08 15. A revised model of visual range in fish.-Sarsia 82:137-147. Bergen.lSSN 0036-4827.Models of visual range and location distances are crucial for quantification of vision based feeding opportunities and predation risk in the pelagic habitat. We compare an earlier published model with measurements of the reactive distance ofGobiusculusjlavescens relative to two species of copepods. Although this model gave reasonable predictions at low light intensities, the measurements of reactive distance at higher intensities were much lower than those predicted by the model. We modified the model to account for saturation at high light intensities. With this additional feature, the correspondence with the G.jlavescens observations was significantly improved. Furthermore, the revised model is consistent with earlier published data on fish contrast thresholds obtained over a wide range of target sizes and irradiance levels. Given the values of only two parameters, one sensitivity threshold and one saturation parameter, the model is capable of predicting visual ranges for relatively large intervals of light intensity, prey size and turbidity. Other published visual range models are briefly reviewed and compared with our model.

Both reduced illumination and increased turbidity caused a significant reduction in reaction distance of Gobiusculus flavescens. The longest reaction distance, 18.9 cm for larger prey (Calanus finmarchicus), occurred at a light level of 80 μmol m −2 s −1 compared to 12.9 cm for a smaller prey (Acartia clausi) at 8 μmol m−2 s−1. Above a light saturation level of 10 μmol m−2 s−1, additional light had little influence on reaction distance. In the turbidity experiments, the longest reaction distances were measured at turbidity levels of 10–20 JTU. Prey size influenced reaction distance at all tested light levels. Search time was influenced by prey size only at low illumination. With increasing turbidity, reaction distance to a group of prey was longer than to one prey.

Based on existing models of fish vision and turbulence‐mediated ingestion rates, we develop a model of ingestion rates in larval fish that combines several physical properties of the environment (turbulence, irradiance, light attenuation) and visual characteristics of predators and prey. The model of visual range was calibrated with observed estimates in larval herring (Clupea harengus) and cod (Gadus morhua). The improved visual ability with length of larvae was predicted to be the most sensitive part of the model. Both turbulence and light had strong impacts on the ingestion rate of larval fish. The optimal level of turbulence increased with larval length. Also, due to the exponential decay of light with depth, it was predicted that larvae will have higher ingestion rates near the surface, even at high wind velocities. It is also proposed that larvae (particularly the smallest larvae) should concentrate on larger prey in turbulent environments. We suggest that biophysical models of larval growth and survival in field situations should include these factors to account for environmental effects on growth, survival and recruitment processes in the early life stages.

No abstract

Perch, Perca fluviatilis, of different sizes and ages were allowed to choose between two feeding patches differing in food ratio and (or) predation risk (presence versus absence of a pike, Esox lucius). The aim of our study was to test whether food-patch choice and risk-taking behaviour (time with the predator present) were related to size, phenotype (fast- versus slow-growing) and (or) stage of development (mature versus immature). Movements of individual fish were continuously monitored with an automatic system employing passive integrated transponder (PIT) tags. Food distribution did not seem to be important in determining perch distribution. Predation risk, however, had a significant influence. The presence of a predator in a patch significantly reduced the proportion of time spent in that feeding patch by all perch. However, the strength of this effect was dependent on individual characteristics, both the slow-growing phenotype and the mature fish spending a significantly lower proportion of time in the risky patch than the fast-growing and immature fish. Switching rates, which differed between character groups, also had an effect on growth rate, varying with predation risk.