To lay the groundwork for subsequent chapters in this monograph of multiple primary cancers in Connecticut and Denmark, we present a description of the historical significance of previous studies, focusing on key surveys that have enhanced our understanding of the origins of multiple cancers. Case reports, hospital series, and cancer registry studies have progressively sharpened our perspective on the patterns and causes of multiple cancers. These findings in turn have generated hypotheses about host and environmental determinants of various combinations of cancer and have provided clues to the actual mechanisms of carcinogenesis. The registries of Connecticut and Denmark which began in the 1930s and 1940s, respectively, afford investigators a unique opportunity to analyze the cancer experience of well-defined populations, followed for long periods. The major contribution of this monograph is the evaluation of second cancer risks among long-term survivors of cancer, including relatively rare tumors about which little information currently exists. For patients with a particular cancer, the number of observed second cancers are tabulated over time and compared with those expected if the patients experienced the same rates prevailing in the corresponding general population. We have discussed problems in distinguishing statistical artifacts from biologically plausible associations in light of the potential biases inherent in follow-up surveys of cancer patients; for example, heightened medical surveillance and mistaken metastases could result in false indications of elevated risk. Several differences in the reporting, follow-up, and coding practices between the Connecticut and Denmark registries are described and probably account for many differences in the reported findings.(ABSTRACT TRUNCATED AT 250 WORDS)