The second type is the agricultural college and experiment station libraries combined into a single agricultural library and kept separate from the university library, as at Wisconsin. This type may be considered as belonging to the departmental type of library. Other states which have adopted this plan are California, Iowa, Minnesota, Missouri, Nebraska and Virginia.

In the third type the collections of agricultural literature including the experiment station collections, are consolidated with the college or university collections and administered as one unit. Examples of this type are the libraries of the University of Illinois, the Oregon agricultural college and the Kansas agricultural college.

Under certain conditions the advantages of one type may far outweigh the disadvantages and leave little doubt that this is the best for the particular institution concerned.

In the library of the first type,—namely, the experiment station library kept separate from the college library but under its control, the collections are composed principally of the following classes of literature:

1. As complete a collection as can be had of publications (a) of the U. S. Department of agriculture; (b) of state experiment stations in the United States; (c) of agricultural, horticultural, dairy and live stock and kindred societies; (d) the publications of departments of agriculture, of agricultural schools and societies in foreign countries, all of which literature may be designated as the "official agricultural literature."

(2) Files, at least current ones, of the leading agricultural periodicals of the United States, together with the best of those published in the interest of each of the special branches of agriculture,—live stock, dairying, horticulture, etc.

(3) A collection of reference works both general and agricultural, as well as standard works on agriculture and its various branches and allied sciences.

Few if any of the separate experiment station libraries can be said to have notably complete collections, aside from the "official agricultural literature." Scientific books and periodicals are expensive and most of the agricultural colleges have not felt able to duplicate expensive sets of periodicals and scientific reference works. Therefore, since the college needs such works as well as the stations, the result has been in most cases that they have been filed in the college or university library and the station collections have been limited principally to the "official agricultural literature" described above.

That the experiment station workers should have readily available as complete a collection as possible of the "official agricultural literature," both American and foreign, seems most desirable if not imperative. Whether this material should be filed in the station library or in the college library and to what extent it should be duplicated is a matter for each institution to decide, according to its needs and local conditions. In the case of an experiment station located on the college campus and near enough to the college or university for the station workers to use the general library, there is still much to be said in favor of a separate reference and reading room for the experiment station staff with an assistant in charge, the collection consisting principally of the "official agricultural literature," a selected list of current periodicals and a good selection of reference books of special interest in experiment station work. The ideal plan would be for this room to adjoin the university library like a seminar room. If it is not feasible on account of distance for the experiment station workers to have the collection next to the general library, then it should of course be in the experiment station building or agricultural hall.

Libraries of the second or departmental type,—namely, where the college of agriculture and the experiment station collections are combined, contain in general all the library resources of the institution along purely agricultural lines, including the "official agricultural literature," and in addition a fairly complete collection in the sciences relating to agriculture. Such libraries have a two-fold purpose. They must supply the needs of the professors and scientists in connection with their investigations and in addition must serve the students of the agricultural college. If the college of agriculture and the experiment station are some distance from the university,—so far as to make frequent consultation of the university library impracticable, there is no question but that the college of agriculture and the experiment station ought to have a separate library for their especial needs. If on the other hand they are near enough to the university library to make it feasible for the professors and scientists to use it frequently, it is an open question whether it is wise to separate the agricultural collections. It is then a question of a central library versus a departmental or special library. The nearer the college of agriculture library is to the university library, the more intensive should its collections become.