Defining visual acuity in the common smelt (Retropinna retropinna) through anatomy and feeding behaviour

Abstract

A fish retina is composed of eight interrelated layers. The unique topography of two of these layers (cone and ganglion cell) is thought to reflect the role that vision plays in the behavioural ecology of any given fish species. Visual acuity is a function of the density of the cone and/or ganglion cells within these layers and high density regions define areas of greater visual acuity. The present study used a fluorescent dye (Dil) to visualise the retinal cells within the eight retinal layers of the common smelt (Retropinna retropinna) and the confocal laser scanning microscope to map the density contours of both the cone and ganglion cell layers simultaneously within a whole-mounted retina. The results confirm a general correspondence in density contours between the two layers in later juvenile through to adult stages with a region of highest density in the ventro-temporal quadrant of the retina. However, the two layers do not correspond to each other in the early juvenile stage. The convergence ratio of cone:ganglion cells remained relatively constant across the retina. Estimates of visual acuity in the smelt were made using two histological measures (density of both cone and ganglion cells in the retina) and two behavioural measures (optomotor response and reaction distance). Maximum visual acuity determined histologically using cone cell density was 9.7' and 14.2' in adult and juvenile fish, respectively, while it was 9.9'-13.2' and 12.0'-16.0' when determined behaviourally using reaction distance. Estimates of maximum visual acuity obtained using the optomotor response or ganglion cell density were considerably lower and it was concluded that histological acuity based on cone cell counts and behavioural acuity based on reaction distance provided equally accurate estimates of the fishes' visual acuity. Visual acuity increased with growth and the region of highest resolving power became narrower and more constrained within the ventro-temporal quadrant of the retina. This was reflected in the adult smelt's prey reaction field determined during feeding experiments in that food particles were taken predominantly forward and above the fish's forward-directed axis. Thus, when results from both the behavioural and histological studies are compared they clearly illustrate that the foraging behaviour of R. retropinna is predictable from a topographic analysis of the density of cone and ganglion cells within their respective retinal layers.