Sunlight exposure is strongly indicated as one of the important etiologic agents in human cutaneous malignant melanoma. However, because of the absence of good animal models, it has not been possible to estimate the wavelengths or wavelength regions involved. We have developed a useful animal model from crosses and backcrosses of platyfish (Xiphophorus maculatus) and swordtails (Xiphophorus heUeri).Two strains of these fish are susceptible to invasive melanoma induction by exposure to filtered radiation from sunlamps in the wavelength ranges A > 290 nm and A > 304 nm. Multiple exposures on 5-20 consecutive days beginning on day 5 after birth or a single exposure of =200 J/(m2 day) of A > 304 nm result in a tumor prevalence of 20% to 40% at 4 months of age compared with a background rate of 12% in one strain and 2% in another. Exposure of the fish to visible light after UV exposure reduces the prevalence to background. The melanomas are similar in many respects to mammalian melanomas, as judged by light and electron microscopy. The genetics of the crosses determined by others and the high sensitivity of the hybrids to melanoma induction indicate that the UV radiation probably inactivates the one tumor repressor gene (or a small number of tumor repressor genes) in the hybrid fish. The small size of the animals and their high susceptibility to melanoma induction make them ideal for action spectroscopy.Agents that cause a decrease in stratospheric ozone would cause an increase in UV-B (A = 280-320 nm) intensities at the earth's surface without appreciably changing the longer UV or visible intensities of light. Melanoma among the white population of the United States and Europe is increasing dramatically as a function of time, probably as a result of changing lifestyles (1). The relation between latitude and melanoma mortality suggests that there is a correlation between the average solar radiation and mortality from malignant melanoma. However, it is not known which parts of the solar spectrum can plausibly be related to the increasing mortality because, although sunlight exposure seems to be an essential component in melanoma incidence, it is not the only one. Body areas most exposed to light are not the primary locations of melanomas as they are for basal and squamous cell carcinomas. There is good evidence that UV-B is tumorigenic in animals (2) and can cause neoplastic transformation in vitro (3). Four types of experiment indicate that light energy absorbed in DNA can cause cellular damage leading to tumors: (i) the tumorigenicity of fish cells as a result of UV-irradiation in vitro can be photoreactivated (4)-a process that monomerizes UV-induced cyclobutane pyrimidine dimers in DNA; (ii) the wavelengths effective in neoplastic transformation in vitro are those absorbed by DNA (5); (iii) transformation in vitro by UV is photoreactivable (6); and (iv) xeroderma pigmentosum individuals are defective in the ability to repair UV damages in their DNA and are extraordinarily sensitive to cancer induction, inc...