Survival, i.e. the time from report to cancer registry to death was studied for 591,456 cases of cancer diagnosed in vivo from 1960 to 1984. Ten years survival increased from 35 to 40%. Survival rates for women were higher than for men. Since 10 years survival almost suggests cure, lead time bias is assumed not to be a major factor. Nor are relaxed histological criteria, detecting non-fatal tumors, intensified microscopic examination, changes in the relative frequency of cancer types, or increasing numbers of elderly patients assumed to be major artefacts. In contrast, improved socio-economic and health status are. Early detection also improves survival in some cancer types.