1. The discharge patterns of 359 single neurons in the hand representation of primary somatosensory cortex (SI) of two monkeys (Macaca mulatta) were recorded during the performance of a passive texture discrimination task with the contralateral hand (104 in area 3b, 149 in area 1, and 106 in area 2). Three nyloprint surfaces were mounted on a drum that was rotated under the digit tips. One surface was entirely smooth, whereas the other two were smooth over the first half and rough over the second half (smooth/ rough) (raised dots, 1 mm high and 1 mm diam, in a rectangular array; spatial period of 3 mm across the rows and columns for most recordings; 9 mm between columns for selected recordings). The monkeys were trained to distinguish between the smooth and smooth/rough surfaces. After the surface presentation, the monkey indicated the texture of the second half of the surface by pushing or pulling, respectively, on a lever with the other arm. For most recordings an average tangential speed of 49 mm/s was tested. For selected recordings motor speed was incremented (63, 75, or 89 mm/s). 2. Two hundred eighty-three neurons had a cutaneous receptive field (RF) on the hand (96 in area 3b, 120 in area 1, and 67 in area 2). Thirty-five neurons had a deep RF (4 in area 3b, 15 in area 1, and 16 in area 2). Seven neurons had mixed cutaneous and deep RFs (4 in area 1, 3 in area 2). Thirty-four neurons had no identifiable RF (4 in area 3b, 10 in area 1, and 20 in area 2). 3. The discharge of 185 of 359 neurons was significantly modulated during the presentation of one or both surfaces compared with the discharge at rest. Cells with a cutaneous RF that included part or all of the distal phalangeal pads of the digits used in the task (usually digits III and IV) were more likely to be modulated during surface presentation (132 of 179, 74%) than those with a cutaneous RF not in contact with the surfaces (24 of 104, 23%). The remaining neurons (mixed, deep, or no RF) were also infrequently modulated (29 of 76, 38%). 4. Of the 185 modulated units, 118 cells were classified as texture related because there was a significant difference in the discharge rate evoked by the smooth/rough and smooth surfaces. Cells with a cutaneous RF that included the digital pads in contact with the surfaces were frequently texture related (100 of 132, 76%). Texture sensitivity was less frequently observed in the remaining modulated neurons (18 of 53, 34%: cutaneous RF not in contact with the surfaces, deep RF, mixed cutaneous and deep RF, no identifiable RF). 5. Texture-related neurons were found in areas 3b, 1, and 2. Two patterns of texture-related responses were observed in the 100 cutaneous units with an RF in contact with the surfaces. Thirty-one units were classified as showing a phasic response at the time the digits encountered the leading edge of the rough half of the surface. Fifty-eight cells were classified as phasic-tonic (or sometimes tonic at the slowest motor speeds) because the response lasted for the duration of the presentation of the rough portion of the surface. The remaining 11 neurons could not be readily classified into one or the other category and, indeed, generally showed clear texture-related responses only at higher motor speeds (> 49 mm/s, 9 of 11). 6. Speed sensitivity was systematically evaluated in 41 of 100 texture-related units with a cutaneous RF in contact with the surfaces. The discharge of 66% of the units (27 of 41) varied significantly with the speed of surface presentation, with discharge increasing at higher speeds. Speed sensitivity was found in all three cytoarchitectonic areas (6 of 6 cells in area 3b, 11 of 22 in area 1, and 10 of 13 in area 2). 7. Contact force was also systematically monitored in these experiments (69 of 100 texture-related cells with a cutaneous RF in contact with the surfaces). Linear regression analyses indicated than 22% (15 of 69) of the texture-related units were sensitive to contact force (13