The objective of my program of research is to shed light on how texture is encoded in the nerve, to investigate texture representations in somatosensory cortex -- including Brodmann's areas 3b, 1 and 2 --, and to assess how these neural representations give rise to percepts of texture. To this end, we combine psychophysical experiments with human observers and neurophysiological recordings, both from first-order somatosensory afferents and from somatosensory cortical neurons of Rhesus macaques. First, we show that, in the peripheral nerve, the neural code for roughness -- the dominant perceptual dimension of texture -- relies on the integration of signals from all three major classes of tactile fibers, and that roughness information is encoded in spatial patterns of activation in one population of nerve fibers and in temporal patterns of activation in two others. Second, we show that these two streams of information -- spatial and temporal -- are integrated in somatosensory cortex resulting in a representation of texture that is distributed over a large population of cells with heterogeneous response properties. Combining data from our peripheral and cortical recordings, we determine which tactile submodalities drive the responses of individual cortical neurons, and show that neurons driven by afferents that specialize in spatial signaling more effectively encode coarse textural features, and neurons driven by afferents that specialize in temporal signaling more effectively encode fine textural features. Third, we show that, while texture signals at the periphery are highly dependent on the speed at which the surface moves across the finger, texture signals in cortex are nearly independent of speed and can account for the documented speed-invariance of texture perception.