The evaluation of potential alternative protein sources for the formulation of least-cost optimal diets to satisfy the nutrient requirements of South African abalone (Haliotis midae) has been hampered by the absence of a suitable, practical, replicable and reliable digestibility technique. A suitable lowcost faeces collection technique was developed in this study to obtain suitable quantities of excreta for analysis from this species. Acid-insoluble ash was identified as a reliable, replicable and safe internal marker in comparison to chromic oxide and crude fibre for use in nutrient digestibility studies with H. midae. This was validated by the consistency and repeatability of the results and by comparison to total collection of faeces. The traditional substitution method used in digestibility studies with fish to evaluate protein digestibility of feed ingredients was found to be unsuitable for H. midae. Apparent protein digestibility values exceeding 100 % derived through this method could be attributed to associative effects between feed ingredients, differential diet and faecal nutrient leaching, and mathematical artifacts in calculations when using substitution versus single protein diets. An ingredient particle size of less than 450 μm in comparison to particle size classes of above 450 μm was shown to enhance nutrient (dry matter, organic matter, protein, fat) digestibility and minimise dry matter leaching from diets. The dietary inclusion level of both pre-gelatinised maize starch and a-cellulose did not influence (P > 0.05) apparent nutrient (protein, fat, fiber, starch) digestibility. Using the above digestibility protocol amino acid availability of all plants ingredients currently used in the South African animal industry was evaluated for H. midae. Soybean meal (96.86 %) and lupins (96.51 %) presented the highest apparent mean amino acid availability of all plant protein ingredients evaluated with H. midae. Canola meal (94.21 %), faba beans (92.87 %) sunflower meal (92.77 %), peanut meal (87.39 %) and cottonseed meal (85.15 %) presented higher apparent mean amino acid availability values than fish meal (82.75 %). Apparent protein digestibility was highly correlated (r = 0.99) with mean apparent amino acid availability, while true amino acid availability was 1.88 % units higher than apparent amino acid availability for all ingredients tested. Predicted apparent protein digestibility in compound diets was within 1.1-6.5 % of determined values. Calcium phosphate mono dibasic presented the lowest (P < 0.05) dietary phosphorus leaching (51.51 % maximum) and highest apparent phosphorus digestibility (66.27 %) in comparison to other inorganic phosphorus sources. Based on the method of direct experimentation to determine the optimal dietary protein level using graded levels of dietary protein 28.1-35.9 % dietary protein from good quality sources is recommended for maximum growth of juvenile H. midae. This study provides a scientifically sound research tool including a faecal collection technique, suitable marker and assay technique that could be use in further studies to improve least-cost diet formulation for H. midae. Future nutritional studies in H. midae should primarily concentrate on reducing dietary nutrient leaching and improving the intake of nutrients in order to properly evaluate responses of this species to different dietary regimes.

The potential to reduce feed formulation costs by the replacement of existing proteinsources in artificial feeds for Haliotis midae was assessed. A comparison between the efficacy of the direct method and chromic oxide marker techniques in determining apparent protein digestibility coefficients, revealed that while the former was not a suitable methodology for use with this species, the latter produced replicable and reliable results. It was established that the chromic oxide marker is inert, is not absorbed by the abalone, does not interfere with the digestive processes and moves through the intestine at the same rate to the protein. However, as this method was time consuming and expensive to implement, a multienzyme pH-stat in vitro protein digestibility technique using a three enzyme system was employed to rapidly assess the protein quality of 34 protein sources. The efficacy of the technique was established by correlating the in vitro digestibility estimates with in vivo digestibility coefficients obtained using the chromic oxide marker technique. The highest degree of in vivo predictability was attained when protein sources were separated according to origin, and significant correlations between either animal (r²=0.89, P<0.004) or plant (r²=0.79, P<0.04) protein sources were found. The effect of animal size on the qualitative protein requirements of two size classes of H.midae was assessed by feeding 12 isonitrogenous and isoenergetic single protein diets to juvenile and young adult animals (10-20 and 40-50mm shell length). The criteria for protein source selection were their bioavailability as determined using the pH-stat in vitro digestibility technique, and their cost and availability within South Africa. The protein sources identified for the trial comprised four fishmeals, casein, spirulina, abalone viscera silage, brewery waste, torula yeast, carcass, sunflower and cotton meals. The results indicated that in terms of growth and feed efficiency, the fishmeals and spirulina were the most suitable candidates as primary protein sources in formulated feeds for H.midae, and with the exceptions of the carcass meal and brewery waste, the remaining protein sources demonstrated promise as partial primary protein source replacements. The smaller size class of abalone displayed significantly reduced growth, feed and protein efficiency than their larger counterparts. In terms of feed conversion and growth response, two-way analysis of variance revealed significant interactions between protein source and animal size, suggesting that qualitative differences exist between the dietary protein requirements of the juvenile and young adult abalone. The commercial implication of thisfinding was discussed. An evaluation was undertaken to determine the effects of the partial and total replacement of dietary fishmeal with selected plant protein sources on growth and nutritional indices of juvenile H.midae. A commercial “Abfeed” formulation in which 100% of the protein component comprised LT-fishmeal was employed as a control. Fifteen isonitrogenous experimental diets were formulated in which the LT fishmeal was substituted at either 10, 15, 30, 50, 75 or 100% with either spirulina, semolina, ground maize, torula yeast, soya, sunflower or corn gluten meals or combinations thereof. No significant differences were found in the growth rates between the control diet and those diets in which 30% of the fishmeal component had been replaced by either soya, sunflower meal, or torula yeast. In addition, 50% of the fishmeal component could be substituted with either soya meal or spirulina without affecting growth. Replacement of either 75 or 100% of the fishmeal had a significant negative affect on growth. Pearson product moment correlations between dietary lysine levels and either growth rates or protein efficiency ratios revealed positive correlations (r=0.77, P=0.0005; r=0.52, P=0.04 respectively) suggesting that lysine may have been the first limiting amino acid in these diets. Carcass analysis revealed that dietary protein source had no significant effect on body composition. An assessment of the dietary arginine requirement of juvenile H.midae using whole proteins to supply graded levels of dietary arginine did not promote a growth response. It was concluded that arginine is probably not the first limiting amino acid in formulated feeds for H.midae. An assessment of the dietary lysine requirement of juvenile H.midae using two micro-encapsulation techniques (gelatin/acacia colloid or cellulose acetate phthalate) was not successful. The efficacy of the encapsulation techniques were established and an assessment of the degree of lysine supplementation undertaken. Failure of the crystalline lysine enriched diets to promote growth suggests that the prospects for using crystalline amino acids in essential amino acid requirement studies in H.midae is low. The results of the present study indicate that the prospects for replacing fishmeal with cheaper alternative protein sources in artificial feeds for H.midae is promising. Furthermore, while the technical difficulties such as the determination of the essential amino acid requirements of the abalone precluded the application of “least cost” programming, the prospects for its future application are promising.