Economically active male parasuicides admitted to the Regional Poisoning Treatment Centre, Edinburgh in 1984 were divided into two groups--those with a previous history of parasuicide and those without. The two groups were compared on a number of psychosocial and clinical characteristics to examine which were related to repetition. A further analysis (of economically active male admissions in 1985 and 1986) tested the predictive value of these discriminating characteristics. By stratification of the population, the interaction of unemployment (a discriminator and a predictor of repetition) with certain other variables was examined. A final analysis examined the risk of repetition by duration of unemployment. Unemployment discriminated those with a past history of repetition from those with no past history and also predicted repetition within a calendar year of the initial episode. The 35-64 age group were particularly vulnerable to repetition if unemployed as were those in social classes I-III and those with a past psychiatric history (inpatient or outpatient). The relationship between unemployment and parasuicide was no longer significant when the population was stratified by presence or absence of personality disorder and history of previous parasuicide, suggesting that these might be confounding variables in the relationship. The particular vulnerability of the long-term unemployed (greater than one year) to repetition was confirmed.