Injury and poisoning episodes at the Regional Poisoning Treatment Centre, Edinburgh, were compared for the years 1980-1986. Injury was more likely to be associated with youth, male sex and single status. When these demographic differences were controlled for, injury was found to be associated with the following clinical variables: personality disorder, previous psychiatric inpatient status, use of alcohol at the time of parasuicide and alcohol problems, and the social variables-violence against others and received from relatives in the preceding 5 years, criminal record and unemployment. These have been consistently associated with repetition of parasuicide. A history of injury as a putative marker of repetition and the clinical implication of its association with alcohol problems are discussed.