Comedy maestro Bill Bailey has a song about zebras, in which he casts their black and white stripes as a message of racial harmony. (“In a world of confusion/ We all need a sign/ If only we could live side by side/ Like the stripes down a zebra’s spine.”) Which, honestly, makes about as much sense as the most commonly cited hypothesis about zebra stripes—that they're a form of camouflage.

“The zebra is conspicuously striped,” wrote Darwin, on one of his less insightful days. The idea that its black-and-white coat might help it blend in rather than, say, stand out seems preposterous, but there are two ways in which this could work. First, the black stripes could match dark tree trunks while the white ones match shafts of light between the trunks. Alternatively, the stripes break up the zebra's outline, making it harder to identify as a juicy piece of horse-shaped steak. Both ideas have been around for a while, but neither has been tested well.

The problem is that we've always looked at zebras through the wrong eyes—ours. Human eyes are exceptionally good at resolving detail in daylight, so “we have a very odd appreciation of the coat of a zebra,” says Tim Caro from the University of California, Davis. By contrast, their main adversaries—lions and hyenas—have eyes with poorer resolution, but greater sensitivity at dawn, dusk, and darkness. So Caro, together with Amanda Melin from the University of Calgary worked out what zebras look like to these predators.

The team measured stripe widths from different body parts on all three zebra species, and used published data to estimate the acuity of lion and hyena eyes. They then calculated how good those predators are at resolving zebra stripes at different distances and light levels.

They found that in daylight, humans with 20/20 vision can resolve zebra flank stripes from around 180 meters away. By contrast, lions can only do so at 80 meters, and hyenas at 48 meters. Those values get much worse for Grevy's zebra (the species with the thinnest stripes), for leg stripes (which are also thinner), and at darker times of day. At dawn and dusk, lions, and hyenas can only resolve zebra stripes at 46 meters and 26 meters respectively.

“At most distances, the zebras are going to look to a lion like a gray waterbuck,” says Caro. “Those stripes are going to fuse together and be indistinguishable.”

That rules out both the blends-among-trees idea and the breaks-up-outline one—neither can possibly be true if the predators can't see the stripes. “If the stripes are doing something exciting, they’ll be doing it close up, by which point the predators have probably realized the zebra is there, because they can smell or hear it,” says Caro. Zebras, being very noisy browsers, are hardly stealthy.

“It’s the first proper test of a very longstanding and prominent idea,” says Martin Stevens from the University of Exeter, who studies camouflage. Its only flaw is that the team didn't specifically measure how closely a zebra matches its background environment, in either color or brightness. Still, “I very much doubt zebra stripes do work in concealment,” adds Stevens.

So, if not camouflage, then what?

Caro, who has been studying zebras for a decade and has written a forthcoming book about their stripes, thinks he knows the answer. I’ve come to the conclusion that really, it just has to be biting flies,” he says.

In Africa, horses are plagued by horseflies and tsetse flies. Collectively, these insects can drink up to half a liter of blood a day, and they spread deadly diseases like sleeping sickness, equine influenza, and African horse sickness. And for some reason, these blood-sucking insects don't like to land on black-and-white stripes. Gabor Horvath from Eotvos University showed that through several experiments, including one where his team stuck several painted horse models in a fly-infested field.

Caro found more evidence to support this idea. First, he mapped the geographic ranges of all seven species of wild horse—the three striped zebras, the African wild ass with thin stripes on its legs, and the uniformly colored Asiatic wild ass, Przewalski's horse, and kiang. Then, he compared these ranges to other factors, including lion and hyena distributions, habitat, temperature, herd size, and the presence of biting flies.

“Every time we ran the analysis, nothing showed up except biting flies, particularly the tabanids—horseflies and deerflies,” says Caro. “You don’t find striping where there’s lots of trees or hyenas, but you do find it where there are lots of biting flies.”

Why do stripes deter flies? Perhaps they create some kind of optical illusion. “We don't know whether the flies don't see a black-and-white striped animal, or misinterpret it as something else,” says Caro. “But the bottom line is that they don’t land.”

Okay, but flies are surely a nuisance to all kinds of savannah animals. Why have zebras gone to such evolutionary lengths to avoid fly attacks while other savannah animals have not? Put it another way: If the fly hypothesis is right, why aren't impalas or wildebeest striped? Caro says that zebras have much shorter coats than other hoofed mammals that they live with, making them especially vulnerable to the probing snouts of flies. He also suspects that they and other horses are uniquely susceptible to the diseases carried by the flies. Domestic horses are very difficult to keep in parts of Central Africa for precisely this reason.

Case closed? Hold your horses.

Brenda Larison from the University of California, Los Angeles says that Caro's team don't have any actual data for the numbers of biting flies; they just estimated those numbers based on temperature and humidity. Also, the experiments from Horvath and others “haven't been done using realistic targets or under well-controlled conditions,” she says. It's unclear how stripes would affect the attractiveness of live animals, which also give off heat and odors.

In her own study, Larison studied the variations in the plains zebra's stripes—in terms of number, thickness, and definition—and found they correlate most strongly with temperature. She reasons that black bands heat up faster than white ones, and so alternating rows would create circulating currents of air that cool the animals down. “This is the one hypothesis that has been completely ignored by researchers,” she says.

Caro is dubious. “There isn't good evidence in the physics literature that you get these convection currents from black and white stripes,” he says. "And those currents wouldn't cool the animal in windy conditions, or when it's moving.” Larison counters that plains zebras “stand still during the hottest part of the day” and when they walk, they do so slowly. “Zebras spend almost two hours more per day in the sun grazing than solid-colored ruminants,” she notes. “What allows them to do so?”

Both of them could be right. There's no reason to think that only one factor drove the evolution of zebra stripes. Temperature and biting flies might both have played a part; the stripes might also dazzle predators at close range when moving, or provide some other weird benefit. The only way to tell is to do more work—on convection currents, on fly eyes, on fly diseases, and more. As Dr. Seuss said, “There’s no end to the things you might know, depending how far beyond Zebra you go!”