In the state college of Washington, the experiment station library is said to be separate from the college, but under its control. Our college library building occupies a central position on the campus, not over two hundred yards from the offices of most of the members of the station staff. Part of the lowest tier of the college library book stack is set aside for the accommodation of the station library, the point of division being marked by a gate. A specially designated member of the college library staff acts as station librarian, under the direction of the college librarian. Her salary is paid largely but not wholly out of the station funds. Her duties as station librarian occupy about one-third of her time, but these duties have precedence over her college library work. In the absence of the station librarian, members of the station staff are served by the college library staff. Our station library is made up almost entirely (1) of publications of the U. S. Department of agriculture; (2) of publications of the state experiment stations, and departments of agriculture and horticulture; (3) of the agricultural, horticultural and kindred periodicals. The college library contains duplicate collections of the first two classes of material. The third class in our experience is relatively of less importance, as it consists chiefly of the popular "farm-papers." The station library, like the college library, has its own card-catalog of U. S. Department of agriculture publications, and its own card-index of experiment station literature. Students and practitioners of the science of agriculture seem to be specially fortunate in that so much of the valuable material on their subject is published and freely distributed by the federal and state governments. There is perhaps no other science in which the unofficial literature is so relatively unimportant. It is true, the technical journals of the allied sciences contain much that is of value to the experiment station worker. But so far as my experience goes, the use of this is not constant and continuous, as is the case with governmental material. Let us have separate and distinct sets of state and federal "Bulletins," for our college workers and for our station workers, as both classes need to refer to them so frequently. But is not this going far enough? Is it not the wisest policy to confine our station library collection principally to these well-thumbed publications, and to place the less constantly used and more expensive unofficial material in the college library, where it can be of service to a larger public? MISS MARGARET HUTCHINS, of the reference department of the University of Illinois library described type =(c) Experiment station library consolidated with the university library.=