JOSEPH F. FRAUMENI, JR., M.D. of the National Cancer Institute to analyze cancer mortality for all counties of the contiguous United States, 1950-69, by age, race, sex, and site of cancer. This report presents, for each councy in the United States, the coral with high rates, and determining if these areas share certain demographic or environmental variables. In approaching from the opposite direction, it should be possible to evaluate ideas or hypotheses by identifying counties the origin of certain cancers by identifying counties and groups of counties logic purposes. It should be possible to generate clues or hypotheses to The mortality data in this report may be used for various epidemio- have populations which are too small to obtain reliable mortality rates for certain cancers, but sufficient numbers are generally available to calculate meaningful values. Furthermore, the mortality data can be correlated with demographic and environmental variables collected by governmental and other agencies, and available on a county level. demographic characteristics and environmental exposures. Some counties As the geographic unit for study, the county has advantages over larger areas in the greater homogeneity of its population with respect to available recently from the 1970 census. 1950-67. The current report has the advantage of two additional calendar years of mortalicy data, and improved population estimates which became publication (National Cancer Institute Monograph No. 33) age-adjusted fates were computed for each state and nationwide for the 18 year period, standards for comparison, age-adjusted cancer death rates are given also for the 48 contiguous states and for the total United States. In an earlier and race (whites and nonwhites) over the 20-year period, 1950-69. As number of cancer deaths and age-adjusted death rates according to sex National Center for Heaith Statistics enabled the Epidemiology Branch smaller geographic scale. Recently, a series of magnetic tapes from the with unusual environmental or demographic characteristics, and determining whether these areas have peculiarities in the frequency of particular cancers. To help visualize the distribution of counties with significancy high rates throughout the country, a series of illustrative maps is being prepared by the Epidemiology Branch, and will be published as an atlas to supplement this report. Finally, the informacion herein should be useful to public health and administrative officials concerned with the impact of cancer in various communities in the United States. Information on geographic distribution of cancer has been particularly useful in developing and testing etiologic hypotheses. International variation is often striking, and has provided many leads to environmental and ethnic factors involved in the origins of cancer. Less well known is the geographic variation for cancer within countries. In the United States, cancer mortality statistics have been analyzed for geographic divisions and states, but no comprehensive surveys of cancer have been made on a INTRODUCTION