Incidence rates for childhood cancers (ages 0-14 years) were produced for the Denver Standard Metropolitan Statistical Area for the period 1976-1983. Some unique features of the Denver area, especially the unusually high background levels of ionizing radiation, warrant comparisons with rates generated from other regions. A total of 348 childhood cancers were identifled in white (including Spanish surname) and black children, with the 200 cancers among males resulting in higher rates than for females (148 cases) for total cancers and selected major cancer types (leukemias, glii brain cancers, and non-Hodgkin's lymphomas). Compared with other geographic areas, Denver rates were very similar, with no marked excesses or deficits except for an increase in male endocrine cancers. Rates for Spanish surname children were very similar to the rates for other whites. Cancer 591539-1542,1987. HILDHOOD CANCERS are a diverse group of malig-C nant diseases which are grouped strictly by age of occurrence. Nonetheless, several lines of evidence justify viewing these cancers as an entity for etiologic studies. The common cancers of childhood (leukemias, lympho-mas, brain tumors, soft tissue sarcomas, and Wilms' tumor) are markedly different from those of adults (lung, colon, female breast, and prostate).' Secondly, the generally accepted temporal model for the etiology of adult cancers in which many years or even decades of exposure affect the incidence of neoplasms cannot be applicable to cancers which develop so early in life. Finally, most of the established exogenous determinants of adult cancers (tobacco use, sexual behavior, and occupational exposures) are not pertinent to cancers in children. Although childhood cancers are rare compared with adult cancers: they constitute one of the most important causes of death and serious morbidity in children, with a profound emotional impact on survivors and their families. Very little is currently known about the etiology of cancers in children3 justifying further descriptive analyses of cancer incidence in children. The current report s u p plements existing literature"8 with incidence data for childhood cancer in the Denver, Colorado area. Gold for their work in reviewing the diagnoses; the registrars at the hospital and state cancer registries for their cooperation; and Ms. Esther John for her work on the computer analysis.