The Special Problems of Serials

has never itself been brought successfully under control. Librarians have never collectively given definite answers to several questions: Should serials be cataloged by monograph catalogers? Should the same rules for choice of entry be applied to serials as to monographs? Should a serial be cataloged from the first issue or only after a complete volume? When a serial changes its title (or entry), is it to be treated as a new serial, or treated as one serial under the new title-in fact, what constitutes a serial? Within an individual library it is not unusual for the same serial to be entered one way in the card catalog, another way in the serials check-in file, and perhaps even a third way in a computer-produced listing. In the wider world outside of individual library control, citations to the same serial, whether referenced by authors or by abstracting and indexing services, often vary considerably. Union lists of serials add a further dimension of confusion by amalgamating varying entries from different holding libraries into one listing, often with a single entry per serial, to conserve space rather than to maintain biblio-graphic accuracy. Thus, a user may have to go from an author’s indexing journal citation to a library catalog or serial record (or both) and then to a union list to locate the library which may actually have the desired issue-cataloged under still another choice or form of entry. In addition to all this variation in entry, there is wide variation in the bibliographic description of a serial and in the identification and formatting of the bibliographic record elements in machine-readable form. Few serials listings are produced today by manual means and the variations in machine formats are a serious deterrent to producin