Search

The decreasing supply and high cost of fishmeal, as well as the sustainability issues of this finite raw material, will continue to push the industry to concentrate their efforts on finding alternative sources of protein to substitute fishmeal in aquafeeds. From all the possible alternatives, e.g. animal by-products, fishery by-products, single-cell protein and the recent insect meal options, plant-based meals seem to be one of the major contributors to reduce reliance upon marine sources. However, when considering plant-based meals for aquafeeds it is commonly agreed that one of the negative aspects is the presence of anti-nutritional factors (ANF’s; e.g. cyanogens, saponins, tannins, etc.) that are detrimental to fish and shrimp (Krogdahl et al., 2010). Although there are processes to remove or inactivate many of these ANF’s, the same does not apply for mycotoxins which are reasonably stable to processing conditions (Cheli et al., 2013). The effects of mycotoxins in fish and shrimp are diverse, varying from immunosuppression to death, depending on toxin-related (type of mycotoxin consumed, level and duration of intake), animal-related (animal species, sex, age, general health, immune status, nutritional standing) and environmental-related (farm management, biosecurity, hygiene, temperature) factors. Therefore, it is often difficult to trace observed problems back to mycotoxins. The main goal of this thesis is to increase the awareness of mycotoxin contamination in aquafeeds and its consequences to aquaculture species, especially characterizing the impact of deoxynivalenol in rainbow trout (Onchorhynchus mykiss). In Chapter 1 the known mycotoxin occurrence and co-occurrence in aquaculture finished feeds are described and critically reviewed by correlating the extent of mycotoxin contamination in aquaculture feeds to its impact in aquaculture species. This chapter also contributes to understanding the risk of mycotoxin carry-over to aquaculture seafood products. Additionally, this chapter aims to expose the scientific community, the regulatory authorities and the aquaculture industry, to the main challenges and myths that the industry faces in developing mycotoxin management strategies. Chapter 2 presents the results the impact of deoxynivalenol (DON) in rainbow trout and the difficulties to diagnose DON ingestion. Here was explored two different DON contamination scenarios, i.e., the effect of short-term feeding of high levels of DON and the effects of long-term feeding of low levels of DON (a more common scenario in aquaculture industry). It was concluded that the ingestion of DON by trout over short-term periods at high dosages (50 days; 1,166 μg kg -1 and 2,745 μg kg -1) impacts growth performance, especially feed intake, with minor or variable biochemical changes in trout blood and histopathological changes. In this case, some fish did exhibit clinical symptoms (i.e., anal papilla), which could be attributed to the DON treatment (reported for the very first time). This chapter also reports, for the first time, the effects of the long-term exposure of rainbow trout to low concentrations of DON (168 days; 367 μg kg -1 DON). Ingestion of DON in the long-term study was asymptomatic; however, the fish started to reduce their growth performance 92 days after ingesting DON. Such low contamination levels, which might be unnoticed by farmers, may have economic consequences for aquaculture. Here it was concluded that the short-term effect of DON ingestion cannot be extrapolated to other contamination scenarios, e.g., long-term exposure. Chapter 3 aimed to elucidate the impact of DON on rainbow trout and study the reasons behind the apparent lack and/or high variability of clinical signs during DON ingestion (as reported in chapter 2). With this experiment, we further confirmed that ingestion of DON by rainbow trout is mainly characterised by impaired growth performance and reduced feed intake, with the total absence of any visible clinical signs (up to 11,412 ± 1,141 μg kg -1). Proteolytic enzyme activities (pepsin, trypsin and chymotrypsin) in trout were altered by DON ingestion. The impact of DON on the abundance of specific measured mRNA transcripts was unexpected with higher expression levels for insulin-like growth factors, igf1 and igf2, which are directly related to elevated insulin levels in plasma. This can also in part be influenced by the trypsin activity and by npy, given its higher mRNA expression levels. The apparent digestibility of dry matter, protein and energy was not affected by dietary levels of DON, however, nutrient retention, protein, fat and energy retention were significantly affected in animals fed DON. Adenylate cyclase-activating polypeptide (PACAP) expression seems to play an important role in controlling feed intake in DON fed trout. In the present study, we have shown for the first time that DON is metabolized to DON-3-sulfate in trout. DON-3-sulfate is much less toxic than DON, which helps to explain the lack of clinical signs in fish fed DON. Being a novel metabolite identified in trout makes it a potential biomarker of DON exposure. It was suggested that the suppression of appetite due to DON contamination in feeds might be a defence mechanism in orderto decrease the exposure of the animal to DON, therefore reducing the potential negative impacts of DON. In Chapter 4, this thesis further explored the current knowledge of DON toxicokinetics and rainbow trout DON metabolism, accessing the organ assimilation rates, excretion and possible biotransformation kinetics. Here we found that an hour after tube-feeding, [3H]-DON was detected in the liver samples of fish, indicating a rapid absorption of DON. In the first hour, [3H]-DON was present in the GIT (20.56 ± 8.30 ng). However, 6.19 ± 0.83 ng was also detected in the water at this sampling point. The fast excretion of [3H]-DON (faster than the average trout gastric emptying time) suggests that DON, as a water-soluble compound, is excreted with the liquid phase of the chyme. The presence of [3H]-DON in the GIT was stable during the first six hours. This long residence time of DON in the GIT may compromise the health of the GIT and favour absorption. Our data suggest that an effective DON detoxifying method should have a period of action of ≤ 6 h. Furthermore, as most of the excretion can be expected to happen after six hours, the detoxification should be irreversible at GIT conditions. Results of this PhD study contributes to better understand the importance and the risk of mycotoxin occurrence and co-occurrence in aquaculture finished feeds, highlighting concerns regarding the unvalued risk of mycotoxin carry-over to aquaculture seafood products. Here is highlighted that long-term exposure to DON is normally asymptomatic which might be unnoticed by farmers, however, representing economic consequences for aquaculture (reduced growth performance). Suppression of appetite due to DON contamination in feeds seems to be a defence mechanism in order to decrease the exposure of the animal to DON, therefore reducing the potential negative impacts of DON. The biotransformation of DON to DON-3-sulfate helps to explain the lack of clinical signs in fish fed DON and may be used as a novel biomarker of DON exposure.

The development of a cost-effective off-the-shelf micro-particulate diet with a nutritionally optimal formulation for larval North African catfish, Clarias gariepinus (Burchell, 1822) is one of the most important requirements for the aquaculture sector in Uganda. North African catfish contribute ~50% of Uganda's production of farmed fish, both in terms of tonnage and economic value, however, studies conducted thus far to develop a micro-diet have taken no account of the digestive physiology of the fish itself, but rather have dwelt more on the composition of the diet and its apparent performance in terms of fish growth. Under these conditions, however, large mortalities are experienced during early development. Only a few studies have been conducted concerning the digestive physiology of the larvae or have characterised aspects of the key digestive enzymes produced by this species. Hence a major mismatch has existed between the general approach to creating micro-diets and the level of knowledge of the functional development of the sensory and digestive system in larval fish, including their theoretical digestive capabilities. This study sought to characterise the ontogeny of key digestive enzymes in North African catfish larvae from hatching to 21 days post-hatch (dph) with special emphasis on the period 2-4 dph where live feed is the more accepted diet for the fish larvae. In real-world terms, however, extensive use of live-feed makes farming extremely labour-intensive and it is costly to operate a catfish hatchery providing a consistent product and at the same time generate profit for farmers. The work described in this thesis was conducted at the University of Stirling's Tropical Aquarium, a facility of the Institute of Aquaculture, which houses a population of North African catfish introduced into the facility over 30 years ago and assumed to be, but not fully verified as, Clarias gariepinus Burchell 1822. In order to ensure that the results of the research presented in the thesis could be effectively applied in Uganda, the identity of this species needed to be confirmed. From a morphological and meristic characterisation, in addition to use of a molecular taxonomic classification, the species held at the Institute of Aquaculture and employed for the present study was confirmed to be Clarias gariepinus, (Burchell, 1822) rather than a hybrid with, Clarias anguillaris which is a situation that is common in African populations. Studies were undertaken to determine the point at which when exogenous feeding should be commenced and to establish when fish could be weaned on to an inert formulated diet in order to improve growth and survival during larval rearing. From the growth studies conducted, findings revealed that at 2Dph a mixed diet of (live) Artemia and (inert) Coppens diet gave better growth than both Coppens and Artemia fed alone, which gave similar growth rates. The histomorphology of both adult and larval C. gariepinus was examined in order to determine the course of development of digestive structures in the larva. In addition, new in situ labelling procedures were developed to highlight the localisation of expression of transcripts for key digestive enzymes comprising two alkaline proteolytic enzymes, namely chymotrypsin and trypsin, one acidic proteolytic enzyme, pepsin, and one chitinolytic enzyme, gastric chitinase. Larval development followed a time-course similar to that described by other researchers with the pancreas being discrete in the larval stages but being more dispersed throughout the intestinal region in the adults. The stomach was present by 4 to 5 Dph, supporting a suggestion that alkaline protease digestion dominates early in development, switching to a more acidic digestive mode following development of the stomach. In situ labelling protocols showed trypsin and chymotrypsin to be localised in the pancreas by as early as 2 Dph, while pepsin and gastric chitinase were localised in the stomach. More specifically, pepsin was localised to oxynticopeptic cells and gastric chitinase to the gastric pits with the former being present by 4Dph and the latter being present by 6Dph. Quantitative real-time PCR (qRT-PCR) was employed to investigate the timing and level of transcript expression for trypsin, chymotrypsin, pepsin and gastric chitinase. First expression was observed, perhaps unsurprisingly, to occur largely at a time when the tissue representing the site of production was first observable by histological examination and expression increased from this time. Trypsin and chymotrypsin expression were detectable from 1Dph, with clear expression by 2Dph, coinciding with the appearance of the first anlage of the pancreas in histology. Pepsin and gastric chitinase mRNA was detected at 3Dph and was highly expressed from 4dph onwards. This coincided with the appearance of the first anlage of the stomach in histology. Important new resources for North African catfish, including the first assembled and annotated transcriptome are now available for this species. These were created from a mixed tissue / stage transcriptome library. Using this transcriptome resource as a basis, an oligo-microarray was constructed, which allowed broad-scale transcriptome analysis to be conducted on samples from early larvae fed using two diet regimes. During development on the standard diet, a number of pathways were observed to be over-represented in terms of transcript expression between 1 Dph and 4 Dph. These included protein digestion and absorption, bile secretion, complement and coagulation cascades, glycolysis, gluconeogenesis, PPAR signalling pathways, and ribosomal pathways. Together these pathways denoted changes in elements of growth, protein production, digestion, the immune system and synaptic transmission inter alia. From the research presented in this thesis, it is clear that the key digestive enzymes studied here are rapidly switched on during the first few days of development and first feeding. Different dietary regimes had important effects on expression of digestive enzymes and knock-on effects for digestion, growth, and immune system development, which may have important consequences for larval health and survival under aquaculture conditions. The results of this work have the potential to assist the Ugandan feed industry in that they can be used as a baseline to make improved diets for early weaning in larval fish, using increased knowledge of which enzymes are involved in larval digestion and when they are produced by the fish. Clues about the course of development of larval immunity, particularly innate immunity, may help to provide strategies to mitigate the high losses experienced by farmers in early development and may assist in the management of high intensity catfish hatcheries. The findings of the work presented in this thesis can help to better formulate inert feeds and optimise timing of introduction to larval stages to maximise growth and survival. Public transcriptomic resources produced in the course of this work can be used by other researchers as a basis for larger broad-scale transcriptomic investigations using microarray, RNAseq or allied technologies to characterise different states including disease, nutrition, physiology and development which could be used to e.g. improve diets, develop vaccines, or increase production values such as growth characteristics.

With increasing world population, the demand for fish is growing thus there is a need to identify products with potential to increase the efficiency of fish production. Phytogenics are among the products being investigated as potential naturally derived growth promoters. The aim of this study was to identify phytogenic compounds and doses with growth-promoting effects in Nile tilapia and investigate relevant pathways underlying their growth promotion effects. The phytogenic compounds limonene, carvacrol and thymol, major constituents of essential oils from the plants citrus, oregano and thyme, respectively, were evaluated. Six Trials (Trials I, II, III, IV, V and VI) were carried out using diets supplemented with varying concentrations of the phytogenic compounds. In Trials I, II and III (Chapter 3), the effects of either limonene (Trial I), carvacrol (Trial II) or thymol (Trial III) on growth performance of Nile tilapia were investigated (objective 1) and performance parameters including final fish weight, daily growth coefficient, growth rate per metabolic body weight, percentage (%) weight gain, % survival, feed intake, feed conversion ratio and protein efficiency ratio were evaluated. Results from Trials I, II and III indicated that dietary supplementation of 400 and 500 ppm limonene and 750 ppm thymol had growth-promoting effects in Nile tilapia but the somatic growth was not associated with enhanced feed intake and feed utilisation efficiency. Trials IV and V (Chapter 4) investigated growth and nutritional physiology pathways in Nile tilapia regulated by individual phytogenic compounds (objective 2). This was accomplished by analysing the effects of limonene (Trial IV) and thymol (Trial V) supplemented diets on the expression of key genes participating in selected pathways of somatotropic axis-mediated growth, appetite regulation, nutrient digestion, absorption and transport, lipid metabolism, and antioxidant enzyme defence system. Limonene was supplemented in the diet at 0, 200, 400 and 600 ppm while thymol was supplemented at 0, 250 and 500 ppm. Trials IV and V found that growth-promoting effects of limonene (400 and 600 ppm) in Nile tilapia involved up-regulation of key genes within pathways including somatotropic axis-mediated growth, nutrient digestion, absorption and transport, lipid metabolism and antioxidant enzyme defence system. Dietary thymol at 250 and 500 ppm did not significantly enhance growth of Nile tilapia nor regulate the nutritional physiology pathways listed above. In Trial VI (Chapter 5), the effects of combined phytogenic compounds (limonene and thymol) on growth and nutritional physiology of Nile tilapia was tested (objective 3) to establish if the compounds had synergistic or additive effects on the growth of the fish as well as complementary effects on the selected nutritional physiology pathways. A candidate gene approach was also used for the selected pathways. Results from Trial VI showed that a diet supplemented with a combination of limonene (400 ppm) and thymol (500 ppm) has neither synergistic nor additive effects on the growth performance of Nile tilapia, with limonene mainly influencing the attained somatic growth. The analysed candidate genes involved in the pathways of nutrient digestion, absorption and transport, lipid metabolism, antioxidant enzymes and somatotropic axis growth also showed no synergistic or additive effects of a dietary combination of limonene and thymol in Nile tilapia. Overall, results from the study suggest approaches for developing functional diets for Nile tilapia using limonene and thymol growth promoters.

Intensive and semi-intensive aquaculture systems are dependent on nutrient input either in the form of supplemental or complete feeds. Most complete diets still include high fish meal (FM) levels (≥10%). However, as the industry attempts to reduce its reliance on FM, feeds must now be formulated with much lower levels especially for omnivorous species such as tilapia. By 2015, mean FM inclusion in tilapia diets was projected to fall below 3% and be further reduced to 1% by 2020. In the global context of competition for crops, finding suitable plant-based replacers for FM and meeting the increasing demand for seafood, lower-cost and under-utilised plant feedstuffs are now receiving greater attention. The study was divided into three distinct components – field survey, growth experiments, and life cycle assessment. Field surveys were used to contextualise the growth experiments and assess commercialisation opportunities for multi-enzyme inclusion in tilapia feeds. Two sets of digestibility and growth experiments were designed to evaluate the feasibility of using high inclusions of plant-based ingredients sourced from locally available feedstuffs in Thailand to substitute FM at low inclusion levels (0 – 5%). The research evaluated the hypothesis regarding the potential of exogenous enzymes (protease, xylanase and phytase) to minimize anti-nutritional effects on nutrient digestibility of proteins, polysaccharides and phosphorus in tilapia. The research also assessed the secondary effects of enzyme supplementation on economic efficiency and life cycle environmental impacts. Tilapia is the second most cultured finfish globally and Thailand is the sixth largest producer. Based on the findings of the field survey, feeding practices of Thai tilapia farmers were confirmed to be diverse. Feed inputs included, but were not limited to, agro-industrial by-products (e.g. rice bran, corn bran etc.) and commercial diets. Commercial diets contained 15 – 30% crude protein and lower protein livestock diets (i.e. pig ration) were often used for supplemental feeding or “fattening”. The experimental low FM diets were therefore formulated as grow-out or “fattening” diets for semi-intensive green-water systems, a prominent feature (>60%) of Thai tilapia farming. In Phase 1, the digestibility experiment assessed the digestibility and growth in tilapia fed 0%, 3% and 5% FM diets with and without xylanase (0.385 g kg-1) and phytase (0.075 g kg-1). Performance decreased significantly with declining FM levels. No differences in feed intake, feed conversion ratio (FCR), specific growth rate (SGR) and weight gain were observed between the enzyme and control diets. Nevertheless, tilapia fed the enzyme supplemented 3% FM and control 5% FM performed similarly (P < 0.05). No enzyme-related effects were noticed for protein digestibility but phosphorus (P) digestibility improved by 9%, except at 0% FM level (P > 0.05). The enzymes had no apparent influence on nitrogen (N) retention contrary to previous studies, however, higher retention for P was observed. Villus length decreased with declining FM levels yet no improvements were seen in tilapia fed enzyme diets. In a simultaneous grow-out experiment, the six experimental diets were compared to an industry 10% FM standard. Conversion ratio was the lowest (1.66) in adult tilapia fed 10% FM diet however the enzyme supplemented 0% FM fed fish had a comparatively low FCR of 1.67. There were no significant enzyme-related effects on weight gain¸ SGR and protein efficiency. Proximal villi results were inconsistent. The cost of feed decreased with declining FM levels but increased with enzyme inclusion. Nevertheless, the economic returns per kg of whole fish produced were better using enzyme supplemented diets compared to the controls. Though the size of the effects on growth and nutrient utilisation were modest, the findings suggested that xylanase and phytase had some level of synergistic action on the targeted anti-nutrients. However, further research was required. In Phase 2, two control diets (2% FM, negative control (NC) and 10% FM, positive control (PC)) were compared with three enzyme supplemented 2% FM diets (NO-PRO, 0.385 g kg-1 xylanase and 0.075 g kg-1 phytase only; LO-PRO, xylanase + phytase + 0.2 g kg-1 protease and HI-PRO, xylanase + phytase + 0.4 g kg-1 PRO). Growth performances improved with enzyme supplementation compared to the NC (P < 0.05). Of the enzyme supplemented diets, the LO-PRO diet showed the highest improvements in weight gain (26%) and feed intake (19%), the latter comparing statistically to the 10% FM PC diet. The HI-PRO diet had the best FCR (1.88), again comparable to the PC (1.73). The NO-PRO diet had the highest protein, P, lipid and energy digestibility, suggesting no additive effect of protease on these coefficients. In terms of gut histomorphology, the LO-PRO and PC diets had the highest measurements and were statistically similar which may have explained similarities in feed intake. Compared to the NC, the HI-PRO diet produced the highest level of change in net profit due to gains in feeding efficiency however, the LO-PRO showed better improvements in terms of growth. Based on these findings, the ternary combination of protease with xylanase and phytase (LOPRO) has potential in limiting FM use for tilapia grow-out feeds, however, the economic efficiencies were still below that of a 10% FM diet. Future considerations for research should target the indigestible dietary components in order to optimise enzyme dosages and maximise the benefits of each enzymes. In conclusion, a comparative life cycle assessment (LCA) was used to evaluate the environmental impacts of low FM diets and commercial feeds associated with tilapia production in Thailand. The study showed that the low FM enzyme supplemented diets had lower impact potentials and were environmental superior to the average (10% FM) commercial standard. LCA modules are recommended for least-cost formulation programmes as an option going forward. Additionally, LCA can be used as a predictive tool to guide farmers, especially small-scale producers, on the potential impacts of feed input choices and feeding practices. This will ensure higher product quality but also demonstrate environmental responsibility on the part of aquafeed and fish producers to final seafood consumers.

There are increasing concerns regarding the environmental impacts and sustainability of intensive fish farming. In particular, criticism has centred on the use of fish meal and fish oil in the diets of farmed carnivorous fish species such as Atlantic salmon. If the industry is to continue to expand, reduction in the levels of fish meal and fish oil incorporated into diets and/or the use of alternative terrestrial sources of protein and oil must occur. The present study assesses several diet types containing different inclusion levels and /or sources of protein and oil in terms of growth and performance as well as assessing the diets in terms of sustainability and their potential to impact on the marine environment. In two nutritional studies, Atlantic salmon fed a low protein (LP) diet achieved similar growth and performance compared to fish fed a normal commercial diet (control diet) in both studies. Growth rates of fish fed a diet containing partial replacement of fishmeal with corn gluten and fish oil with rapeseed oil (SUS) were better than those of fish fed the control and LP diet and were similar to those of fish fed a high energy; nutrient dense (ND) diet. The amount of wild fish required to produce 1 Kg farmed salmon based on fish meal and fish oil inclusion levels were lowest for fish fed the SUS diet (1.3 kg) whilst fish fed LP diets had a lower conversion value compared to both the ND and control diets based on fish meal inclusion levels only. These results suggest there is potential for aquaculture to be more environmentally sustainable by reducing the amounts of marine fishmeal and oil used in diets fed to intensively farmed Atlantic salmon. In terms of dissolved wastes, fish fed a nutrient dense diet had higher feedrelated concentration peaks of ammonia detected which occurred earlier compared to fish fed other diet types. In contrast, fish introduced to low protein diets at different sizes throughout the marine phase of production had consistently lower concentration peaks of ammonia detected compared to fish fed a standard commercial ration. All groups fed a low protein diet had between 17 and 28 % less ammonia detected as a feed related concentration peak compared to the control group. The area under each concentration peak of ammonia ranged from 43.6 to 88.8 % ± 60 min of total ammonia detected over time for all diets. These results show that feeding fish diets containing lower inclusion levels of fish meal resulted in lower feed related concentration peaks of ammonia being detected. However most of the ammonia was excreted over a short time period and its potential to impact on the marine environment was assessed using mesocosm studies. In the first of two studies, mesocosms were fertilised with NH4CI based on ammonia concentration peaks from either 500 or 1000 T rainbow trout production (LN and HN enclosures respectively). These enclosures had consistently more phytoplankton present than the control enclosures. There was evidence of rapid uptake of excess nutrients with the HN enclosures having more cells present than the LN enclosures. In the second experiment, enclosures were fertilised based on ammonia concentration peaks detected from Atlantic salmon fed a low protein (LP enclosures) or a nutrient dense diet (ND enclosures). Both had consistently more cells present than the control enclosures. These results suggest that increases in phytoplankton communities may occur as a result of a single pulsed release of ammonia simulating discharge from intensively farmed fish. The predicted rate of dissolved nitrogen production was calculated from fish introduced to low protein diets throughout a complete marine production phase using a mass balance model. Fish introduced to low protein diets at 330, 800 and 1600 g had lower dissolved N discharge rates (22.51, 22.02 and 21.07 Kg'1 NT'1 Production respectively) compared to fish fed a standard commercial ration (23.32 Kg-1 NT-1 Production). These results show that there is potential to maximise use of low protein diets, which would result in less ammonia excretion and reduce the potential risk to impact on the marine environment. In an attempt to accurately quantify waste outputs from intensively farmed Atlantic salmon a custom-made tarpaulin was designed. Initial studies have shown that there is potential to collect data on waste outputs from fish reared in the marine environment whilst taking in to account seasonal and daily fluctuations in water temperature and salinity. The system can be used to directly compare different feed types and feeding strategies. It has been shown that changing the macronutrient inclusion level and sources results in differences in the physical characteristics of extruded feeds. High energy diets and substitution diets have similar settling velocities compared to a standard commercial diet but produce lower environmental impacts when modelled for solid waste impacts. Low protein pellets have slower sinking rates but generate more waste due to a higher FCR. The data from these results should be used in conjunction with other data for a range of feeds and environmental conditions to employ a “look-up table” approach to differentiate between diets when modelling waste dispersion.

In order to investigate the antinutritive effects produced by the intake of saponin and tannin (hydrolysable and condensed), ten experiments were conducted on Oreochromis niloticus and Clarias gariepinus utilising casein/gelatin purified diets. Fish were fed (3% body weight/day) four isoenergetic isoproteic diets containing 0.02/0.01, 0.08/0.09, 0.42/0.47 and 0.87/0.91% saponin, (for O. niloticus and C. gariepinus respectively), 0.05, 0.08, 0.27 and 0.71% tannic acid or 0.06, 0.16, 0.73 and 1.66% catechin which represent diets I, II, III and IV respectively in each study. In all studies fish fed diet I showed the maximum weight gain, highest specific growth rate (SGR), the best feed conversion (FCR) , the best protein efficiency ratio (PER) and the best apparent protein digestibility (APD) in comparison with all other diets; whereas no significant differences (P< 0.05) were observed in dry matter digestibility among the diets in all studies. Fish fed diet IV showed the poorest values (SGR, 0.31/0.21; FCR, 9.51/4.75; PER, 0.39/0.69; PPV, 4.37/-0.35 and APD, 84.04/80.63 with saponin), (SGR, 1.89/2.01; FCR, 1.36/1.05; PER, 2.31/2.97; PPV, 36.93/54.76 and APD, 86.49/95.33 with tannic acid) and (SGR, 2.16/1.85; FCR, 1.19/1.03; PER, 2.55/2.92; PPV, 37.88/50.93 and APD, 89.59/91.31 with catechin) in O. niloticus and C. gariepinus respectively in comparison with others. In terms of carcass composition, there were an increases in moisture and ash contents in fish fed diets III and IV of the study on saponin whereas lipid content was markedly decreased in proportion to the level of saponin in the diets. However, fish in all treatments did not show any significant difference (P< 0.05) in their body protein content. In the studies with tannic acid and catechin, carcass composition of fish showed no significant difference (P< 0.05) in all treatments. Histological examination showed cytoplasmic vacoulation of intestinal cells, disruption of straited borders and infiltration of lamina propria especially in O. niloticus and fatty degeneration in liver of fish fed diets III and IV of the study of saponin. However, fish in other studies (tannic acid and catechin) showed no abnormalities. Digestive enzyme activities were inhibited by the inclusion of dietary tannic acid or catechin, and the inhibition was in proportion to the level of the antinutrients. However, both forms of tannins had no effect on lipase activity of O. niloticus while the opposite was observed in C. gariepinus. It was shown that the activities of proteolytic enzymes were more inhibited by the inclusion of tannic acid or catechin than the activities of a-amylase and lipase. The results of the present thesis show that saponin and tannin (hydrolysable and condensed) acted as neutral or negative factors with respect to growth and feed performance, digestive enzyme activities and histology of intestine and liver.

A purified diet, the raw material zein and two diets composed of complex foodstuffs (fishmeal, soy, wheat), were tested over a seven and a half (7.5) month period, to obtain the in vivo Apparent Digestibility Coefficients (ADC) for dry matter, crude protein, lipid, carbohydrates, ash, energy and the ADCs of their amino acids, for the African catfish Clarias gariepinus (Burchell 1822). Faeces were collected by sedimentation and digestibility coefficients calculated by use of chromium oxide as external indicator. Apparent digestibility coefficients ranged between 54% and 96%. Crude protein ADCs (81-90%) recorded for diets in this experiment were similar to those previously recorded. Ash digestibility (54-89%) varied significantly between diets probably as the result of the reduced bioavailability of minerals and trace elements in their constituent raw materials. ADCs for individual amino acids ranged between 82% and 99%. Zein, as an experimental feed ingredient for African catfish, showed acceptable ADCs for both gross nutrients and amino acids. When the nutrient level in the test ingredient was very different from that in the reference diet, calculation of ADCs based on relative nutrient contributions appeared necessary. Dextrin (an intermediate product of starch hydrolysis) was the prevalent carbohydrate source in the purified diet. The high carbohydrate digestibility for this diet (72%) suggests high starch digestibility in African catfish and confirms previous hypothesis of good starch utilization in the species due to elevated amylase levels in the anterior part of the intestine. A 26-day experiment was conducted on triplicate groups of sub-adult catfish (102.2g) using two feeding frequencies, twice per day and three times per day with fish fed to appetite on a purified diet. Fish fed twice ate 1.42 % of b. w. d'1, those fed three times consumed 1.27% of b. w. d'1. Fish fed twice also exhibited better growth and food conversion. Despite the use of purified diets, performance indices for the groups fed twice per day were regarded as good compared to previously reported data. Administration of seven diets differing in protein: energy levels over a seventy-six day period, indicated that C. gariepinus (120-233g) performed best when fed diets containing gross energy between 22-24 kJ.g'1, P:E ratio of 21.5-23 mg protein.kJ1, crude protein at 46%, crude lipid at 10-17% and carbohydrate at 26-32%. These data served to confirm the carnivorous nature of C. gariepinus compared to the less carnivorous North American catfish Ictalurus punctatus. However, performance indices for the fastest growing groups where lower than those achieved in experiments by previous researchers, possibly due to the weight range of the experimental animals used, the available tank surface and the employed stocking densities. Carbohydrate levels (26-32%) of the best performing diets during these experiments were much higher than used by previous researchers (16-18%) for the same species and still higher than those employed for other carnivores (15-25%) (salmonids, sea bass, sea bream). Taking into account the high carbohydrate digestibility of all the diets used in these experiments more comprehensive use of carbohydrates in Clarias catfish diet formulations is suggested. The increase of carcass lipid as a result of increased dietary non-protein energy demonstrated for other species has also been demonstrated for African catfish. Blood parameters (haematocrit and total hemoglobin) proved not to be valuable performance indices in relation to general nutritional parameters associated with growth and food utilization. Low carcass incorporation values for arginine and methionine resulted in low requirements (1.97 and 0.26 g.lOOg'1 dietary Crude Protein, respectively) as determined by the carcass deposition technique. The same technique revealed the following values, as g required Amino Acid.lOOg'1 dietary Crude Protein: Histidine 1.39, Isoleucine 1.56, Leucine 4.87, Lysine 4.49, Phenylalanine 4.56, Threonine 2.04, Tryptophan 2.59, Valine 2.08. Diets used in these experiments were characterized by lower levels of arginine, lysine, methionine, cystine, threonine and tryptophan to those proposed by most recent research. It is suggested that the sequence of pre-experimental adaptation and starvation might have exerted an irreparable loss for some of those amino acids and consequently lower values as determined requirements. Furthermore, mediocre oxygen levels prevailing in the employed experimental system in conjunction with sampling manipulations (prolonged emmersion) might have created a shift to ureotelism with a result in low arginine tissue levels. Results of this study could be used to optimize formulations based on the “economically optimal protein level” and the peculiarities of individual culture systems. Such formulations could be based on both conventional and non-conventional foodstuffs of a standard supply and price for the selected country-area, which would standardize production practices, production levels and costs. Use of labelled substrates coupled with further experimentation on the required dietary carbohydrate: lipid ratios and the optimum dietary fatty acid profile will clarify the underlying metabolic pathways and probably lead to better elucidation of the deposition of essential amino acids. More information on the digestibility of various raw materials in conjunction with the determined dietary requirements, will lead to the manufacture of more balanced and cost-effective diets for the African catfish C. gariepinus.

The supply of juvenile Atlantic halibut, Ilippof'lossus hippoglossus, has been sporadic and until recently, has fallen short of expectations, due to difficulties associated with the hatchery phase. This thesis focuses on some specific aspects of intensive hatchery production which needed to be addressed in the areas of egg production and quality, triploidy induction and weaning. A quality assessment technique for halibut eggs, based on observations of morphological anomalies occurring during early blastomere divisions was devised. The degree of abnormality in the appearance of five features was quantified and a strong relationship between these characteristics and hatch rate of eggs incubated in microtitre plates was revealed. It was concluded that this method is of great potential use as an early predictive indicator of egg viability. The effect of temperature on egg production was studied in two broodstock groups held either at ambient or low stable temperature during spawning. High temperatures caused a delay and shortening of the spawning period as well as a reduction in egg quantity and quality. The results indicated that temperature control is a necessary feature of broodstock management at sites where ambient temperature profiles are unsuitable. The efficacy of hydrostatic pressure shocking for the induction of triploidy was tested on newly fertilised eggs. A 5 min pressure shock of 8500 psi, administered around 15 min after fertilisation resulted in high triploidy yields. This treatment had little effect on survival to hatching. However, a preliminary experiment indicated that triploid halibut may be prone to higher mortality through the hatchery cycle. Six weaning experiments were conducted to determine the influence of size, age, and developmental stage of halibut larvae on diet uptake, survival, growth and fry quality. Gradual replacement of Artemia with dry diet, co-feeding, and the use of intermediate diet types, were evaluated. In the absence of live prey, pre-metamorphic larvae (< 100 mg) would accept non-living feed particles but total replacement with a conventional dry diet was unsuccessful. However, following good growth prior to weaning, 700 "day old larvae could be successfully weaned over a nine day period. Improved growth and substantial savings in live feed were among the positive benefits resulting from early weaning.

The pathway for the biosynthesis of docosahexaenoic acid (22:6n-3) fromlinolenic acid (18:3n-3) was investigated in rainbow trout (Oncorhynchus my kiss)liver in vitro, using primary cultures of hepatocytes and liver microsomes toinvestigate the products of desaturation and elongation of [l-14C]-18:3n-3 and[l-14C]-20:5n-3. Argentation thin-layer chromatography and radio gaschromatography were employed to analyse the metabolic products of the radiolabelledfatty acid substrates and to determine the rate of conversion of 18:3n-3 to 22:6n-3. The recovery of radioactivity in various polyunsaturated fatty acids (PUFA) of trouthepatocyte lipid, including 20:5n-3 and 22:6n-3, established that juvenile trout werecapable of converting 18:3n-3 to 22:6n-3. To establish the extent to which theformation of 22:6n-3 was enhanced in the absence of dietary PUFA, particularly20:5n-3 and 22:6n-3, and therefore, facilitate the investigation of 22:6n-3biosynthesis, rainbow trout were fed a diet based on olive oil and deficient in (n-3) PUFA. Feeding the diet deficient in (n-3) PUFA efficiently reduced the endogenouslevels of (n-3) PUFA in trout hepatocytes and at the same time markedly increased therate of 22:6n-3 formation from both [l-14C]-18:3n-3 and [l-14C]-20:5n-3. When thedesaturation and elongation of [1-14C]-18:3n-3 and [l-14C]-20:5n-3 were investigatedin microsomes isolated from trout liver, no radioactivity from either substrate wasrecovered in 22:6n-3. High proportions of radioactivity from [l-14C]-20:5n-3 were,however, recovered in 24:6n-3 and 24:5n-3. These radiolabelled C24-PUFA producedby the microsomal incubations were separated by argentation chromatography andused as substrates in incubations with hepatocytes isolated from trout liver. Docosahexaenoic acid (22:6n-3) was generated from both radiolabelled C24-PUFAsubstrates by trout hepatocytes. The results establish that the biosynthesis of 22:6n-3in rainbow trout hepatocytes does not involve A4-desaturation of 22:5n-3 but ratherproceeds via the A6-desaturation of 24:5n-3 with the subsequent chain shortening ofthe 24:6n-3 produced. Cyclopropene fatty acids and curcumin significantly inhibited the desaturation and elongation activity of hepatocytes and liver microsomes fromrainbow trout. The A6 and A5-desaturase and elongase substrate specificities wereinvestigated; it was shown that (n-3) PUFA substrates were always preferred by theenzymes over (n-6) PUFA.

A series of nutritional experiments were carried out to evaluate the utilization of dietary protein and energy by gilthead sea bream (Sparus aurata L.). Optimum dietary protein requirements and the sparing effect of dietary lipid upon protein were investigated for different fish sizes. Optimum dietary protein requirements were 55% and 42% for 0.8g fry and 60g juveniles, respectively. In 5g fingerlings, dietary protein levels could be reduced from 52% to 45% when lipids were increased from 9% to 15%, best protein to energy ratio (P;E) being 21.9 g protein/MJ of gross energy. These results suggest that dietary protein level could also be reduced from 55% to 50% in fry diets containing 15% dietary lipids. The optimum proportions of dietary protein and lipid level found for 90g growers were 54% and 11%, respectively, the high requirements for protein could be due to an increased protein demand during sexual maturation for gonad development. The increase of dietary lipids produced an increment in carcass lipid deposition, both in visceral and non-visceral tissue, but these levels were in all cases well below reported carcass lipid contents in wild S.aurata in the Mediterranean. When different dietary carbohydrate sources were evaluated with 42g S.aurata juveniles, the ability of fish to digest carbohydrate was limited in general. Apparent Digestibility Coefficient (ADC) values being lower than 85% regardless of carbohydrate source. Increased amounts of fibre in diets produced lower protein and lipid digestibility, this effect being even more pronounced on carbohydrate digestibility. Corn starch was the most effective carbohydrate source in terms of "energy-yielding", although wheat bran appeared to be a suitable fedstuff for practical diets. Increased dietary lipid levels supported increased metabolizable energy (ME), recovered energy (RE), energy retention efficiency (ERE) and better protein conversion in 46g juveniles, indicating improved utilization of protein and energy.

The bioavailability of nutrients in commercial feeds available in Portugal was evaluated in three size groups (40g. l00g and 180g) of rainbow trout (Oncorhynchus mykiss) on two private farms. Dietary crude fibre level was used as the digestibility marker. A series of laboratory trials was also carried out to investigate the influence of various biotic and abiotic factors on apparent digestibility coefficients and to provide corroboration for the field survey. The principal factors influencing apparent digestibility values during the field survey were diet quality, temperature and water-dissolved oxygen. The apparent digestibilities of crude protein, organic matter and dry matter were strongly inversely correlated with dietary fibre level in the three size groups of farmed trout. In addition, significant positive correlations were obtained between lipid level in commercial feeds and the organic matter digestibility coefficients. Under controlled conditions, dietary lipid level (7%, 14% and 21%) had a marked effect on digestibility, the highest digestibility values being obtained with the 21% lipid diet (87.77%, 76.55% and 70.46%) compared to 84.59%, 63.25% and 57.74% with the 7% lipid diet for crude protein, organic matter and dry matter, respectively. A signification correlation was also obtained between all apparent digestibility coefficients and the feed protein level for all size groups of fish studied in the field. In the laboratory, however, only at 21®C was there a marked effect and significantly higher digestibility values were obtained with the higher protein (45%) diet (85.98%, 83.74%, 80.37% and 84.54% for protein, organic matter, dry matter and energy, respectively) as compared to the lower protein (30%) diet (74.16%, 71.66%, 65.54% and 68.70% for protein, organic matter, dry matter and energy, respectively). General increases in apparent digestibility values of more than 10% were obtained between the lower (10°C or 15°C) and the higher (21 °C or 22°C) experimental temperatures, much higher than values previously reported. Food intake was shown to be directly dependent on dissolved oxygen (D.O.) and for one unit decrease of D.O. food intake decreased by about 0.5% or 0.25% body weight/day whether fish were subjected to an abrupt decrease of water O2 level or to prolonged hypoxia. Furthermore, apparent digestibility values were significantly increased when fish were subjected to prolonged hypoxia. Feeding frequency (1, 2 or 4 meals/day) did not influence digestibility, whereas time of day and fish size did have a significant effect. Mean digestibility values estimated either from chromic oxide or crude fibre analysis were significantly different. The difference was, however, very small and bearing in mind the advantages of an internal marker as compared to an added marker, this study has shown that crude fibre is a useful tool for feed evaluation, especially in practical situations. The need for, and importance of, digestibility evaluations is discussed and summarised in the form of a conceptual model.

Tissue vitamin E in farmed Atlantic salmon (Salmo salar L.) fed commercial diets, were measured by high-performance liquid chromatography. In healthy fish, fed diets containing about 20 mg lOO g*'supplementary all-rac-a-tocopheryl acetate, vitamin E (a-tocopherol) accummulated in the liver and plasma. Over a 20 month period, beginning four months prior to seawater transfer, mean a-tocopherol concentrations ranged between 51 and 754 |tg g*' in liver and 7 and 68 pg nl*' in plasma. Sharp increases in the a-tocopherol content of these tissues were related to a rise in the proportion of lipid as polyunsaturated fatty acid in response to transfer to seawater. There was no marked difference between tissue vitamin E concentrations in farmed and wild salmon. In contrast, vitamin E depletion was found to be a consistent feature of pancreas disease (PD) in farmed salmon. The degree of depletion was found to vary from case to case. A serial study indicated that the onset of vitamin E depletion occurred early in the course of the outbreak. Soma information was collected with regard to the cause and possible pathological effects of vitamin E depletion in PD, but definitive studies are still required. Parental administration of all-rac-a-tocopherol in molten cocoa-butter reduced plasma pyruvate kinase activity in affected fish from one outbreak of PD, suggesting that muscle degeneration in PD may respond to treatment with vitamin E. However, no effects on the histological signs of PD were noted. In a second outbreak, supplementation of the diet with 100 mg all-rac-a-tocopherol 100 g'* produced no statistically significant elevation of plasma vitamin E or reduction in plasma pyruvate kinase activity. In a third case, dietary supplementation with a water-miscible preparation of all-rac-a-tocopheryl acetate produced no significant rise in plasma vitamin E concentrations and no histological evidence of any therapeutic effect was obtained. Relationships between husbandry practices and vitamin E status were examined experimentally. Acute handling and crowding had no observable effect on plasma a-tocopherol concentrations. Chronic crowding stress produced a slight, but statistically insignificant, reduction in plasma and liver a- tocopherol concentrations. Chronic crowding stress did not precipitate signs of vitamin E deficiency in salmon fed diets with no supplementary source of a- tocopherol. Starvation caused a significant depletion of a- tocopherol within 3 weeks in tbe plasma and liver of post-smolts in experimental tanks, but had no effect on tissue vitamin E concentrations in one sea-winter salmon held in sea-cages over a 4 week period. Differential feeding rates of l-3t body weight day* over a six week period had no measurable effect on plasma vitamin E concentrations. The dietary level of a-tocopheryl acetate had no effect on the magnitude of the cortisol response to acute handling stress and no beneficial effect of vitamin E supplementation was determined in fish under conditions of chronic crowding.