Recently, Kappers and Koenderink (1999 Perception 28 781-795) showed that what subjects haptically perceive as parallel often deviates greatly from what is actually physically parallel. In their experiment, subjects had to rotate a test bar in such a way that it felt as though it was parallel to a reference bar. Their data were obtained with the right hand on a table plane to the right side of the median plane of the subject. The present study extends that work in a number of ways: (1) the locations of the stimuli cover the total reachable table plane; (2) distances between stimuli can also be large (more than 1 m); (3) experiments are done both unimanually (with the right and left hand) and bimanually. Like in the previous study, the results show large systematic deviations that correlate significantly with horizontal (left-right) distance between the two bars but not with vertical (forward-backward) distance. Thus we have established that a description of the results in terms of a horizontal gradient in the deviations is valid over a large part of haptic space, over large distances, and in both unimanual and bimanual conditions. The subject-dependent horizontal gradients ranged from -12° m-1 to -27° m-1 in the present experiment. In all conditions a significant haptic oblique effect can be demonstrated.