In early spring, people walking through the deserts of California might be able to hear a high-pitched whistle. That noise comes from a male Costa’s hummingbird, but not from his throat—it’s all in his tail.

Males woo females in a number of ways. They sing. They spread the iridescent feathers of their throats, transforming their heads into shiny, violet octopuses. And they fly up to tall perches to plummet into acrobatic dives, careening downward before pulling up at the last second. When they hit a critical speed, the back edges of their outermost tail feathers start to flutter. That’s the source of the whistle.

Christopher Clark likens that whistle to the Stuka dive bombers that Germany deployed in World War II. As these planes dove, propeller-driven sirens on their wings produced a wailing sound that will be instantly familiar to anyone who’s watched old war movies. The Stuka’s wail was intended to weaken morale and intimidate the enemy. The Costa’s hummingbird’s whistle, by contrast, is intended to attract a mate.

By placing isolated feathers from the hummingbirds in a wind tunnel, Emily Mistick, a graduate student who worked with Clark at the University of California at Riverside, showed that the frequency of the whistle they produce is directly related to the airspeed around them. In other words, the faster a male dives, the higher the pitch of his whistle. For a female, that’s a potentially useful piece of information. By listening to the sound as it passes overhead, she could conceivably judge how fast the male is traveling, and how fit and athletic he is. The whistle, then, seems like an honest signal, which accurately reflects the male’s quality, in a way that weaker individuals simply cannot fake.

But not quite.

By filming the males, Mistick and Clark found that they dive in a way that dilutes some of the information that could be gleaned from their whistles. “At first, I thought it would be easy to take a recording of a bird in the wild and work out how fast it was going based on the sound it was making,” says Clark. “But I gave this problem to an undergraduate and they couldn’t do it. I gave it to Emily and she didn’t figure it out either. And I realized that if this is hard for a biologist to measure, it might be hard for a bird to measure.”

As the male dives toward the female and speeds up, the pitch of his whistle rises. From the female’s perspective, the whistle also sounds even higher than what the male actually generates. This is because of the Doppler effect, which makes an approaching noise sound higher-pitched, and a retreating one sound lower—the same phenomenon that creates the changing wail of a siren as an ambulance drives past.

For the siren of a passing ambulance, if the downshift in its wail is very sharp, it’s a sign that the vehicle is zooming by. If the transition is more drawn out, you’d guess that the ambulance is in less of a rush. The same thing is true for the hummingbird’s whistling tail, so the rate at which his pitch changes also clues the female in to his speed.

Based on all of this, a male that’s trying to show off his athleticism should dive straight at a female and pull up just above her. That trajectory produces the most dramatic whistle with the highest perceived pitch, and the sharpest downshift between high and low pitch.

But instead, Mistick and Clark found that males dive to the side. This trajectory still produces a pretty high maximum pitch, but it also flattens out the ambulance effect, drawing out the whistle. For some reason, the males are hiding a potential indicator of their true speed. “They’re not showing off exactly how fast they’re going,” says Clark. It’s as if a peacock were graying out its tail feathers.

It’s not that the male doesn’t cater to what the female hears—quite the opposite. His fluttering tail feather is like a speaker; if he’s diving to the side, the speaker would normally be aimed away from the female, reducing the volume of the sound she hears. But Mistick and Clark found that the males fix this problem by spreading out half of their tails and twisting them toward the females, essentially aiming their speakers. They’re clearly working to shape the female’s experience, but just not in a way that reveals their speed.

To Clark, this suggests that the whistle might not be an honest signal after all. If the dives and whistles are all about showing off, why the modesty? Why hide a cue that would show how good you really are? The dive might be a way for females to find mates of the right species—the closely related Anna’s hummingbird produces dives with a much lower pitch, so perhaps Costa’s has evolved to have as different a sound as possible.

The more intriguing alternative is that the dives are purely aesthetic displays that evolved for arbitrary reasons. Females might develop a random preference for certain male traits, whether an acrobatic move, a set of shiny feathers, or a whistling tail. Their offspring inherit those sexy traits, and also the preferences for them. Desires and the objects of those desires evolve together, producing displays that aren’t objectively informative, but subjectively pleasing. In other words, males dive not to show off their qualities, but because females like it. (This hypothesis, rooted in subjectivity and aesthetics, has been gaining steam of late, thanks to books by the evolutionary biologists Richard Prum and Michael Ryan.)

These ideas aren’t mutually exclusive, says Gail Patricelli from UC Davis. “I don’t think every detail of a display needs to support maximum information transfer for us to think that there is information involved in courtship. My guess is that males still have to be in pretty good health to pull off a sexy dive. Nonetheless, I think that Clark makes a good case that aesthetics are a key driver of the form of this display.”

“I agree that it’s unlikely to be an honest signal, and the alternatives are purely speculative at this stage,” adds Laura Kelley at the University of Exeter. She wonders if the dive could be a kind of audition in which males that do well enough move onto the next stage of courtship, where they get to show off their shiny throat octopus. “Males might cheat by making it harder for females to accurately assess their quality in the early stages,” she suggests.

What’s missing in all of this is any information about the females. No one knows how good they are at hearing high-pitched sounds, at discriminating between subtle changes in pitch, or at judging the difference between two very similar dives. Clark has tried using electrodes to record their responses to various noises, but their heads are so small that he couldn’t get any useful data. Instead, he is now trying to condition the females to certain sounds and see if they change their behavior in response to something different. It’s arduous work, but necessary. Perhaps the story of Costa’s hummingbird, like many others in evolutionary biology, will turn out to be less about male quality than about female choice.