Fig. 3. Coupling strength of rules 1 and 2 between ipsilateral legs. Likelihood of protraction in anterior legs (receiver legs) relative to t_trans10 (see Fig. 2), i.e. time of touch-down in posterior legs (sender legs). Coupling strength is plotted for the straight walking sequence of the pre-stimulus period (solid lines) and the curve walking sequence of the second half of the stimulus period (dotted lines). Error bars indicate 95% confidence intervals. Inserts indicate the leg pair concerned (see Fig. 1B). Horizontal line segments marked with arrowheads indicate the baseline, i.e. the level expected without any coordinating influences (solid arrowheads: straight walking; open arrowheads: curve walking). The latter was evaluated from step sequence pairs taken from independent trials. Rule 1 is always acting strongly, as values prior to touch-down are close to zero and significantly below the baseline. Values are similar for straight and curve walking. Coupling strength associated with rule 2 changes with the behavioural context of the animal. During curve walking it increases in outer legs and decreases or remains similar in inner legs.