lead to confusion. It is far better to accept the system merely as an arbitrary set of numbers corresponding to certain topics and resolutely dismiss rigid ideas of logical sequence and consistency. There are certain permissible modifications, however, which may be made without violating the integrity of the system. To avoid the writing of long numbers a single letter may be used for the first three of four digits. Thus an Electrical Engineer would naturally have most of his material under 621.3 (Electrical Engineering), and for this number he could substitute the single letter E. Likewise a railroad man might use R for 625 (Railroad Engineering). Another modification consists in the use of an alphabetical arrangement for certain subsections combined with the decimal arrangement for main sections. This is sometimes useful in minute subdivisions. For example, under 621.728 (Material and Supplies for the Foundry), these various materials may be arranged in alphabetical order. The use of form divisions is a modification that may often be employed to advantage. There are certain set forms that are used throughout the whole range of the Dewey classification. These are: 01 Philosophy or theory 02 Compends, text-books, etc. 03 Cyclopedias, dictionaries, etc. 04 Essays, addresses, etc. 07 Education, teaching. Schools, colleges, universities 08 Tables, calculations. Miscellanies 09 History. Progress and development. These forms may be further extended; thus: 064 Exhibits, etc. (under Societies) 072 Laboratories (under Universities) Other form divisions that apply particularly to Engineering are the following: 002 Quantities and costs 003 Contracts and specifications 004 Designs and drawings 005 Executive 006 Working and maintenance 007 Laws 008 Patents 009 Reports These four divisions may be enclosed in parentheses and annexed directly to the usual. class number. Thus, 62(07) indicates Engineering Education; 621.32(09), Progress in Electric Lighting; 621.57 (008), Patents on Icemaking Machinery, etc. The object of this parenthesis separation of the form division is convenience in cross-references. For example, if one is interested in Patents he may write his class numbers as follows: (008)62 Patents-Engineering (008)66 Patents-Chemical Technology In this way all cards on Patents are grouped together. Other modifications will suggest themselves to the user as he becomes more familiar with the system. To index and file an item, it is marked with its classification number and placed in the numerical order thus indicated in the collection which it joins. To find an item, the subject index is consulted for the classification mark, which gives the numerical place in the collection of all the items the collection contains under the heading sought. The items under each head may be arranged alphabetically by authors or titles, and if the collection is large enough to require it, the items may be alphabetically indexed under each head, though generally the items themselves, properly arranged, furnish their own index. XX LITERARY CRITICISM CRITICISM OF BOOKS BY REA Literary criticism is an art in it branch of literature as clearly defi of the creation of the work critici many cases more so. A book of fi tinctly a creative work and its cr posed to possess an extensive knowle ting literature and of the rules composition, a trained taste in select and a quick imagination capable of comparisons and of setting forth vi sions. On the other hand, a techni as a rule, a compilation and sifting ting literature on a specific subje thorough and intelligent criticism ca ten only by one who possesses a knowledge of the subject and an ac at least with the principal literature to it. In the cases of individual revi or both of these requirements are missing. Partly for this reason, "E News," in an editorial a few years a plea for similar team work in the c books as in their writing. 340 It has already been said that any scientific book is rarely the work of a single man; this is even more rare in the case of engineering books. A book dealing with any branch or ramification of engineering work must be in the main a summary of the experiences and views of a great number of men. The more completely it represents those views and experiences, the greater is the value of the book, and the more truly is it a "storehouse of knowledge," a source of instruction for the young, and a manual of reference for the old engineer. The mere writing of a book is, of course, usually the task of one man. But the actual writer's share in the production of the book is only that of collecting, averaging, and arranging the opinions and results of those who have labored in the field of which it treats. In addition, it is also his province and duty to forecast, from the results of their compilation, the development of the immediate future. If in this work he is able to point the way to improvement, indicating, for example, how the results of the calculator and the work of the builder may be harmonized, he will have produced a book of maximum value. In appreciating this, however, it must not be forgotten that the writer of the book is only one of its many authors. Books of engineering, like all other things, are more or less imperfect, approaching perfec |