Intracellular [Ca²⁺] ([Ca²⁺]_i) was imaged in atrial and ventricular rat myocytes by means of a high-speed Nipkow confocal microscope. Atrial myocytes with an absent t-tubule system on 8-di- ANEPPS staining showed an initial rise in Ca²⁺ at the periphery of the cell, which propagated to the interior of the cell. Ventricular myocytes showed a uniform rise in [Ca²⁺]_i after electrical stimulation, consistent with a prominent t-tubular network. In atrial myocytes, there was a much shorter time between the peak of the [Ca²⁺]_i transient and the peak contraction as compared to ventricular myocytes. A regional release of Ca²⁺ induced by an exposure of one end of the myocyte to caffeine with a rapid solution switcher resulted in a uniform propagation of Ca²⁺ down the length of the cell in atrial myocytes, but we found no propagation in ventricular myocytes. A staining with rhodamine 123 indicated a much greater density of mitochondria in ventricular myocytes than in atrial myocytes. Thus the atrial myocytes display a lack of “local control” of Ca²⁺ release, with propagation after the Ca²⁺ release at the periphery induced by stimulation or at one end of the cell induced by exposure to caffeine. Ventricular myocytes showed the presence of local control, as indicated by an absence of the propagation of a local caffeine-induced Ca²⁺ transient. We suggest that this finding, as well as a reduced delay between the peak of the [Ca²⁺]_i transient and the peak shortening in atrial myocytes, could be due in part to reduced Ca²⁺ buffering provided by mitochondria in atrial myocytes as opposed to ventricular myocytes.