After coronary occlusion, myocardium originally supplied by the occluded vessel ultimately separates into infarct and surviving muscle. To clarify this process, evolution of collateral blood flow to infarct and to surviving myocardium was retrospectively analyzed after permanent left anterior descending occlusion in 24 closed chest dogs. Microspheres were injected before occlusion and 5 and 20 minutes and 4 hours after occlusion. Ten minutes after occlusion, dogs received either verapamil, 0.4 mg/kg, followed by 0.6 mg/kg per hour for 6 hours (n = 10) or equivalent saline solution (n = 14). These dogs were sacrificed 3 days later, the distribution of the occluded artery was defined by dye perfusion and infarcted myocardium was determined by triphenyltetrazolium staining of heart slices. Surviving muscle within the distribution of the occluded artery was divided into medial regions adjacent to the infarct (medial adjacent) and remote from the infarct (medial remote) and lateral regions adjacent to infarct (lateral adjacent) and remote from the infarct (lateral remote). In both control and verapamil groups, collateral flows in all regions increased significantly by 140 to 400% over 4 hours. However, the relative magnitude of collateral flow to different regions showed a consistent order: infarct less than medial adjacent less than medial remote less than lateral remote. There were no significant differences in regionally matched flows or size of infarction between control and verapamil-treated groups. Collateral perfusion begins to show distinctive patterns of change in infarct and surviving muscle very soon after coronary occlusion. Collateral flow within subdivisions of the occluded coronary artery bed increases as distance from the infarct increases, with lateral segments having higher flows than medial segments. This relation persists during the first 4 hours after occlusion. In this study, verapamil neither increased collateral flow to the occluded bed nor altered minimal flow requirements for myocardial survival.