Reflected reentry was produced in canine false tendons, mounted in a three-chamber bath, in which the central fiber segment was partially depolarized with an "ischemic" solution to provide an area of impaired conduction. Lidocaine (3-5 mg/l) added to the central chamber further impaired conduction; consequently, reflections were obtained at lower frequencies and over a wider range of stimulation cycle lengths. Complete block followed drug exposure in some preparations. At moderate stimulation rates, the drug-induced shift of the frequency-dependence of manifest reflected reentries was either arrhythmogenic or antiarrhythmic. The action of lidocaine can be explained on the basis of effects on depressed fast or slow response activity at the boundary regions of the ischemic gap. Under slow response conditions, in false tendons homogeneously exposed to a solution containing 20 mM K+ and 9 mM Ca++, lidocaine delayed propagation or induced complete block. In fibers mounted in a single sucrose gap, the drug increased the current required to reach threshold without significant changes in resting potential, threshold voltage or input resistance. The effect of lidocaine on threshold current was lost in Na+-deficient solutions. Thus, lidocaine impairs conduction through K+-depolarized false tendons, an effect that may be related to the Na+ background current.