Continuous measurements of anterior-posterior leg position recorded from stick insects walking on a wheel were tested for relationships among spatial and temporal parameters of leg coordination. This analysis revealed that the protraction of middle and rear legs is guided by the ipsilateral front and middle legs respectively. Protraction endpoint for each rear leg shows a significant positive correlation with the simultaneous position of the ipsilateral middle leg (Figs 1,2; Table 1). An analogous, but somewhat weaker, correlation exists between the protraction endpoint of each middle leg and the position of the ipsilateral front leg.
This coordination of spatial parameters was tested experimentally by manipulating the position of the forward leg. When a middle leg is restrained in various positions, the ipsilateral rear leg adjusts its protraction endpoint accordingly (Fig. 3). However, its retraction endpoint does not undergo parallel shifts; consequently, step amplitude, protraction duration, and step frequency all change as a function of middle leg position. When a sinusoidal movement is imposed on either a middle or front leg, the adjacent, caudal leg continuously adjusts its protraction endpoint according to the momentary position of the forward leg (Fig. 4). This adjustment is again accompanied by changes in step amplitude and step period, changes which may affect all five unrestrained legs.
The anterior-posterior leg position measured in our experiments primarily reflects the angle of the coxo-thoracic joint; this angle is monitored by hair rows and hairplates located on the coxa (Wendler, 1964; Baessler, 1965). Modifying these external proprioceptive inputs revealed both inter- and intrasegmental control functions. The caudally situated hair rows are important for measuring the small variations in the position of the target leg which occur during normal walking. Immobilization of these hairs on a middle leg causes the mean protraction endpoint of the ipsilateral rear leg to shift forward (Fig. 5: o1 versus C) and reduces or eliminates the step by step correlation of this protraction endpoint with middle leg position (Table 1). The additional immobilization of the cranially situated hairplates usually leads to a caudal shift in the protraction endpoint of the ipsilateral rear leg (Fig. 5: o2 versus o1) and reduces any residual correlation (Table 1). The actual position of the protraction endpoint reflects an integration of intersegmental signals representing the position of the target leg and intrasegmental signals from the sensory hairs on the protracting leg. Both operations may affect the duration of protraction in both the operated target leg and the adjacent, caudal leg.