Muscle activation patterns of embryonic chick hind limb muscles were determined from electromyographic (e.m.g.) recordings in an isolated spinal cord/hind limb preparation of stage 34-36 embryos, and were compared with in ovo e.m.g. activity from similarly staged embryos. Muscle activity in ovo consisted of periodically recurring sequences of bursts during which antagonistic muscles often alternated and synergistic muscles were co-active, as compatible with their mature function. However, more variable behaviour was also observed. Burst sequences in ovo were often initiated by a short-duration, high-amplitude discharge that occurred synchronously in all muscles studied, and which was followed by a period of electrical silence that was longest in the flexor muscles. This type of activity has not been described previously in mature animals. In ovo movement sequences were generally initiated by extensor activity which progressively declined in duration and intensity throughout the sequence, while flexor activity progressively intensified. The onset of activity in extensor muscles was accompanied by an abrupt decrease in flexor activity, whereas the converse was not observed. Spontaneous movement sequences also occurred when the spinal cord and hind limb were isolated and maintained in oxygenated Tyrode solution for several hours. Deafferentation experiments indicated that the motor pattern in this preparation was generated centrally by circuits within the spinal cord. Activity from the isolated cord was less variable than that occurring in ovo, consisting of sequences of highly regular recurring bursts. Each burst began with a brief high-amplitude discharge that occurred synchronously in all muscles and which was similar to that observed in ovo. This was followed by a silent period, which was longest in the flexors, and then by a more prolonged burst. Although its behaviour differs from that in ovo in some respects, it is concluded that the isolated cord maintained in vitro produces a spontaneous and patterned motor output.