It had long been known that the decapitated cock can cross a yard. During the last century an automatic mechanism controlling stepping movements has also been found in other vertebrates. The system controlling locomotion has many features similar to these systems controlling other natural movements: respiration (28), micturition (98), scratching (154), mastication (33), etc. Today we know that there are spinal automatisms for each limb generating its stepping movements. Activity of these automatisms depends essentially on the afferent inflow from the moving limbs. There also is interaction of the limbs during locomotion that promotes their coordination. The existence of two descending systems with different functions in the control of locomotion (Fig. 1) also can be considered as an established fact. Activity of a number of neurons involved in the control of locomotion has been studied directly during locomotion in decorticate, thalamic, and mesencephalic cats. To explain the experimental data at hand, several hypotheses of organization of the spinal automatism of stepping have been forwarded: a chain-reflex hypothesis, a hypothesis of two reciprocal half-centers, and a ring hypothesis (Fig. 2). Although general features of the system controlling locomotion are more or less clear, many questions are not yet answered. It is unknown what relative contributions to motoneuronal activity are made by proprioceptive reflexes versus influences from the automatism of stepping. Furthermore the structure of the spinal stepping automatism is not known. It is not clear if the spinal stepping automatisms of the forelimbs are as potent as those of the hindlimbs. The descending system responsible for activation of the spinal automatism of stepping has not yet been identified in direct experiments. The inputs and outputs of the subthalamic and midbrain "locomotor" regions have not been found, and we know almost nothing about intrinsic interaction of neurons in these regions. The role of inhibitory thalamic influences is scarcely known. Finally, we have no data concerning the influence of either cortical (42, 186) or visual mechanisms in locomotor control.