The time-course for the development of ultraviolet A-induced erythema in psoralen-sensitized skin differs from that caused by ultraviolet B or ultraviolet A but objective data are not available. During psoralen ultraviolet A therapy, the minimal phototoxic dose is determined 72 h after exposure, when psoralen ultraviolet A erythema is assumed to be maximal. This measurement is of fundamental importance in optimizing the therapeutic regimen. We examined a detailed time-course for development of psoralen ultraviolet A erythema in 16 subjects. The erythemal responses to ultraviolet B, ultraviolet A and psoralen ultraviolet A were assessed visually and using a reflectance device. Ultraviolet B erythema was maximal 24 h after exposure compared with subsequent time-points. Psoralen ultraviolet A erythema was evident at 24 h, with reduction in the median ultraviolet A minimal erythema dose from 14 to 5 J per cm2 in the presence of psoralen (p < 0.01; n = 9). Peak psoralen ultraviolet A erythema, assessed by minimal phototoxic dose, did not occur until 96 h or later in 75% of subjects. Using individual dose- response curves, we determined that only 67% of mean maximum psoralen ultraviolet A erythemal intensity had developed by 72 h. Furthermore, at the time of maximal erythema, the slope of the psoralen ultraviolet A dose-response curve was approximately 2-fold shallower than that for ultraviolet B-induced erythema. If assessment of psoralen ultraviolet A erythemal sensitivity had been made at 96 h instead of the conventional 72 h time-point, peak erythemal responses would not have been missed in any of the subjects. Based on these findings, it seems appropriate to consider whether psoralen ultraviolet A minimal phototoxic dose measurements should be performed 96 h after exposure.