June 2, 1916] 



SCIENCE 



767 



manufacturing scale is, indeed, one of the 

 most difiScult parts of a research resulting 

 in a new product, and the importance of it 

 is shown by the fact that all the large indus- 

 trial research laboratories, however con- 

 cerned they may be with the theory of the 

 subject, have, as parts of the laboratory and 

 under the direction of the research staff, 

 experimental manufacturing plants which 

 duplicate many of the processes employed 

 in the factory itself. 



AU these arguments tend to show that an 

 industrial research laboratory must neces- 

 sarily be of considerable size, but this re- 

 quirement is much accentuated by another 

 consideration altogether. 



Except in a few branches of pure sci- 

 ence small research laboratories are rela- 

 tively inefficient, in the technical sense of 

 the term, that is, they require more time 

 and cost more money for the solution of a 

 given problem. 



When considering this subject it is neces- 

 sary first to dismiss from the mind com- 

 pletely the idea that any appreciable num- 

 ber of research laboratories can be staffed 

 by geniuses. If a genius can be obtained 

 for a given industrial research, that is, of 

 course, an overwhelming advantage which 

 may outweigh any disadvantages, but we 

 have no right to assume that we can obtain 

 geniuses; all we have a right to assume is 

 that we can o'btain at a fair rate of recom- 

 pense, well trained, average men having a 

 taste for research and a certain ability for 

 investigation. The problem, then, is how 

 can we obtain the greatest yield from a 

 given number of men in a given time? In- 

 vestigation of the subject shows that the 

 yield per man increases very greatly as the 

 number of men who can cooperate together 

 is increased. The problems of industrial 

 research are not often of the type which can 

 be best tackled by one or two individual 

 thinkers, and they rarely involve directly 



abstract points of theory, but they contin- 

 ually involve difficult technical and mechan- 

 ical operations, and most of the delays in 

 research work arise because the workers en- 

 gaged on the subject do not know how to do 

 some specific operation. In my own experi- 

 ence, I have seen a good man stick for six 

 months in an investigation because he did 

 not know and could not find out how to 

 measure a conductivity with a precision 

 higher than one part in a thousand, a point 

 which was finally found to be perfectly well 

 known to several scientific workers in the 

 country. Again, it took another good man 

 three months to learn how to cut a special 

 form of section, but having learned the 

 trick he can now cut sections for all the 

 workers in the laboratory with no delay 

 whatever. 



In this connection the advantage of per- 

 manent set-ups of apparatus may be pointed 

 out. Among a large number of chemists 

 some one will continually be wanting to 

 photograph an ultra-violet absorption spec- 

 trum or to take a photomicrograph, and if 

 the apparatus for these purposes is erected 

 and in charge of a competent man who 

 understands its use, the work can be done 

 without any delay at all, the photography 

 of the absorption spectrum of an organic 

 liquid by a man who is used to the work ta- 

 king only an hour ; but if this point is vital 

 to the research and the chemist is quite un- 

 acquainted with the technique of the sub- 

 ject and has no apparatus available, it may 

 easily take him six months to find out what 

 has been done on absorption spectra, to buy 

 and erect the apparatus, and become skilled 

 in its working. 



From these causes, then, the efficiency of 

 a laboratory increases very greatly with its 

 size provided that there are good arrange- 

 ments for cooperation between the different 

 workers of the laboratory so that they are 

 kept informed of each other's problems. 



