It has been frequently found that cross-modal matching between vision and kinaesthesis is asymmetric such that the kinaesthetic—visual matching is more accurate than the reverse condition (e.g. Connolly & Jones, 1970; Millar, 1972). The present article points out that vision and ‘kinaesthesis’ should not always be regarded as equivalent sensory channels. It is shown that if kinaesthesis is taken as passive movement (the subject's relaxed arm was moved by the experimenter) then the passive movement matching of a visual standard corresponds very well to the reverse condition. However, matching a visual standard with a voluntary excursion of the limb (active movement) differs from the other two conditions both in absolute accuracy and in proportion of undershooting to overshooting. It is argued that some operations usually regarded as kinaesthetic depend only upon central monitoring of efference while others which can also be called kinaesthetic are based upon peripheral feedback. Some implications for cross-modal matching of shapes and for theories of visualkinaesthetic memory are briefly discussed.