The Construction and Government of Lunatic Asylums and Hospitals for the Insane

be applauded. It is enough to say that in appearance and arrangement it resembles the Index Medicus. One has to be less sanguine over the Beaumont index. The first question it raises is whether a computer is really necessary to produce a guide to a small archive of 400 documents; though one must in fairness recognize that the index is experimental in nature and intended to stir up comment. It is claimed that the system by which it was produced can be applied to any other archive. My own response is clouded with scepticism. Here, in 165 closely packed double-column pages, bulging with entries, are recorded the contents of a 'small, homogeneous and important' historical collection. Five indexes tell us in minute, and often ludicrous, detail how many letters were exchanged, for instance, between St. Martin and Beaumont; where and when they were written, subjects discussed and the actual location of the documents within the collection. On the face of things this might seem admirable, but in fact I fear that confusion is the main result. It helps no one, least of all the historian, to have an archive of this nature microscopically dissected into its smallest constituents. Who, but someone transfixed by trivia, will be helped by index entries beginning with abstractions such as 'thank you', 'suggest', 'frustration', 'proposed' and so on? The computer has enjoyed its joke. Will the poor researcher? Obviously a lot of hard and expensive work was put into the original indexing. Could not the results have been pruned and arranged by a person practised in the art of archive description? Old-fashioned methods are sometimes still the best; they could certainly have been turned to producing a worthwhile and economic piece of work in this instance. The computer is important enough to be kept in its due and proper place. E. GASKELL