768 



SCIENCE 



[N. S. Vol. XLIII. No. 1118 



"When considering the efficiency of re- 

 search work it must be remembered that 

 the efficiency is necessarily extremely low 

 since it is very rarely possible to arrange 

 any research so that it will directly proceed 

 to the end required. 



It is the common opinion of those who 

 have to deal with the organization of re- 

 search that only a small percentage of all 

 the investigations started are likely to be 

 successful, the great majority being either 

 dropped before they come to an end, or 

 being carried through, and filled simply as 

 records, without any results having been 

 obtained which would justify the expense 

 of the investigation; that is to say, indus- 

 trial research is justified only by the great 

 value of the successful attempts, and these 

 must bear the burden of a great number of 

 unsuccessful attempts which may have been 

 quite as costly as the successful ones 

 themselves. The object of organization 

 is to attempt to reduce the proportion of 

 unsuccessful investigations which will be 

 undertaken, as has already been shown. 

 This can be done by increasing the size of 

 the laboratory, by increasing the specializa- 

 tion of the workers, and especially by in- 

 creasing cooperation between workers in 

 different fields. 



Naturally, the most important step which 

 could be taken to increase the efficiency of 

 industrial research would be to increase the 

 likelihood of correct choice of a promising 

 investigation, but, unfortunately, very little 

 can be done in this direction. Those with 

 the most experience in research work are all 

 agreed that it is almost impossible to say 

 whether a given investigation will prove 

 remunerative or not. The only general con- 

 clusion that can be drawn is that the deeper 

 a given investigation goes towards the fun- 

 damentals of the problem the more likeli- 

 hood there is that the results will be of 

 value, and the more superficial an investi- 



gation is, even although it appears more 

 promising at first sight, the less likelihood 

 there is that it will finally prove of real 

 worth, so that the choice of investigations 

 must necessarily be made largely at random 

 and will be influenced to a great extent by 

 the ideas of the scientific workers them- 

 selves ; if any worker has a desire to take up 

 any particular line of work, provided that 

 it is associated with the general trend of 

 work in the laboratory, it is usually wise to 

 let him do so, but the expedition with which 

 a decision can be reached as to the probable 

 value of the investigation after it has been 

 started is very greatly enhanced by com- 

 plete cooperation of workers in the different 

 branches of science in consultation on the 

 problem. 



At this point it might be well to discuss 

 the organization of a large research labora- 

 tory. Such a laboratory should be estab- 

 lished in charge of a director who has had 

 some actual manufacturing experience in 

 the work processes but at the same time he 

 must have a considerable sympathy with 

 purely scientific work and an interest in 

 the advancement of scientific theory. Both 

 these qualifications are desirable, but if 

 such a director combining the two can not 

 be found, then a man of full scientific train- 

 ing should be chosen and put into a position 

 of responsibility in the manufacturing side 

 of the industry until he has become fully 

 acquainted with the technique of the in- 

 dustry. It is most inadvisable to take a 

 man from the industry who has not had a 

 full scientific training, including advanced 

 research work in academic problems, since 

 he will generally be lacking in sufficient 

 knowledge of and sympathy with the more 

 academic investigations of which he will be 

 in charge and if the two necessary qualifi- 

 cations can not be found united in one man, 

 it will be necessary to take a man with the 

 scientific qualifications and give him the 



