A hypothetical model is given for a National Serials Data System based on the best educated guesses of what the system should do and how, therefore, it should function. The model focuses attention on the ultimate goal rather than on the decision-making processes relating to choice of data elements, unique identification codes, etc. This conceptual presentation proceeds from the following set of propositions: (1) the primary function of a National Serials Data System should be to provide for the identification of serials titles and to provide a means of locating their holdings among the participating libraries; (2) the provision of data relating to or supporting the processing of serials in libraries; (3) provision of searchable data relating to the intrinsic characteristics of serials publications for analysis or study; (4) prompt development of a strategy controlling the initial and subsequent definitions of scope of coverage for the system; (5) a decentralized, rather than centralized, service network is preferable for the system; (6) deciding among alternative strategies; (7) standardization of the data elements, tape format and character set required for the core identification and location module and (8) development of a strong central management component to ensure the system's performance, technical controls and further development. (NH)