The rising proportion of melanomas diagnosed at an early pathologic stage is commonly ascribed to better public education. However in the US SEER program of cancer registration it has been found that the rates for in situ melanomas are closely related by a log linear relationship to the incidence of invasive melanomas and that this relationship is unrelated to calendar year or gender or patient age. This relationship is sufficiently strong to leave little room for other factors. The relationship may be different in populations with different melanoma rates and responses to them. It is suggested that the results are due to variations within populations of individual response to melanoma cell proliferation.