The sound of death can take many forms: the retort of a gun, the screech of tires, the hack of a cough. But for many moths, death sounds like a series of high-pitched squeaks.

Moths are hunted by bats, which track them down by releasing high-frequency calls and analyzing the rebounding echoes. This skill, known as echolocation, allows them to view their world—and their prey—even in total darkness. Bats evolved the ability to echolocate tens of millions of years ago, and in the intervening time, moths have developed their own countermeasures. Some evolved ears, which allow them to eavesdrop on the calls of hunting bats and take evasive action. Others play the bats at their own game, releasing their own ultrasonic clicks to jam the radar of their predators, or to feign the echoes of distant objects.

Bats, in turn, have evolved their own tricks for circumventing the moths’ defenses. Some, for example, use stealth.

Bat calls are too high-pitched for us to hear, and we should all be grateful for that because they’re also some of the loudest sounds produced by any animal on land. If we could hear them, it would be like listening to a passing ambulance, a jackhammer, or a rock concert. But Townsend’s big-eared bat—a North American species with a foot-long wingspan—is an exception. Aaron Corcoran and William Conner from Wake Forest University have found that when it hunts, it does so at a whisper, with very quiet calls that moths can’t hear. It has evolved into a winged ninja—silent and undetectable, until it’s too late.

This discovery helps to settle a question that’s been bugging Corcoran for a while. He and other evolutionary biologists often talk about evolutionary arms races, in which predators and prey evolve ever-more sophisticated measures and countermeasures to outwit each other. But while scientists have documented myriad examples of prey adapting around their predators, there are surprisingly few strong examples of predators doing the reverse. (These examples include the shocking powers of the electric eel, the snakes that have evolved to resist the poisons of newts, and the mouse that turns scorpion venom into a painkiller.)

Partly, that’s because it’s hard to show that predators are adapting to their prey specifically. For example, for decades, scientists have suggested that bats could shift their calls to higher or lower frequencies that moths can’t hear—and indeed, there are bats that do this. But higher-pitched calls give them a sharper view of their surroundings, and lower-pitched ones travel further and give a wider view of the world. It’s possible that such calls evolved to help bats navigate, and were only incidentally useful for subverting the ears of prey.

In 2010, Hannah ter Hofstede from the University of Bristol found a more unambiguous countermeasure. She showed that the barbastelle, a small European bat, is also a whisperer, with echolocation calls that are 10 to 100 times quieter than those of other moth-hunting bats. This seemed like a clear-cut case of an anti-prey adaptation, since there’s really only one advantage to quieter sonar: catching sharp-eared moths.

Ter Hofstede demonstrated this by tethering moths in small arenas, and hooking electrodes to the neurons in their ears that detect bat calls. These neurons typically fire when bats are around 19 meters away, giving the moths plenty of time to react. But those same neurons only detect barbastelles when they are 2 meters away, giving them just half a second to dodge. By then, it’s probably too late.

Probably. In these experiments, the bats never actually got to attack the moths—and Corcoran wanted to know what would happen if they did. He chose moths that are known to jam bat echolocation and tethered them to fishing lines, hanging them in large outdoor arenas that were surveilled by cameras and microphones. Then he waited for bats to approach.

The results were clear. Compared to the long-legged myotis, a similarly sized bat and a fairly typical echolocator, the stealthy Townsend’s big-eared made calls that were 20 to 40 decibels quieter. And as a result, they catch moths on 80 percent of their attacks. “That’s pretty unheard of for a bat attacking well-defended prey,” says Corcoran. “And the moths almost never exhibited their normal defenses. They very rarely do diving maneuvers, and they never made their jamming clicks.” And when the moths did try to dodge, they did so at a third of the distance for the big-eared than the myotis.

But these whispers come at a cost: The big-eared bat can only detect moths at half the range of the louder myotis. This might explain why the former species isn’t very common. “Maybe you’re doubling your chance of capturing a moth, but you’re reducing your ability to find that moth by half,” says Corcoran. “The math may or may not work in your favor.”

There might be other costs, too. Many of the moths that make jamming clicks also do so to advertise the poisons in their bodies, says ter Hofstede. She wonders whether stealthy bats might be more likely to attack toxic prey. “Bats that catch these kinds of moths generally let them drop without eating them, so it is not dangerous to the bats, but it would represent wasted effort,” she says.

For now, it seems that the moths haven’t evolved a response to the stealthy sonar—and that’s probably because it’s not worth it. “At the sites where I work, there’s 15-plus species of bat, and this one is taking advantage of the fact that all the others echolocate very loudly,” says Corcoran. “Being a rare enemy that uses this odd trick, there’s not enough pressure for the prey to evolve a counter. If the moths focused on this one predator, their response would be all off for all the other bats.” In other words, Townsend’s big-eared bat is a hipster ninja: Its stealth only works because it’s the only bat that uses it.