The mitochondrial respiratory chain is a major site at which oxygen free radicals are generated, especially under conditions in which the respiratory chain is inhibited in the presence of oxygen. Increased mitochondrial radical production is thought to play a major part in the pathway to cell death seen at reperfusion of anoxic heart and brain. We have modelled mitochondrial-specific oxidative stress using the lipophilic cationic fluorescent dye tetramethyl rhodamine methyl ester (TMRM), which accumulates within mitochondria in response to mitochondrial membrane potential. The dye is widely used as a potential sensitive probe, but it is also highly phototoxic as illumination generates increased singlet oxygen (Duchen, 2000; Jacobson and Duchen, 2002). As the dye is localised to the mitochondria, so is the source of radical species. Illumination of astrocytes or cardiomyocytes loaded with TMRM reveals a stereotypical progression of events consisting initially of an increase in spontaneous focal transient and reversible mitochondrial depolarisations. After a period of time that varies between cell types and illumination intensity, this is followed by a global mitochondrial depolarisation. In freshly isolated ventricular cardiomyocytes this is seen as a wave that progresses along the length of the myocyte over a period of minutes. This global mitochondrial depolarisation is followed by ATP depletion, seen in myocytes as the onset of rigor, and eventually by cell death. We have found that cells can be protected by i) depleting intracellular calcium stores with thapsigargin or by buffering intracellular calcium with EGTA-AM or BAPTA-AM or ii) by cyclosporine A or sanglifehrin, that inhibit opening of the mitochondrial permeability transition pore (mPTP). Measurements of intracellular and intramitochondrial calcium concentrations suggest that oxidative stress increases the probability of SR calcium release, seen as an increased frequency of calcium sparks and waves in cardiac myocytes, followed by an increase in mitochondrial calcium loading, measured using dyes such as rhod-2. Both ryanodine and IP3 gated channels have been shown previously to show an increased probability of calcium release in response to oxidative stress. We therefore propose that these events can be readily explained by a scheme in which oxidative stress increases local ER/SR calcium release. The calcium is taken up by mitochondria which become progressively calcium loaded. The combination of mitochondrial calcium loading and mitochondrial oxidative stress increases the probability of permeability transition (Jacobson and Duchen, 2002). This may initially be transient and reversible, but eventually becomes irreversible, causing global mitochondrial depolarisation. This is followed by ATP depletion and an inevitable progression to cell death. We suggest that the basis for this response is provided by the proximity of SR/ER and mitochondria, by the mitochondrial capacity to accumulate calcium, and by the sensitivity of three channel types: the ryanodine receptor, the IP3 gated channel and the mPTP – to ROS (see Missiaen et al., 1991; Kawakami et al., 1998; Petronilli et al., 1994), largely reflecting the roles of critical thiol groups in regulating the opening probability of each class of channel.