A questionnaire was mailed, in 1986, to 1,345 grocery checkers who belonged to one union local in south-central California. Among the 1,058 female respondents (82% response rate), we found a 12% prevalence of hand/wrist symptoms characteristic of carpal tunnel syndrome (CTS). Symptom prevalence was positively associated with age, average work hours per week, years worked as a checker, and use of diuretics. The estimated effect of years worked as a checker was greater for younger subjects than for older subjects, in whom the association reversed, suggesting the selective loss of symptomatic workers from their jobs (a form of the healthy-worker effect). Although no effects were found for specific job-related activities, we estimate that CTS symptoms in at least three of every five symptomatic workers were attributable to occupational exposures.