We should not forget that by this plan the reader is usually made the judge of whether a book is worth keeping.  Why do we preserve by continual reprinting Shakespeare and Scott and Tennyson and Hawthorne?  The reprinting is done by publishers as a money-making scheme.  It is profitable to them because there is a demand for those authors.  If we cease to care for them and prefer unworthy writers, Shakespeare and Scott will decay and be forgotten and the unworthy ones will be preserved.  Thus a great responsibility is thrown upon readers; so far they have judged pretty well.

Just now, however, we are confining ourselves to the use of books for information; and here there is less preservation than elsewhere.  Especially in science, statements and facts quickly become out of date; here it is not the old but the new that we want—­the new based on the accurate and enduring part of the old.

Before we leave this part of the subject it may be noted that many persons have no idea of the kinds of information that may be obtained from books.  Even those who would unhesitatingly seek a book for data in history, art, or mathematics would not think of going to books for facts on plumbing, weaving, or shoe-making, for methods of shop-window decoration or of display-advertising, for special forms of bookkeeping suitable for factories or for stock-farms—­for a host of facts relating to trades, occupations, and business in general.  Yet there are books about all these things—­not books perhaps to read for an idle hour, but books full of meat for them who want just this kind of food.  If Book-taught Bilkins fails, after trying to utilize what such books have taught him, it is doubtless because he has previously failed to realize that books plus experience, or, to put it differently, the recorded experience of others plus our own is better than either could be separately.  And the same is true of information that calls for no physical action to supplement it.  Books plus thought—­the thoughts of others plus our own—­are more effective in combination than either could be by itself.  Reading should provoke thought; thought should suggest more reading, and so on, until others’ thoughts and our own have become so completely amalgamated that they are our personal intellectual possessions.

But we may not read for information at all—­recreation may be what we are after.  Do not misunderstand me.  Many persons have an idea that if one reads to amuse himself he must necessarily read novels.  I think most highly of good novels.  Narrative is a popular form of literary expression; it is used by those who wish to instruct as well as to amuse.  One may obtain plenty of information from novels—­often in a form nowhere else available.  If we want exact statement, statistical or otherwise, we do not go to fiction for it; but if we wish to obtain what is often more important—­accurate and lasting general impressions of history, society, or geography,