I So CARNEGIE INSTITUTION OF WASHINGTON. 



It is through this extension, rather than through adding to the 

 already accumulated mass of facts, that progress is most to be hoped 

 for in the future. 



II. 



A consideration which I wish most respectfully to urge upon the 

 Institution is the great advantage which comes from mutual discus- 

 sion and attrition between men engaged in contiguous fields of work. 

 My own work would have been much more effective could I have 

 enjoyed this advantage more fully, and I am profoundly impressed 

 by the waste of labor shown in an important fraction of current 

 scientific researches through the authors not being acquainted with 

 the best methods of work. , 



III. 



Under these conditions it still seems to me, as it has almost from 

 the day the Institution was founded, that the most effective way in 

 which it can promote research in exact science is by the organiza- 

 tion of an institute or bureau of exact science in general. If I had 

 only my special field in view, I might suggest simply an astronom- 

 ical institute ; but it seems to me that this would be too restricted to 

 get the best and most desirable results. I can not but feel it most 

 important that exact methods should be extended into other branches 

 of science than astronomy. 



In defining the field of work in such a bureau or institution a di- 

 vision of physical and natural science into three great fields may well 

 be borne in mind. One of these fields is that of the old-fashioned 

 natural science, which is concerned very largely with morphology, 

 physiology, and vital processes which do not admit of reduction to 

 mathematical forms. 



Another field is that of purely experimental science. 



The third field which really needs development is that of obser- 

 vation, which I propose shall be now occupied. The work required 

 is, in brief, the development of mathematical methods and their 

 application to the great mass of existing observations. Doubtless 

 suggestions as to experiment would frequently come in. These 

 would be carried out by others. 



IV. 



The Organization. — The first requirement for the organization is a 

 managing head in whom the Institution has entire confidence, who 

 should be required to devote all his available energy to the work, 



