The representation of shape in the responses of slowly adapting mechanoreceptive afferent fibers (SAs) in monkeys was investigated. A series of flat plates was used, each having an increase in thickness (a step) in the middle so that one-half of the plate was thicker than the other. The cross-sectional shape of the step approximated that of a half-cycle of a sinusoid. The height of the step was fixed at 0.5 mm, while its width (half-cycle wavelength) was varied from 0 to 3.13 mm, resulting in step shapes that varied in steepness and curvature. The steps fell into 2 categories, characterized as "steep" and "gradual." A servocontrolled mechanical stimulator stroked each step across the distal fingerpad from the high to the low side of the step and back, while maintaining the contact force at 20 gm wt. Evoked action potentials in single SAs innervating the fingerpads of anesthetized monkeys were recorded. Each SA's response to a step provided a spatial response profile (discharge rate as a function of step position) that reflected the distribution of curvature across the step shape. All the major features of the SA response could be consistently explained as being due to the sensitivity of the SA to the amount and rate of change in skin curvature. The response profile was altered by changes in stroke direction, step shape, and stroke velocity. Differences in stroke direction (back and forth) were indicated by differences in pattern of response: a "burst-pause-burst" for strokes from high to low, and a "pause-burst-pause" for strokes from low to high; a greater discharge rate in response to the step for low to high strokes, and for some SAs, the reduction or absence of basal discharge in one of the directions. The discharge rate during the burst for either direction of stroking was greater for steep than for gradual steps, and increased, for a given step shape, with increases in stroke velocity. Regardless of differences in stroke velocity, steep steps were distinguished from gradual steps by having narrower burst widths for low-to-high strokes and narrower pause widths for high-to-low strokes. The same stimuli were delivered to the human fingerpad, and the capacities of humans to discriminate between the steps were measured. It was concluded that the spatial features of SA responses, representing the widths of regions of active and inactive SA populations, as well as the intensive feature of discharge rate, accounted for the gross sensory discriminations of shape.(ABSTRACT TRUNCATED AT 400 WORDS)