Batesian & Mullerian Mimicry

What mimicry is, in one picture

Picture a bird that has just thrown up an orange-and-black monarch butterfly. For the next several weeks it will refuse anything orange-and-black on sight. Now two very different things can ride on that lesson. A harmless butterfly can evolve the same orange-and-black livery and get a free pass it never earned — that is Batesian mimicry, named for Henry Walter Bates, who noticed in the Amazon in the 1850s that palatable butterflies were wearing the colors of distasteful ones. Or a second genuinely toxic butterfly can adopt the same livery, so the two species jointly teach the bird and split the price of the few individuals it sampled before learning — that is Mullerian mimicry, named for Fritz Muller, who in 1879 gave the first quantitative argument in all of evolutionary biology.

The whole phenomenon rests on a single fact about predators: they are not born knowing which prey are dangerous. They learn it, by sampling. A warning signal — a bright color, a rattle, a foul smell — is only useful because predators can learn to associate it with punishment and then generalize that avoidance to anything that looks similar. Mimicry is evolution exploiting that generalization. A Batesian mimic exploits it parasitically (it pays nothing and dilutes the lesson); Mullerian mimics exploit it cooperatively (they all pay, and sharing one signal lowers everyone's tuition).

How the mechanism works, step by step

Strip mimicry down to its moving parts and it is a four-stage feedback loop between prey signal and predator brain:

1. A real defense exists. Some species is genuinely costly to eat — a monarch loaded with milkweed cardiac glycosides (cardenolides) that block the Na⁺/K⁺ ATPase and trigger vomiting, a coral snake with a neurotoxic venom, a wasp with a sting. Without a real cost, there is nothing for a predator to learn and nothing worth copying.
2. That defense is advertised — aposematism. The defended species pairs its toxin with a loud, contrasting, easily remembered signal: high-contrast reds, oranges, and yellows against black; bold repeated stripes or spots. Conspicuousness and regularity are not accidents — they make the signal fast to learn and slow to forget.
3. A predator learns by sampling. A naive bird, lizard, or mammal eats a few aposematic prey, gets sick or stung, and forms an association. In blue-jay experiments, a single emetic monarch can be enough; for milder defenses it takes a handful of trials. The predator then generalizes: it avoids not just the exact individual it ate but everything that looks close enough.
4. A copy slips into the protected zone. Any mutant — palatable (Batesian) or defended (Mullerian) — that resembles the established warning pattern inherits a slice of the predator's learned avoidance. Selection then sharpens the resemblance generation after generation, because the closer the copy, the more of the avoidance it captures.

The split between Batesian and Mullerian falls out of one question: does the copy also punish the predator? If not, it is a parasite on a finite reputation, and its success is self-limiting (step into the frequency section). If yes, it is a partner, and every added defended individual makes the shared signal stronger — Muller's insight that two defended species each lose fewer individuals together than either would alone.

The players and the conditions

Mimicry systems have a fixed cast of three roles, and which role a species plays determines everything about the outcome:

• The model. The genuinely defended species whose signal is being copied — the monarch, the velvet ant, the coral snake. The model carries the real cost of the defense and, in Batesian systems, pays a hidden price too: every harmless mimic erodes its reputation.
• The mimic. The copier. In a Batesian system the mimic is palatable and contributes nothing; in a Mullerian system the "mimic" is itself a model — the labels become symmetric, because each species is simultaneously copying and being copied.
• The dupe. The predator whose learning and memory make the whole thing work. No learning predator, no mimicry. The signal is tuned to the predator's vision (birds see into the near-UV and into the red, which is why so many warning patterns are red-on-black), its memory span, and its tendency to generalize.

For Batesian mimicry to be evolutionarily stable, four conditions must hold: the model must be common enough and defended enough that predators keep the lesson fresh; the mimic must stay relatively rare; predator and mimic must share the same habitat and active period as the model; and the resemblance must be close enough to fool the predator's generalization. Break any one — let the mimic outnumber the model, or let the model go locally extinct — and the protection collapses. Mullerian mimicry is far more robust: because every participant is defended, there is no dilution problem, and the system tolerates large numbers and even imperfect resemblance.

Batesian vs Mullerian mimicry compared

Property	Batesian mimicry	Mullerian mimicry
Mimic's defense	None — palatable	Real — toxic/stinging/distasteful
Relationship type	Parasitic (mimic exploits model)	Mutualistic (cost shared)
Effect of mimic on model	Harmful — dilutes the warning	Beneficial — reinforces the warning
Frequency dependence	Negative — rare mimic wins, common mimic loses	Positive — more individuals strengthen the signal
Stable mimic abundance	Must stay rarer than the model	No upper limit; numbers help everyone
Resemblance pressure	Strong — must precisely fool predators	Relaxed — even rough convergence pays
Classic example	Hoverflies mimicking wasps	Heliconius erato ↔ H. melpomene
First described	H. W. Bates, 1862	Fritz Muller, 1879

The numbers: how strong is the payoff

Mimicry is one of the few areas of evolutionary ecology where the selective advantage has been measured directly, because you can paint butterflies and release them.

• Learning is fast. Hand-reared birds typically need only 1–5 sampling encounters to form a durable aversion to a conspicuous pattern; a single intensely emetic monarch (carrying roughly 100–800 µg of cardenolides) can produce one-trial avoidance lasting weeks.
• Resemblance pays in survival. In mark-release-recapture transplant experiments with Heliconius, butterflies translocated into a region where the local predators have learned a different pattern suffer markedly higher attack and lower recapture than residents wearing the locally familiar pattern — the survival penalty for wearing the "wrong" warning livery is large and repeatable.
• The dilution threshold is real. Predator-learning experiments with artificial prey show that as the fraction of palatable mimics rises past roughly half the warningly colored population, predators resume attacking the signal and the mimics' protection erodes — a direct demonstration of negative frequency dependence.
• The genetic switch is small. Much of Heliconius convergence is controlled by a handful of large-effect loci — optix (red), WntA (black pattern boundaries), and cortex (yellow/white). A few major genes flipping the same way in unrelated lineages is what lets distant species snap onto an identical pattern.
• Timescale. Bates collected the Amazonian butterflies that seeded the idea over an 11-year expedition (1848–1859); the convergence those species show has been built over millions of years, yet pattern shifts between adjacent mimicry rings can be sharp enough to map across a few kilometers of forest.

Where it shows up in the wild

• Heliconius butterflies (Mullerian). The flagship system: dozens of co-occurring, independently cyanogenic species in Central and South America converge into geographic "mimicry rings," with H. erato and the distantly related H. melpomene tracking each other's red-black-yellow patterns across their whole range.
• Monarch and viceroy (now Mullerian). The viceroy (Limenitis archippus) was the textbook Batesian mimic of the toxic monarch (Danaus plexippus) until a 1991 experiment showed viceroys are themselves unpalatable — reclassifying the pair as Mullerian co-mimics and becoming the standard warning against assuming a look-alike is harmless.
• Hoverflies and wasps (Batesian). Stingless, harmless hoverflies (Syrphidae) wear the yellow-and-black banding of stinging wasps and bees. Their resemblance ranges from near-perfect to laughably crude, which has made them a model system for asking why imperfect mimicry persists at all.
• Coral snakes and kingsnakes (Batesian). Harmless scarlet kingsnakes (Lampropeltis elapsoides) copy the red-yellow-black banding of venomous coral snakes (Micrurus). Field experiments with plasticine replicas show the mimicry only protects where coral snakes actually occur — and breaks down at the geographic edge of the model's range, exactly as Batesian theory predicts.
• Drone flies and honeybees, robber flies and bumblebees, ant-mimicking spiders and bugs. Batesian mimicry of stinging Hymenoptera is one of the most repeatedly evolved disguises in the animal kingdom.
• Beyond color. Some tiger moths produce ultrasonic clicks that mimic the acoustic warning of genuinely toxic moths to deter bats — Batesian mimicry in the acoustic channel rather than the visual one.

Common misconceptions and pitfalls

• "A look-alike must be a harmless cheat." The monarch–viceroy reversal is the cautionary tale. You cannot classify a system as Batesian without actually testing the mimic's palatability — many supposed Batesian mimics turn out to be defended, making them Mullerian.
• "The mimic is consciously imitating." Nothing is imitating anything on purpose. Mimicry is the cumulative result of predators differentially killing the worst copies; resemblance is built by selection over generations, not by an organism choosing to look like another.
• "Mimicry needs near-perfect resemblance." Plenty of Batesian mimics are crude, and they still gain protection because predators generalize and because attacking even a possible wasp is risky. Mullerian convergence can be rougher still, since every participant punishes mistakes.
• "More mimics is always better." True for Mullerian systems, false for Batesian ones. A Batesian mimic that becomes too common destroys its own protection by teaching predators that the pattern is often safe — the defining negative-frequency-dependent trap.
• "Camouflage and mimicry are the same thing." Camouflage (crypsis) hides you from being detected; aposematic mimicry does the opposite — it makes you maximally conspicuous so you will be recognized and avoided. They are opposite anti-predator strategies. (Masquerade — looking like an inedible object such as a twig or bird dropping — is a third, separate category.)
• "Aposematism is free to start." The hardest part of the whole story is the origin of the warning signal itself: the first conspicuous mutant is easier to find and gets eaten before any predator has learned. Mimicry can only get going once some species has already paid to build a recognizable warning.