Night vision in numbers: optical sensitivity of animal eyes

The values behind who sees best in the dark: the optical sensitivity S of one rod or rhabdom to an extended white scene, for 102 species, against the dark-adapted human eye (S = 0.93 µm² sr). 22 are calculated from aperture, focal length and receptor size all measured in the species; 80 (sharks and lanternfishes) use a measured receptor size with an F-number taken from a group constant or the Matthiessen ratio, labelled "estimated F-number".

Method

• S = (π/4)² × (A/f)² × d² × (k l / (2.3 + k l)) µm² sr (Land 1981; Warrant & Nilsson 1998), as each source computes it; A aperture, f focal length, d and l receptor diameter and length, k absorption coefficient. The species value is the median of its single-receptor values.
• Night dial = log10(S / S_human). Render gain in stops = log2(S / S_human) = dial × 3.32, clipped to ±8 stops. The slider's brightness is min(1, 2^(gain − darkness)), with darkness counted in halvings of the light from the point where human eyes start to lose detail; the app's night mode uses the same gain.
• Cone S and S values summed over ganglion cells are not used: they are not comparable with single-receptor S. Behavioural, ERG and VEP thresholds are not mapped onto the dial.
• S is a property of the optics and receptors. It leaves out neural summation, integration time and receptor noise. In frogs, the scotopic threshold tracks S closely (Rosencrans et al. 2018), which supports S as the first-order measure for vertebrates.
• Species without a measured S keep the engine's activity rule (activity pattern, reflective layer, pupil shape), which is not a measured sensitivity.
• 3 species in the verified table are not in the catalogue, so they are left out here: Camponotus consobrinus, Onitis aygulus, Onitis belial.

Back to the night-vision slider · How it works · All sources