The relationship between body size (adult height and weight) and cancer incidence was investigated in an international ecological study of 24 populations. Site-specific and total cancer incidence rates (age standardized) from 1973 to 1977 were correlated with body size data generally obtained between 1954 and 1974. All-sites cancer incidence was highly correlated with height among both men (r = 0.50; p less than or equal to 0.01) and women (r = 0.70; p less than or equal to 0.001). Among men, there were significant correlations between height and cancers of the central nervous system (r = 0.72), prostate (r = 0.66), bladder (r = 0.65), pancreas (r = 0.59), lung (r = 0.47), and colon (r = 0.46). Significant correlations were observed for cancers of the rectum (r = 0.76), pancreas (r = 0.75), ovary (r = 0.73), central nervous system (r = 0.68), breast (r = 0.65), uterine corpus (r = 0.50), and bladder (r = 0.48) in women. Adjustment for weight altered these correlations only minimally. Weight was significantly correlated to all-sites cancer only among women (r = 0.44; p less than 0.05), and site-specific correlations were significant for the same sites as for height, but the magnitude of the correlation coefficients was somewhat diminished. In addition, adjustment for height greatly reduced the correlations with weight. These findings support previously observed associations between height and specific cancers (e.g., breast and colon) and identify several additional cancer sites that may be similarly related.