Fig. 11. Model for achieving take-off performance. (A) We hypothesize that a minimum of four independent pathways are required to coordinate take-off behavior: one coordinating wing opening on either side of the body and two coordinating different types of leg extension. In our model, take-off performance is determined by both the latency between activation of wing and leg pathways and the choice of leg pathway. In our diagram ₁ represents the delay due to sensory and/or central processing before one or both wing pathways are activated, and ₂ represents the delay before activation of one of the leg pathways. The difference between these two delay times is the observed wing–leg interval, . We propose that which leg pathway is activated for a given take-off determines the speed of that take-off. Alt., alternate. (B) The latency between wing and leg pathway activation determines take-off steadiness. This model is supported by our data: the graph shows the time between first wing motion and first leg motion plotted against the resulting take-off steadiness S for each fly observed (N=43). Upward-pointing triangles represent voluntary take-offs, while upside-down triangles mark escape responses. The fill color of the upside-down triangles (escapes) indicates the conditions of the wings during take-off, as defined in Fig. 10: black, ^**Esc. (N=5); white, ^*Esc. (N=5); gray, Esc. (N=17). The green line is a best-fit linear regression to the data. The line has a positive slope, indicating a direct correlation between and steadiness. (C) Summary of how coordination of the hypothesized pathways leads to the observed differences in take-off performance.