Atrial myocytes, lacking t-tubules, have two functionally separate groups of ryanodine receptors (RyRs): those at the periphery colocalized with dihydropyridine receptors (DHPRs), and those at the cell interior not associated with DHPRs. We have previously shown that the Ca(2+) current (I(Ca))-gated central Ca(2+) release has a fast component that is followed by a slower and delayed rising phase. The mechanisms that regulate the central Ca(2+) releases remain poorly understood. The fast central release component is highly resistant to dialysed Ca(2+) buffers, while the slower, delayed component is completely suppressed by such exogenous buffers. Here we used dialysis of Ca(2+) buffers (EGTA) into voltage-clamped rat atrial myocytes to isolate the fast component of central Ca(2+) release and examine its properties using rapid (240 Hz) two-dimensional confocal Ca(2+) imaging. We found two populations of rat atrial myocytes with respect to the ratio of central to peripheral Ca(2+) release (R(c/p)). In one population ('group 1', approximately 60% of cells), R(c/p) converged on 0.2, while in another population ('group 2', approximately 40%), R(c/p) had a Gaussian distribution with a mean value of 0.625. The fast central release component of group 2 cells appeared to result from in-focus Ca(2+) sparks on activation of I(Ca). In group 1 cells intracellular membranes associated with t-tubular structures were never seen using short exposures to membrane dyes. In most of the group 2 cells, a faint intracellular membrane staining was observed. Quantification of caffeine-releasable Ca(2+) pools consistently showed larger central Ca(2+) stores in group 2 and larger peripheral stores in group 1 cells. The R(c/p) was larger at more positive and negative voltages in group 1 cells. In contrast, in group 2 cells, the R(c/p) was constant at all voltages. In group 1 cells the gain of peripheral Ca(2+) release sites (Delta[Ca(2+)]/I(Ca)) was larger at -30 than at +20 mV, but significantly dampened at the central sites. On the other hand, the gains of peripheral and central Ca(2+) releases in group 2 cells showed no voltage dependence. Surprisingly, the voltage dependence of the fast central release component was bell-shaped and similar to that of I(Ca) in both cell groups. Removal of extracellular Ca(2+) or application of Ni(2+) (5 mM) suppressed equally I(Ca) and Ca(2+) release from the central release sites at +60 mV. Depolarization to +100 mV, where I(Ca) is absent and the Na(+)-Ca(2+) exchanger (NCX) acts in reverse mode, did not trigger the fast central Ca(2+) releases in either group, but brief reduction of [Na(+)](o) to levels equivalent to [Na(+)](i) facilitated fast peripheral and central Ca(2+) releases in group 2 myocytes, but not in group 1 myocytes. In group 2 cells, long-lasting (> 1 min) exposures to caffeine (10 mM) or ryanodine (20 microM) significantly suppressed I(Ca)-triggered central and peripheral Ca(2+) releases. Our data suggest significant diversity of local Ca(2+) signalling in rat atrial myocytes. In one group, I(Ca)-triggered peripheral Ca(2+) release propagates into the interior triggering central Ca(2+) release with significant delay. In a second group of myocytes I(Ca) triggers a significant number of central sites as rapidly and effectively as the peripheral sites, thereby producing more synchronized Ca(2+) releases throughout the myocytes. The possible presence of vestigial t-tubules and larger Ca(2+) content of central sarcoplasmic reticulum (SR) in group 2 cells may be responsible for the rapid and strong activation of central release of Ca(2+) in this subset of atrial myocytes.