The use of the folio and quarto size is now generally restricted to such books as could scarcely be printed in octavo, as dictionaries and similar books of reference. The parliamentary blue book, which long resisted the progress of octave civilization, is now beginning to shrink into a more manageable size. With regard to separate volumes, the most convenient practice is to consider them as a mere printer's division, which may vary in different editions; and to number them consecutively, without reference to their contents. The Germans have a very inconvenient practice of dividing a volume into parts, each of which is a volume in the ordinary meaning of the word; so that a work consisting of nine volumes, for example, may be divided into four volumes, one of which consists of three parts, and the other three of two parts each. The result is, that every reference must specify both the volume and the part: thus, Band II. Abtheilung III. S. 108. Frequently, too, this mode of numbering misleads the bookbinder, who (unless properly cautioned) numbers the volumes in the ordinary manner.
Volumes, as I have remarked, are merely a printer's division. Every literary composition ought, however, to have an organic division of its own. The early Greeks seem indeed to have composed both their poems and prose works as one continuous discourse. The rhapsodies of Homer and the muses of Herodotus were subsequent divisions introduced by editors and grammarians. But literary experience pointed out the commodiousness of such breaks in a long work; and the books of the Æneid and of the History of Livy were the divisions of the authors themselves. Since the invention of printing, the books of the prose works of the classical writers have been subdivided into chapters; while for the books of poems, as well as for the dramas, the verses have been numbered. The books of the Old and New Testament have likewise been portioned into chapters, and into a late typographical division of verses.
In making a division of his work, an author ought to number its parts consecutively, without reference to volumes. The novels of Walter Scott are divided into chapters, the numbering of which is dependent on the volume; so that it is impossible to quote them without referring to the edition, or to find a reference to them in any other edition than that cited. For the same reason, an author ought not to quote his own book in the text by a reference to volumes.
The division most convenient for purposes of reference is that which renders a quotation simple to note, and easy to verify. Divisions which run through an entire work (such as the chapters of Gibbon's History) are easy to quote, and the quotation can be easily verified when the chapter is not long. The numbering of paragraphs in one series through an entire work, as in the French codes, in Cobbett's writings, and in the state papers of the Indian government, is the simplest and most effectual division for purposes of reference. The Digest can now be referred to by book, title, and paragraph; nevertheless the Germans (who, notwithstanding their vast experience in the work of quoting, seem to have a predilection for cumbrous and antiquated methods) still adhere to the old circuitous mode of quotation, against which Gibbon long ago raised his voice (Decl. and Fall, c. 44. n. 1.).
Some works have been divided by their authors into chapters, but the chapters have been left unnumbered. Niebuhr's Roman History is in this state.
The internal division of a work by its author is not, however, merely for purposes of reference. It may likewise be a logical division; it may follow the distribution of the subject, and assist the reader by visibly separating its several parts. This process, however, may be carried so far as to defeat its purpose (viz. perspicuity of arrangement) by the intricacy of its divisions. Here again we must recur for an example to the Germans, who sometimes make the compartments of their writings as numerous as a series of Chinese boxes all fitted into each other. First, there is the part, then the book, then the chapter, then the section, then the article, and then the paragraph, which is itself subdivided into paragraphs with Roman numerals and Arabic numerals; and these again are further subdivided into paragraphs with Roman letters, and Greek letters, and sometimes Hebrew letters. To refer to a work divided in this manner by any other means than the volume and page, is a labour of as hopeless intricacy as it is to follow the logical cascade down its successive platforms.
It is a considerable convenience where the book or chapter is marked at the head or margin of the page; and in histories, or historical memoirs, chronological notation is very convenient.
In general no book (not being a book arranged in alphabetical order, as a dictionary, encyclopedia, &c.) ought to be printed without a table of contents. The trouble to the author of making a table of contents is very small, and the expense to the publisher in printing it is in general imperceptible. Modern English books rarely sin in this respect; foreign books, however, both French and German, are frequently wanting in a table of contents. The invaluable collection of the fragments of Greek historians lately published in Didot's Series—a work indispensable to every critical student of ancient history—has no table of contents, referring to the pages, prefixed to each volume. The Poetæ Scenici Græci of Dindorf is without a table of contents; and a similar want is a serious drawback to the use of the cheap and portable edition of the Greek and Latin classics published by Tauchnitz at Leipsic.
Lastly, an index adds materially to the value of every work which contains numerous and miscellaneous facts. The preparation of a good index is a laborious and sometimes costly task; the printing of it, moreover, adds to the price of the book. Many of the indexes to the English law-books are models of this species of labour; the indexes in the Parliamentary Reports are likewise prepared with great care and intelligence. Even a meagre index, however, is better than no index at all; and where the publisher's means, and the demand for the book, do not admit of the preparation of a copious index of subjects, an alphabetical list of names of persons and places would often be an acceptable present to the reader of an historical or scientific work.
L.