It was only with the closure of the Guardbook for 1978 imprints, and the introduction of a new cataloguing code accompanied by Library of Congress subject headings, that serious attempts were made to analyse the subject content of each item acquired by the University Library. Up until that point subject analysis had been minimal – access points for material about a named individual, and for grammars, dictionaries, encyclopaedias and volumes of conference proceedings, without using a controlled vocabulary. For much of its earlier history, the only consideration of subject which took place was in determining where to place each item on the shelves.

A portion of the Library’s intake has never been classified by subject. From the 1870s onwards acquisitions under legal deposit which were judged of secondary importance – so-called Upper Library material once the books were stored in the tower – were arranged solely by date and size. This material apart, during the 19^th and early 20^th centuries an attempt was made to classify academic titles, albeit very crudely and using a variety of home-grown classification schemes. As late as 1968 the readers’ guide issued by the Library stated that –

No unified system of classification has ever been applied to the whole Library.  As the books gradually spread into room after room of the Old Library many ad hoc schemes of arrangement were devised; these have resulted in what are essentially chronological blocks.  The older literature on any subject is thus likely to be scattered over many classes.

When the stock was transferred to the new building in the mid-1930s, much of this material remained on open access. The published guide to subjects and classes in the open library was consequently complex. German literature, for example, could be found in classes U, LN, OA, XXVII, 746-750.

Much but not all of the classification scheme which is still in use in the open library was already in place by the beginning of the 20^th century. 1-179, the range of numbers for theology, did not exist in 1934, so was presumably introduced when the Library relocated to its present site in the mid-1930s. The scheme was clearly developed piecemeal by successive generations of librarians over half a century, which accounts for some of its idiosyncrasy. The history of Holland since 1648 has just two numbers, 601:4 for the 17^th and 18^th centuries, and 601:5 for the 19^th century onwards. In contrast Portugal gets 4 divisions for the same period, Sweden 5 and Serbia 6.

The scheme has been amended and expanded over time. In 1905 there was just one number for German literature (748). A distinctive number for each genre was in place by 1919, and chronological subdivisions were introduced in the mid-1930s, reflecting the significant expansion in our collections of German literature at this time.  The scheme expanded again in 1984, when a post-1945 subdivision was introduced. The generic number for biographies (459) was phased out in the early 1990s, and biographies of an individual were henceforward placed with a subject. Cinema was only given its unique identifier (415:3) in 1999, but 5360 items have been added to the class since that date, an interesting indication of how quickly open shelf classes can expand and why the shelving crisis in the Library has become so acute.

Classification of closed access material also took place for a long time, though the schemes used were sometimes as crude as the 19^th century ones had been. Most basic was the “unbound” books sequence Ua-Ue, which had only 9 subject divisions. The four-figure classes, for less substantial academic items, whilst much less detailed than the open shelf arrangement with which they have close parallels, did allow for a level of specificity. All German literature was placed in 7746, 8746 or 9746 depending on the century of publication. Nevertheless material was often arranged in very broad categories. The histories of Russia, Poland and Finland, for example, were grouped in one single number.

S3-figures, the equivalent class for more special and valuable material, often highly illustrated, permitted a marginally more detailed level of subject analysis than in the four-figure classes. German literature all went into S746. But here again idiosyncrasy was rife. Spanish literature, for reasons now lost, was alone amongst European literatures in having 9 subdivisions. Erasmus was given his own unique number (S61:29).

Once keyword and subject searching was possible, classification of closed access material was gradually abandoned.  Classification of unbound material went first, in the mid-1990s.  The four-figure classes each became a simple numerical sequence, dropping all subject reference in the classmark. 20^th century publications on German literature, like everything else, were placed in a general 9000 class, followed after the millennium by a C200 equivalent. S3-figure classmarks were abandoned in 1998, replaced by a S950 category.

Classification for the open shelves still continues, enabling readers to browse a percentage of our modern academic collections.  However, given the obvious problems caused by lack of space, less of the new intake is currently sent to the shelves than at any point since the 1930s. That situation will presumably change again once our off-site storage is up and running, and we have more flexibility in how the Library’s stock is arranged.

David Lowe