Hypotheses concerning the influence of changes in the design of the human musculoskeletal system on performance cannot be tested experimentally. Computer modelling and simulation provide a research methodology that does allow manipulation of the system's design. In the present study this methodology was used to test a recently formulated hypothesis concerning the role of the biarticularity of the gastrocnemius muscle (GAS) in vertical jumping [Bobbert and van Ingen Schenau, J. Biomechanics21, 249-262 (1988)]. This was done by comparing maximal jump heights for a model equipped with biarticular GAS with a model equipped with a monoarticular GAS. It was found that jump height decreased by 10 mm when GAS was changed into a monoarticular muscle. Thus, the hypothesis formulated by Bobbert was substantiated, although quantitatively the effect is small. Our result differs from that of Pandy and Zajac [J. Biomechanics24, 1-10 (1991)], who performed similar model calculations. It is shown that the results described by these authors can be explained from the moment-arm-joint-angle relation of GAS at the knee in their model.