The aim of a registry is to obtain information from all new cases in a well defined geographic area. They represent a source of data of great scientific quality. They have the double interest of describing and monitoring the cancer risk and to conduct research from collected data or punctual investigations. In the field of descriptive epidemiology, they allow to estimate national incidence, to predict times trends in incidence, to estimate the prevalence or to analyse the geographic distribution of cancers. It is important to dispose of representative health care practices information, really in use, those of specialised centres series being affected by selection bias. Population-based survival, as computed by cancer registries is an indicator of the effectiveness of health systems in cancer control. In the field of analytic epidemiology, information is measured at an individual level in the aim of identifying risk factors. For such studies a cancer registry is not indispensable but the constituted net correspondent make them easier in a rapid identification of new cases and in avoiding many bias. Population-based registries are also an essential tool for the evaluation of organized mass-screening programmes.