nothing to the value or significance of its contents. Such luxuries, the darlings of the genuine bibliographer, the tests of his professional taste and the chevaux de bataille of his collection, are nevertheless only to be indulged in by a conscientious man when he is certain that such an indulgence is compatible with the ends for which national libraries exist. Even the ideal of rendering the library a representative of the thought and knowledge of the age must either be moderated, or pursued at the risk of incurring comparatively expenditure. A new periodical gives pause: it must be taken, like a wife, for better or worse; for once commenced it can seldom be dropped. New editions of scientific works occasion much perplexity: it is equally vexatious to be behind hand with the latest results of discovery, and to spend money over something which is certain to be soon superseded by something better still. In such cases compromise alone is possible, and compromise can never be quite satisfactory. Such difficulties press less heavily on the curators of academical libraries, where the demand for universality is not preferred, and even an accidental circumstance may legitimately impart a bias to the entire collection. The acquisition of Professor De Morgan's books, for instance, has made it imperative upon the University of London to be always strong in logic and mathematics, at all events. The principle of specialisation, indeed, admits of being carried very far in a large community, where it is possible to conceive groups of libraries working in different directions to a
