SUMMARY1.In vitrotime course studies on the effects of hexamethonium (7 mmol/L), pancuronium (5 μmol/L) and decamethonium (220 μmol/L) on nerve‐evoked (2 Hz for 2 s every 20 s) maximal twitches (T1, T2, T3, T4) of the rat hemidiaphragm were conducted. All three drugs progressively depressed all four twitches in a given train but at different rates (T4>T3>T2≫ T1).2. The response‐time profiles for T1and T4varied widely for the three drugs such that, for the same degree of T1‐block, each drug produced a different magnitude of T4‐block during the onset of and recovery from neuromuscular blockade.3. Analysis of the T1versus T4/T1plot showed that, at 50% T1‐block, the corresponding T4/T1(i.e. train‐of‐four ratios) during the onset (and recovery) phase were 0.16 (0.29), 0.46 (0.40) and 0.66 (0.53) for hexamethonium, pancuronium and decamethonium, respectively. Thus, for the same degree (i.e. 50%) of twitch (T1) tension depression, the three drugs differed widely in their ability (hexamethonium ≫ pancuronium>decamethonium) to produce fade as reflected in the respective train‐of‐four ratio.4. Our results therefore show that the train‐of‐four ratio (T4/T1) at 50% T1‐block obtained from suchin vitrotime course studies is a useful quantitative index of the potential of various drugs to cause train‐of‐four fade. Based on this index a classification of various compounds already studied is proposed as follows: hexamethonium ≫ pancuronium ∼ (+)‐tubocurarine>decameth