The regional cerebral blood flow (rCBF) was measured with 15O-butanol and positron emission tomography (PET) in 10 healthy subjects in order to compare cerebral activation involved in the somatosensory discrimination of microgeometric features with cerebral activation associated with the discrimination of macrogeometric features. Subjects performed two-alternative forced choice (2-AFC) discriminations of pairs of stimuli from a series of quantified standardized stimuli that differed in roughness (microgeometry), and a separate 2-AFC task of smooth tactile stimuli that differed in length (macrogeometry). Results are presented from three conditions: (1) a roughness discrimination task; (2) a length discrimination task; and (3) a control trial in which subjects were required to reproduce similar exploratory finger movements only, but without a specific stimulus to feel. Mean subtraction images were computed using the computerized adjustable brain atlas of Greitz et al. (1991, J. Comput. Assisted Tomogr., 15, 26-38) and areas of significant blood flow change were identified. Both the roughness and the length discrimination tasks activated overlapping cortical fields contralaterally in the anterior and posterior lip of the postcentral sulcus. However, in the length discrimination, activation of the posterior lip of the postcentral sulcus extended deeper into the sulcus and there was also a separate additional area of activation in the anterior part of the precentral gyrus. Furthermore, the length discrimination task activated fields in the overt part of the supramarginal gyrus bilaterally as well as fields in the angular gyrus bilaterally. Thus roughness discrimination uses only a subset of the cortical regions that are needed for the recovery of length information, which requires more extensive somatosensory processing. This finding may be partly explained in that length perception needs both edge detection of the stimuli used, as well as integrated information of surface length and velocity, which is not necessary for roughness perception. Specific differences in the acquisition of necessary tactile information between the two discrimination tasks was reflected in different sampling strategies.