In this experimental study, we investigated the varieties of excitability of gastrocnemius muscle via sciatic nerve as per different death models (asphyxia, abundant-blee ding and gradual-bleeding) on rats and the significance for the estimation of postmortem interval was evaluated. For this purpose, the rats were applied different stimulus intensities (5, 20, 40 mA) with 0.1 ms duration, before, during and every 5 min after death, using rectangular impulses, and the mean amplitude, onset latency and area values for each compound muscle action potential (CMAP) were elicited. It was detected that amplitude and area increased and onset latencies prolonged in the first postmortem 15 min. From the 15 min, CMAP area and amplitude showed an ever-increasing decrease and the prolongation of onset latencies became apparent. The decrease rate of area and amplitude was found to be statistically significantly different in asphyxia and abundant-bleeding models compared with in gradual-bleeding model, at 30 min measurements. However, there was not any significant difference in onset latency increase rates of three groups. Separately, any significant correlation between the agony and excitability periods among the groups could not be detected. The fact that the increase rate of onset latency did not show a significant difference as per death models indicated that onset latency ratios would be more appropriate criteria in determination of postmortem interval, regardless the reason of death. (C) 2002 Elsevier Science Ireland Ltd. All rights reserved.