The second grade of expert aid is that which pronounces on concrete cases, which decides whether a given book (either from inspection of the mere title or of the volume itself) is suitable for the library. This kind of aid is not difficult to obtain, and there are persons in almost every place qualified in some degree to give it. It requires, however, a preliminary selection and generally the obtaining of books on approval, which is easier in a large place than a small one. The library is only one of various institutions that must use expert aid of this kind. The same limitations apply to all. Take, for instance, the work of reference, the cyclopedia, we will say. Its editor cannot write of his own knowledge the articles on Venezuela, and open-hearth steel, and Plato. He must rely on the information, direct or secondhand, of experts. But he cannot allow his experts to write his cyclopedia. Some cyclopedias are written very nearly in that way, and they are not the best. The expert must be coached before he does his work and the work must be edited when finished. It is on the proper combination of expert and editorial work that the value of the finished volumes will depend. So it is with library selection. The librarian is the editor of a big cyclopedia of thousands of volumes. He must have expert aid in selection, but he must not allow his experts to select the library uncontrolled. They must be instructed beforehand, and their advice must be carefully considered after it has been given. It must, in short, be edited. This brings us to the consideration that we have ultimately to face in discussing any phase of human activity--the question of personality. If the librarian and the book committee are incompetent and believe themselves to be competent--then the collection, in spite of all efforts, will reflect their faults--it will be intolerant, or trivial or ill-balanced. Much, therefore, depends upon the actual book selector for the library. Should this be the librarian, or a committee of the trustees, or the board itself, or an advisory committee of outsiders? Probably the best results are obtained through a preliminary selection made by the librarian with the aid of lists and the advice of individual experts--not committees--as suggested above, and then submitted to some person or committee representing the Board of trustees. This places the final responsibility where it belongs--on the trustees; but with a satisfactory librarian, the duties of the reviewing committee would consist chiefly of deciding on matters of policy--rarely of considering individual titles. It would decide, for instance, on how closely fiction is to be censored, on how far the library is to go in the purchase of recent fiction, on the extent to which foreign languages are to be recognized, on the purchase and duplication of text-books, on the policy of the library with regard to denominational religious works or of controversial books generally--and so on.