Summary

Objectives

1. To identify potential feed ingredients to replace fish meal in aquaculture diets for silver perch.
2. To evaluate promising ingredients in terms of their in vitro and in vivo digestibility and assimilation.
3. To develop and evaluate methods of improving the usefulness of ingredients through processing (e.g. extrusion or cooking) and the use of enzymes and supplements.
4. Identify areas where inadequate knowledge of nutritional requirements may restrict fish meal substitution and determine these requirements for silver perch.
5. To formulate and evaluate diets with reduced contents of fish meal for silver perch.

Non Technical Summary:

If aquaculture is to continue to expand in Australia cost-effective diets based on Australian agricultural ingredients urgently need to be developed. The replacement of fish meal as the protein source of choice is a global research priority driven by static or declining supply of fish meal and rapidly expanding aquaculture and aquaculture feed industries. Australia has very poor supplies of fish meal and other aquatic meals but fortunately has abundant supplies of agricultural ingredients with potential for use in aquaculture diets.

In recognition of the need to develop diets for Australian aquaculture species, with reduced contents of fish meal, a number of institutions independently commenced this type of research in the early 1990's. The Fisheries Research and Development Corporation (FRDC) was approached by a number of institutions to financially support this research. In response, FRDC created their first 'Sub-Program' with the aim of coordinating research to develop Australian aquaculture diets. The two primary objectives were to replace fish meal and obtain an early commitment from commercial aquaculture feed manufacturers to adopt results.

Six separate projects were formed; four on species considered to represent most 'types' of species being farmed in Australia, one on feed processing and one on a technology audit for amino acids. The four 'model' species were prawns Penaeus monodon, silver perch Bidyanus bidyanus, barramundi Lates calcarifer and Atlantic salmon Salmo salar. Each project involved a number of collaborating scientists from different institutions and all projects were coordinated through a Sub-Program Steering Committee. Regular meetings with investigators from all projects as well as feed manufacturers, ingredient suppliers and R&D corporations were held twice each year.

This report describes the progress achieved with the silver perch project: Replacement of Fish meal on Aquaculture Diets for Silver Perch. Silver perch are an omnivorous, freshwater species endemic to eastern Australia. They have shown outstanding potential for culture in static earthen ponds and are one of the few species being cultured, or considered for aquaculture, in Australia which might replace some of the more than 55 000 t/yr of imports of white-fleshed fish. Growth and production potential of silver perch are similar to channel catfish in the USA and carp and tilapia in south-east Asia.

Objective 1: To identify potential feed ingredients to replace fish meal in aquaculture diets for silver perch

Literature and data base searches were conducted and a comprehensive list of ingredients currently available for use in animal feeds in Australia was compiled. Available data on biochemical composition, price and availability of these ingredients was obtained. This was used to select ingredients for further evaluation.

Additional ingredients were identified following discussions with the Grains Research and Development Corporation, the Grain Pool, the Australian Wheat Board and the Meat Research Corporation. The specific features making ingredients worth considering for aquaculture were discussed and the various agencies recommended ingredients their stakeholders produced which might be suitable. Descriptions of ingredients and data on composition can be found in Sections 6.2, 6.3, 6.4, 6.7, 6.8, 6.10 and 6.12.

Objective 2: To evaluate promising ingredients in terms of their in vitro and in vivo digestibility and assimilation

Initially, a series of three experiments were done to determine the most appropriate techniques for in vivo digestibility determination. Results from these experiments demonstrated that collection of faeces by settlement over 18 h was a suitable method for determining digestibility in juvenile silver perch (see Section 6.1 of this report). As the sum of digestibility coefficients calculated separately for individual ingredients was similar to that calculated for a diet comprised of those ingredients, the assumption that digestibility coefficients are additive was validated (see Section 6.1 of this report).

Methods for in vivo digestibility were developed or evaluated by Dr Alex Anderson at QUT. Results from these experiments are reported separately (see Fish meal Replacement in Aquaculture Feeds: in vitro studies on feed ingredients for aquaculture species. Final Report to FRDC Sub-Program 93/120). In summary, in vitro methods were shown to be useful for ranking but not for accurately determining digestibility coefficients for use in diet formulation.

In vivo digestibility determination involves measuring the amount of dry matter, energy or a specific nutrient which is ingested and then subtracting what is excreted. This involves collecting and analysing faeces. It is the first critical stage in evaluating the potential of an ingredient for use in a formulated diet. Following method development and verification (see Section 6.1 of this report), digestibility coefficients for over 60 ingredients, including some processed in different ways, were calculated. Digestibility coefficients for dry matter, energy, protein and, in most cases, individual amino acids (except tryptophan) are presented (see Sections 6.2, 6.3, 6.7, 6.8, 6.10 of this report).

Once digestibility coefficients are available, the next step is to determine the maximum amount of an ingredient which can be used in formulated diets. Many ingredients, especially those derived from plants, have anti-nutrients, some of which affect utilisation of the ingredient but not digestibility. In addition, excessive amounts of some ingredients may reduce the attractiveness of the diet or suppress palatability. Information on how well ingredients are utilised is also critical for effective diet formulation using new ingredients.

Growth experiments were conducted to provide this information for meat meals, poultry offal meal, feather meal and dehulled lupins. This added to earlier research results to estimate maximum contents of soybean meal, canola meal, peanut meal and lupins. One experiment was also conducted where all fish meal was replaced with specially modified wheat gluten meal. The most promising ingredients evaluated are meat meal, especially low ash meat meal, poultry offal meal and dehulled lupins. The specially-modified wheat gluten meal also deserves further evaluation. These results are reported in Sections 6.4, 6.6, 6.9 and 6.12.

Objective 3: To develop and evaluate methods of improving the usefulness of ingredients through processing (e.g. extrusion or cooking) and the use of enzymes and supplements

Most ingredients with potential for use in aquaculture diets are inferior to fish meal in terms of their nutritional composition (especially total protein content and amino acid profile), carbohydrate content or presence of anti-nutrients. Some of these deficiencies can be overcome. Digestibility and utilisation of an ingredient can be improved by processes such as grinding, cooking, and removal of less digestible components such as carbohydrate (e.g. through dehulling and removal of starch and non-starch polysaccharides) and ash (eg through removal of bone).

We found grinding diets below a particle size of between 710 and 1 000 m m was unnecessary but that steam conditioning or extruding practical diets containing starch improved gelatinisation of starch, digestibility, gustatory characteristics and fish growth (see Sections 6.8 and 6.10). Removal of hulls, by dehulling, improved dry matter and energy digestibility of two species of lupins, field peas, chick peas and vetch but not faba beans (vetch was poorly accepted by silver perch). Further protein concentration, through the removal of starch and/or non-starch polysaccharides further improved energy and dry matter digestibility of lupins, field peas and faba beans (other protein concentrates were not available). Protein digestibility of most pulses was high (see Sections 6.3 and 6.7 of this report).

Removal of part of the ash fraction from meat meals increased total protein content and improved the value of meat meals for use in diets for silver perch (see Section 6.4 of this report).

Some of the most common supplements used to overcome nutritional deficiencies in ingredients and diets are crystalline amino acids. During this study, crystalline lysine, methionine and/or threonine were added during several experiments but there is no conclusive evidence that silver perch responded to these supplements at any time. This may be due to problems with utilisation of crystalline amino acids, and such problems have been widely reported with some species of fish, or indicate that the diets supplemented were not deficient in those amino acids. This area requires further evaluation.

Objective 4: Identify areas where inadequate knowledge of nutritional requirements may restrict fish meal substitution and determine these requirements for silver perch

When we formulated the first reference diets for silver perch we set nutritional specifications using published requirements for other species as a guide. In particular, species such as channel catfish and tilapia, which are omnivorous, were used.

Given the high cost of protein, the major initial task was to estimate requirements of this nutrient for silver perch. As energy might be able to spare requirements for protein, the interaction between energy and protein was also important to quantify.

Preliminary research indicated that growth increased with both protein and energy but that as early diets were formulated before we had accurate information on protein and energy digestibility, results were confusing and difficult to interpret. We were able to conclude that energy could spare requirements for protein but high lipid content diets led to high lipid content of fish tissue, a negative from a marketing perspective.

Recent research with pigs and poultry has introduced the concept of a protein dependant phase and an energy dependant phase for maximum protein deposition. With this concept, at a certain energy content protein deposition increases during the protein dependant phase and then plateaus out. Further increases in protein deposition require additional energy.

We applied this approach to determine optimum protein and lysine requirements for silver perch. We used a single digestible energy content, one used successfully in practical diets gaining wide commercial acceptance and which we knew did not produce excessive lipid deposition in the fish carcass. During this experiment we determined that for silver perch fed a diet with approximately 14-15 MJ/kg digestible energy the minimum digestible protein content which produced maximum growth was only 25.2% (much lower than most estimates of optimum protein for cultured fish). We also estimated that 14-15 MJ/kg digestible energy diets do not need to contain more than 1.5% digestible lysine for optimum growth (see Section 6.11 of this report).

Objective 5: To formulate and evaluate diets with reduced contents of fish meal for silver perch

Most nutritional research is done using small juvenile fish and small tanks. The applicability of results generated using these methods is often questioned by farmers who wish to grow fish through to a market size in large ponds or tanks. Nutritional research needs to be validated in commercially relevant facilities.

In this study we utilised the results for ingredient digestibility and utilisation efficiency and the effects of diet processing to formulate two 'least-cost' diets for a large-scale farming experiment. Cost of ingredients was estimated as ingredient cost only, not including transport costs as this component is clearly dependant upon where the feed mill is located.

Our least-cost diets contained only 5 or 10% fish meal with the remainder replaced with meat meal, dehulled lupins or dehulled field peas. The least-cost diets out-performed the earlier fish meal/soybean meal reference diet and the cost of producing fish, based on ingredient costs and food conversion ratios, was lower (see Section 6.5).