This thesis is based on ten published articles. The experimental work was carried out at the Faculty of Health Sciences, University of Copenhagen. The aim was to investigate and describe a number of basic mechanical and physiological mechanisms behind human walking. The methodologies used were biomechanical movement analysis and electrophysiology. The walking experiments were carried out in a gait lab, where the subjects were video recorded while they walked across two force platforms, which measured the ground reaction forces. Net joint moments about the hip-, knee- and ankle joint were calculated by combining the movement data and the external reaction forces (inverse dynamics). Muscle activity and sensory input to the spinal cord were measured by electromyography (EMG) and electrical stimulation of peripheral nerves. The results showed that the gait pattern varies to a great degree between individuals. Some people choose to exert the highest forces about the ankle joint while others prefer to use the knee joint. By use of a cluster analysis, fifteen healthy subjects could be divided into two groups. The extensor moment about the knee joint was the main factor for separating the two gait patterns, but the group with the highest extensor moments about the knee joint also walked with more flexed knee joints, higher EMG activity in the quadriceps muscle and higher bone-on-bone forces. This may lead to development of osteoarthritis over the years. Walking on high-heeled shoes reduced the ankle joint moment significantly either because of reduced muscle fiber length and/or increased co-contraction about the joint. On the contrary, the extensor moment about the knee joint was almost doubled in the high-heeled condition compared to bare footed walking at the same velocity. Also the EMG activity increased in the leg muscles. This could be an explanation pertaining to the higher incidence of osteoarthritis in women than in men. Patients with a drop-foot cannot put the foot to the ground with the heel first. Moreover, they have to increase flexion of the hip joint during the swing phase because the foot hangs in a plantar flexed position. It was shown that the ankle joint plantar flexor moment increased in the healthy leg and that the knee joint extensor moment increased significantly in both the affected and the healthy leg. The latter is most likely due to the patients trying to avoid an asymmetrical gait pattern. It is recommended to use an orthosis with drop-foot patients in order to keep the ankle joint dorsiflexed prior to touchdown, otherwise bone-on-bone forces in both knee joints will increase and probably lead to osteoarthritis. The hip joint moment varies less between individuals. However, both during walking and running an unexplained hip joint flexor moment is present during the last half of the stance phase. The moment appears to oppose the speed of progression and it has been suggested that it serves to balance the upper body. This was investigated in a group of healthy subjects who were asked to walk with their upper body in a reclined, inclined and normal position, respectively. It was shown that the hip joint flexor moment was similar in the reclined and the normal position but lower when walking in the inclined position and it can be concluded that the upper body is not balanced by hip joint flexor muscles but rather by accelerations of the pelvis and activity in abdominal and back muscles. These experiments also showed that the trailing leg is brought forward during the swing phase without activity in the flexor muscles about the hip joint. This was verified by the absence of EMG activity in the iliacus muscle measured by intramuscular wire electrodes. Instead the strong ligaments restricting hip joint extension are stretched during the first half of the swing phase thereby storing elastic energy, which is released during the last half of the stance phase and accelerating the leg into the swing phase. This is considered an important energy conserving feature of human walking. The gating of sensory input to the spinal cord during walking and running was investigated by use of the Hoffmann (H) reflex in m. soleus and m. gastrocnemius medialis. This reflex expresses the central component of the stretch reflex, i.e. the transmission from Ia afferents to α-motoneurones in the spinal cord. The soleus H-reflex was shown to be strongly modulated during the gait cycle. In general, it was facilitated in the stance phase and suppressed in the swing phase. However, as it was the case with the biomechanical parameters, inter-individual H-reflex modulations were found and they were highly reproducible between days. One group of subjects had an almost completely suppressed H-reflex during the entire swing phase, while another group showed a gradually increasing reflex excitability during the swing phase. This group also walked with a lower extensor moment about the knee joint and higher plantar flexor moment about the ankle joint and it is speculated that this gait pattern highly relies on reflexes to deal with unexpected perturbations. The subjects with the suppressed reflex during the swing phase also showed a higher EMG activity in the anterior tibial muscle, so it is likely that the suppression of the H-reflex was at least partly due to reciprocal antagonist inhibition. All subjects showed complete suppression of the H-reflex at toeoff. This seems necessary to avoid a stretch reflex being elicited in the soleus muscle as the ankle joint undergoes a fast dorsiflexion just after toeoff. The reflex modulation is clearly an integrated part of the human gait pattern and is absolutely necessary for normal gait function with smoothe movements. Furthermore, it is anticipated that the afferent input from the muscle spindles is used to drive the motor output from the α-motoneurones together with descending activity from the motor cortex. During running the H-reflex increased in both the soleus and the gastrocnemius already before heel strike and before the onset of EMG activity in the same two muscles and with a relatively high activity in the anterior tibial muscle, but this was most pronounced in the soleus. The H-reflex was always higher in the soleus also when expressed as percentage of the maximal M-wave. This is due to the difference in muscle fiber type distribution between the two muscles. The H-reflex increased from walking to running in both muscles and further with increasing running speed. Unexpectedly, there were no signs of the faster gastrocnemius becoming more important at higher running speed. During walking it is not possible to observe a stretch reflex in the form of a synchronized activation of a large number of muscle fibers as this would disturb the movement pattern. It is rather likely that the input from Ia afferents directly contributes to activate the α-motoneurones. However, during running the stance phase is much shorter, which enables the possibility of a stretch reflex to contribute to a strong contraction during push-off. EMG peaks in the soleus with an appropriate latency were observed in the soleus during running. This was not the case with the gastrocnemius and the explanation is most likely that the gastrocnemius is biarticular and not stretched to any great extent during running.