Executive Summary:

Performance and taste of silver perch were compared for fish fed meat meal diets and grown in earthen ponds to market size. Silver perch fingerlings (58.5 g average, 56.3-59.6 g range) were stocked at a density of 15 000 fish ha-1 earthen ponds (o.1 ha) and cultured for 187 days to a market size of >400 g fish-1. Fish were fed the best diet evaluated in an earlier experiment (95LC2; 5% fish meal) or one of two test diets formulated using a least-cost linear feed formulation program and digestibility coefficients for a range of Australian agricultural products (rendered animal protein meals and wheat were the main ingredients). In the test diets, all of the fish meal was replaced. GRC2 was formulated to similar nutrient specifications to 95LC2 whereas GRC3 contained similar digestible energy but a lower digestible protein which was found to sufficient in previous research.

Ponds have not yet been harvested but based on large samples of fish (~200/pond), there were no significant difference between the performance of silver perch fed any of the three diets. The mean weights (461 g, 453 g, 433 g), growth rates (mean 2.4, 2.4, 2.3) and assumed FCRs (1.6, 1.7, 1.7) based on expected survival of 95% (as has been recorded in three previous experiments in the same ponds) were similar to the best performance from earlier large-scale trials in the same facilities and well above average performance recorded on commercial farms.

A diet ingredient cost of AUD$0.74 to produce 1 kg of fish using GRC3 is the lowest yet recorded and compares well to that for fish fed GRC2 (AUD$0.77/kg) or 95LC2 (AUD$0.89/kg) during this trial or AUD$1.79/kg for SP35 which is a commercial diet based on fish meal and soybean meal still in use by some farmers.

Sensory evaluation of fish fed all three diets (95LC2, GRC2 and GRC3) during this experiment was conducted to determine whether adult fish consumers will taste a difference in the cooked fish. In addition, sensory evaluation aimed to: 1) determine that if three diets produce a discernible difference in flavour, on what sensory attribute they differed? 2) what relative preference, if any, was shown by the consumers for the three differently treated fish and 3) what sensory attributes drive consumer attributes of the cooked fish? The consumer sensory evaluation was carried out in Sydney in April, 19999 and consisted of a controlled sensory evaluation procedure administered to 45 adults.

The three fish tasted significantly difference on 5 out of the 21 sensory attributes. These attributes and the diets ranking (best to worst) are: smell strength GRC3<95LC2; weedy flavour strength GRC3<95LC2; metallic taste strength GRC3<95LC2; aftertaste strength GRC2<95LC2; aftertaste liking GRC2>GRC3>95LC2.

The sensory attributes driving the acceptance of the fish and the diets with the highest or best score were: Flesh colour liking 95LC2; Overall appearance liking GRC2; Smell liking GRC3; Flavour liking GRC3; Mud flavour strength GRC3 (lowest); Fresh flavour strength GRC3; Aftertaste liking GRC2.

There was no significant difference between the three fish in terms of the consumers' overall liking of the fish, however, 95LC2 was the poorest performer and GRC3 the best performer on the full sensory profiles.

In conclusion, the independent sensory evaluation indicated that the diet GRC3 should be the diet of preference for highest quality production and GRC2 is a reasonable substitute but produces a less favourable sensory profile on two attributes (weedy and metallic).