Two-photon molecular excitation microscopy has several advantages over conventional confocal fluorescence microscopy, including the ability to section deeper into scattering samples and to allow spatially resolved flash photolysis. We describe and examine the benefit of incorporating non-descanned fluorescence detection in our microscope system. In a scattering sample where almost no signal could be obtained at a depth of 50 microm with confocal detection, non-descanned detection resulted in an improvement of signal strength by more than an order of magnitude at depths >40 microm. The spatio-temporal properties of stationary spot two-photon excited flash photolysis (TPEFP) in drops of test solutions and cardiac myocytes were also examined. At input powers that produce >10% of the maximum rate of DM-nitrophen photolysis, serious photodestruction of the reporter fluorochrome (Fluo-3) at the photolysis spot occurred. At power levels of approximately 4 mW for periods <50 ms, we were able to produce small repeatable calcium release events using DM-nitrophen in cardiac myocytes, which were similar to naturally occurring calcium sparks. The properties of these artificial calcium sparks were very similar to signals obtained from drops of test solutions, suggesting that the apparent rate of calcium diffusion in myocytes is similar to the rate of diffusion of Fluo-3 in solution. Using TPEFP, we also examined the ability of a combination of EGTA and a low-affinity calcium indicator to track the time course of calcium release. Although the addition of EGTA improved the temporal fidelity of the rise of the calcium signal, it did not significantly reduce the spread of the fluorescence signal from the photolysis spot.