On the basis of these considerations, the possible action spectrum for
melanoma can be narrowed considerably, but not confined to any one solar emission band. The physical factors discussed eliminate all but UV, visible, and NIR radiation as possible solar agents. Ionizing radiation fits neither the epidemiologic data nor first-order physical considerations. Wavelengths longer than the NIR wavelengths, although they could conceivably account for the occurrence of
melanoma under clothed parts of the body, carry so little energy that they are probably unimportant. Epidemiologic evidence regarding the effects of skin pigment favors UV or visible radiation. A distinction between these two components is not obvious; UV-C and UV-B photons carry greater energy and are more likely to induce biochemical cutaneous effects, but the total flux in the UV-A and visible radiations is far greater. That UV-B radiation may play a role in
melanoma is supported; at the same time, one cannot exclude the possibility that the action spectrum for
melanoma is, instead, the UV-A, the visible, or even the NIR portion of the sunlight spectrum. The strong differential effect of altitude on the transmission of light of different wavelengths might serve as an important discriminating variable. If solar UV radiation is implicated in the development of
melanoma, then altitude should emerge as a significant factor in epidemiologic studies. If visible or IR radiation is the active agent, then differences on the basis of altitude should be small or negligible. Intrinsic solar variations that follow the annual sunspot number appear inadequate in either the UV or the visible band to account directly for the apparent 11-year modulation of
melanoma incidence found in some registries. Secondary atmospheric effects brought about by the action of solar UV changes on the
ozone layer may be adequate to explain a weak 11-year modulation in
melanoma incidence, although continuous measurements of UV-B flux made at sites in the United States through a full solar cycle have shown no such effect. Nor do these early measurements reveal the long-term increase in UV-B intensity expected from the destruction of
stratospheric ozone by industrial
pollutants over the last 10 years.