Thirty-two jobs at a pork processing plant were semi-quantitatively analyzed in terms of their ergonomic characteristics, then classified as "hazardous" or "safe" in terms of potential risk for elbow or hand/wrist disorders. The spectrum, number, and incidence of such disorders occurring during the preceding 20 months were then compared to the job analyses and hazard classifications. There were 104 disorders associated with 15 job categories. The disorders included epicondylitis (24), nonspecific hand/wrist pain (41), carpal tunnel syndrome (CTS) (21), trigger finger (12), trigger thumb (3), and De-Quervain's tenosynovitis (3). The strength demands of the jobs associated with morbidity were significantly greater than those of jobs without morbidity. Differences in wrist posture were less significant. Type of grasp and repetitiveness were not significantly different. Practically all morbidity (96%) was associated with the hazardous job categories and occurred with a characteristic pattern of co-morbidity. The hazard classification scheme correctly predicted risk of upper extremity morbidity for 13 (87%) of the 15 job categories associated with morbidity and correctly predicted a lack of risk of morbidity for 16 (94%) of the 17 job categories not associated with morbidity. Significantly elevated relative risks were observed for any upper extremity disorder (11.4), any disorder excluding CTS (39.4), all specific disorders (6.9), and all specific disorders excluding CTS (19.4). The relative risk for CTS was 2.8 and not statistically significant. The results of this study provide additional epidemiological evidence that upper extremity musculotendinous disorders and some cases of CTS may be causally associated with work. The exertional demands of a task best explained the occurrence of morbidity.