14. Sub-headings and sub-sub-headings should be alphabetically arranged. They should not be systematic or chronological.
15. Sub-headings should be chosen by the same rules and principles as main headings and thus make a duplicate list. There may be practical limits to this but principle is clear.
16. The arrangements of titles under main subject or sub-headings need not be alphabetical. Much is to be said for the chronological order of authorship or publication, but almost the only use for alphabetical arrangement by authors under heading is a poor duplication of author catalog use. It might be a real advantage to break the bad habit of using subject catalog for author purposes and on the other hand, the chronological arrangement of titles in the vast number of cases would save turning all the cards as required in the alphabetical order. Nevertheless the alphabetical is now the common method.
17. Complex books may be analyzed for the subject catalog. This is the distinctive advantage of the subject catalog over the shelf list that it can put different articles in the same volume or various subjects involved in one title under all their effective headings. It is obvious, however, that this principle must be limited—to apply in a wooden way would involve all periodicals and essays, a rock on which more than one attempt at subject cataloging has been wrecked.
18. The subject catalog should not be overloaded with references. The principle of economy of attention requires this. Few things are more aggravating in working under subjects than to have to finger over a large number of irrelevant cards. Some of the remedies for this are subdivision, the arrangement in chronological order of publication as above suggested, limiting analysis by excluding all works analyzed in accessible indexes and, where there is more than one edition of the same work, indicating one only and referring to the author catalog for the others.
19. The card should not be overloaded with details. The principle of economy of attention involves reducing the amount of material in a title to its lowest terms (whether on card or printed book) a matter greatly helped by typographical distinctions or corresponding distinction in the breaking of written lines, the location of certain details on certain lines or certain fixed places on the card, the use of red ink, underscoring, and similar details enabling the user to get the essential facts as to the identity of the work and its location in the building in the shortest possible time.
20. The indications on the cards of either catalog should be as brief as may consist with clearness and so displayed on the card as to catch the eye quickly.
21. Subject cataloging is a practical art, not a science. Names will be changed from time to time and a part of the art is therefore to develop a method of record on cards which shall cost the least possible effort for making changes.
Dr. G. E. Wire, of Worcester, continued the discussion of subject headings, with special reference to medical headings in the third edition of the A. L. A. List of subject headings.
Dr. Wire said a lack of knowledge of medical and surgical terms had led the compilers of nearly all the library catalogs into using erroneous headings, "Sees" and "See alsos" and that these errors had been continued in the third edition of A. L. A. subject headings.