Fruit flies rely on a plethora of sense organs to keep them airborne and prevent crashes: if they see an object looming in their visual field on a collision course, they execute a 90° turn, called a saccade, in a lightning fast 70 ms to get themselves out of trouble. They also depend on short, stubby, modified hind wings called halteres to detect rapid changes in their body's orientation in space. Researchers already know that feedback from the eyes and halteres complement each other to keep a fly aloft - the eyes respond to slower and smaller changes in a fly's body orientation while the halteres respond rapidly to larger and quicker changes. John Bender and Michael Dickinson at Caltech want to know more about how feedback from sense organs affects a fly's saccades. Knowing that halteres respond to rapid changes, they set out to discover whether feedback from the halteres affects the size and speed of fly's saccades, and whether vision also plays a role(p. 4597).

First the team had to create an apparatus that would allow flies to perform saccades freely. Modifying an idea to tether a fly attached to a wire filament with magnets, first tried out by Martin Heisenberg and Reinhard Wolf,Dickinson and Bender built a simpler magnetic tether. Each fly had a steel pin attached to its back, and the pin's end sat in a small depression in a minute block attached to a magnet fixed directly above the fly. Flies could turn freely in the low friction set-up, and their every move was captured by a camera filming at over 500 frames s-1.

Before testing the feedback from halteres, Bender says, the team had to confirm that visual feedback did not affect the speed and size of saccades. They placed tethered flies inside a cylinder, lined with light-emitting diodes(LEDs) displaying a striped pattern on the inside. Each time a fly turned, the team moved the stripes by turning LEDs on and off in a specific sequence. The stripes were moved either in the same direction of the turn or in the opposite direction, to trick the flies into thinking they had turned more or less than they really had, and adjust their saccades accordingly. They found that the flies made no adjustment to their saccades when the stripes were rotated in either direction. This confirmed that visual feedback did not play a role, so the team then turned to the halteres.

To change feedback from the halteres, they stuck a small blob of glue on the end of each one, which approximately doubled their weight. `This was pretty tricky', says Bender, as the tip of each haltere is similar in size to the sharp end of a pin. When the halteres were heavier, the angle of saccades to the left and right were smaller, and saccades were a slower speed. Removing the left haltere to reduce its mass increased the angle and speed of saccades,both to the left and right, showing that feedback from the halteres does play a role in controlling a fly's saccades. Although, Bender adds, they're not sure of the exact mechanism yet.

Finally, Bender and Dickinson wondered if flies' halteres could compensate for changes to saccades caused by clipping the wings, affecting flight aerodynamics. While tethered flies with clipped wings altered their wing movements in order to maintain a stable heading, saccades were smaller. When they plunged flies with clipped wings into darkness, their flying got a lot worse, which also occurs in flies with intact wings. The team suspect that in this situation feedback from the eyes is needed to keep flies on the straight and narrow, and that feedback from the halteres is not precise enough to compensate for the effects that changes to the wings have on saccades, as they can't respond to slow changes in body orientation.

Bender, J. A. and Dickinson, M. H. (
). A comparison of visual and haltere-mediated feedback in the control of body saccades in Drosophila melanogaster.
J. Exp. Biol