Mammalian cells attached to membrane filters or deposited on filters without attachment were exposed to gas-phase singlet oxygen (O-1(2)) in the absence of any other reactants. Cells were exposed in a monolayer or less, in the absence of external medium, during steady-state O-1(2), generation, ensuring that singlet oxygen impinged directly and equally on all cells simultaneously. The current methodology for cell exposure ensures that O-1(2) is initially the only reactive species to which the cells are exposed. Results seen with this system can therefore be attributed solely and unambiguously to events initiated by O-1(2). Further, all cells in the sample receive the same magnitude of exposure per surface area per time interval, which supports calculations of the amount of O-1(2) required for irreversible cell damage, based on measured O-1(2), flux and exposed cell surface area. Exposure to pure O-1(2) irreversibly damaged a variety of cell types, including rat basophilic leukemia, human squamous carcinoma and Chinese hamster lung fibroblastcell lines,and murine primary hepatocytes. Cell survival curves following exposure to O-1(2) followed apparent first-order kinetics. A large number of singlet oxygen collisions (approximately 10(12)-10(13)) were required to inactivate a cell, on average, indicating a low probability that singlet oxygen collision will reduce cell survival. Regardless of cell type or the survival endpoint measured, lethal toxicity required a fairly constant number of O-1(2) collisions per cell. This poses a serious caveat in the assignment of causality in correlating O-1(2)-initiated cellular damage with mechanism of death, i.e. most damage observed will not be related to death. The importance of various toxic effects of O-1(2), whether lethal or nonlethal, will depend on the magnitude of exposure and therefore on the context in which exposure occurs.