Groves F - The researcher used population-based data from census tracts in the Denver, Colorado metropolitan area during 1976-1983 in order to test the premise that high erythropoietin inhibits the development of leukemia and other childhood cancers. For this purpose, the link between cancer incidence and elevation was examined. The corresponding elevation was determined by using the latitude and longitude coordinates for the centroid of each census tract. Stratification of cases (N=316) into quintiles was done based on elevation of the centroid of the census tract in which they resided. Findings supported the erythropoietin-suppression hypothesis. For children residing at the first (<=5284 feet), second (5285-5353 feet), third (5354-5415 feet), fourth (5416-5570 feet), and fifth (>5570 feet) quintiles of elevation, the estimated standardized cancer incidence ratios were 80, 101, 100, 76, and 68.