A total of 2528 workers in flax mills in Northern Ireland were followed up for 16 years. Follow-up was 97% complete. Deaths were identified and date and cause ascertained. Expected deaths were calculated on the basis of age and sex specific rates for Northern Ireland. Both male and female workers had fewer deaths than expected, and mortality showed no clear relationship with type of work. There was a small excess in the mortality of the workers who had had byssinosis at the time of the original survey, but there was no evidence that the more severe grades of byssinosis were associated with higher mortality than the less severe grades. Workers who smoke are known to have an increased risk of developing byssinosis, and cigarette smoking may be responsible for most of the excess deaths in the workers with byssinosis.