Fifteen popular claims about how animals see, each checked against the published research. Every answer
names its papers, and each one links to a scene where you can test it yourself.
No. Dogs see color, with one channel fewer than you.
MEASURED A dog has two kinds of color cone where you have three: a dog's peak at about
429 nm and 555 nm, yours at about 419, 531 and 558 nm. So reds and greens fall together for a dog, much as for a
red-green colorblind person, and in tests dogs acted just like one.
MEASURED A wedge-tailed eagle told stripes apart down to 132 to 143 stripe pairs per
degree in bright light. People in the same kind of test reached 60. That is 2.2 to 2.4 times.
NEEDS SOURCE "4 to 8 times sharper": we could trace it only to a popular book. A 2020 review
says the 8-times claims "are definitely not true".
MEASURED At sunset, an eagle's sharpness is about a third of its best.
No. No eye can see without light. A cat needs about 5.5 times less light than you.
MEASURED After an hour in the dark, 6 cats still picked the faintly lit door of two at
a light level that 2 people in the same box couldn't use. The paper: "lower by a factor of almost 6". A modern
check with 2 cats and 3 people found the same: 5.5 times.
Sources: Gunter (1951), Journal of Physiology; Kang, Reem, Kaczmarowski & Malpeli (2009), Journal of Neurophysiology.
No. 130 is how much more light it reflects, not how well it sees.
NEEDS SOURCE The 130 is how much more light the back of a cat's eye reflects than yours.
We could not open the book it comes from.
INFERRED The mirror (the tapetum) sends light that missed the rods back through for a
second pass. Worked out from measurements, that gives them about 29% more light.
Sources: Kang, Reem, Kaczmarowski & Malpeli (2009), Journal of Neurophysiology; Banks, Sprague, Schmoll, Parnell & Love (2015), Science Advances.
No. 12 kinds of color receptor, and they aren't cones.
MEASURED 12 kinds of color receptor, from just below 300 to above 700 nm. In some
species, 16 kinds of receptor in all, counting ones for ultraviolet, polarization and fine detail.
NEEDS SOURCE "16 color cones" comes from a popular comic.
INFERRED More receptors didn't mean finer color. Trained to strike one of two colored
lights, a mantis shrimp failed once the two were closer than 15 to 25 nm. People: 1 to 8 nm. (We read these numbers
secondhand, in a later paper from the same lab.)
Sources: Thoen, How, Chiou & Marshall (2014), Science, as reported by Wang & Marshall (2025), Journal of Experimental Biology; Cronin, Bok, Marshall & Caldwell (2014), Philosophical Transactions B.
NEEDS SOURCE "350°" traces only to a popular book about trail riding.
INFERRED Textbook figures (each eye 190° to 195°, the two overlapping by 55° to 65°)
add up to 315° to 335°, and come with no method we could find.
MEASURED In the one behavior test we found, 3 horses noticed objects at every position
tested, out to nearly straight behind them, but told them apart only out to about 138°.
Sources: Hanggi & Ingersoll (2012), Behavioural Processes; Timney & Macuda (2001), Journal of the American Veterinary Medical Association.
MEASURED Away from where you're looking, clutter jumbles things together: an object gets
lost if anything else sits within about half its distance from where you're looking. This crowding, not blur, is
most of what side vision loses.
MEASURED A famous demo of "what your eye really sees" blurs the edges about 4 times too
much, by one researcher's estimate.
Sources: Rosenholtz (2016); Bouma (1970); Pelli et al. (2004).
INFERRED It comes from a review that says: "If one assumes from all these studies that
visual acuity in the typical dog is about 20/75".
MEASURED In a 2017 test, 6 dogs in bright light told stripes apart down to 5.5 to 19.5
stripe pairs per degree (the sharpest was a pug). People in the same test: 32 to 44.
Sources: Miller & Murphy (1995), Journal of the American Veterinary Medical Association; Lind, Milton, Andersson, Jensen & Roth (2017), PLoS ONE.
INFERRED A famous 1995 finding said kestrels track voles by the ultraviolet in their
urine. Checked later against how raptors' eyes were measured to work, the urine was found "unlikely to provide a
reliable visual signal". Raptors' eyes let less than 30% of near-ultraviolet light through.
Sources: Lind et al. (2013), Journal of Experimental Biology; Lind, Mitkus, Olsson & Kelber (2014), Proceedings of the Royal Society B; Potier, Mitkus & Kelber (2020).
Not in any test we found. Yet their camouflage matches color.
MEASURED The common cuttlefish's eye has a single light-catching pigment, most
sensitive at 492 nm. On a checkerboard whose two colors were matched in brightness for its eye, it stayed plain,
as if the board were one flat color. It did react to squares that differed in brightness by about 15%.
NEEDS SOURCE How it matches colors it cannot tell apart: we found no test in a living
animal that settles it. One idea, that an off-center pupil lets it tell colors apart by how blurry they are, is a
computer model.
Yes: the only insect known to. But its 3D vision runs on movement.
MEASURED So far, the praying mantis is the only insect known to have stereo vision:
judging distance by comparing what its two eyes see. Its stereo vision looks for where the light is changing in
each eye, not for matching patterns.
MEASURED "There is no evidence that mantis stereopsis works at all with static
images." That is an absence of evidence, not proof it can't. A mantis can still see a still fly, and it judges how
far away still things are by swaying its head from side to side.
Sources: Rossel (1983), Nature; Nityananda et al. (2018), Current Biology; Poteser & Kral (1995), Journal of Experimental Biology.
Yes. It is one of a bee's three colors, not a secret extra layer.
MEASURED A honeybee has three kinds of color receptor, like you, but shifted: they
peak at about 344 nm (ultraviolet), 436 nm (blue) and 544 nm (green). Yours peak at about 419, 531 and 558 nm.
Ultraviolet is a real color to a bee: it tells it apart from every other light.
INFERRED But it is not a hidden layer on top of the others. One review calls
ultraviolet, in most cases, no more important than blue, green or red. And you can't see it mostly because the
lens of your eye absorbs it.
Sources: Peitsch et al. (1992), Journal of Comparative Physiology A; Kevan, Chittka & Dyer (2001), Journal of Experimental Biology; Douglas & Jeffery (2014), Proceedings of the Royal Society B.
Nearly: about 330° to 350°, and only with their coarse side eyes.
MEASURED The side eyes cover about 350° together (31 spiders of one species, mapped
from their turns toward a moving dot).
INFERRED Those side eyes are coarse; they catch movement. The sharp front pair sees
only a narrow strip, and it is about 5 times coarser than your best: still the sharpest known for an animal that
size.
Sources: De Agrò, Rößler & Shamble (2024), Journal of Experimental Biology; Shepeleva (2022), Vision.
Not in the test we found. Their heat sense is real, and very blurry.
MEASURED Western rattlesnakes watched a warm target cross a striped wall whose average
temperature equaled the target's. They showed no clear reaction on striped walls, and reacted strongly on plain
ones: "no evidence" that the pit can make out details of 9° or less.
MEASURED In a field test, sidewinders with their eyes covered wandered more and came
across landmarks less often; with their pits covered, they were unaffected.
Sources: Clark, Bakken, Reed & Soni (2022), Journal of Experimental Biology; Schraft & Clark (2019), Animal Behaviour.
Each answer above is taken from the Sources panel of the scene it links to, where every claim is labeled
MEASURED, INFERRED, MODELED or
NEEDS SOURCE, with the full reference. Found a mistake?
Tell us, and we'll fix it and list the correction.