We developed a three-dimensional cylindrical diffusion-reaction model of a single amphibian myofibril in which Ca(2+) release occurred only at the Z-line. The model incorporated diffusion of Ca(2+), Mg(2+), and all relevant buffer species, as well as the kinetic binding reactions between the buffers and appropriate ions. Model data was blurred according to a Gaussian approximation of the point spread function of the microscope and directly compared with experimental data obtained using the confocal spot methodology. The flux parameters were adjusted until the simulated Z-line transient matched the experimental one. This model could not simultaneously predict key parameters of the experimental M- and Z-line transients, even when model parameters were adjusted to unreasonably extreme values. Even though the model was accurate in predicting the Z-line transient under conditions of high [EGTA], it predicted a significantly narrower Ca(2+) domain than observed experimentally. We modified the model to incorporate a broader band of release centered at the Z-line. This extended release model was superior both in simultaneously predicting critical features of the Z- and M-line transients as well as the domain profile under conditions of high [EGTA]. We conclude that a model of release occurring exclusively at the Z-line cannot explain our experimental data and suggest that Ca(2+) may be released from a broader region of the sarcoplasmic reticulum than just the T-tubule-sarcoplasmic reticulum junction.