The physical strain and mechanical efficiency of manual wheelchair propulsion using handrim and hubcrank propelled racing wheelchairs were studied during a submaximal wheelchair exercise test on a stationary roller ergometer. Ten healthy male able-bodied subjects conducted two exercise tests in a random order and measurements of physical strain (oxygen uptake, minute ventilation, respiratory exchange ratio, heart rate) and gross mechanical efficiency were obtained. During the experiment torque data, speed and power output were determined at a sample frequency of 0·1 Hz. Analysis of variance for repeated measures (p<0·05) was used to establish differences. The hubcrank propulsion mechanism showed a significantly lower physical strain and higher gross mechanical efficiency in comparison with the handrim propulsion mechanism. The lower strain and higher efficiency in propelling the hubcrank partly seems to be due to the continuous biphasic cyclic propulsion movement, which allows both push and pull forces to be exerted, This involues flexor and extensor muscles around elbow and shoulder, leading to a reduced tendency to fatigue in individual muscles in the upper extremity. The more natural and neutral wrist-hand orientation also seems to diminish finger flexor activity and wrist-stabilizing muscle activity, and will thus reduce physical strain both with respect to the cardiorespiratory and musculoskeletal systems. The latter may influence the tendency to develop carpal tunnel problems positively. The reduced strain of the hubcrank propulsion mechanism clearly has a number of advantages over handrims for the human engine in the short and long run. However, technical innovation should address current practical problems of steering and braking. Clearly, hubcrank can be used in low-seated wheelchairs (i.e. racing wheelchairs) only, and in subjects with a sufficiently large range of motion in the upper extremity. Moreover, the increased width is a drawback of hubcranks. Care should be taken while negotiating door posts.