irst things first: David Pfennig likes snakes. Grew up hunting them in South Texas, in fact. "The big thing to catch there was the Mexican milk snake," Pfennig says. "One night my cousin John grabbed one off the side of the road."

Or so John thought. He'd actually picked up a deadly coral snake.

Apart from his beautiful colors—alternating rings of black, bright yellow, and deep red, which are a heads-up DANGER! warning for predators—the eastern coral snake looks perfectly unremarkable. His rounded head is hard to distinguish from his body, which is the same slender diameter all the way down to his tail. His eye is round and dark. He looks like a snake that a kid would make from Play-Doh.

But the coral snake doesn't play. He's an elapid, in the same family as cobras and mambas. His venom is neurotoxic, meaning it attacks his prey's central nervous system, eventually paralyzing the diaphragm. But he tends to be nocturnal, secretive, and nonaggressive, so fatal bites to humans are rare.

Now, Pfennig and his cousin John knew a thing or two about snakes, so John dropped that coral snake in a hurry. But why did he mistake it for a harmless milk snake?

Well, the milk snake is a con artist. It looks a lot like the dangerous coral snake, a ruse that gets it a bit of respect from predators. In North Carolina, we have a similar set of sneaky serpents: the decidedly innocuous scarlet kingsnake bears a striking resemblance to the venomous eastern coral snake. In fact, the two are so similar that we have a little ditty to help us tell them apart:

Red touching yellow, kill a fellow
red touching black, venom lack.

hese snakes are prime examples of Batesian mimicry. In 1862, British naturalist Henry Bates proposed that some unprotected species closely resemble unpalatable or harmful species and are therefore avoided by predators. Biologists widely accept Bates' theory and even use it to illustrate natural selection, Charles Darwin's mechanism of evolution. But Bates' idea has never been experimentally proven.

Which brings us back to Pfennig, now an associate professor of biology. He'd been thinking for a couple years about Batesian mimicry. Along came William Harcombe, a Carolina biology undergraduate, who told Pfennig he wanted to get involved in a research project. And Karin Pfennig, who is married to David, is a bioinformatics postdoc at the University of Texas-Austin. The three put their heads together and devised an experiment to test Bates' idea.

Their plan: find out how scarlet kingsnakes fare when they live far from where any coral snakes are found. Batesian mimicry predicts that these particular kingsnakes shouldn't be able to rely on their looks for protection from predators.

Why? Because in Batesian mimicry, the predators call the shots. Let's say there's a hungry possum near Pinehurst, North Carolina, where both scarlet kings and coral snakes are found. Batesian mimicry holds that generations of evolutionary selection to avoid the poisonous coral snake—and anything that resembles it—have gone into the making of this possum. So the Pinehurst possum sees the scarlet kingsnake as a potential threat.

But does that hold true in Rocky Mount, North Carolina, where no coral snakes live? Batesian mimicry says no: Rocky Mount predators wouldn't have been under selection to avoid coral snakes. So to a Rocky Mount possum, a kingsnake is probably just dinner.

The team's idea: plant fake snakes at dozens of sites around North and South Carolina, some in areas where both corals and scarlet kings live, and others far outside the coral snake's range. The sites would be in game lands and federal wildlife refuges, in hopes that the fake snakes would be molested by real predators, not by say, house cats or people. And the sites should have sufficiently snakey spots—the right kind of rotting log, for instance—"to make sure," Pfennig says, "that this was a site where a predator would actually find a snake."

ach site would get three "snakes." First, a replica of the scarlet kingsnake—a mimic of the mimic. Then two replicas would serve as experimental controls: one, a generic snake, was plain brown. The second control used the coral and kingsnake's colors arranged in lengthwise stripes instead of rings. "If we used only ringed and plain brown replicas, we wouldn't know whether predators simply have a generalizable avoidance of bright colors," Pfennig explains. "So we wanted to make striped replicas, in the same colors as the ringed, just to make sure that wasn't happening."

After a month, the team would collect the replicas and determine whether any predators had messed with them. If attack rates on the kingsnake replicas proved significantly higher outside the coral snake's range, then Bates might have been on to something.

Sounds easy, right? Just tell that to Will Harcombe. He had to make 1,600 snakes.

Harcombe made them by molding plasticine—a nontoxic modeling clay that would retain bite and scratch marks—onto lengths of heavy wire. He rigged up a caulking gun and the clay snakes just squirted out. Okay, so it wasn't quite that easy. But the snakes got made.

And the snakes got attacked. "Some were just mangled," Pfennig says, pulling what used to be a brown control snake replica out of a numbered paper bag. The replica has been squished beyond recognition, and there are some very large teeth marks in one end of it. "The only animal that could do that would be a black bear," Pfennig says, "or a wolf. But we don't have wolves where we're doing this. So that's probably a black bear."