1. Isolated rabbit hearts were perfused with a modified Tyrode solution containing noradrenaline in concentrations increasing stepwise from 5.9 nM to 5.9 muM at 5 min intervals. This dose regime was applied twice before and once 20 min after starting perfusion with one of 9 tricyclic drugs. Ventricular rate and right atrial and ventricular tensions were recorded using the transverse method. 2. Infusions of noradrenaline evoked ventricular arrhythmias in hearts perfused with amitriptyline 4.8 muM, chlorpromazine 5.0 muM, desipramine 5.0 muM, dibenzepine 34.7 muM, doxepin 4.7 muM, imipramine 4.7 muM, noxiptiline 9.1 muM and opipramole 9.2 muM. The incidence of arrhythmias increased with the concentration of noradrenaline applied and the dose of tricyclic drug administered. Whenever arrhythmias had started they continued as long as noradrenaline was infused. Noradrenaline failed to produce arrhythmias in hearts not exposed to drugs and after iprindole 4.7 muM or cocaine 2.9-18 muM. 3. Propranolol 0.1 muM inhibited the incidence of arrhythmias after doxepin 4.7 muM plus noradrenaline 5.9-190 nM. 4. Neuronal uptake of exogenous noradrenaline in the rabbit heart was inhibited by the tricyclic drugs in the following order of declining p potency: doxepin, noxiptiline, amitriptyline, desipramine, chlorpromazine, imipramine, dibenzepine, opipramole and iprindole. 5. Among tricyclic drugs the potency to inhibit amine uptake is related to the incidence of arrhythmias evoked by a submaximal concentration of noradrenaline. It appears, however, that these two parameters are not causally linked. 6. The isolated rabbit heart perfused with noradrenaline might be used as a model for testing the arrhythmogenic actions of tricyclic drugs and the treatment of such arrhythmias.