(Inside Science) -- In the darkness, the fly's legs dangled high above open space. It hung from the ceiling of an empty box, where an electromagnet held on to a pin glued to the fly's back. But then the electromagnet turned off, dropping the fly into a free fall.
Soon, though, it would beat its wings and take flight. After all, flying is what flies do.
That's what biologist Stéphane Viollet thought would happen, at least.
By watching the tumbling flies -- specifically, hoverflies -- he was trying to learn how they orient themselves and whether they have an internal gravity sensor. For flying insects that zigzag through the air in wild, gravity-defying loops, it's crucial to know which way is up. The question is how they know.
Now, from their new experiments, Viollet and his colleagues may have found an answer for the hoverfly: It relies primarily on its vision.
That's in contrast with humans and most other vertebrates, which can sense gravity with built-in accelerometers in the inner ear. In humans, for example, three small, liquid-filled chambers in the ear move the instant you feel acceleration. Tiny hairs inside those chambers detect the sloshing and send signals to the brain, helping you keep your bearings.
But it's not fully known what happens in flying insects. "It's still an open question for many, many insects because insects don't have an inner ear," said Viollet, who works at the National Center for Scientific Research and Aix-Marseilles University in France.
Some studies have suggested that cockroaches might orient themselves using tiny organs on their rears that act as pendulums. But otherwise, Viollet said, it's unclear whether insects can sense gravity in flight.
If any bug could, though, it might be the hoverfly, whose hovering flight patterns would require good vertical orientation. So to see whether it could sense gravity, the researchers studied how it responds to a free fall, when it momentarily feels weightless.
If the fly could track gravity, it should be able to detect the sudden drop and quickly take flight. But if it didn't have a built-in accelerometer, it would still have to distinguish up from down -- perhaps by relying on visual cues. To test this, the researchers shut off the lights for the drop.
Surprisingly, in the pitch-black the fly couldn't orient itself and kept falling for more than a foot until it slammed into the bottom of the box (Not to worry -- because of the fly's low mass, it wasn't hurt).
"It was very unexpected," Viollet said.
As the fly plummeted in the dark, it realized its predicament and started flapping its wings. "They know they're falling -- they can perceive the airflow," he explained. "But it's too late. The airflow cannot give information about the orientation with respect to gravity. So finally, they crash."
And they kept crashing -- about 70 percent of the time in the experiments.
But it was different with the lights on. The researchers dropped the fly in a uniformly white box, and one lined with horizontal black-and-white stripes, which provide more visual cues as to which way is up.
With the lights on, the flies crashed only 30 percent of the time in the white box. In the striped box, only 10 percent of them hit the floor. The flies also started flapping their wings sooner than the ones in the dark. Sight, it seems, is essential for not only orientation, but also a timely reaction.
Still, in every case, it took more than 100 milliseconds before they started beating their wings -- a reaction time much slower than cats, humans, and other animals that have built-in accelerometers. Such a lethargic response further suggests hoverflies don't have gravity-tracking organs that would immediately alert them of sudden accelerations.
The findings, published today in the Journal of Experimental Biology, may also apply to other flying insects, Viollet said.
For airborne bugs that take twisting flight paths, not having built-in accelerometers could actually be an advantage, he said. Otherwise, the constant zigzagging would be overwhelming, like being on a perpetual roller coaster.
Still, not everyone is convinced yet. "To conclude from this one experiment that these insects don't have a gravitational sensor at all is a bit extreme I feel," said Mandyan Srinivasan, a bioroboticist at the University of Queensland in Australia.
He points out that many insects do exhibit some kind of accelerometer, so it's likely the hoverfly does too. For example, hawk moths and fruit flies may use their antennae for orientation and sensing gravity. Because insects like hoverflies have to make sharp, banked turns, they must be able to detect some acceleration to control the centrifugal force.
Perhaps the hoverfly is simply not using its accelerometer while being dropped in the dark, Srinivasan said. The experimental setup is so unnatural that maybe the fly doesn't know how to react.
But he lauds the experiment and its surprising result. "If it really turns out that hoverflies don't have accelerometers at all, that would be a world first," he said. "If that's the case, it would be wonderful."