Supernormal stimuli

Why will a bird abandon its own egg to sit on an absurd oversized fake?

Put a plaster egg the size of a football beside an oystercatcher's own clutch and the bird will clamber onto the monstrosity and try to brood it, sliding off, climbing back, ignoring the perfectly good eggs it laid. Tinbergen and his colleagues found this sort of thing repeatedly, and it looks like straightforward stupidity.

But I want to resist that reading, because it assumes the bird was trying to identify an egg and got it wrong. What if it was never doing that at all? What would the bird have to be responding to, for the giant fake to be the correct answer?

Think about what recognition would actually cost. To identify "my egg" the bird would need to represent an object, compare it against a stored model, and reject mismatches. That is expensive machinery, and in the environment where the behaviour evolved it would be solving a problem that never arises. Nothing in a salt marsh produces a speckled ovoid except an oystercatcher.

So what is the cheap alternative? Pick out one or two features that in practice only eggs possess, and respond to those. Ethologists call such a feature a sign stimulus or releaser, and the machinery that fires on it an innate releasing mechanism. The bird is not asking "is this my egg". It is running something closer to "in the nest scrape, curved, speckled, larger is stronger — sit".

Now hold that last clause up to the light, because everything follows from it. If the response is tuned by size, in which direction is it tuned? Within the natural range, a bigger egg genuinely is a better bet: more yolk, a larger chick, and no risk of confusion with anything else. So selection builds a response that increases with size. And here is the crucial point — what would ever have caused selection to build in a ceiling? A cut-off at the top of the natural range is extra machinery that never once earned its keep, because nothing that large ever existed.

Do you see the consequence? The rule is not merely fallible at the extreme; it is unbounded at the extreme, and the experimenter walks straight into the gap.

The same shape appears wherever anyone has looked. Herring gull chicks peck at a red spot on the parent's bill to solicit food; Tinbergen found that a thin rod bearing red and white bands — not remotely gull-shaped — drew more pecking than a realistic model head, because the response keys on contrast and shape of the marked stimulus rather than on "parent". Male jewel beetles in Western Australia were recorded by Gwynne and Rentz in 1983 attempting to mate with discarded brown beer bottles, whose size, colour and dimpled reflective surface exaggerated the cues of a female. Graylag geese roll a stray egg back into the nest with a stereotyped bill movement, and will do it for objects no goose could have laid.

So the bird is not failing at perception. It is succeeding at a proxy, and a proxy is only as good as the correlation between the cue and the thing. Break the correlation — which a laboratory, or a bottle factory, does effortlessly — and the behaviour continues to be executed perfectly while pointing at nothing.

Does that not invert the usual framing? The absurdity is not in the bird's response. It is in the mismatch between the world the rule was built for and the world it now runs in.

Three refinements are worth making, and one caution.

First, "the bird has no concept of an egg" is an inference from behaviour, not an observation of a mind. Many birds do learn their own eggs — notably hosts under pressure from brood parasites like cuckoos, where the cost of a mistake is enormous and egg-recognition abilities are demonstrably sharp. The oystercatcher's indifference tells us that in its ecology the pressure to discriminate was weak, not that discrimination is impossible for birds.

Second, the effect is not universal or unlimited even where it occurs. Responses generally rise with the cue and then fall away when the stimulus becomes too different in other respects, and several classic ethological demonstrations have proved harder to replicate cleanly than their textbook retellings suggest. It is a real and repeatable phenomenon, not a universal law of nervous systems.

Third, the popular extension of the idea — that engineered foods, cosmetics or media are supernormal stimuli hijacking human releasers — is a suggestive analogy, not an established finding. The mechanism is plausible and the evidence is much thinner than the confidence with which it is usually asserted, and it is worth keeping those apart.

• How brood-parasite hosts evolve genuine egg recognition, and what that costs them.
• Whether peak shift in learning produces the same unbounded preference by a different route.
• How sensory bias shapes the evolution of signals, when senders exploit a receiver's existing tuning.

Every term the collection defines is gathered in the glossary.