duty to train men for research work and to that end should provide 

 high-grade graduate instruction as rapidly as possible. The appropria- 

 tion of funds to this purpose is no less necessary and consistent 

 than providing material equipment. Moreover, further progress in 

 agricultural efficiency is fully as dependent upon the acquisition of 

 new knowledge as it is upon the diffusion of information that we now 

 possess. The body of knowledge that we call agricultural science needs 

 clarifying of error on the basis of the most severe inquiry and many 

 difficult problems must be solved before certain phases of agricul- 

 tural practice can proceed on an established basis. 



Universities and non-agricultural colleges might well offer special 

 training along lines related to agricultural science, and devote more 

 attention to some of the scientific problems important to agricul- 

 ture, while the provision of some ready means by which graduate 

 students of agriculture may learn what facilities are now offered by 

 this class of institutions would be a useful service. No field is more 

 promising of useful results. 



B. The organization and policy of research agencies should be 

 such as to render the career of investigation attractive and to ensure 

 adequate returns of scientific truth for the efforts put forth. 



(i ) The immediate executive officer of a research agency should be a 

 broadly trained scientist whose tenure of office should be permanent, 

 subject only to the requirements of efficiency, and absolutely indepen- 

 dent of political considerations, and whose time and thought should not 

 be seriously absorbed by other duties. Such expert direction is essential 

 to securing proper unity of work and the efficient coordination of 

 the efforts of individuals. ■ 



If in any department or institution the administration of laws and 

 the management of research are combined under the same head, who 

 is a political officer, some form of organization should be adopted that 

 will give continuity and coherence to the work of investigation so that 

 its direction will not be subject to the influence of political changes. 



(2) Wherever possible, the organization of research agencies 

 should be on the basis of problems to be solved and not of processes 

 used in their solution. In other words, the subdivisions of an agri- 

 cultural research agency for administrative purposes should corres- 

 pond with the recognized subdivisions of the subject of agriculture itself 

 and not with the classification of the pure sciences. 



The science of agriculture is an exceedingly complex one, and the 

 investigation of a problem in any one of its subdivisions may require 

 that the methods and processes of various primary sciences be concen- 

 trated upon it ; and in a large organization each of its bureaus or divi- 



