We previously reported that the kinematics of reaching movements reflect the superimposition of two separate control mechanisms specifying the hand's spatial trajectory and its final equilibrium position (Scheidt and Ghez In revision). We now asked whether the brain maintains separate representations of the spatial goal for planning hand trajectory and final position. One group of subjects learned a 30° visuomotor rotation about the hand's starting point while performing a movement reversal task ("slicing") in which they reversed direction at one target and terminated movement at another. This task required accuracy in acquiring a target mid-movement. A second group adapted while moving to - and stabilizing about - a single target ("reaching"). This task required accuracy in final positioning. We examined how learning in the two tasks generalized both to movements made from untrained initial positions and to movements towards untrained targets. Shifting initial hand position had differential effects on the location of reversals and final positions: Trajectory directions remained unchanged and reversal locations were displaced in slicing whereas final positions of both reaches and slices were relatively unchanged. Generalization across directions in slicing was consistent with a hand-centered representation of desired reversal point as demonstrated previously for this task whereas the distributions of final positions were consistent with an eye-centered representation as found previously in studies of pointing in 3-dimensional space. Our findings indicate that the intended trajectory and final position are represented in different coordinate frames, reconciling previous conflicting claims of vectorial and eye-centered representations in reach planning.