Experiments were performed on canine superfused ventricular epicardial tissue slices to determine the effects of 1.0-2.0 mM heptanol, an uncoupling agent, on conduction longitudinal and transverse to myocardial fiber orientation. Conduction velocities were measured between proximal and distal pairs of epicardial electrodes oriented transverse and longitudinal to the direction of a conducted wavefront evoked by pacing at a basic cycle length of 2,000 msec from one margin of the tissue before and after the addition of heptanol. In a separate group of tissues, the dual bipolar orthogonal electrode was used to sequentially map epicardial activation at 40 to 45 sites in a 1 cm x 2 cm area before and 30 minutes after the introduction of heptanol. In a third group of tissues, transmembrane potentials were recorded with standard microelectrode techniques to determine the effects of heptanol on action potential characteristics. Heptanol did not significantly effect action potential amplitude or maximum rate of depolarization. After 1.0 mM heptanol, conduction velocity began to decrease in 1-2 minutes and reached a steady state in 15-20 minutes. Conduction velocity in the longitudinal direction decreased from a control value of 0.56 ± 0.13 to 0.46 ± 0.10 M/sec (± SD) at 30 minutes after heptanol (p = 0.005). In the transverse direction, it decreased from 0.24 ± 0.09 to 0.17 ± 0.05 M/sec (p = 0.002). The ratio of longitudinal to transverse conduction velocities increased from 2.54 ± 1.00 to 2.94 ± 0.82 (p = 0.042). Thus, heptanol preferentially slowed conduction in the transverse direction. Because heptanol did not greatly influence active membrane properties, we used cable equations to calculate the time course of the change in effective junctional resistivity, which rose from 133.2 Ω · cm before heptanol to 312.2 Ω · cm 30 minutes after heptanol administration. We conclude that heptanol slows conduction velocity by selectively increasing junctional resistivity. The preferential slowing of conduction in the transverse direction is mose likely due to the fact that more junctional resistances are encountered per unit distance in the transverse than in the longitudinal direction.