Human subjects were required to differentiate grating surfaces of alternating grooves and ridges by moving a finger back and forth across the surface. Their discriminative capacities were measured, as well as the movement and force profiles that they selected. To measure discrimination, a forced choice paradigm was used in which three surfaces were presented on each trial. Two surfaces were the same (standards) and the subject was required to indicate which of the three surfaces (the comparison) differed from the other two. Two series of surfaces were used with standards whose spatial periods were 770 and 1002 mu, respectively. Subjects were able to discriminate, at the 75% correct level, two gratings which differed in spatial period by the order of 5%. When tangential movement between the surface and the finger was eliminated, and only radial contact permitted, discrimination was degraded and the 75% correct levels increased to the order of 10%. Subjects were free to choose their own patterns of finger movement and of contact force between finger and surface. Movement was measured cinematographically. For all subjects movement patterns were close to sinusoidal, with frequencies in the range of 4.0 Hz and with mean velocities of the order of 160 mm/s. Patterns of contact force were measured by a force transducer. For all subjects the force varied rhythmically in synchrony with movement, but the patterns and magnitudes varied with the subject. Gratings were scaled for perceived roughness by a magnitude estimation technique: the relationship between perceived roughness and grating period was monotonic.