Growth in early life stages of fishes : an explanatory model

Although fish larvae grow very fast, little is known about to which extent their metabolism and nutritional requirements differ from larger fish. Modelling can be a powerful too] to promote understanding and optimisation of growth in fish larvae. The main objective of this study was to develop an explanatory model that can simulate growth in fish larvae. The study was conducted in three steps, i.e., a design, an experimentation and a modelling step. In the first step, the framework of the model was developed using yolksac larvae of the African catfish, <em>Clarias gariepinus</em> . The model is nutrient driven, and based on the biochemical reactions underlying the growth process. The model predicted growth rather accurately until complete yolk absorption, and suggested what information was needed for its further development. In the second step, experimental data on protein metabolism and the energetics of growth were collected. They constituted the basis for the design and parameterisation of a model for fed larvae. In the African catfish, the cost of growth decreases with increasing growth rates, down to the theoretical minima for protein synthesis. Fast growing larvae of turbot, <em>Scophthalmus maximus</em> , also have a cost of growth close to the theoretical minimum for protein synthesis. The amino acid (AA) profile of the free pool in turbot larvae is highly variable, being very sensitive to AA coming from the diet and protein turnover. Both in the African catfish and turbot, the larval AA profile changed during ontogeny, especially before the start of exogenous feeding. The AA profiles of the diets differed considerably from the larval ones, suggesting high unavoidable AA losses. In yolk-sac larvae of catfish, there is little regulation of catabolism of AA, and no sparing of essential AA. In contrast to this, turbot larvae may be able to spare essential AA towards the end of the larval stage. In the third and final stage, the model was parameterised and validated for African catfish and turbot. With a sensitivity analysis of the model the key parameters governing larval growth were identified. These results are discussed in view of elucidating the dietary requirements of fish larvae, evaluating the present feeding practices, and proposing research trends for the future.