Evidence of violation of a basic assumption in models currently used for food consumption estimation in fish

Abstract

The quantification of food consumption in fish is essential for the development of quantitative trophic models of aquatic ecosystems and in aquaculture. Four models and their derivatives have mainly been used to estimate daily ration by modelling change in stomach contents over time, namely the Bajkov, Elliott-Persson, Olson-Mullen and MAXIMS models. In spite of their widespread application, none of these models have so far been adequately tested to ascertain whether the assumptions they are based on are fulfilled and the daily ration predictions are accurate. Two original datasets, one obtained from milkfish, the other from tilapia hybrids, were used here to apply these four models and compare the consumption estimates with the quantities known to be ingested. In both species, all the model predictions were close together (1.76 to 1.81 % Body Mass Equivalent (% BME) for milkfish; 0.79 to 0.82 % BME for tilapia) but distinctly lower than the known consumption levels (3.15 % BME for milkfish, 1.90 % BME for tilapia). A comparison of the stomach contents observed in the feeding phase with those expected on the basis of the evacuation rate calculated in the non-feeding phase and assuming complete consumption of the food provided strongly suggested that the rate of stomach evacuation in the feeding phase was significantly higher than that when the fish were no longer feeding. When the consumption estimates were recalculated with all models on the assumption that the evacuation rate doubles in the feeding period compared to the non-feeding phase, the model predictions increased to 2.73 to 3.58 % BME for milkfish and 0.98 to 1.58 % BME for tilapia, supporting the validity of this assumption. The present findings cast doubt on one of the fundamental assumptions of all food consumption models and therefore also on the results obtained with the aid of these models.