Objective: The aim was to examine the ability of several previously proposed criteria and of “vector mapping” to distinguish slow conduction from conduction block in canine myocardium. Methods: Two different in vitro preparations of canine epicardium were used. In 10 tissues, an anatomical barrier was simulated by a cut (five longitudinal and five transverse to fibre orientation). Eleven tissues removed two weeks after occlusion-reperfusion infarction were also studied. Isochronal activation maps were constructed from extracellular and intracellular recordings and vector loops formed by summing two orthogonally recorded bipolar electrograms were used to indicate the direction of cardiac activation. Results: In the cut model, electrograms recorded from over the anatomical barrier were usually normal or showed double potentials but could occasionally be fractionated. Isochronal activation maps obtained from extracellular recordings were able to identify five of five anatomical barriers transverse to, but only one of five barriers longitudinal to, fibre orientation. The direction of cardiac activation indicated by vector loops identified conduction block in all 10 preparations. In 11 tissues removed from canine ventricles with experimental myocardial infarction, microelectrode recordings were used to characterise regions as either slow conduction or conduction block. Isochronal activation patterns obtained from extracellular recordings generally showed impulses proceeding through zones of conduction and around zones of conduction block but disagreed with the results of microelectrode impalements in two of 11 cases. Electrogram morphology was also not always able to distinguish slow conduction from block. The direction of cardiac activation determined by vector mapping accurately characterised all regions of tissue as showing either slow conduction or conduction block. Conclusions: Limited regions of conduction block or slow conduction are frequently present in epicardial tissues removed from experimental myocardial infarction. The morphology of extracellular electrograms and isochronal activation mapping performed from extracellular recordings is often but not always able to distinguish slow conduction from conduction block. Vector mapping is useful in distinguishing slow conduction from conduction block in these situations and may help evaluate myocardial conduction patterns. Cardiovascular Research 1994;28:259-269