This study investigated the role of different types of discontinuities in tissue architecture on the spatial distribution of the transmembrane potential. Specifically, we tested the occurrence of so-called 'secondary sources,' ie, localized hyperpolarizations and depolarizations during the application of extracellular electrical shocks (EESs). Changes in transmembrane potential relative to action potential amplitude (ΔV(m)/APA) were measured in patterned cultures of neonatal rat myocytes, stained with voltage sensitive dye (RH-237), by optical mapping (96-channel photodiode array, 6- to 30-μm resolution) during the application of EES (field strength, 8 to 22 V/cm; duration, 6 ms). Across narrow cell strands (width, 218±59 [mean±SD] μm), EES applied during the relative refractory period produced a linear and symmetrical profile of ΔV(m)/APA (-65±23% maximal hyperpolarization versus +64±15% maximal depolarization). In contrast, the profile of ΔV(m)/APA was asymmetrical when EESs were applied during the action potential plateau (-95±32% versus +37±14%). At high magnification, no secondary sources were observed at the borders between cells. In dense isotropic cell monolayers or in monolayers and strands showing intercellular clefts, secondary sources were frequently observed. Intercellular clefts of the size of one to several myocytes were sufficient to produce secondary sources of the same magnitude as those that elicited action potentials in dense cell strands. There was a close correlation between the location of secondary sources during EES and localized conduction slowing during propagation. Thus, densely packed cultured cell strands behave as an electrical continuum with no secondary sources occurring at cell borders. Small intercellular clefts can create secondary sources of sufficient magnitude to exert a stimulatory effect.