Among 4,184 patients with cancer of the esophagus, 55 second primary cancers were observed, whereas 64 were expected [relative risk (RR) = 0.86]. The absence of an excess risk of alcohol- and tobacco-related cancers was not anticipated. A significant 19% deficit of second cancers was found among 30,843 patients with stomach cancer. Cancer of the rectum, kidney, and lung all occurred significantly below expectation. An excess risk of ovarian cancer (RR = 1.9) was seen in women. Reasons for these findings are not entirely clear. Cancer of the small intestine is rare, and despite a relatively short survival expectation, a moderate excess of second cancers was seen among 868 patients (36 vs. 26.8). Only cancers of the liver and gallbladder were significantly elevated, and the possibility of misclassified metastases is discussed. Colon cancer is one of the most common cancers in Denmark, and 29,490 patients with this disease were at slightly lower risk for development of second cancer (RR = 0.96; 95% confidence interval = 0.9-1.0) than the general Danish population, excluding secondary colon cancers. Esophageal, stomach, and liver cancers occurred less frequently than expected. That cancers of the uterine corpus and ovary were significantly increased supports the notion that common risk factors, such as diet and endogenous hormones, influence the development of these cancers. A significant 23% deficit of second cancers was also found among 26,597 patients with cancer of the rectum, excluding secondary rectal cancer. Significant deficits were seen for cancers of the stomach (RR = 0.5), lung (RR = 0.8), and brain (RR = 0.5), and for multiple myeloma (RR = 0.4). The likelihood of underreporting of second cancers, especially of the digestive system, is discussed. However, cancer of sites previously reported to be associated with rectal cancer, e.g., the colon, breast, and uterus, did not occur below expectation. Cancers of the liver and biliary tract occurred in 4,453 patients; their average survival was only 1 year. Except for a slight excess of cancer of the ovary (5 vs. 1.6), the risk of second cancer development for all sites was consistent with unity (RR = 0.90). The risk of second cancers among 7,752 persons with cancer of the pancreas was not greater than expected (88 vs. 85.2). Males were at significant risk of kidney cancer (RR = 3.2), whereas females showed elevated rates of cancers of the uterine corpus (RR = 3.2) and ovary (RR = 3.1). No site occurred significantly below expectation.