Hip flexor (SART) activity precedes activity in ankle flexors (TA) during fictive locomotion by an average ~20ms (Quevedo et al, SFN 2000). Is this sequential activation a reflection of the intrinsic organization of the CPG or does it change as locomotion changes? Detailed analysis of the ENGs reveals variability in the onset delay of TA activity. In locomotor bouts with a frequency of 1 - 1.5Hz (normal) the start of activity in SART almost always preceded that in TA. However, in a few steps, the onset of activity in SART followed that in TA. This change in the sequence of flexor activity onset was more prevalent when the locomotion was irregular, or the frequency not constant. The variation in the onset delays is not associated with the frequency of locomotion, nor the burst duration of TA or SART activity. The onset delays for the normal locomotor bouts ranged from 9 to 31ms. The burst durations of TA and SART were similar in most of the bouts. The variability in the onset delay might be associated with ...