The purpose of this study was to determine if significant arsenic exposure was occurring at a Superfund site with elevated surface soil arsenic concentrations. A second objective was to determine the statistical relationship between the various methods of measuring arsenic exposure in humans. Random urine, 24-hr urine, hair, and fingernail samples were collected at the end of the workweek from 40 employees at an active pesticide manufacturing facility which had formerly produced arsenical pesticides. There was no indication of adverse health effects among the employees attributable to arsenic exposure. Mean urinary, hair, and fingernail concentrations of arsenic were well within normal values and indicated that significant arsenic exposure was not occurring among the employees. Random and 24-hr urine measurements were significantly correlated. Hair and fingernail results also were significantly correlated. Urine results did not correlate well with hair or fingernail results. Results of this study suggest that although there may be some individual variation, random and 24-hr urine arsenic results are not substantially different. For the purpose of screening for arsenic exposure, random urine samples may be an adequate and preferable test for those populations in equilibrium with their environment.