There is much to be said in favor of the third type of agricultural library,--namely, where the agricultural collections are incorporated with the college or university collections. When the topography of the campus and the location of the buildings are such as to make it feasible for the station workers and the agricultural professors to use the college or university library, the balance of the arguments seems to be in favor of this arrangement, both as regards economy of funds and efficiency of service, if the special needs of the station workers can be and are given proper consideration by providing the really necessary duplicates and an assistant especially qualified to aid in the bibliographical research connected with the investigations of the experiment station. There is a decided tendency toward unity in modern science. This is especially true in the sciences relating to agriculture. The entomologist needs to use botanical books, the botanist must use chemical books, etc., etc. This has an important bearing on library problems and as far as agricultural libraries are concerned, is an argument for centralized collections. As it is probable that it will be a long time in the future, if ever, before the experiment stations will have sufficient funds to build up complete collections for their special use, independent of the colleges, and since it is a question whether, if funds were actually available, it would be wise to expend them in duplicating to such an extent the college library collections, it seems evident that some compromise arrangement is inevitable. In attempting to work out a satisfactory library plan, every institution should make a careful survey of local conditions, such as the size of the collections, the size of the library staff, funds available, location and architecture of the college and experiment station buildings, and then attempt to work out the best possible policy under its peculiar conditions. In working out such a policy, there are three important points to consider,--the question of administration, the question of convenience and the economy of funds. As regards administration, attention has already been called to the fact that the station is by law a department of the college and under its control. If fully lived up to, this fact would seem to decide many vexed questions of administration. Sooner or later, it is believed, the colleges and experiment stations will find that there is less to be gained by standing alone than they had supposed and they will realize the advantages of a unified library administration for the institution as a whole.