Two major problems have to be solved by a flying animal or machine. (i) On the time average, flight force has to be produced which is sufficient to keep the body airborne and to propel it through the air. (ii) To stabilize a given position or trajectory, the vector of the generated flight force has to be controlled in its magnitude, orientation and position relative to the body. In the present study, the response of wing-beat kinematics to wind and visual stimuli was investigated in tethered flying Drosophila melanogaster. When the fly is subjected to an air stream in a wind tunnel, or to striped patterns moving in its frontal field of view, the overall shape of the wing path is altered, including variations of the wing-beat amplitude and the angles of attack. The aerodynamic forces were calculated from the kinematic data according to the quasi-steady aerodynamic theory, to investigate whether this approach is sufficient to describe the control mechanisms of the fly. The stimulus-induced changes of kinematic and aerodynamic variables were compared with control reactions expected in free flight or measured during tethered flight under similar stimulus conditions. In general, the calculated flight forces are too small to account for the measured lift, thrust and torque responses to the particular stimuli, or would even increase the input stimulus instead of being compensatory. This result supports the notion that unsteady aerodynamic mechanisms are likely to play the major role in flapping flight. Following this line of thought, some kinematic responses can be qualitatively understood in terms of unsteady aerofoil action.