Rethinking Bad Taste

How much mimicry is outright cheating?

Talking to evolutionary biologists Hannah Rowland and Mike Speed can shake your faith in a supposedly settled area of science. Generations of textbooks have presented animal mimicry as one of the marvels of evolution, allowing two species to confound their predators by looking alike. Marvel of evolution it is, but surprisingly for such a high-profile example, researchers still have a lot of questions about how mimicry works.

CHOOSE ONE. A Pseudoceros flatworm (top) has some chemical defenses of its own but also scares away predators by looking like the more noxious sea slug Chromodoris preciosa (bottom). Gosliner

NOVEL LUNCH. A great tit surveys “novel world,” set up here to test how good- and bad-tasting mimics affect each other’s chances of being eaten. In the novel world, prey “species” are tiny envelopes. L. Lindström/Univ. of Jyväskylä

FAKERS. In an illustration from the 1862 article that launched the modern study of defensive mimicry, Henry Walter Bates noted that butterflies that weren’t closely related could look like each other. The top and third rows show Dismorphia species and the other rows show Ithomiini. Wikipedia

LOOKING BAD. Three different sea slug species in the Phyllidia genus surround a slimmer sea cucumber (center), all with a similar look. Each of the sea slugs, about 2 inches long, can release a toxin. The sea cucumber (Bohadschia graeffei) only resembles them when it’s young, and according to evidence so far, doesn’t seem to be very palatable either. Species that look alike and have their own defenses have inspired a theoretical debate about whether less-dangerous ones undermine the protection of the others. Gosliner

NOVEL PREY. To see how birds react to new kinds of creatures, researchers made their own species of prey out of tiny square envelopes, each stuffed with an almond slice (bottom three squares). The envelopes marked with the black square or sagging square stand out against the background of random X’s and taste awful. The envelope-creature to their right “hides” among the Xs and tastes fine. The Xs on the two raised squares above make the birds’ hunt trickier. Ihalainen/Univ. of Jyväskylä

NEW SPECIES. Researchers invented a new kind of prey for their tests, little envelopes that hold an almond slice. Creator Johanna Mappes says the researchers themselves glue together thousands of prey for each experiment because the job takes such attention to detail that only someone with high stakes in the outcome can sustain the effort. Ihalainen/Univ. of Jyväskylä

In the usual classroom explanation, there are equal partners and fakers among the mimics. The equal partners are, for example, two butterflies that look like each other and that both carry a foul-tasting toxin. A bird that bites either one of them gets a lesson in what not to eat. Since there are two species, not as many of either one have to get killed or injured to educate the latest generation of local birds. The two species share the cost of training the predators.

Then there’s the toxic-and-cheater pair. One of them carries a foul-tasting toxin while the other tastes just fine but looks like the toxic species. The cheater gets benefits: Birds that associate the warning colors with a disgusting mouthful avoid the tasty species. But some birds without an adequate culinary education catch the cheater and get a confusing message: that the butterfly’s colors say, “Come and get it.”

These birds are more likely to attack the genuinely bad-tasting species. The cheater thus undermines the protection that the other species is creating for itself. Both cases are in line with conventional mimicry standards. But Speed, of the University of Liverpool in England, has raised questions about pairings that don’t fit either scenario: One species is vile, but the mimic is neither toxinfree nor totally bad tasting. Are these mimicry pairs still helping each other, or is the milder-tasting one cheating? And do any two species ever really have the same degree of bad taste? Other biologists have proposed that following that line of argument to the end might mean that in the real world, there aren’t any truly equal mimicry partners.

Testing these questions has been tricky. Speed, Rowland, also of Liverpool, and other researchers have turned to a whole novel world of artificial prey to get at the story behind mimicry.

Copycat classics

In 1862, British entomologist Henry Walter Bates published a discussion of distinct butterflies with similar wing markings that he had seen in the Amazon. Today, biologists use the term Batesian mimic for a clear-cut cheat—a tasty creature that borrows warning colors from a foul one.

In the late 1870s, Fritz Müller described how vile species that look alike could share the costs of repelling predators. Today, biologists call these look-alike nasties Müllerian mimics.

Butterflies inspired early ideas about defensive mimicry, but biologists have spotted apparent mimics among many other creatures. For example, Terry Gosliner of the California Academy of Sciences in San Francisco has discovered marine flatworms with the colors of better-defended sea slugs. Birds and certain snakes may also be deploying mimicry.

The idea applies to more than just visual similarities. For example, a North Carolina research team has argued for both Batesian and Müllerian forms of mimicry in the clicking sounds that certain moth species make as a bat swoops toward them to attack.

Speed’s work in the early 1990s didn’t specify any particular species. He developed a mathematical model that attempted to update mimicry theories to include new data on learning. In Speed’s Pavlovian world, for instance, a predator can learn to avoid an extremely nasty-tasting kind of prey after just one encounter. However, the predator continues to eat prey that tastes less disgusting in small amounts.

For the extreme case of two terrible-tasting prey, Speed’s model confirmed the classic theory that equally foul-tasting mimics benefit from each other’s presence. His model also showed that as one of the mimics’ repellent qualities diminish, protection erodes. In 1993, Speed proposed the category of quasi-Batesian mimics.

His work set off volleys of criticism. Two biologists studying mimicry, James Mallet of University College London and Mathieu Joron of the University of Exeter in England, questioned whether quasi-Batesian mimicry was likely to exist in the real world. They said that Speed’s model didn’t correctly account for the way that changing prey abundance would affect prey survival rates.

In the late 1990s, Angus MacDougall and Marian Stamp Dawkins of Oxford University in England created a computer model of a predator that would sometimes misidentify various species of prey. Based on studies of learning and behavior, the model assumed that an animal tends to better discriminate among distinct items, including prey, the fewer categories it has to cope with.

Since mimicry reduces the number of apparent prey species available, the predator makes fewer mistaken attacks. The model indicated that improving the predator’s accuracy in this way should save the lives of some of the prey and compensate for the effects of mimicry cheats.

Testing, testing

In spite of a century of scientists’ theorizing about mimicry, there had been few experiments testing the idea. The hang-up was that to test basic ideas about how a predator learns, researchers needed prey that no bird had ever seen before. They needed to observe predators attacking something for the first time—a tough assignment to pull off in nature or even in the lab.

In the early 1990s, Speed and his colleagues decided to construct fake species for real predators to sample in the wild. The mimics would be made of colored cards and pastry dough.

The researchers dyed portions of dough with yellow, green, blue, or red food coloring and placed them on cards of contrasting colors. To make nasty-tasting species, the team dosed some of the pastry with mustard and quinine. To make the less-repellent and even yummy creatures, they simply reduced the additives’ doses.

An invented prey species was a fat slug of one of the doughs at the center of a triangular card. The study thus included a variety of fake species ranging from fully edible to really foul.

In the spring of 1995, the researchers set out the dough creatures on two lawns in Liverpool. On each of 40 consecutive days, they distributed 85 of the artificial prey at each site and watched blackbirds, sparrows, robins, and starlings drop by to dine. The team repeated the experiment the following spring.

In a 2000 publication, Speed’s team reported that during one period, when the array included equal numbers of a highly nasty dough creature and its half-nasty mimic, the nasty ones suffered more attacks than did another nasty species without a look-alike. Abundant mimics that are only semirevolting can weaken more-noxious prey’s protection, the researchers concluded.

Novel world

Results depended on what kinds of birds happened to drop by, and some unexpected effects during certain weeks might have come from a surge of birds new to the area. To get more control, Speed needed an artificial world, and while he’d been working with pastry, a research team in Finland had been inventing one.

In the mid-1990s, Johanna Mappes and Rauno Alatalo of the University of Jyväskylä developed what they call a “novel world” with artificial prey species that birds could hunt inside a large room. A paper-covered floor creates an artificial environment in which artificial prey can nearly disappear or stand out (see “Making a Novel World,” below).

The Jyväskylä researchers have used the setup to experiment with predator behavior of birds captured from the surrounding forest. Before Speed’s team tested its hypothesis in the novel world, the Finland-based researchers investigated what birds might do when confronting Speed’s nasty and half-nasty pairs.

Eira Ihalainen set out dozens of invented prey made of almond slices inside small paper envelopes that blended into the paper-carpet background. But also included were boldly marked envelope species dosed with a bitter solution and, sometimes, look-alikes with a less-terrible taste. The half-nasty species didn’t seem to slow birds’ learning to avoid the mark for nasty flavor, Ihalainen and her colleagues found.

The work, reported in the March Journal of Evolutionary Biology, suggested that the half-nasty mimics weren’t the menace that Speed had discussed, but the Ihalainen team’s study focused on the predator, not the prey.

Rowland and Speed joined Ihalainen and others at the Jyväskylä facility to set up an explicit test of whether half-nasty mimics undermine the protection cultivated by nasty-tasting species.

The combined team set out various numbers and kinds of mimic-envelope species and tallied the first 50 envelopes that a bird chose. When they mixed the worst-tasting species with a milder-tasting look-alike, both species enjoyed a benefit, the researchers report in the July 5 Nature. So, Speed didn’t find his quasi-Batesian effect.

Circumstances may still exist where the effect would turn up, says Rowland. Says Speed: “There are different ways of running the experiment, and these seem to give us different results.”

For example, the dough tests in Liverpool held the number of total prey constant, but the almond-envelope-prey experiment in Finland used variable numbers.

The debate over mimicry, especially that of the quasi-Batesian type, has become “obfuscated,” says Speed. So he’s going back to design new experiments and to review the theory. The question about whether half-nasty mimics are cheaters “seems a good one, and one worth answering,” he says. “If my initial hypotheses were dead wrong, then that’s actually fine with me: We’ll have understood the system.”

Making a Novel World

Use lots of glue

Johanna Mappes and Rauno Alatalo got interested in experiments that they couldn’t do in this world, so they invented another one. In the mid-1990s, Mappes, of the University of Jyväskylä in Finland, had become interested in the warning colors, smells, or sounds that a foul-tasting animal typically displays. In theory, these signals are easy to notice and remember.

Testing ideas about the evolution of the signals is complicated by the fact that actual species have had a chance to evolve reactions to warning signals. Newly hatched chicks, for example, show an innate hesitation to peck at any object showing yellow and black stripes, which are common warning signals of creatures such as wasps and bees.

So Mappes and Alatalo, also of Jyväskylä, invented new prey species. They had to search for markings that would be unlikely to arouse any innate bias in birds. “Circles were out of the question since they resemble an eye,” says Mappes.

Eventually, the researchers decided to go with an X for the basic prey that would be hard to spot against their artificial background. The “warning-signal” prey, intended to stand out, get bolder marks, such as black diamonds or squares.

The team creates prey by soaking sliced almonds in a bitter solution. Researchers tuck each slice between paper squares and glue the paper to form a little pouch. Each experiment takes thousands of these envelopes.

The Mappes team covers the floor of a 57-square-meter room with paper. To create a background for the prey to hide in or stand out from, the scientists mark the paper with a random scattering of Xs. To add a little depth to the background, they also glue on some cardboard Xs.

The predators that forage on this carpet of Xs are great tits caught in nearby forests for a research interlude. In a typical experiment, the birds get several days to adjust to indoor life and to learn how and why to open little paper envelopes. Finally, they get one session in the big room before returning to the forest from novel world.

Susan Milius

Susan Milius is the life sciences writer, covering organismal biology and evolution, and has a special passion for plants, fungi and invertebrates. She studied biology and English literature.

From the Nature Index

Paid Content