(b.) ENTRIES CONSIDERED AS PARTS OF A WHOLE.

118. The systematic catalogue undertakes to exhibit a scientific arrangement of the books in a library in the belief that it will thus best aid those who would pursue any extensive or thorough study. The dictionary catalogue sets out with another object and a different method, but having attained that object—facility of reference—is at liberty to try to secure some of the advantages of classification and system in its own way. Its subject-entries, individual, general, limited, extensive, thrown together without any logical arrangement, in most absurd proximity—Abscess followed by Absenteeism and that by Absolution, Club-foot next to Clubs, and Communion to Communism, while Christianity and Theology, Bibliography and Literary history are separated by half the length of the catalogue—are a mass of utterly disconnected particles without any relation to one another, each useful in itself but only by itself. But by a well-devised net-work of cross-references the mob becomes an army, of which each part is capable of assisting many other parts. The effective force of the catalogue is immensely increased. {58}

119. Make references from general subjects to their various subordinate subjects and also to coördinate and illustrative subjects.

Cross-references should be made by Full from Classes of persons (Merchants, Lawyers, Artists, Quakers, etc.) to individuals belonging to those classes; from Cities to persons connected with them by birth or residence, or at least to those who have taken part in the municipal affairs or rendered the city illustrious; from Countries to their colonies, provinces, counties, cities, etc. (unless their number is so great or the divisions are so well known that reference is useless); also, under the division History to rulers and statesmen, under Literature to authors, under Art to artists, and so on; from other Subjects to all their parts, and to the names of persons distinguished for discoveries in them or knowledge of them. Short and Medium will make such of these references as seem most likely to be useful.

The construction of this system may be carried on simultaneously with the ordinary cataloguing of the library, each book as it goes through the cataloguer’s hands not merely receiving its author- and subject-entries, but also suggesting the appropriate cross-reference; but when all the books are catalogued the system will not be complete. References are needed not merely to the specific from the general but to the general from the more general and to that from the most general. There must be a pyramid of references, and this can be made only by a final revision after the completion of the cataloguing. The best method is to draw off in a single column a list of all the subject-headings that have been made, to write opposite them their including classes in a second column and the including classes of these in a third column; then to write these classes as headings to cards and under them the subjects that stood respectively opposite to them in the list, to arrange the cards alphabetically, verify the references, and supplement them by thinking of all likely subordinate headings and ascertaining whether they are in the catalogue, and also by considering what an inquirer would like to be told or reminded of if he were looking up the subject under consideration. In this way a reasonably complete list may be made.

It will, however, often happen that there is no entry under the including subject. Take a simple instance. The catalogue, we will suppose, contains twenty histories of towns belonging to seven counties in Connecticut. In the revision described above references have been made both from Connecticut to these counties and to the towns from the counties, but only three of the counties have any titles under them. The others would not make their appearance in the catalogue at all if there were no cross-references. And as this will happen continually, it follows that the system will very greatly increase the number of headings and therefore the length of the catalogue. Such fullness may be allowable in regard to the state which contains the library, which, of course, should be treated with exceptional completeness. It may possibly be worth while for all the States of the Union and for England, but to attempt to do the same for all countries and all subjects is too much. A modification of the plan must be introduced which will make it much less complete but still useful. With many subjects the next heading in the ascending series must be skipped, and the references massed under one still higher; in the supposed case, for example, the references to all the towns will be made under Connecticut and under those counties alone which have any other entry under them.

120. Make references occasionally from specific to general subjects.

Of course much information about limited topics is to be found in more general works; the very best description of a single plant or of a family of plants may perhaps be contained in a botanical encyclopædia. This fact, however, must be impressed upon the inquirer in the preface of the catalogue or in a printed card giving directions for its use; it is out of the question to make all possible references of the ascending kind. From Cathedrals, for example, one would naturally refer to Christian art and to Ecclesiastical architecture, because works on those subjects will contain more or less on cathedrals. But so will histories of architecture and {59} histories of English, French, German, or Spanish architecture; so will travels in England, France, Germany, Italy, Spain. And anyone who desired to take an absolutely complete survey of the subject, or who was willing to spend unlimited time in getting information on some detail, would have to consult such books. Yet the cataloguer may very excusably not think of referring to those subjects, or if he thinks of it may deem the connection too remote to justify reference, and that he should be overloading the catalogue with what would be generally useless.

There are many things that are seldom used, and then perhaps but for an instant, and yet their existence is justified because when wanted they are indispensable, or because they make useful what is otherwise useless: a policy of insurance, life-preservers in a steamer, the index of a book, large parts of the catalogue of a library, among others the cross-references. Of such a nature, but much less useful, more easily dispensed with, is a

121. Synoptical table of subjects.