A second follow-up of 27,011 diagnostic X-ray workers in China revealed a 21% greater incidence of cancer than expected based on the experience of 25,782 physicians who did not routinely use X-rays (RR = 1.21; 95% CI: 1.08 to 1.35). This risk is lower than the 50% excess reported previously and reflects, in part, the reduced risk among workers first employed after 1965, when hospital exposures to radiation probably were lower than in earlier years. The X-ray workers were employed between 1950 and 1985 and followed for an average of 16.1 years. Significantly elevated risks were seen for leukemia (RR = 2.4, n = 34 cases), and cancers of the esophagus (RR = 5.2, n = 19), liver (RR = 1.8, n = 65), and skin (RR = 2.8, n = 9). Cancers of the breast (RR = 1.5, n = 20), thyroid (RR = 1.7, n = 8), and bone (RR = 7.6, n = 4) also occurred more often than expected. Non-significant deficits were observed for cancers of the oral cavity and pharynx (RR = 0.6, n = 16), colon and rectum (RR = 0.8, n = 20), stomach (RR = 0.8, n = 36), and lung (RR = 0.9, n = 45). Excess risks for leukemia and esophageal cancer were seen among men but not among women. The RR for leukemia was higher for X-ray workers who began employment before 1970 than for those who started more recently and also for those who were young when employment began. The patterns of risk associated with duration of work, and with age and calendar time of initial employment, suggest that the excesses of leukemia and skin cancer, and, possibly, cancers of the breast and thyroid, were due to occupational exposure to X-rays. However, there was little evidence that radiation contributed to the increased occurrences of liver or esophageal cancers.