Abstract. In vivo antiarrhythmic effects of diltiazem hydrochloride and nifekalant hydrochloride, a pure class III antiarrhythmic drug (Vaughan Williams' classification), on adrenaline induced ventricular arrhythmias were examined in halothane anesthetized guinea pigs. Continuous adrenaline infusion (12.5 m g / kg per min) induced ventricular arrhythmias. Arrhythmogenicity was significantly increased with vagotomy and higher concentration of halothane. After injection of diltiazem at 0.5 mg / kg, the arrhythmic ratio (the number of ventricular ectopic beats divided by the total heart beats) was significantly reduced compared with the predrug control value (0.69 vs 0.04, P<0.05). No significant change of arrhythmic ratio was observed after injection of nifekalant (0.57 vs 0.61, ns). After administration of nifekalant, the mean minimum adrenaline infusion rate that induced ventricular arrhythmia decreased from 9.29 to 6.43 m g / kg per min. On the other hand, before administration of diltiazem, the mean arrhythmogenic rate of adrenaline was 8.50 mg / kg per min, but ventricular arrhythmias were no longer induced during continuous infusion of diltiazem at 0.5 mg / kg per min. These results were qualitatively consistent with previous experiments using the canine halothane-adrenaline model. In conclusion, the halothaneadrenaline arrhythmia model using the in vivo guinea pig is useful for screening drugs with potential anti-or pro-arrhythmic properties.