Motor conduction velocity (MCV) was measured in the median, ulnar, and peroneal nerves of 52 referents and 95 workers from a shoe factory, in a search for dose-effect and dose-response relationships between exposure to organic solvents and subclinical impairment of the peripheral nervous system (PNS). According to the environmental concentrations, the exposure was expressed as the sum of the products between the median hygienic effect (ranging from 0.08 to 2.89) and worktime (from 1 to 29 a) in every job. The motor action potential (MAP) amplitudes, durations and shapes of the exposed workers were significantly affected even if to a different extent and significance level when compared to those of the MCVs of the referents. The MCVs of the median (t = 3.17, p less than 0.01) and peroneal (t = 2.11, p less than 0.05) nerves were reduced as compared to reference values. In the exposed group, the MCV of the median nerve was negatively correlated with exposure score (r = 0.45, p less than 0.01) but not with age. No relationship was found between MVCs of the ulnar and peroneal nerves and exposure. The MCV of the median nerve was particularly slow in a subgroup of workers with an exposure score of greater than 20 (t = 2.30, p less than 0.05 vs the other exposed workers; t = 5.56, p less than 0.01 vs the referents). This exposure score represents only 50% of the maximum allowable score for a worktime of 40 a. Even if none of the examined workers showed clinical signs of polyneuropathy, evidence of subclinical effects on the PNS were found among subjects with long-term exposure to hexacarbon mixtures. Therefore, our results suggest the need for a lowering of current threshold limit values at least for n-hexane.