrected milk, milk fat yield and milk fat percent were not affected by treatment and values were 17.3 kg, 18.4 kg, 0.76 kg and 4.42 percent; 19.0 kg, 20.0 kg, 0.82 kg and 4.35 percent; and 18.1 kg, 19.1 kg, 0.79 kg and 4.37 percent for the 2.4, 4.8 and 7.2 kg cow-1 day-1 concentrate treatments, respectively. Milk protein percentage of cows on the 7.2 kg concentrate cow-1 day-1 was significantly higher than that of cows on 4.2 kg concentrate cow-1 day-1 feeding level. Live weight increased significantly as the level of concentrate feeding increased and values were 17.9 kg; and 28.9 kg on the 2.4 and 7.2kg concentrate treatment, respectively. There was a significant increase in the live weight of cows that were fed 7.2 kg cow-1 day-1 (as-is) in comparison to those cows that were fed 2.4 kg concentrate cow-1 day-1 (as-is). This may have resulted from more nutrients being partitioned to live weight gain rather than milk production. No further response in milk production was observed when concentrate daily feeding was increased from 4.8 to 7.2 kg cow-1 day-1. It is postulated that the higher concentrate allowance resulted in a higher substitution rate and lower DMI intake from pasture. 6 There was no significant decline in the rumen pH (6.2 ± 0.4 and 6.2 ± 0.5) when the concentrate level was increased from 2.4 to 7.2 kg cow-1 day-1 (as-is). The total VFA (118.1 ± 45.9 and 139.4 ± 45.6 mmol L-1) and isovalerate (0.009 ± 0.07 and 0.248 ± 0.52 mmol L-1) increased significantly when the concentrate was increased from 2.4 to 7.2 kg cow-1day-1. No other rumen parameters were affected by treatment. Ruminal DM and NDF degradability of the Kikuyu/ryegrass pature were not affected by the level of concentrate supplementation. An increase in the concentrate level from 2.4 to 7.2 kg cow-1day-1 did not reduce degradability of either DM (94.67 ± 5.97, 94.49 ± 5.09; P = 0.919) or NDF (92.15 ± 8.69, 94.4 ± 11.73; P = 0.451), respectively. Results of rumen parameters and PD values were within the range reported by Bargo et al., (2003), viz. pH 5.76 – 6.29, NH3-N concentration 8.7 – 32.2 mg dl-1, total VFA concentration 90.3 - 151.4 mmol L-1 and PD values 89.5 – 93.5 % reported by Bargo et al. (2003). According to these authors, there is no simple relationship between any amount of the concentrate supplemented, and the ruminal pH and concentrate feeding only affects the in situ ruminal digestion of pasture when it is fed, at quantities higher than 8 kg DM cow-1day-1 (Bargo et al., 2003).

I investigated the effect of turbidity on the foraging behaviour of juvenile Chinook salmon (Oncorhynchus tshawytscha) in the laboratory. Specifically, I examined a behavioural "tradeoff" between visual ability and "perceived" risk. I assessed visual ability by measuring the reaction distance of juvenile Chinook to planktonic Artemia prey. I found Chinook exhibited a log-linear decline in reaction distance with increasing turbidity. To determine the effects of turbidity and microhabitat on foraging rate, I conducted separate experiments for surface (Drosophila), planktonic (Artemia), and benthic (Tubifex) prey across a range of turbidity levels (0, 25, 50, 100, 200, 400, 800 mg•L⁻¹). Foraging rates were reduced at higher turbidity conditions for all three prey. However, for surface and benthic prey, foraging rates were also low in clear water; highest rates were attained at intermediate turbidity levels (50-200 mg•L⁻¹). The degree to which intermediate turbidities were associated with higher foraging rates was size-dependent. Smaller individuals (150-57 mm FL) exhibited relatively higher foraging rates in clear conditions than did larger individuals. However, planktonic foraging rates by juveniles were consistently high in clear water, regardless of fish size. In experiments manipulating light level independent of turbidity, I allowed salmon to forage under conditions which were either turbid, or clear but with light intensity correspondingly reduced. Foraging rates were similar between the two treatments for planktonic prey, but differed for benthic and surface prey. Generally, foraging rates exhibited by juvenile Chinook salmon could not be explained on the basis of visual ability alone. I suggest that young salmon also exhibited foraging behaviour consistent with their perception of risk to predation. In arena experiments, juvenile Chinook distributed themselves randomly in turbid conditions; in clear conditions they associated with the bottom. When bird and fish predator models were introduced the fish altered their spatial distribution, occupying deeper regions regardless of turbidity. However, their response in turbid conditions was less marked and lasted for a shorter time. Turbidity apparently mitigated the perceived risk of predation in juvenile Chinook. I developed a conceptual tradeoff model that predicted the general effect of turbidity on foraging behaviour. Assuming differences in either prey quality or perceived risk of predation in three microhabitats (surface, water column, bottom), the model resolved the apparent dissimilarities between planktonic and other foraging behaviours. Perceived risk of Chinook to predation was significantly different between surface and water column microhabitats. When exposed to a non-visual "fixed-risk" stimulus (sound), salmon apparently perceived less risk as turbidity increased. I conclude that in turbid waters juvenile salmon exhibit foraging behaviour in a manner consistent with a tradeoff between their visual ability and perceived risk. / Science, Faculty of / Zoology, Department of / Graduate