The aims of this paper were to identify patients with tactile neglect, to corroborate its independence from visual neglect, to ascertain its presence along the radial dimension of space and to detect a gradient of performance, as recently reported for visual neglect. To this end eight patients with vascular right cerebral lesions and 28 healthy controls have been tested with a modified version of the tactile maze test of De Renzi et al. [Cortex, vol. 6, pp. 191-203, 1970] and a seven-test clinical battery to detect the presence of visual USN. Four patients presenting tactile USN were selected: two showed left tactile neglect while the other two showed tactile neglect for far space. The distribution of omissions showed a clear spatial gradient. Since our patients were blindfolded, neglect behaviour in the tactile domain cannot be attributed to the attractive power of visual information. However, 25% of the normal subjects were also significantly faster in the 'near' space pointing towards the possibility that neglect for 'far' space can reflect the exaggeration of a normal asymmetry.