Systematic communication about methods, theories, findings, and the state of the disciplines is central to modern science. Patterns of communication between scientists have changed substantially, not only since the founding of the first journals and scientific societies in the 17th century, but even in the last few decades. Scientists' Knowledge Networks (KNs) are formed of nodes (forums) that include meetings, data archives, and documents, such as journals. The structuring of KNs varies from one field to another: "working paper" collections play a much more important role in high energy physics than in molecular biology; conference proceedings are much more important in computer science than in most of the social sciences. The structure of KNs in a discipline changes over time, with the rise and fall in the importance of particular nodes, such as specific conferences or journals. This era is marked by significant experimentation in the use of electronic media in creating and restructuring nodes of scientific KNs, with excitement about the expansion of scientific communications that these developments enable. Electronic journals, on-line conferences, and disciplinary corpora of shareable data are examples of new nodes. The nodes are not just technologies; they involve varied, but complex and subtle social practices for filtering and reviewing acceptable communications. These developments are reshaping the KNs of various fields in ways that we do not effectively comprehend, and the viability of these new and altered forums has varied across the disciplines. Prior studies indicate that scientists' willingness to trust the value of specific nodes as sources of legitimate and reliable knowledge can be pivotal. We know too little about the ways that scientists come to trust and value new forums, to guide effective experimentation. A laissez-faire evolutionary approach risks needless failures in the development of new electronic forums or in the restructuring of existing forums, leading to wasted resources and knowledge losses. On the other hand, the "best practice" approach risks indiscriminate promotion of communications practices that fit only certain fields onto all. This multi-disciplinary study will provide an empirically-grounded analytic understanding of why various disciplines structure their KNs as they do. It entails a comparative empirical study of the changes between 1970-1999 in the structure of KNs in six fields that are at different levels of maturity, that have different KN structures, and that have also adopted different forms of electronic media as important nodes in their KNs: molecular biology - model organisms, molecular biology - protein crystallography, particle physics, astrophysics, human-computer interaction, and information systems. Data will be drawn from published and on-line sources, as well as extensive interviews with editors, forum organizers, and active scientists. Results will be disseminated in the scholarly literature and at appropriate cross-disciplinary forums on scholarly publication practices.
