Now I need not waste time in arguing here that this process can not be made to apply universally or be used indefinitely. To standardize a work of art would be to kill it. Standardization is valuable where interchangeability is necessary rather than adaptation to local conditions. Portable houses, for instance, with interchangeable parts, have been standardized to a certain extent, but only within the bounds of uniform climatic conditions. The standard houses for Michigan and Alabama would have to be different. It is important, therefore, as I have said, to know, when standardization is being carried out, the limits of its advisability and the conditions under which it becomes useless or injurious. This is of interest to us librarians because our methods and processes, our buildings, our book collections and the use of both have long been undergoing this very process. And it is surely desirable that almost all the routine processes of library work, and the others to some extent, should be standardized. This standardization has been going on ever since librarians began to meet together and began to issue their own professional literature; in other words, ever since the formation of the A.L.A. in 1876 and the establishment of _The Library Journal_ about the same time. The subsequent formation of State Library Associations and local library clubs, as well as the establishment of other library periodicals, has greatly multiplied the opportunities for librarians to talk over their work with each other, to learn of other and better ways of doing things, to compare existing methods and to determine, if possible, which of them best serves the purpose for which it was devised. These things having in some measure been decided, they were then crystallized and fixed by the rise and success of Library Schools, summer-schools and training classes, which selected the methods that had stood the test of time and had emerged from the crucible of discussion and formulated them into standards which were thenceforth taught to their students. This, I think, is a fair statement of the way in which our present library standards came to be standards. It is a good way to select the best and to ensure that the best shall not be departed from. If the best always remained best, we should have no quarrel with it. Unfortunately there is flux and change all about us. A method is best when it best corresponds to the conditions. We can ensure that the method shall not be changed, but we have no control over a large proportion of the conditions. They change, in spite of us; and then the methods ought to change with them. In some instances we have erred, possibly, by making it a little hard to change them. We are now ready to consider some of the cases where standards ought not to obtain--where one library ought to try to be different from another instead of exactly like it.