There was a report that midazolam-thiopental coinduction reduced ED99 of thiopental for hypnosis from 5.75 mg to 2.37 mg. This study was designed to test whether the former study can be used clinically and to compare midazolam-thiopental coinduction with thiopental or midazolam induction. After 120 patients of ASA calss I, II classified into 3 groups randomly, 5.75 mg/kg of thiopental was injected intravenously to each patient for Group I while 0.23 mg/kg of midazolam was injected intravenously to each patient for Group II. For Group III, 2.37 mg/kg of thiopental was injected intravenously to each patient one minute after 0.02 mg/kg of midazolam was injected. Induction rate, induction time, and the changes of cardiovascular response for each group have been compared and the results are as follows: 1) Induction rate was 100%, 75%, 75% for each group. 2) Time for the spontaneous eye closure and loss of the eyelid reflex for Group I were 12.2±8.0 seconds, 20.36.9 seconds respectively, while for Group II 51.5±26.5 seconds, 69.3±29.5 seconds, and for Group III, 25.3±7.4 seconds, 37.0±10.8 seconds (G. I< G. III < G. II). 3) However, heart rate was relatively higher in Group I at the point of induction but there were no significant differences between groups statistically. 4) It was also impossible to prove that the changes of blood pressure were more stable in one group than others statistically. By judging from the results above, midazolam-pentothal coinduction can be a useful means of induction but coinduction is not superior to thiopental or midazolam induction.