Forty-four limbs from 11 healthy volunteers were examined. Spinal and scalp somatosensory evoked potentials to median and peroneal nerve stimulation were recorded and the peripheral (wrist-Erb, Erb-cervical, knee-thoracic spine) and central (cervical-scalp, thoracic-cervical spine, spine-scalp) conduction times and velocities (CTs, CVs) were calculated. Sensory and mixed trunks of median and peroneal nerves were also stimulated and their motor and sensory CVs in mid-distal districts were measured. Motor responses to scalp (motor areas for hand and leg muscles) and spinal cord stimulation (cervical and lower thoracic levels) were carried out through skin rectangular plate electrodes delivering high voltage (880-1870 V) brief anodal pulses. The intracranial (scalp-cervical) and intraspinal (cervical-thoracic spine) CTs and CVs of motor pathways were measured. The elbow-cervical and knee-thoracic spines CTs of motor fibres were also calculated through the F wave method, which gave values almost superimposable on those obtained through direct spine stimulation. Nerve propagation was faster in sensory than in motor fibres in peripheral nerve mid-distal districts, while this difference was reduced or reversed in more proximal segments, including nerve roots. The scalp-cervical CT was slightly shorter in motor than in sensory fibres after subtraction of synaptic delays (6.12 vs. 6.18 msec). The scalp-lower thoracic spine, as well as the intraspinal, CVs were 7-12% faster in sensory than in motor pathways (45.3 vs. 38.7 m/sec for the former; 62.65 vs. 55.4 m/sec for the latter). The reported method allows the evaluation of fast conducting motor and sensory pathways along 'central' and 'peripheral' nerve structures of the entire body. Preliminary findings on scalp stimulation of brain motor areas with low voltage pulses are also included.