ciencies caused by the negligence of one of the professors of to-day will have been made good. And it is a heavy burden that the negligent professor is leaving for his successors. Suppose, to take an illustration at random, that the members of the mathematical department do not care whether the library buys any mathematical books or not and hand in almost no orders. This goes on perhaps for ten years, when new members enter the department, and the new blood infused begets new activity and the desire to order books once more. The new instructors will find that they have not only to buy the important current publications but to make up the arrears of a decade. While the science of mathematics has been advancing the library has been stagnant, and it will be years before it can recover from the temporary paralysis. Thus the iniquity of the professor of today shall be visited on his successors unto the third and fourth generation. Then there is the professor who orders by fits and starts, who one year orders everything that appears on his subject and the next year almost nothing, and who is sure to fail to order the most important book. This danger of important publications being overlooked by the professors is a serious matter that can only be guarded against by watchfulness on the part of the librarian. The danger increases where there are several instructors ordering on one appropriation, or when a book might fall under any one of several appropriations.

But if the negligent professor is an injury to the library, so in but slightly less a degree is the hobby-horsical professor the man who allows his hobby to gallop away with the whole of his appropriation, leaving nothing for the riders of other and perhaps better horses. But to drop the metaphor, which to tell the truth is almost running away with me, the professor with a specialty may do decided harm to the library, or at least to the portion of it under his direction. He will almost invariably develop its resources on this specialty at the expense of the other sides of his subject. Perhaps the French professor is an enthusiastic Molierist, and has gathered many editions of Molière, supplemented by a great mass of biographical and critical material. The collection is of undoubted value, and contains everything needed for a most profound study of the master of French comedy; but the student of modern French lit

erature and criticism will find that his wants have been scantily provided for. Or again it is the professor of history whose hobby is the history of the Slavic countries; he has stocked the library well with histories of Russia and Bulgaria in all sorts of unreadable languages, but left England and France and Germany to look out for themselves. Or perhaps the professor of physics is engaged in investigating the velocity and length of light waves, and orders only books that will help him in his own researches; the students anxious to explore other fields of physics, the whole realm of electricity for instance, must go away unsatisfied. imaginary cases are perhaps somewhat exaggerated, although I have in mind one at Harvard that furnishes a fairly close parallel. Yet they do serve to illustrate the grave danger of leaving the matter of ordering too much in the hands of the professors. The specialties mentioned above are all valuable in themselves, but if allowed to develop without due regard to other branches of the same subjects, the inevitable result will be a lamentably one-sided library.

These

Another serious defect of this system is the failure to provide sufficiently for subjects not covered by the college curriculum. To a certain extent the librarian can purchase the more important of such books out of the reserve fund under his control; but that sum, usually small, is apt to be pretty well exhausted by the demands upon it for general works and referencebooks. The result is that when, as not infrequently happens, a new subject is introduced among the college courses, the new instructor finds the library contains few of the books that he needs.

But before we consider the remedies for this undesirable state of things, it is but just that we should glance at the other side of the picture. Too much cannot be said in praise of the priceless aid that the conscientious and scholarly professor often gives in the selection of books. Busy with a thousand other things, he devotes his time and his skill to advancing the interests of the library in ways that are beyond the reach of the librarian, no matter how accomplished he may be. For the latter cannot have the special knowledge that the professor has attained by a lifetime of study in one line. Thus in the course of years of earnest effort may the professor create special collections that shall be unrivalled

and shall make the library a Mecca for scholars. The recompense for his unstinted labors will be the pleasure of seeing the gradual completion of his ideals, and of having his own knowledge increased by the use of the books he has helped to gather. And in passing I cannot refrain from paying a tribute to the late Professor Child, of Harvard. No one outside of the Harvard Library can realize the great service he did tow ards building up its collections. A student of the highest rank and profoundest knowledge, he gave to the library for many years the benefit of his ripe scholarship and wide acquaintance with books. The amount of time and energy that he devoted to this work of selecting books can hardly be appreciated; its value cannot be over-stated. The result of his labors is that the Harvard collection of folklore is probably second to none in the world, while in the other subjects in which he took especial interest ballads, mediæval literature and romances, and Scandinavian and Slavic literature the collections certainly equal any in this country. Nor were these books gathered by any excessive or lavish expenditure of money, but rather by careful purchasing and patient waiting. It is when one considers what such a man as Professor Child can do for a college library that one hesitates to condemn utterly this system of allowing the professors to choose the books. However serious one may deem its defects, the possibilities of the plan stand out so clearly that one is inclined to find modifications rather than to seek a substitute.

In the way of modifications and improvements much can be done by the librarian without making any radical change in the system. By personal interviews with the professors who are not doing their part in the ordering or who are ordering books too largely on one subject, he may bring them to see the error of their ways and (perhaps) to reform. He can frequently send them for approval titles of books that he thinks ought to be in the library in their department. At Harvard we have a printed return envelope that we send with such suggestions. I have found one obstacle to the use of this scheme in that the very professors that we wish to reach - that is, those who do not order enough books are the very ones who fail to approve the suggested title. Either they ignore the hint entirely or they decline to sign the order for one or another reason - usually be

cause they have a list of books they mean to order when they have time, but that time never comes. On the other hand, the men who always order more than their appropriations will pay for and whose orders are always waiting on the deferred list for better times or a new year

these men are ready to approve every order sent them. A friendly hint from the librarian that an appropriation is not being used up and is therefore in danger of being reduced by the committee another year on the ground that it is unnecessarily large is often effectual. Even the most indifferent professor is disturbed at that prospect; he likes to have his full allowance whether he uses it or not. If the librarian can ask the professor to examine the book itself instead of merely sending him the title of it, it is often a great gain. Arrangements can be made with the booksellers to send packages of new books from time to time on approval; these can be kept open to the inspection of the professors for a week or so, and it is surprising how much they will be looked over and how many good books will be bought from them that would not otherwise be ordered.

-

Encourage the professors to order by using as little red-tape as possible — at least keep it out of their sight. Do not insist on their writing out a regular order-card for every book they want. Take the orders gratefully in whatever way they prefer to give them - checked in a publisher's or bookseller's catalogue, written in a letter or scrawled on a page torn from a pocket diary, or delivered by word of mouth take them in any form; the librarian and his assistants can copy them on the order-cards and tie them up with as much red-tape as the ordering system may require. Even when the absentminded professor comes in and says he wants a book he saw noticed a few weeks ago, he has forgotten the author's name and can't recall the title, but he is sure it is a good book because the Nation or the Saturday Review, he really can't remember which, said so, try to find out what he wants and get it for him. Make him and all the rest of the faculty understand that they are doing the library a favor when they order books.

A remedy that I should like to see triedperhaps some of you may have tried it? is to allow the librarian to fill up the order lists under such appropriations as are not nearly exhausted by a given time, say two months before the end

of the year.

This would throw a heavy task on the librarian, but would tend toward a more even development of the library than if the appropriations were allowed to lapse, or than if the unused balance, as has sometimes been suggested, were turned over to some other department where it would be more appreciated. A similar remedy, and one that we have tried several times at Harvard, is to give the librarian a share in appropriations that experience had shown were not used up by the professors in charge.

The evil of injudicious and one-sided ordering is harder to cope with than insufficient ordering. Moral suasion is about the only means available, and that is often of little use in stopping the mad gallop of the professional hobby-horse. The professor accustomed to this system is apt to resent any interference with his full freedom of ordering, and to think himself a better judge of the needs of the library in his particular department than the librarian. Still with a little diplomacy much may be done to change the current of ordering into better and wider channels. If there are other and especially younger men in the affected department induce them to take their share in the work of ordering. And let the librarian himself continually exert his influence to counteract the evil.

A more radical remedy would be to grant the professors only the right of suggestion instead of a practically absolute control over the appropriations. Let their orders be considered on their individual merits by the librarian or a book committee, as would those asked for by the patrons of any other library. But if the professors found their orders were liable to rejection they would be likely to take less interest than they do now, and the library would thus lose much of the benefit derived from their special knowledge. While the growth of the library as a whole would undoubtedly be more systematic and even, I cannot but think that in many departments the books would be chosen with less skill and good judgment. In other words, in spite of the many drawbacks which I have pointed out, it seems pretty clear that the best interests of the college library are promoted by depending largely on the aid of the professors in selecting the books. But in order to secure the best results the librarian must not feel that he is relieved of the responsibility for

the proper increase of the library. His duties are rather added to, for he must see that each member of the faculty does his proper share of the ordering and does it well. He must urge Prof. This to order more books, must keep Prof. That from ordering too many, and must make Prof. The Other dismount from his hobby-horse. And all the while he is managing this staff of specialists he must spend wisely his own appropriation for general books. Thus with constant watchfulness and ready action the librarian can do much toward bringing this system nearer to its ideal perfection.

I have not tried to distinguish between the needs of the large and the small college library. The principle of selection must be the same for both; the details will differ in every library. No matter how ample the income may appear, the demand for books will always exceed the supply of money. The larger the library the greater its needs. No professor in a college with an annual income of perhaps $1000 for books would think of suggesting that $700 of it be spent for Wadding's Annales Ordinis Minorum, or $200 for Bulæus' Historia Universitatis Parisiensis; but his fellow in the university with its annual expenditure for books of $15,000 or $20,000 may demand such works as a right. What particular books to buy will be a different problem in each library, and the solution of it will depend on the amount of income, the class of readers, the presence of a considerable body of graduate students and professors doing special work, the proximity of other libraries, and many other varying factors. In short, generalization would be futile. For this reason I have preferred to consider the methods of selection which in principle do not differ materially for great or small libraries.

There are several other matters which I hope will be discussed in connection with the general topic of the selection of books. In the few minutes left at my disposal I can only suggest them in the hope that you will take them up for discussion.

The first is the question of periodicals. How large a portion of its funds is the college library justified in spending for periodicals? Should we allow a professor to use the whole or a great part of his appropriation for them? The demands of the professors in this direction are almost insatiable. In order to keep abreast of

the advance of any science it is necessary to have the costly and constantly increasing periodical literature of that science. The aggregate expense of providing these indispensable tools of the professor's trade is enormous, and the value of them is often in the main temporary, for the more important results are sure to appear sooner or later in the form of monographs, and the original tentative form will retain mainly an historical interest. Every periodical subscribed to constitutes a permanent liability against our funds, and cripples the library's purchasing power in other and more lasting directions. At Harvard we are constantly resisting the pressure to add new ones to our lists, yet in spite of ourselves we are spending over a third of our income for periodicals and the publications of learned societies, and the burden is growing from year to year.

Another question is the matter of providing duplicates of books in special demand. A professor recommends his class of 50 or a 100 men to read a certain book before the next lecture. Even if the book is on our reserve shelves it is impossible for the whole class to read it in two or three days. Shall we buy a second copy? a third? or even more? The Harvard practice has been against this duplication, although we have made an occasional exception. The need is usually a comparatively temporary one, and it has not been thought a wise policy to spend our funds in gratifying it. The most important exceptions have been in the class-room libraries, where sometimes a dozen copies of a much-used book are provided. And this brings me to my last point for discusion the selection of books for class-room libraries.

Class-room or seminary libraries have now become a prominent feature in most colleges and universities. Should the choice of books for them be left entirely to the professors in charge of the particular department? Should the books in them be duplicates of those in the central library or should the aim be to supplement the main collection? My own opinion is very strong that if a college owns but one copy of a book it should be in the main library, and that these subordinate libraries should contain merely a working collection of the more necessary books, duplicating those in the college library proper. Yet I find some class-room libraries managed on precisely the opposite theory of providing only books that are not in the college library, while others are run haphazard without any theory at all, the books being purchased by the professor in charge without regard to whether or not they are duplicates.

Let me sum up in a few words my main conclusions on this subject of selecting books for college libraries. The system usually in vogue of leaving the matter chiefly in the hands of the members of the faculty is by no means ideal; and I have tried to indicate some ways of avoiding its principal defects. No system can be perfect, and on the whole the advantages of this one outweigh its faults. With a few slight modifications, and most important of all with a librarian alive to his responsibilities and ready to amend by every means in his power the recognized shortcomings of the system, the books of the library under this plan will be well and wisely selected, and the library will thus be helped to do its proper share in the furtherance of the higher education for which our colleges and universities stand.

IN

N considering this subject let me remark, in the first place, that the paper has nothing whatever to do directly with the question of moral fitness of books, but relates to that fitness to survive in the struggle for life which may or may not be affected by the question of moral, intellectual, or imaginative worthiness.

Again, and in a similar line, let me say that this paper does not concern primarily the question of the evolution of the ideas themselves, or their relative fitness to survive. No doubt there is a struggle for life among ideas, but this paper has to do with the idea in the definite form of a book, and the thesis is this: That among books as among men, animals, and plants, there is a struggle for life in which the fittest survive by reason of some favorable variation which gives them some advantage in the struggle, and the aim of the paper is to indicate some of the factors which tend to survival.

When the paper was first outlined it was discouraging, the facts seemed so obvious that it appeared as if the paper would not be interesting even to its author, but since then I have come across the following, signed J. E. in Mr. Aflalo's "Literary year-book":"But our point here," he says, "speaking calmly, is that in the propagation of books alone of Nature's creations, we can discover no tendencies at work to insure the survival of the fittest through the extermination of the unfit."

Here, at least, was some one to whom the matter was not obvious. In fact something quite otherwise was obvious. The issue could hardly be more clear-cut. It is a concrete exexample of the struggle for life among ideas. These two ideas of J. E. and E. C. R. are mutually exclusive. They cannot both survive, and yet in the long run the fittest will survive because it is the fittest, and then they will know which is the fittest because it has survived. In the meantime I beg to state the reasons for thinking that the fittest survive in this struggle for life in the race of books, and that there are definite tendencies at work to insure this survival.

The fact of the struggle and perishing is clear enough. All books do not live forever. We see moth and rust, weather and pet dogs at their destructive work all the time. There are, in fact, practically no autographs of the ancients. not a ms. of the New Testament earlier than the fourth century. Some of the ancients have utterly disappeared; but though so many copies of each have perished, Plato and St. Paul, for example, do survive. Why is it, then, that one book survives while another perishes? Is it the result of chance or law? I answer law in both cases.

It is clear in the first place that, apart from accidental causes, there is a direct tendency to cause the extermination of some books in the very fact of multiplication of books. The reading public does not provide food enough, so to speak, to support an indefinite book population. When this feeding-ground is overstocked some of the population must perish, and the less strong, the less beautiful, the less clever, fail to get the needed dollars, or readers, on which they feed, and so they perish; and books once strong are continually elbowed out of their natural feeding-ground by stronger ones with more favorable variations.

Metaphor aside, we know that few books are long lived; that the number of books which can keep human attention at any given time is so small that there must be a continual tendency of the less valuable to fall away, and if it is said that three books are written to every one that finds a publisher, and that three out of four among books published fail to pay, we know that large numbers actually are exterminated through the action of their enemies, the publishers, the critics, and pre-eminently the indifferent public, who, however, do not slaughter but only fail to nourish. This makes the analogy with animal survival the more striking, for, as Mr. Darwin had to keep constantly pointing out, the struggle among plants and animals does not so often mean slaughter as starvation, the weaker is not destroyed by the stronger, but the stronger is better able to secure food and escape fire and flood, and therefore sur