The Butterfly That Can’t Fly Straight
Watch a white cabbage butterfly (Pieris rapae) hover over a flower and it looks like it’s fighting gravity and losing. The body jolts up, then down, then up again, in a stutter that reads less like flight and more like a malfunction. Park a hoverfly over the same bloom and it barely moves; a hummingbird moth holds dead level while its wings do the work at a blur. The cabbage butterfly, wingspan 1.8 to 2.6 inches (4.5 to 6.5 cm), white with black-tipped wings, does neither. It bobs the whole time, and it looks, frankly, like it’s struggling.
The natural read is that this insect just wasn’t built for precision hovering the way bees, dragonflies, and hoverflies were. It looks like it’s muddling through. It isn’t.
It’s Not the Wings. It’s the Whole Body.
A 2025 study in Physics of Fluids, vol. 37, issue 5, article 051904, from a team at Beihang University, filmed wild-caught white cabbage butterflies hovering under high-speed cameras. Rather than attaching tracking markers to the wings, which risks tearing the delicate membrane and altering how the insect actually flies, the researchers trained a deep-learning model to follow specific points on the body and wings straight from the footage.
What they found flips the usual hovering story. In bees and hoverflies, the wings do almost everything: fast beats, precisely angled strokes, control at the wingtip. The cabbage butterfly’s wings never manage that. Instead, the tracking data show the body itself pitching, tilting up and down continuously through every wingbeat cycle, as the main driver of the hover, not the wings. During a steady hover, the average body angle sits close to 90°, effectively standing the insect on end to cancel out the forward push each stroke produces. The swing of the abdomen and the flap of the forewing both feed into that pitching motion in a noticeable way. Wing pitch angle still shifts too, but as a secondary adjustment riding on top of the body’s own tilt.
Why Bother Tilting at All?
A slow, light wingbeat doesn’t send much force straight upward on its own. Upstroke and downstroke are close to symmetrical here, each contributing roughly equally, and the study found that only about 50% of the total aerodynamic force from a wingstroke ends up pointed vertically. The hindwings alone contribute only about 10% of the weight support needed. That’s because the swirl of air that briefly clings to a wing’s trailing edge and adds lift sheds early on the hindwing and never builds the low-pressure region that would help hold the insect up. Meanwhile the leading-edge vortex on the forewings stays attached and strong through the stroke, which is where most of the real lift comes from, generated through the same rapid acceleration and delayed-stall tricks other flapping fliers rely on. Tilting the whole body redirects more of the total force upward, stacking just enough of it against gravity to hold position in the air. It’s a workaround: an insect that never evolved the wingbeat frequency or fine wing control of a dedicated hoverer gets there by moving something else entirely.
Hovering is rare for this species compared with insects that treat it as routine business. Study author Yanlai Zhang described hovering as an essential survival tool tied to two behaviors with zero margin for error: visiting flowers and getting away from whatever wants to eat the butterfly. Working out the aerodynamics behind it, Zhang added, gives real insight into how butterfly flight evolved in the first place.
Jerky Doesn’t Mean Clumsy
This is the myth the study kills: that the stutter looks erratic because the flight really is erratic, a pretty insect with clumsy aerodynamics bolted on as an afterthought. The tracking data say otherwise. The body’s pitching motion is continuous and repeatable, not random twitching, and it is the entire mechanism by which the animal hovers at all. What reads as instability from the outside is a working control system, just one built on body angle instead of the wing angle everyone expects to see. Next time a video like that crosses your feed, you’ll know the wobble isn’t a glitch, it’s the whole trick.
What the Study Doesn’t Tell Us
The paper doesn’t give an exact figure for how many degrees the body swings through, an exact wingbeat frequency, or how many individual butterflies were filmed and analyzed. It also stops short of a direct, side-by-side comparison with known hoverers like honeybees or hummingbird moths, so the mechanism is clearly different, but not yet measured against them for efficiency or control. Whether that roughly 50% force-conversion figure holds for other butterfly species, or is particular to Pieris rapae, is untested. Exactly why butterflies hover so rarely compared with dedicated hoverers isn’t spelled out either.
The researchers point to flapping-wing micro aerial vehicles as a place this could matter eventually: insect-scale drones running on low-frequency, light wingbeats instead of the fast, power-hungry rotors most designs use now. Nobody has built or tested one that pitches its whole body the way a cabbage butterfly does. For now, that idea sits in the paper’s discussion section, not on a workbench.