Gap junctions play a pivotal role for the velocity and the safety of impulse propagation in cardiac tissue. Under physiologic conditions, the specific subcellular distribution of gap junctions together with the tight packaging of the rod-shaped cardiomyocytes underlies anisotropic conduction, which is continuous at the macroscopic scale. However, when breaking down the three-dimensional network of cells into linear single cell chains, gap junctions can be shown to limit axial current flow and to induce ‘saltatory’ conduction at unchanged overall conduction velocities. In two- and three-dimensional tissue, these discontinuities disappear due to lateral averaging of depolarizing current flow at the activation wavefront. During gap junctional uncoupling, discontinuities reappear and are accompanied by slowed and meandering conduction. Critical gap junctional uncoupling reduces conduction velocities to a much larger extent than does a reduction of excitability, which suggests that the safety for conduction is higher at any given conduction velocity for gap junctional uncoupling. In uniformly structured tissue, gap junctional uncoupling is accompanied by a parallel decrease in conduction velocity. However, this is not necessarily the case for non-uniform structures like tissue expansion where partial uncoupling paradoxically increases conduction velocity and has the capacity to remove unidirectional conduction blocks. Whereas the impact of gap junctions on impulse conduction is generally assessed from the point of view of cell coupling among cardiomyocytes, it is possible that other cell types within the myocardium might be coupled to cardiomyocytes as well. In this context, it has been shown that fibroblasts establish successful conduction between sheets of cardiomyocytes over distances as long as 300 μm. This might not only explain electrical synchronization of heart transplants but might be of importance for cardiac diseases involving fibrosis. Finally, the intriguing fact that sodium channels are clustered at the intercalated disc recently rekindled the provocative question of whether gap junctions alone are responsible for impulse propagation or whether electric field mechanisms might account for conduction as well. Whereas computer simulations show the feasibility of conduction in the absence of gap junctional coupling, a definite answer to this question must await further investigations into the biophysical properties of the intercalated disc.