The techniques introduced by Kimura and Ingram et al. were applied to assess the distribution of motor refractory periods (DMRPs) in peroneal fibres of 28 diabetics with symptoms indicating polyneuropathy and in 28 controls. Results were compared with conventional motor nerve conduction velocity (MNCV) and compound muscle action potential (CMAP) measurements. MRP95 and MRP90 (the 5% and 10% slowest recovering fibres) obtained with Ingram's technique were prolonged in the diabetic patients. In the 26 patients with a value of MRP95 and MNCV, a prolonged MRP95 and a decreased MNCV were found in 12 patients. Thus conventional MNCV had a sensitivity of 46% to detect neuropathy; addition of MRP95 obtained with Ingram's technique raised the sensitivity to 73%. Specificity was 100% in both cases. With Kimura's technique or with the fast recovering fibres in Ingram's technique, it was not possible to discriminate the patients from the controls. This study indicates that measurement of the DMRP with the technique introduced by Ingram et al. improves the electrophysiological diagnosis of diabetic polyneuropathy.