It has been hypothesized (Lederman et al, 1990 Perception & Psychophysics 47 54-64) that sighted people adopt a visual translation process when attempting to identify 2-D raised images by touch they employ a visual image as a mediator between haptic sensory information and the object representation. If this hypothesis is correct, the performance in identifying pictures by touch (with eyes closed) ought to be better when the head is facing the picture than when facing in a very different direction. In this study, thirty-six blindfolded participants were required to identify raised pictures of common objects with their head facing either in the same direction as the raised picture or in an orthogonal direction. Identification performance was measured in terms of accuracy and response latencies. Overall, participants were more accurate and faster when their heads faced in the same direction as the picture. This finding is discussed in terms of spatial congruency between haptic representations of pictures and visual (or spatial) imagery created during exploration of haptic pictures.