valuable literature of all former centuries." This is very true; but I think that catalogues can only form the groundwork for bibliographies, and are not sufficiently satisfactory to supersede them. Moreover, each country should produce its own national bibliography.
Mr. Cutter divides libraries into (1) those for study, and (2) those for reading; and this division must always be kept in view. We shall chiefly consider the first division, although it will not be right altogether to pass over the latter. Libraries for reading have been rightly considered in the light of educational institutions; and the various points connected with the information to be given to readers, as to what they should read, and how they should read, perhaps belong more properly to Education than to Bibliography.
As to the order in which the catalogue should be disposed we have considerable choice, and Mr. Cutter has given in the United States Special Report (pp. 561-67) a most elaborate classification of the different species of catalogues, but the main divisions are the classified and the alphabetical. Years ago the classified was considered the ideal; but when this ideal was brought down to practice it usually failed, and the result was almost useless. The late Professor De Morgan made the following pertinent remarks on this point:—
"A classed catalogue is supposed to be useful to those who want to know what has been written on a particular subject. Now, in the first place, who are the persons who look at a book list with any such view? Not beginners in a wide field of research. Did any one in his senses ever go to a library to learn geometry, for instance, and take the subject in a classed catalogue, and fall to work upon some author because he was therein set down? This attempt to feed the mind à la carte would certainly end in an indigestion, if, which is rather to be hoped, it did not begin in a surfeit."1
Again:—
"Any one who is willing to trust the maker of a catalogue, however highly qualified, with the power of settling what books he can want in reference to a given subject, is either a person who consults only the most celebrated works, and has nothing to do with research, or one who is willing to take completeness upon trust, and to content himself with blaming another person if he do not reach it."2
It is a common mistake to speak of a classified catalogue as a Catalogue Raisonné. A Catalogue Raisonné is a catalogue with bibliographical details and notes, in which the merits or demerits of the books are discussed. Therefore a Catalogue Raisonné can be alphabetical as well as classified. An alphabetical catalogue can be either one of authors, or of subjects, or what the Americans have styled the Dictionary Catalogue. A catalogue of authors will contain the description of anonymous books under headings in the same alphabet, and it may either have an index of subjects, or subject cross-references included in the general alphabet. But as the rules to be considered later on relate chiefly to the catalogue of authors, it is not necessary to say more on this point here. Again, De Morgan has made some excellent remarks on the catalogue of authors:—
"An alphabetical catalogue has this great advantage, that all the works of the same author come together. Those who have had to hunt up old subjects know very well that of all lots which it is useful to find in one place, the works of one given author are those which occur most frequently. Again, those who go to a library to read upon a given subject generally know what authors they want; and an alphabetical catalogue settles the question whether the library does or does not contain the required work of the author wanted. We believe that of those who go into a place where books are collected, whether to read, buy, borrow, (or even steal), nineteen out of twenty know what author they want; and to them an alphabetical catalogue is all-sufficient."3
Mr. Cutter has written the history of the Dictionary Catalogue in the United States Special Report (pp. 533-39), and he traces it back in America to about the year 1815.
Mr. Crestadoro, in his pamphlet, The Art of Making Catalogues of Libraries, 1856, recommended an inventorial catalogue of unabridged titles arranged in no order, but numbered, and an alphabetical index to the numbers of this inventory. The index thus formed was somewhat similar to the Dictionary Catalogue (United States Special Report, p. 535). Mr. Bradshaw held very strongly the view that an alphabetical catalogue was an index, and that a full shelf catalogue was the real catalogue; and few things he enjoyed more than to read through a list of the books as they stood on the shelves.4 In a letter to me, dated September 9th, 1879, he wrote:—
"It is a cardinal point with me that an alphabetical catalogue of a library is really an index, or should be so, to any other kind of catalogue you choose to make; while if you once lose sight of this fact you are quite sure to cumber the catalogue up with bibliographical details which are entirely out of place."
Scientific cataloguing is of modern invention, and to the British Museum it is that we owe the origination of a code of rules—rules which form the groundwork of all modern cataloguing. Good catalogues were made before rules were enunciated, but this is accounted for by the fact that bibliographers, like poets, are more often born than made.
Carefulness must be one of the chief characteristics of the cataloguer, for he will frequently find himself beset with difficulties. Mr. W. F. Poole, the author of that most useful work the Index to Periodical Literature, states this very forcibly when he writes:—
"The inexperienced librarian will find the cataloguing of his books the most difficult part of his undertaking, even after he has made a diligent theoretical study of the subject. He will find after he has made considerable progress that much of his work is useless, and scarcely any of it correct."5
The cataloguer must not jump to conclusions upon insufficient authority, or, as some persons have proposed, take a short list from the books and amplify the titles from bibliographies. Such a course will lead to endless blunders, and create confusion like that described by Professor De Morgan:—
"Lalande, in his Bibliographie Astronomique, wrote from his own knowledge the title of the second edition of the work of Regiomontanus on Triangles, Basle, folio, 1561. He knew that the first edition was published about thirty years before, and so he set it down with the same title-page as the second, including the announcement of the table of Sines, Basle, 1536. Now, as it happened, it was published at Nuremberg in 1533, and there was no table of Sines in it. The consequence is that Apian and Copernicus are deprived of their respective credits, as being very early (the former the earliest) publishers of Sines to a decimal radius. No one can know how far an incorrect description of a book may produce historical falsehood; but there are few writers who have the courage to say exactly how much they know, and how much they presume."6
Before concluding this Introduction it may be well to say something about a few catalogues that have been issued in the different styles. One of the best classified catalogues ever published in England is that of the London Institution, which was first printed in 1835, and completed in 1852.7 This has indexes of subjects, and of authors and books. The catalogue is very useful as a bibliography; and as the library was well selected, the reading of its pages is very instructive; but what shows the general uselessness of a classified catalogue for the work of a library is that in actual practice an alphabetical finding index has been in more constant use than the fuller catalogue.
Of an alphabetical catalogue of subjects an example may be found in that of the Library of the Board of Trade, which was published in 1866. Here the authors are relegated to an index, and all the titles are arranged under the main subject. This may be convenient under some circumstances, but it is not satisfactory for general use. The idea of the scheme was due to the late Mr. W. M. Bucknall, then librarian to the Board of Trade; but the catalogue itself was made by the author of this book. The system adopted was to use the subject-word of the title as a heading; but an exception was made in the case of foreign words which were translated. For instance, there is a heading of Wool. Under this first come all the English works; then the French works under sub-headings of Laine, Laines, and Lainière; then German under Schafwollhandel