Prehension involves processing information in two hypothesized visuomotor channels: one for extrinsic object properties (e.g., the spatial location of objects) and one for intrinsic objects properties (e.g., shape and size). The present study asked how the two motor components that correspond to these channels (transport and grasp, respectively) are related. One way to address this question is to create a situation where unexpected changes occur at the input level of one of the visuomotor channels, and to observe how the movement reorganizes. If transport and grasp are independent components, then changing the object location, for example, should affect only the transport, not the grasp component. Subjects were requested to reach, grasp and lift as accurately as possible one of three dowels using the distal pads of the thumb and index finger. On certain trials, upon movement initiation towards the middle dowel, the dowel was made to instantaneously change its location to one of the two other positions, requiring the subject to reorient the hand to the new dowel location. Results consisted of comparing the movement characteristics of the transport and grasp components of these perturbed movements with appropriate control movements. Kinematics of the wrist trajectory showed fast adjustments, within 100 ms, to the change of dowel position. This duration seems to correspond to the minimum delay required within the visuomotor system for visual and/or proprioceptive reafferents to influence the ongoing movement. In addition, these delays are much shorter than has been found for conditions where object location changes before movement initiation (approximately 300 ms). The faster times may relate to the dynamic character of the deviant limb position signals, with the only constraint being the physiological delays for visual and kinaesthetic signals to influence the movement. A spatiotemporal variability analysis of the movement trajectories for non-perturbed trials showed variability to be greatest during the acceleration part of the movement, interpreted as due to control by a relatively inaccurate directional coding mechanism. Control during the deceleration phase, marked by low trajectory variability, was seen to be due to a sensorimotor process, using motor output signals, and resulting in an optimized trajectory supporting a successful grasp. Analysis of the grasp component of prehension showed that perturbing object location influenced the movement of the fingers suggesting a kinematic coupling of the two components. However, forthcoming work shows that, when object size changes, and location remains constant, there is a clear temporal dissociation of the two components of prehension.(ABSTRACT TRUNCATED AT 400 WORDS)