The present study reviewed the modular approach in adaptive motor control by taking cognitive efficiency into account. Three experiments were conducted to compare different visuomotor learning mechanisms (modular adaptation, use-dependent plasticity, and spatial realignment) in response to visuomotor rotations. During exposure, the visual feedback of flicking movements in a single-target scenario was rotated either 30° clockwise (CW) or counterclockwise (CCW) at the left and right starting locations, respectively. Exposure to the CW and CCW rotations was carried out in an alternating order. After adaptation to the rotations, generalization was evaluated by assessing aftereffects from a set of untrained starting locations to the target (Experiments 1 and 2) or from the trained starting locations to a set of new targets (Experiment 3). Predictions made based on the different visuomotor learning mechanisms were compared to the empirical data. In spite of evidential advantages of modular structure, the current work could show a particular case of visuomotor transformation, in which modularity lacks efficiency. Results indicate that the adaptive motor control system employed the spatial realignment to accomplish adaptation more efficiently.