Regeneration of myelinated and unmyelinated sensory nerve fibres after a crush lesion of the rat sciatic nerve was investigated by means of retrograde labelling. The advantage of this method is that the degree of regeneration is estimated on the basis of sensory somata rather than the number of axons. Axonal counts do not reflect the number of regenerated neurons because of axonal branching and because myelinated axons form unmyelinated sprouts. Two days to 10 weeks after crushing, the distal sural or peroneal nerves were cut and exposed to fluoro-dextran. Large and small dorsal root ganglion cells that had been labelled, i.e., that had regenerated axons towards or beyond the injection site, were counted in serial sections. Large and small neurons with presumably myelinated and unmyelinated axons, respectively, were classified by immunostaining for neurofilaments. The axonal growth rate was 3.7 mm/day with no obvious differences between myelinated and unmyelinated axons. This contrasted with previous claims of two to three times faster regeneration rates of unmyelinated as compared to myelinated fibres. The initial delay was 0.55 days. Fewer small neurons were labelled relative to large neurons after crush and regeneration than in controls, indicating that regeneration of small neurons was less complete than that of large ones. This contrasted with the fact that unmyelinated axons in the regenerated sural nerve after 74 days were only slightly reduced.