Normal subjects traced sagittal lines on a graphic tablet using a stylus held in their right hand. The hand was hidden by a mirror in which they saw the lines projected from a computer screen. In normal trials, the line seen in the mirror exactly corresponded to the traced line. In perturbed trials, a bias was introduced by the computer, so that the line appeared to deviate in one direction (right or left) by a variable angle (2, 5, 7 or 10 degrees). Subjects consistently displaced their hand in the opposite direction for producing a visually sagittal line. After each trial, they were asked in which direction they thought their hand had moved. In perturbed trials, they grossly underestimated the hand deviation. In addition, a post-hoc analysis revealed that one group of subjects misperceived the direction of their hand movement in the direction opposite to the perturbation (Group 1, including 9 Ss), whereas the other group gave responses in the correct direction (Group 2, including 4 Ss). In a second session using the same experimental paradigm, a motor response was asked for: subjects had to indicate the perceived direction of their hand during each trial by drawing a line with their eyes closed. Again, responses indicated a poor conscious monitoring of motor performance. These results suggest that normal subjects are not aware of signals generated by their own movements.