Tumors were induced in the skin of SKH hairless mice by exposure to fluorescent FS sun lamps or to a long-arc xenon solar simulator. Tumores developed about equally well with varying amounts of UV-A radiation (lambda greater than 320 nm) given simultaneously. In contrast, incremental changes in the UV-B region (lambda less than 320 nm) led to substantial increases in carcinogenic effectiveness. A tumor-"initiating" dose of UV-B (4-10 wk of daily FS lamp exposures) was rendered less effective by subsequent exposures of the mice to UV-A (6 hr/day, F-40 T12BL lamps). The mechanism for this effect is not known. Most tumors induced by a short course (10 wk) of FS lamp exposure grew slowly or regressed, whereas mice exposed for a longer period (30 wk) developed more tumors, and many of those that appeared early grew aggressively. Effects of daily dose fractionation were less clear, and the subject requires further study. These and other variables are being tested in a program designed to yield useful information on the effects of changing spectrum, dose, and dose delivery rates on sunlight-induced cancer.