A clustering experiment applied co-citation analysis to a cumulation (1972-1974) of the Social Sciences Citation Index (SSCI) tapes in order to outline a citation structure for the social and behavioral sciences. Co-citation analysis is a new technique that establishes a link between two older documents and a crude metric of the strength of that link, by counting the number of times the two documents are jointly cited in new documents. The major finding was that the analysis could cluster documents so that related works appeared together and could display relationships among documents and among clusters of documents which reflect scientific content. In contrast to the natural sciences, the social and behavioral sciences utilized older documents and placed greater emphasis on scholarly monographs, even in those areas most closely related to biological sciences, such as parts of experimental psychology. Generally, published work in the social and behavioral sciences seems to be heavily influenced by exceedingly small groups of researchers, who are represented often by quite old documents and who are not readily displaced by new research. These findings raise questions about the operation and function of the social and behavioral sciences. (LMM)