THE PRODUCTION OF SCIENTIFIC KNOWLEDGE 193 



Great care must be taken in the oriQ;inal selection of the 

 director, since it is very difficult to remove him and his re- 

 moval involves a great disruption of the work of the institute. 

 Occasionally those responsible for the organization will realize 

 that they have made a mistake— that the man they have chosen 

 cannot do the work set before him— and then there should be 

 no hesitation in making a change. This may seem an easy 

 thing to do, but it is really very difficult. The great defect in 

 management of all kinds is the tendency of those in authority 

 to tolerate inefficiency rather than to face the unpleasant task 

 of removing the inefficient. It is commonly believed that 

 business men are harder in their dealings than public officials 

 or executives in other walks of life. Anyone who has had 

 much business experience will, however, agree, I think, that 

 the greatest fault of business management is a tendency in 

 personnel matters to avoid the issue because of weakness and 

 sentiment. The motto for an executive of any kind in the 

 treatment of those responsible to him is that he should be 

 tough and he should be generous. He should demand a high 

 standard of efficiency and endeavor to maintain it by making 

 any changes that seem necessary, but he should be generous 

 to the weaknesses of the inefficient and the misfortunes of 

 the unlucky. It is unlikely, of course, that these principles 

 for the selection and guidance of research directors will be 

 carried out fully by any board of direction, but I believe that 

 their application will be greatest if the controlling body con- 

 sists primarily of scientific men. 



The oreat dangrer is that the institutes misrht fall victims to 

 a system of political jobbery and that the staff and even the 

 director might be appointed for other reasons than their com- 

 petence. This difficulty, however, would supply its own 

 remedy. The institute would simply fail, and the advance 

 of science, locally checked, would proceed elsewhere. 



A problem that will arise if a considerable number of 

 research institutes are supported by public fimds will be the 

 use and application of the results obtained. This will be 

 complicated by the belief held by the public that a new tech- 



