The use of fluorescent Ca2+ indicators to observe [Ca2+]i transients in voltage-clamped single cells has many advantages over previous methods, such as the use of aequorin in multicellular preparations, for studying excitation–contraction coupling. In the studies reviewed in this article, [Ca2+]i in single isolated mammalian ventricular myocytes was observed through the use of the fluorescent Ca2+ indicator, fura-2. Individual cells, loaded with fura-2 either by internal perfusion or by exposure to fura-2/AM, were generally studied with the use of inverted microscopes equipped with ultraviolet epifluorescence illumination, intensified silicon intensifier target cameras (ISIT), and (or) a photomultiplier tube. Analysis of subcellular patterns of fura-2 fluorescence was performed by digital analysis of the images obtained with the ISIT camera. Variation of membrane voltage and exposure of cells to ryanodine (which was assumed to selectively block the release of Ca2+ from the sarcoplasmic reticulum) were used to investigate the cellular processes that determine the [Ca2+]i transient. The main results of these studies are the following. (1) In any population of enzymatically isolated heart cells, there are (i) mechanically quiescent cells in which [Ca2+]i is spatially uniform, constant over time, and relatively low; (ii) spontaneously contracting cells, which have a relatively elevated [Ca2+]i, but in which the spatial uniformity of [Ca2+]i is interrupted periodically by spontaneous, propagating waves of high [Ca2+]i; and (iii) cells that are hypercontracted (rounded up) and that have higher levels of [Ca2+]i than the other two types. (2) In voltage-clamped cells of (i) above, (a) the amplitude (at 100 ms) of ryanodine-sensitive [Ca2+]i transients elicited by pulse depolarization (range, −30 to 80 mV) has a bell-shaped dependence on membrane voltage (maximum at 10 mV). (b) Rapid, ryanodine-sensitive "tail transients" are elicited upon repolarization from membrane potentials greater than 30 mV; their amplitude increases as the amplitude of the preceding pulse increases, (c) The amplitude of slow, ryanodine-insensitive [Ca2+]i transients increases continuously with membrane potential throughout the range −20 to 80 mV. In conclusion, the observed cellular and subcellular heterogeneity of [Ca2+]i in isolated cells indicates that experiments performed on suspensions of cells should be interpreted with caution. The spontaneous [Ca2+]i fluctuations previously observed without spatial resolution in multicellular preparations may actually be inhomogeneous at the subcellular level. The voltage dependence and pharmacology of the rapid transients elicited by pulse depolarization or by repolarization are consistent with their having arisen from Ca2+ released from the sarcoplasmic reticulum, via Ca2+-induced Ca2+ release (CICR). In particular, the "tail transients" are a clear demonstration of CICR in an intact cell under physiological conditions, since they arise from a rapid, spatially homogeneous release of Ca2+ from the sarcoplasmic reticulum that does not depend on depolarization. The Ca2+i transients remaining in the presence of ryanodine may arise from Ca2+ entering via the Na–Ca exchanger. The characteristics of these [Ca2+]i transients are consistent with certain concepts on the Na–Ca exchanger in cardiac muscle.