REPORT OF THE PRESIDENT, 1915. 17 



both from a priori argument and from common experience, 

 namely, that productive research, Hke any other constructive 

 work, requires arduous, persistent, and above all sustained effort 

 under the direction of disciplined experts. Coruscations in 

 science occur frequently enough, but unfortunately most of them, 

 as every investigator knows, are ignes fatui. It is more rational, 

 therefore, in the interests of progress to provide for continuity 

 in research than to give special attention to the excessively rare 

 events of sudden discoveries and inventions which prove to be 

 of permanent value. These advances per saltum will take care 

 of themselves; but the surer and more rapid process of general 

 advance, and the one on which attention should be concentrated, 

 in order to build for the future as well as for the present, is the 

 process of summation of increments of knowledge, each rela- 

 tively infinitesimal in comparison with the possible aggregate. 



These considerations apply in greater or less degree to all 

 branches of the Institution's work, but their apphcation is indis- 

 pensable to the highest success of the departments of research. 

 Without continuity and without sustained activity and support 

 they afford no adequate reasons for existence. It must be appro- 

 priately assumed, therefore, that their work will go on indefinitely, 

 in the same sense, and for much the same reasons, that educa- 

 tional work is assumed to be endless. The incidence of their 

 activities may be expected to change from time to time; their 

 leaders and collaborators will come and go; but there is no 

 prospect that their fields of investigation will become sterile or 

 exhausted. Science is unable to assign an epoch for the begin- 

 ning of research and may not venture to predict an end thereof; 

 it may assert confidently only that its methods, which have proved 

 effective and trustworthy in the past, will prove still more effec- 

 tive and trustworthy in time to come. 



What has been said with respect to the departments of research 



applies with but shght modification to the activities of Research 



Division of Associates. Their work includes in its aggregate 



^socSS ^ wider range of subjects and permits a larger 



degree of administrative freedom, but recognition 



of the time element is no less essential to this work than to that of 



the departments. Then* progress must also be measured by 



