This review summarizes research conducted in our laboratories over the past 5 years aimed at determining the temporal and spatial relationships between eye and hand movements and the amount of central processing that must occur before performing a manual aiming movement, relative to the amount of processing that is done online. All of our research to date points to a two-component model of speed-accuracy control in manual aiming. Several studies have shown that eye and hand movements in manual aiming are inextricably linked both temporally and spatially. Typically, the eye arrives in the vicinity of the target first; this coincides with peak acceleration of the finger during the initial impulse phase of a movement. There is also significant temporal and spatial coupling of the finger, elbow and shoulder in aiming, and movements appear to evolve in a proximal-to-distal fashion. Movements are endpoint driven and variability is reduced with distal approximation to the target. This movement control strategy means that visual information is not only available for use in modifying responses, but there is sufficient time available for its use. In sequential complex aiming movements, the use of visual feedback and on-line processing become even more important. Practice does not diminish the need for on-line processing; rather, its use appears to ensure greater movement efficiency.