Purpose: To clarify the relationship between neuromuscular block and train-of-four fade and to investigate the causes of these drug-dependent differences, we compared the neuromuscular block and TOF fade after pancuronium, vecuronium and rocuronium.Methods: In 24 anesthetized rats, the sciatic nerve was stimulated, and the twitch of left tibialis anterior muscle was recorded. After T 1 (first twitch response) was kept constant at 95% block by administration of pancuronium, vecuronium, or rocuronium (n=8, in each), the TOF fade was measured when T 1 block was decreased to 40% and 20%. In addition, using 24 phrenic nerve-diaphragm preparations, the fade was measured when the T 1 block increased to 20% and 40% by titrating of either one of the three drugs (n=8, in each). Results: In in vivo experiments, the fade produced by pancuronium was greater than that by vecuronium or rocuronium when T 1 block was at 40% (81 ± 9 vs 63 ± 15 and 63 ± 6%, respectively) and at 20% (66 ± 13 vs 34 ± 17 and 40 ± 6%, respectively). In contrast, in in vitro experiments, the differences did not reach significant levels among the three drugs either at 20% (32 ± 19 vs 33 ± 10 and 32 ± 17%) or 40% of block (62 ± 29 vs 65 ± 14 and 55 ± 14%). Conclusions: For vecuronium and rocuronium, the results were similar in vivo and in vitro. For pancuronium, fade was greater in vivo. These results suggest that different neuromuscular blocking agent have different relationships between the fade and the block. In vitro results might not be the same as in vivo, possibly due to pharmacokinetic differences.