224 THE PATH OF SCIENCE 



director in the light of the information he has. They decide 

 not to do it, but then the legal department reminds them: 

 "You have forgotten that we made a contract in which we 

 agreed to carry out this piece of work." The success or fail- 

 ure of the work, then, is a clear test of the validity of the 

 original decision. In three cases from the author's experi- 

 ence, where the decision had been made to abandon a piece 

 of work but where it was carried forward without any change 

 in opinion and for quite other reasons, the work proved en- 

 tirely successful. Experiences of this kind demonstrate how 

 difficult it is to make plans for the conduct of research and 

 even the decisions essential for its operation. 



The experience of the last thirty years suggests that the 

 greatest success has attended those industrial research labora- 

 tories in which the director has been permitted a high degree 

 of autonomy and an assurance of continued support. Indus- 

 trial research is an adventure; it is even a gamble, though 

 one in which the odds are on success, provided that the ^vork 

 is continued in spite of delays and discouragements. Such 

 an adventure demands from its sponsor much courage and 

 much confidence. But if the director and his staff are well 

 chosen, the confidence will not be misplaced, and the re- 

 wards will be commensurate with the risks. 



