Eyes in low-light environments typically must balance sensitivity and spatial resolution. Vertebrate eyes with large "pixels" (e.g. retinal ganglion cells with inputs from many photoreceptors) will be sensitive but provide coarse vision. Small pixels can render finer detail, but each pixel will gather less light and thus have poor signal relative-to-noise, leading to lower contrast sensitivity. This balance is particularly critical in oceanic species at mesopelagic depths (200-1000 m) because they experience low light and live in a medium that significantly attenuates contrast. Depending on the spatial frequency and inherent contrast of a pattern being viewed, the viewer's pupil size and temporal resolution, and the ambient light level and water clarity, an optimal visual acuity exists that maximizes the distance at which the pattern can be discerned. We develop a model that predicts this optimal acuity for common conditions in the open ocean, and compare it to visual acuity in marine teleost fishes and elasmobranchs found at various depths in productive and oligotrophic waters. Visual acuity in epipelagic and upper mesopelagic species aligned well with model predictions, but species at lower mesopelagic depths (>600 m) had far higher measured acuities than predicted. This is consistent with the prediction that animals found at lower mesopelagic depths operate in a visual world consisting primarily of bioluminescent point sources, where high visual acuity helps localize targets of this kind. Overall, the results suggest that visual acuity in oceanic fish and elasmobranchs is under depth-dependent selection for detecting either extended patterns or point sources.

The data analyzed here are primarily derived from a literature synthesis in which we searched for published measures of visual acuity (measured morphologically) in fishes. We also searched for published daytime depth ranges and values of lens diameter for each species. Fish were assigned habitat designations based on information from FishBase.org and confirmed by author TS. Some data are also the result of an optical model created to determine optimal visual acuity under a set of visual parameters. Lastly, phylogenies used to make Figure 3 were derived from Rabosky et al. 2018 (for teleost fishes) and VertLife.org (for elasmobranchs).

All files are .csv files, openable with Microsoft Excel. Custom analysis codes can be run using the open-source R coding software. Funding provided by: National Science FoundationCrossref Funder Registry ID: http://dx.doi.org/10.13039/100000001Award Number: OCE-2154144Funding provided by: National Oceanic and Atmospheric AdministrationCrossref Funder Registry ID: http://dx.doi.org/10.13039/100000192Award Number: NA19NOS4510193Funding provided by: Swedish Research CouncilCrossref Funder Registry ID: http://dx.doi.org/10.13039/501100004359Award Number: 2021-04917