Crossmodal matching of successively presented objects and shapes relies on a comparison of a stimulus representation with a target. Whereas some studies suggest a translation of modality-specific representations of visual and kinaesthetic stimuli into a representation of the target modality, other findings point to an amodal representation of stimuli. To investigate crossmodal short-term memory representations a unimodal (visual – visual, kinaesthetic – kinaesthetic) and crossmodal (visual – kinaesthetic, kinaesthetic – visual) delayed matching-to sample task was employed. Participants compared two angle-shaped movement trajectories presented either visually as a moving light spot on a screen or kinaesthetically as a passively guided hand movement applied via a manipulandum. To test for a possible translation process an interference task was inserted within the delay (6 sec). During the interference task, a visually or kinaesthetically presented ellipse (interference type), which had to be rated as vertical or horizontal, was shown either at the beginning or in the middle of the delay (interference time). An unfilled delay served as control condition. The results showed no effect of interference type on correct responses or reaction times in the interference task. Therefore, interference task difficulty was comparable for visual and kinaesthetic ellipses. Task performance in the unimodal and crossmodal delayed matching-to-sample task measured as correct responses was better in the control condition than in the interference conditions suggesting that both interference types impaired short-term memory representations. There was no effect of interference type or interference time. Thus, the results indicate an amodal spatial representation of visual and kinaesthetic information in crossmodal short-term memory rather than a recoding of modality-specific representations.