October 25, 1917] 



NATURE 



157 



quite possible that at some future date as high as 

 60 or 80 per cent, of the possibilities may be realised 

 without any appeal to arms for the nation or any 

 unusual incentive for the individual. 



(2) The Increase of Organised Knowledge. — The re- 

 search by which organised knowledge is increased will 

 doubtless always be carried on chiefly by three distinct 

 tvpes of research organisations : research by the 

 Government in national laboratories, research by the 

 universities in connection with the work of instruction, 

 and research by industrial laboratories in connection 

 with the interests of manufacturing concerns. Apart 

 from these three main classes of laboratories there 

 will always be large, privately endowed research organ- 

 isations, dealing with neglected fields of remote com- 

 mercial interest, private industrial laboratories sup- 

 ported by consulting fees, and co-operative testing 

 laboratories, also self-sustaining. 



National, university, and industrial research follow 

 three essentially different lines. There is considerable 

 overlap in field, it is true, but each is centred on a 

 different kind of research. The proper function of 

 national research is the solution of such problems as 

 concern the nation as a whole, affecting the general 

 interests of all classes of individuals ; it is the custo- 

 dian of standards, it develops methods of precise 

 measurements and investigation, it is trouble engineer 

 for the solution of very difficult problems or the 

 problems of producing units so small as not to be able 

 to have their own research laboratories. It is the 

 proper guardian of the public health. It solves 

 problems connected with contagious and vocational 

 diseases. It develops methods of making good roads, 

 increasing the fertility of the soil, and stocking waters 

 with fish. National research is of all grades, from 

 that dealing with fundamental principles up to that 

 relating merely to lessening the costs of production. 



University research must always, in the very nature 

 of things, be concerned chiefly with the advancement 

 of the various sciences as such, and with the de- 

 velopment of the fundamental principles of each 

 science. The best university instruction is along these 

 lines, and investigators and students in close touch 

 with them will naturally have most new ideas in close 

 connection with fundamental principles. University 

 research is necessarily one of small jobs and the best 

 minds, and is without very much continuity. The 

 advanced student is interested in a research just long 

 enough to make it acceptable as a doctor's thesis. 

 The instructor is too burdened with teaching to give more 

 than a margin of time to research. But a very small part 

 of the universitjfe research is extended 3'ear after year, 

 covering a wide field. This is quite as it should be, 

 the university looking after those fields of research 

 of little commercial value on one hand, and not 

 directly affecting the interests of the nation as a 

 whole on the other, but of fundamental and far- 

 reaching importance to all. 



Industrial research takes the middle ground and has 

 already become a distinct profession. It is in close 

 touch with practical commercial application on one 

 hand, and with fundamental principles on the other. 

 Its proper field is anything between elimination of 

 works troubles and the investigation of fundamental 

 principles. The staff of the ideal industrial research 

 Iaborator>' is composed of experts of wide experience 

 who can serve the manufacturing departments in a 

 consulting capacity without sacrifice of time. We mav 

 perhaps best summarise the preceding statements b'v 

 describing the ideal research man and the ideal research 

 laboratory. 



Some writers have spoken of the investigator as a 

 rare individual to be sifted out from educational insti- 

 tutions with great care for a particular line of work. 

 My personal opinion is that a large percentage of the 

 NO. 2!;04, VOL. IO0I 



men students are fitted for research work if properly 

 started along the right line. The investigator should 

 have a mind at once fertile and well-trained. His 

 mind should be teeming with new ideas, but he should 

 possess unerring judgment to reject those which are 

 not logical or promising. We are often asked what 

 sort of preparation in physics would be best for men 

 intending to take up research as a life work. It has 

 even been proposed to give courses in "applied 

 physics " for the benefit of those intending to take up 

 industrial research. Our invariable reply is that the 

 best preparation for a research man is a thorough 

 grounding in the fundamental principles of his science : 

 physics, chemistry, or whatever it may be. If he has 

 this thorough knowledge of fundamental principles it 

 is safe to say that in any properly organised research 

 laboratory with the proper leadership and companions, 

 such a student will have many times as many useful 

 ideas as he can himself possibly follow up with re- 

 search. Scarcely anyone who has completed advanced 

 work in a science can read, say, a journal of abstracts 

 without thinking of many problems which he would 

 like to investigate. Fertility of mind is not so much , 

 an inborn quality of the mind itself as of the training 

 and association which that mind has had. 



The ideal industrial research organisation may per- 

 haps be outlined with a knowledge of its development 

 during the last fifteen years. I shall give, frankly, 

 my personal views on the matter, based on an intimate 

 knowledge of four universities, three professional re- 

 search laboratories, and a visiting acquaintance, so to 

 speak, with quite a number of others. The ideal in- 

 dustrial laboratory, to my mind, consists of two quite 

 distinct divisions : one taking the brunt of works 

 troubles and testing or making analyses of the mate- 

 rial used. The other wing is complementary to this, 

 and deals with the larger fundamental problems en- 

 countered, problems requiring skilled specialists and 

 considerable time for their solution. The alternative 

 organisation with a single research laboratory covering 

 both works troubles and fundamental problems is not 

 so successful. The plan in this case is to have con- 

 siderable research in progress of very little interest to 

 the company, but engaging a staff much larger than 

 required to take care of ordinary works troubles. In 

 this case, when works troubles are many and insistent, 

 as they are wont to be at times, the staff engaged upon 

 fundamental research forms a reserve to be called out 

 occasionally to deal with works troubles. The chief 

 disadvantage of this is that the fundamental work is 

 subject to more or less frequent interruption and 

 cannot be so efficiently carried on. On the other hand, 

 when the research is in two quite distinct divisions, 

 fundamental work is not subject to interruption by 

 works troubles. 



Industrial research is pre-eminently fitted to be 

 carried on by team work. This we have developed to 

 a high degree in Pittsburg, and consider very much 

 more efficient than the alternative cell system, w'here 

 each leading man has a room or suite of rooms to 

 himself and keeps his work to himself. In the ideal 

 organisation tw^o or three men work together on the 

 same large problem or group of problems, the aim 

 being to have a good theoretical man and a good 

 experimentalist working together as much as possible, 

 or even a physicist and chemist in some cases. The 

 characteristic of the team-work plan, however, is the 

 conference system. The five or six men most in- 

 terested in each line of research meet for an hour each 

 week to discuss the problem in its various aspects, to 

 plan new work, and to consider various interpretations 

 and applications of the results obtained. The ideal 

 conference is not few^er than four and not more than 

 eight men, and includes an efficient stenographer. To 

 one experienced in such team work the results of 



