In 1985, 30 rotogravure printers exposed to toluene for 4-43 years (median 29) were examined by means of interviews and psychometric testing. They were 33-61 years of age (mean 50). Comparisons were made with a reference group of 72 men aged 27-69 (mean 47). The referents had never been exposed to solvents and were all in good health. The printers were employed by two Swedish companies. The mean exposure levels were 43 and 157 mg/m3 of toluene, respectively, at the two printing shops. Before 1980 the exposure levels had exceeded 300 mg/m3. On Monday mornings, before psychometric testing at the department of occupational medicine, toluene was measured in venous blood samples from most of the exposed subjects. A high proportion of the printers reported fatigue (60%), recent short-term memory problems (60%), concentration difficulties (40%), mood lability (27%), and other neurasthenic symptoms. In the psychometric tests their performance was poorer than the reference group's in most of the tests applied. Even performance on the synonyms test, usually considered resistant to mild brain affliction, was worse in the group of printers. Adjusting for this difference in the group comparisons reduced the group differences substantially. Alcohol consumption above 200 g/week was found to reduce the subjects' psychometric function more than toluene exposure. The printers' sum of neurasthenic complaints correlated negatively with their score in several tests. Exposure variables showed only weak associations with test results. Blood toluene levels were positively correlated with scores in spatial tests. The direction of the correlations suggests that the influence of acute pharmacologic effects is undetectable on Monday mornings before work. In conclusion, we found that exposure to toluene at levels below 157 mg/m3 following long-term exposure did induce neurasthenic problems and might reduce psychometric test performance.