Unmistakable

Kept safe not by blending in — by being recognised at a glance, and remembered

R823’s minnow is safest when it cannot be told from the nine beside it; its whole cover is being unremarkable, and its danger begins the moment it can be picked out. Turn the shape over once more. Some prey run the protection in exactly the opposite direction. They are safest precisely when they can be picked out — instantly, from a distance, by a hunter who met their kind once before and has no wish to repeat the lesson. Their danger does not begin when they are recognised. It begins when they cannot be.

What recognition does. Biologists call it aposematism: a defended prey — bitter, toxic, stinging, spraying — advertises that defence with a signal built to be learned and not forgotten: saturated reds and yellows, hard black bands, a shape unlike anything else on the ground. The signal looks like the opposite of a defence, because it raises the chance of being noticed. What it buys in exchange is a change in what noticing leads to. A cryptic prey survives an encounter by not starting one. An unmistakable prey survives by making the encounter end in a lesson — and every lesson a predator learns is protection for every later individual wearing the same colours, including ones the predator has never met.

Measured cleanly. The bug Lygaeus equestris comes in two forms of the same distasteful species: a normal red one and a grey mutant. Birgitta Sillén-Tullberg gave hand-raised great tits ten consecutive trials apiece, sixteen birds on the red form and sixteen on the grey, one attack allowed per trial, both forms presented at a known place against a background matching the grey — so detection was held equal by design, and colour was the only difference left standing. The birds learned to avoid both forms. But the red ones survived far better, for three separable reasons the paper lists: the birds were more reluctant to attack red prey in the first place, they learned avoidance of it faster, and when they did attack they killed it less often. More birds learned to leave the red form alone without killing a single one. A separate test found no difference in distastefulness between the forms, so the advantage was the colour itself. The author’s summary figure is the one this page’s toy is built on: survivorship of the red form was 6.4 times the grey’s — high enough that, on her arithmetic, the conspicuous form could afford to be detected 6.4 times as often and still come out even.

The field check: recognition is local. A laboratory lesson could be an artefact of cages, so consider the skunk, whose black-and-white is a signal with no cage around it. Jennifer Hunter placed taxidermy mounts at ten California sites — skunk-shaped and fox-shaped, some dyed grey, some dyed black-and-white — and filmed what came. Where wild skunks were common, experienced predators such as coyotes, bobcats, bears and mountain lions gave anything skunk-like a wide berth, whether the warning came from its colour or from its shape; either half of the signature was enough. Where skunks were uncommon, the same kinds of predator approached the mounts, licked them, rolled on them, tried to drag them away. The skunk’s body had not changed between the two places. Its protection had, because the protection was never entirely in the body: it lives in the memory of the local predators, and a place without the memory is a place where being unmistakable buys nothing yet.

The lesson can travel without the bite. The protection also spreads more cheaply than one victim per predator. In a Cambridge wood, researchers dispensed brightly dyed almond flakes — some tasty, some bitter — to wild blue tits and great tits fitted with RFID tags, then tracked who foraged with whom. Birds learned to avoid the bitter colour within eight days, adults faster than juveniles, and the avoidance flowed along the birds’ social links, especially from adults to juveniles: a bird could acquire the lesson by watching another bird’s disgust, without paying for the mouthful itself. Every such transfer is a predator educated at no cost to the prey population — the signal doing its work at a distance, through witnesses.

The honest narrowings. First, the signal must be backed. Work with naïve chicks and patterned artificial prey found that high-contrast patterns deterred first approaches, but prey that were approached anyway were eaten more over repeated encounters when no real defence stood behind the display — an unbacked warning is a lesson in the wrong direction, taught just as memorably. Second, the protection is perishable: it is stored in individual predators, so it thins wherever naïve predators keep arriving — each season’s young birds are, at first, a partial amnesia in the population that was supposed to remember. Third, the first step is the dangerous one: a novel bright form should, on the naive logic, be eaten before any predator has learned it, which is why the origin of warning colour was long treated as a paradox. Part of the resolution is that predators are conservative eaters in their own right: in field experiments with artificial pastry prey, even fully palatable novel colour morphs, starting at one in twenty, regularly persisted and spread to fixation under wild bird predation, protected at first simply by being unfamiliar. Being unmistakable and being novel are neighbours, and novelty itself is a small, temporary shelter on the way to being known.

Run the break-even. A schematic arithmetic toy, stated plainly. It does no simulation and invents no survival rates. It takes the one empirical figure above — the red form’s survivorship was 6.4 times the grey form’s, at equal detection — and asks the question the original paper asked: how much extra detection can that advantage pay for before being conspicuous stops winning? Set the detection multiplier and compare. Above 6.4×, the cryptic form wins on this arithmetic; at exactly 6.4×, they are even; below it, being recognised is worth more than being overlooked. The toy models that single trade-off only — not learning speed, not attack behaviour, not any real population.


A minnow is kept by nine others being exactly as visible as it is and no more interesting. A red bug is kept by the opposite arrangement: one lesson, learned once by one bird, standing guard over every red back that bird will ever see — a protection with no wall and no secret, held in somebody else’s memory, and good only for as long as it is remembered and only where it has been taught.