Knowledge of adaptations to changes in speed and mode of progression (walking-running) in human locomotion is important for an understanding of underlying neural control mechanisms and allows a comparison with more detailed animal studies. Leg movements and muscle activity patterns were studied in ten healthy males (19-29 yr) during level walking (0.4-3.0 m X s-1) and running (1.0-9.0 m X s-1) on a motor-driven treadmill. Movements were recorded in the sagittal plane with a Selspot optoelectronic system. Recordings of EMG were made from seven different muscles of one leg by means of surface electrodes. Durations, amplitudes and relative phase relationships of angular displacements and EMG activity were analysed in relation to different phases of the stride cycle (defined by the leg movements). The durations of the entire stride cycle and of the support phase were found to decrease curvilinearly with velocity. Swing and support phase durations were linearly related to cycle duration in walking, and curvilinearly related in running. The characteristic occurrence of double support phases in walking was also seen in very slow running. Support length increased with speed up to about 1.2 m both in walking and running, but was longer in walking at the same velocity. Increases in net angular displacements were largest for hip movements and for knee flexion-extension during the swing phase in running. With increasing velocity a clear shift in relative rectus femoris activity occurred from knee extension to hip flexion. Gastrocnemius lateralis (LG) was co-activated with the other leg extensors prior to foot contact in running, whereas in walking LG was not turned on until later in the support phase. The ankle flexor tibialis anterior had its main peak of activity after touch-down in walking and before touch-down in running. The same basic structure of the stride cycle as in other animals suggests similarities in the underlying neural control. Human speed adaptation is distinguished primarily by an increase in both frequency and amplitude of leg movements and by a possibility of changing between a walking and a running type of movement pattern.