Pit viper infrared sensing

How can a snake with no light-sensitive pigment for heat still strike an unseen mouse accurately in total darkness?

A rattlesnake in a sealed dark room, blindfolded, still strikes a mouse and hits it. The obvious explanation is that it has a second kind of eye — one whose pigment is tuned to heat rather than to light. That explanation is wrong, and it is worth seeing why it is wrong before asking what the snake actually does, because the reason it cannot work turns out to dictate the shape of the organ that replaced it.

What is actually arriving from the mouse? A body at about 310 K radiates as a warm object does, and Wien's law puts the peak of that emission at 2898 micrometre-kelvin divided by 310 K — roughly 9.4 micrometres. Now ask what one photon of that wavelength is worth. Photon energy is 1240 electronvolt-nanometres divided by the wavelength, so 1240 / 9400 is about 0.13 eV.

Compare that with what vision costs. Bending the retinal molecule inside a visual pigment takes something near 2 eV, which is why the eye's window closes around 700 nanometres. A thermal photon carries roughly a fifteenth of that. No pigment chemistry can be built on it — and here is the deeper problem. Suppose you designed a molecule that could be flipped by 0.13 eV. Ordinary thermal jostling at body temperature supplies kT, which is 8.6 x 10^-5 eV per kelvin times 310 K, about 0.027 eV, in every collision. A detector sensitive to a fifth of that would be firing constantly on its own warmth. The photon route is not merely unused; at these energies it is unusable.

So the snake cannot detect the photons individually. What is left? Detecting them collectively — as heat. Radiation landing on a surface warms it, and warming can be measured. Everything about the pit organ follows from making that measurement fast and sensitive.

Consider what limits it. The temperature rise of an absorbing surface is the absorbed power divided by its thermal mass, less whatever leaks away by conduction. So you want the absorber to be as light as possible and as thermally disconnected from the animal as possible — because the head is a large reservoir at the snake's own temperature, and any absorber bolted to it will simply be held at that temperature. That is precisely what the pit is: a chamber under the eye, spanned by a membrane roughly fifteen micrometres thick — the figure is recalled — suspended with air on both faces, so it hangs thermally isolated in the middle of a hole. Skin on the head would average the signal away; a film with almost no mass, touching almost nothing, does not.

Behind that membrane sit terminal masses of the trigeminal nerve, unusually stuffed with mitochondria. The channel that opens when they warm is TRPA1 — in these snakes a heat-gated channel, in mammals better known as the receptor irritants like mustard oil act on. That matters for the story: the organ is not a new sense invented from nothing. It is an existing temperature-and-irritant channel, retuned, sitting under a modified pit in the skin.

One thing is still missing. Warming tells the snake there is something hot; it does not tell it where. The answer is geometric and rather crude: the pit's opening acts as a pinhole, so a warm object off to one side warms one patch of membrane rather than another, and direction becomes position. The aperture is wide relative to its depth, so the projected image is badly blurred — estimates of the angular resolution depend heavily on the modelling assumptions, and I would not defend a specific number. Sharpening happens afterwards, in the nervous system: pit input reaches the optic tectum and maps there in register with the visual map, so the same neurons can be driven by either channel.

Three refinements are worth stating plainly.

First, the organ measures contrast, not temperature. A membrane in radiative balance with its surroundings reports nothing. A mouse is detectable because it is warmer than the leaf litter behind it — which is why the sense works badly on a hot afternoon and well at night, and why a target at ambient temperature is invisible to it however solid it is.

Second, do not let the phrase "infrared vision" import optics that are not there. There is no lens, no focus, and no pigment. The comparison to a camera is a comparison to the very worst kind of camera, and most of the apparent acuity is a downstream reconstruction rather than a property of the image.

Third, the honest evolutionary statement is narrower than the usual one. Saying the pit "evolved for hunting warm prey" outruns the evidence: the sense is also used in thermoregulation and predator detection, and the underlying channel long predates the organ. What is well supported is that a general-purpose thermoreceptor was co-opted, and that the structure around it — thin, suspended, apertured — is explained by the physics above rather than by any story about intent.

The claim is testable in an unusually clean way. Block both pits and leave the eyes free, and a snake should still strike accurately in the light; blindfold it and leave the pits open, and accuracy in the dark should survive. Do both and it should fail. Experiments of exactly this design on rattlesnakes support the split. The refuting observation is simple: a snake with both pits sealed that still struck a warm target accurately in complete darkness would leave nothing of this account standing, since the only other candidate channels — chemical and vibrational — do not deliver the direction fast enough for a strike.

• How pit and visual maps are brought into register in the tectum, and what happens when they disagree.
• Why boas and pythons converged on labial pits with a different structure but the same channel family.

Every term the collection defines is gathered in the glossary.