There is a nocturnal mammal, about the size of a common red fox, which has evolved (by way of some unspecified-at-this-point selective pressure) the ability to see well in situations involving large contrasts. More specifically, an ability to make out details in areas with large differences in luminosity subtending very small angles as seen from the creature's point of view.

Something like: If you point a flashlight at it at night, it is able to make out not just the lamp in the flashlight, but also your fingers grasping the flashlight, from some non-negligible distance.

Two questions:

  • What adaptations to the creature's eyes (and more generally its visual system) would allow such vision?
  • What other effects on the creature's eyesight might (would) those adaptations have?

2 Answers 2


We already have this. The human eye can see some mighty wide dynamic ranges. That's why scenes can look beautiful, but when we snap a picture, we find the picture is heavily saturated. The dynamic range of the human eye is about 10-14 fstops, or a ratio of about 1,000,000:1. The equivalent level of sensitivity for touch would be being able to lift a car, and then be aware of a feather dropping on top of it!

General purpose dynamic range is very difficult to attain, because you need a very sensitive sensor and a very wide bandwidth channel for the data. Typically animals adapt to the sorts of high dynamic range scenes they expect to need to operate in. One major example of this is the need to be able to see a bright sunlit field and yet still see things resting in the shadows. To support this, human eyes (and all eyes that I am aware of) have a highpass filter built into the retina. As a photo receptor is hit by light, it inhibits nearby receptors so our retina really shows the difference in light across the scene. The brain then stitches this together. You can see this effect by staring directly at a point without letting your eye move. Eventually your entire vision will turn grey as your brain loses track of the overall lightness of the image. In fact, your eyes undergo a small jitter on a regular basis, simply to refresh the image that you were staring at! Eyes are marvelous things.

Another approach would be occlusion. This is actually used on some of our telescopes to blot out a bright star to explore for dim stars or exoplanets around it. In this case, the telescope designers knew that there would generally be one circular bright object and a bunch of dim stuff around it, so they put a physical barrier to occlude the bright star's light from reaching the sensor.

You could do similar if our optic nerve had a reverse path back to the retina that we could use. Your brain could "paint" an occlusion by sending signals to parts of the retina to attenuate themselves. Of course, this would only work in scenes where the brain could accurately predict which areas of the retina needed occluding, but it wouldn't be hard to develop a scenario where your flashlight holding example was realistic.

  • $\begingroup$ "General purpose dynamic range is very difficult to attain, because you need a very sensitive sensor and a very wide bandwidth channel for the data." What if less fidelity in the dynamic range was an acceptable tradeoff? Like, say you'd be able to transmit a constant number of "brightness levels", then your needed bandwidth wouldn't increase in the case of bigger dynamic range, just the steps between the brightness levels. $\endgroup$
    – Vaesper
    Nov 3, 2016 at 20:19
  • 1
    $\begingroup$ @Vaesper You mean like gamma encoding? We do that too =) The truth is, our vision is really a work of art. Improving on it is not easy. You typically have to do a trade off analysis to see what you're willing to give up in exchange for your goals. (For example, the membrane on the eyes of cats that lets them see in the dark better also limits their best-case eyesight during the day) $\endgroup$
    – Cort Ammon
    Nov 3, 2016 at 20:33
  • $\begingroup$ Cat's iris, lens system, and rod distribution (read: optics and sensor) account for their great nightvision; Their nictitating membrane does not. livescience.com/40459-what-do-cats-see.html $\endgroup$
    – The Nate
    Nov 4, 2016 at 3:35
  • $\begingroup$ @TheNate I forgot the name of the membrane I was thinking of, so I left it out. I'm thinking of the silvery one behind the retina which makes cat's eyes glow in the dark. $\endgroup$
    – Cort Ammon
    Nov 4, 2016 at 4:04
  • $\begingroup$ Ah. That's the lucida tapitum (L "Colorful tapistry") Not technically a membrane, but it is a separable layer on there. Many mammals have a version of that, including cows, btw. There are those who think it helps with night vision, but I've never seen research. (Alas, I'm out of touch with my former confederate who actually performed such sensory research... That would be too convenient, right?) I've seen conjecture, which that link references, along those lines, but I don't personally believe it given the mechanics. $\endgroup$
    – The Nate
    Nov 4, 2016 at 4:17

Our eyes have a dynamic gain, taking several minutes to adjust. So, the bright light will overblow the night vision and reset the range to spoil the night vision.

An adaptation to allow simultanious vision at different brightness levels would be to have different pixels perminantly set for different levels, and the dark-range pixels are not hurt by over exposure.

We already have separate rods and cones, and the rods are ignored when they go beyond their brightest possible setting. So it would be a refinement of this system to have the rods stay in a darker setting only, and cones adjust to brightest elements being viewed. Having two different input channels perceived as different would add visual contrast, in the same manner as being different colors: the “rod colored” fingers would be seen clearly against the bright (normal color) light, with only bloom to prevent seeing the edge of the fingers (one pixel width) immediatly adjecent to the light.

You were specifically asking about an adaptation to the real mammalian eye. Further separating rods and cones seems to be the way to go forward with what we have already. If you wanted to design an optical instrument from scratch, there is a wider range of possibilities.

  • $\begingroup$ All correct with the assumptions made, but you might also consider redundant/distinct eyes. This would allow a different iris threshold and rod/cone optimizations and NOT require redefining the calculations to accommodate the refinements, apart from incorporating the extra sensors as a gestalt. (Internally, there are some epithelial cells that build up in response to overexposure, limiting light while the retina heals.If you could beam-split and have a high-exposure section that runs through a deeper filter and let less through to the rest, say, maybe that's something to play with.) $\endgroup$
    – The Nate
    Nov 4, 2016 at 3:43
  • $\begingroup$ I suppose I should clarify: I mean, the genetics could be almost completely duplicated with alternate molecules for certain regulations or for the sensors or both. If all you need is a second set of eyes with a slightly different threshold for pigmentation and some stain in the vitreous humor to cut down on permeability, the changes required are mostly at the fetal development stage. $\endgroup$
    – The Nate
    Nov 4, 2016 at 4:21
  • $\begingroup$ @TheNate I'm having a harder time seeing (haha) how a creature would evolve extra eyes, than seeing how it would evolve changes to their existing eyes, but if you have an alternative answer, then please do post it as an answer. Comments are meant to seek clarification or suggest ways to improve a post, not for proposing entirely new answers. $\endgroup$
    – user
    Nov 4, 2016 at 8:53
  • $\begingroup$ It's not an entirely new answer, far as I understand the question, since it only defines a slightly different way to isolate the dim and bright sensors, but if you think its enough of a difference, I might post it anyhow. $\endgroup$
    – The Nate
    Nov 4, 2016 at 8:56

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .