Visually evoked escape flight initiation in Drosophila, according to the accepted account, involves a rapid extension of the middle legs that propels the fly into the air while the wings are still folded. This description has remained unchallenged and is accounted for in terms of the activation of a simple neural circuit, the Giant fibre (GF) system. The accepted description of escape is however inconsistent with the sequence of events recorded when the GF system is stimulated. Specifically, previous electrophysiological recordings have shown that the wing depressor muscles are activated before the wings are in a position to be depressed because they have not yet been elevated. Here we show that the accepted behavioural description is wrong. Escape flight initiation actually begins with wing elevation. The current model of the GF system is revised to account for the actual sequence of events that occur when a fly escapes.