Aggregate- and individual-level trends in parasuicide and unemployment among men and women resident in Edinburgh during the years 1968-87 are reported. During the first half of the period unemployment and parasuicide rates increased in parallel (r = 0.91, P less than 0.001 among men; r = 0.82, P less than 0.01 among women). However, while unemployment continued to rise after 1977, parasuicide rates tended to fall (r = -0.83, P less than 0.001 among men; r = -0.38, NS among women). It could be shown that the reversed relationship between parasuicide and unemployment after 1976 was not due to changes over time in the validity of the local unemployment rate as a marker of economic conditions. Suggested explanations for the falling parasuicide rate in Edinburgh include the possibility of a "threshold" effect and the influence of other (non-economic) factors. The rate of parasuicide among the unemployed fell, especially after 1973, while the rate among the employed rose until 1976-7, thereafter tending to fall. The relative risk (rate among the unemployed/rate among the employed) declined until 1979, since when it has fluctuated at about 10. The highest rate of parasuicide was consistently found among the long-term unemployed.