Arm swing in human walking is an active natural motion involving the upper extremities. Earlier studies have described the interrelationship between arms and legs during walking, but the effect of arm swing on energy expenditure and dynamic parameters during normal gait, is inconclusive. The aim of this study was to investigate the effect of restricted arm swing during walking on energy expenditure and spatio-temporal parameters. Fifteen healthy men (aged 19-29 years, height 158-192 cm, weight 50-80 kg) volunteered to walk on a Woodway (Hamburg, Germany) treadmill for 10 min at four different, random velocities, customary walking speed (CWS=100%), slow walking speed (80% CWS), fast walking speed (120% CWS) and very fast walking speed (140% CWS). Participants walked twice at each velocity: once with a natural arm swing and once with restricted arms. Sensor Medics V[subscript max]29 (California, USA), a metabolic system machine measured oxygen consumption. Heart rate was also recorded. Concurrently, selected spatial-temporal parameters were taken. Walking with restricted arms resulted in a significant increase in V[subscript O2] consumption (V[subscript O2]C) (8-34%), heart rate (8-17%) and O[subscript 2] cost (8-20%). A significant increase in cadence (F=674.71; P less than 0.000) was accompanied by an equal reduction in stride length (F=748.21; P less than 0.000). Estimating V[subscript O2]C based on changes in cadence while speeding and with restricted arms, indicated an 8-22.7% reduction in V[subscript O2]C. Restricted arm swing seems to increase V[subscript O2]C, whereas an increase in cadence tends to reduce V[subscript O2]C to maintain constrained high velocity. The results of this study are relevant to gait efficiency of patients suffering from upper extremity dysfunction such as upper-limbs amputees and Parkinson's disease or in long distance walking of healthy participants carrying loads in their arms.