Visually-guided, upwind turning behaviour of free-flying tsetse flies in odour-laden wind: a wind-tunnel study

Abstract. To test the hypothesis that tsetse flies use visual input from the apparent movement of the ground to assess wind direction while in flight, Glossina morsitans morsitans Westwood females were video- recorded in a wind-tunnel as they entered, in cross-wind flight, a broad plume of simulated host odour (C0₂ at c. 0.05%). The tunnel (2.3 times 1.2 m wide) generated winds up to 0.25 m s^-1 and had a strongly patterned floor that could be moved upwind or downwind to increase or decrease the visual input due to wind drift. Flight tracks were analysed for speed, direction relative to the wind, and angle of turn. Mean groundspeeds were c. 1.8 m s^-1. In control measurements in still air (with or without odour) flies turned 50:50 'upwind': 'downwind'. With a 0.25 m s^-1 odour-perme- ated wind, 79% turned upwind, and c. 70% left view flying upwind. When the floor was moved at 0.25 m s^-1 upwind (to mimic the visual input from the ground due to a 0.5 m s_^-1 wind), the strength of this response increased. If instead the floor was moved downwind, faster than the wind speed (to mimic the visual input due to a wind from the opposite direction), 59% turned downwind and c. 70% left view flying downwind, and thus away from the source (though progressing 'upwind' in terms of the visual input from apparent ground pattern movement). Upwind turns were on average significantly larger than downwind turns. It is concluded that tsetse navigate up host odour plumes in flight by responding to the visual flow fields due to their movement over the ground (optomotor anemotaxis), even in weak winds blowing at a fraction of their groundspeed.