In a multicentre double-blind, inpatient, placebo-controlled trial the effects on premature ventricular beats (PVBs) of mexiletine in a standard, submaximal dose were studied by Holter monitoring in 144 outpatients. After wash-out, mexiletine was administered for 14 days. The effects were re-tested, after one week of a placebo, in a second 14-day period of mexiletine administration. Of the patients 73% in the first period and 82.5% in the second period responded to mexiletine (a reduction of 75% or more of PVBs/24 h--p less than 0.001 compared with the placebo for both periods). Mexiletine also significantly reduced the Lown class of PVBs and the frequence of paired PBVs, ventricular tachycardia, multiform beats and R on T wave phenomenon. Mexiletine showed an equivalent effectiveness in the four main aetiological groups of arrhythmias. Fifty nine patients complained of adverse effects (gastrointestinal or neurological) the intensity of which led to the stopping of the treatment in 16 of them. These results show that mexiletine is highly effective, even in submaximal doses, in preventing ventricular arrhythmias of whatever origin.