The desirability of being able to link two or more records of vital or medical events occurring to the same individual is largely self-evident. In clinical practice it is often informative to consult past records of hospital or other treatment and it is often useful to be able to locate subsequent records in order to follow up patients treated at some time in the past. Research applications of such record link age are numerous. They arise, for example, when the research worker wishes to examine associations between past and present morbidity experience and where the possibility of biased recollection might invalidate information obtained by interview. For example, the question whether x-irradiation of the pregnant mother raises the risk of leukaemia in the offspring of the pregnancy remained unresolved for several years because of doubts concerning the validity of data on radiography during pregnancy based on interviews with the mothers of affected and unaffected children. If data may be conveniently retrieved from contemporary records such doubts may be eliminated.