January 3, 1918] 



NATURE 



357 



work corresponding with a limited field of science, so 

 that while his special attention is devoted to that one 

 department his field of activity just overlaps that of the 

 departments on each side of him, while his general 

 knowledge of the subject should, of course, cover a 

 much wider range. It is important that each man 

 should have his own special field of work, "and that 

 overlapping should not be complete, since such com- 

 plete overlapping will inevitably produce friction de- 

 structive of co-operation and harmony. The way in 

 which such a subdivision is arranged may perhaps be 

 best illustrated by Fig. 3, which shows the range of 

 the specific investigations of those who in our labora- 

 tory cover the range of research work between sensito- 

 metry and pure physical chemistry. There are five 

 workers in this range; the first, A, being a pure 

 physicist, B a physicist with a considerable experience 

 of chemistry, C a physical chemist who is specialised 

 in photography, D a physical chemist who is 

 specialised in photographic theory, and E a pure 

 physical chemist. The interest of each of these workers 

 overlaps the field of the other workers, but nevertheless 

 each of them has his own specific problem, his own 

 equipment and apparatus. Thus, A and B use sensito- 

 metric apparatus chiefly, C both sensitometric appa- 

 ratus and the thermostatic and electrical equipment of 

 physical chemistry, D microscopic apparatus and chem- 

 ical apparatus dealing with the precipitation of silver 

 salts, and E the analytical and solubility apparatus of 

 chemistry. 



The whole of this range is also connected with colloid 

 chemistry, and especially the overlap of the different 

 sections involves colloid problems, so that we can consider 

 colloid chemistry as dealing with the interrelations of 

 the different sections of photographic chemistry, and 

 can represent its province in the diagram by shading 

 the overlapping areas. The colloid division of the 

 laboratory will therefore be interested in the work of 

 each of the specific investigators, and will be of assist- 

 ance to all of them. 



These charts, prepared for a photographic labora- 

 tory, are equally applicable in form for almost any 

 other convergent laboratory, so that if we have to 

 work out the organisation of a research laboratory 

 which is to study any interrelated group of problems, 

 we can do it by the construction of charts similar 

 to these. Thus, considering Fig. i, we place first at 

 the bottom of the chart the general subject considered 

 and its various branches, and then above these the 

 scientific problems involved, separating out on opposite 

 sides of the chart those problems which would involve- 

 different branches of pure science. Thus, we can 

 place on one side biological problems, then physical 

 problems, then chemical problems, and so on, so recon- 

 structing a chart similar to Fig. i from the bottom 

 up, until at the top we have the various branches of 

 pure science involved, subdividing these branches until 

 each subdivision represents the work capable of being 

 handled by one man in the laboratory. 



It will now be possible to draw Fig. 2, showing on 

 the circumference the different sections of the labora- 



NO. 



2514, VOL. 100] 



tory for which accommodation, apparatus, and men 

 must be provided, and showing the relation of these 

 sections to the problem as a whole, and having worked 

 this out, it is easy to find the amount of space and 

 the number of men which will be required or •w'hich 

 the funds available will allow for each part of the 

 work. 



Specialised laboratories may originate in various 

 ways, but it seems clear that with an increasing total 

 amount of research and with an increasing realisation 

 of the importance of research more laboratories will 

 be developed, and no doubt laboratories which origin- 

 ally were of the divergent type will with their growth 

 tend to split into a linked group of convergent labora- 

 tories. Consider, for instance, a very large industrial 

 research laboratory covering a wide field of research 

 and dealing with many different types of problems. 

 There are two types of organisation possible to such 

 a laboratory. It might be divided according to the 

 branches of science in which the workers were pro- 

 ficient. It might have, for instance, chemical divi- 

 sions, physical divisions, and so on, but if the groups 

 of problems dealt with were reasonably permanent in 

 their character it would more probably develop into 

 a group of convergent laboratories in which men from 

 different branches of science — chemists, physicists, and 

 so on — worked together (and probably even had their 

 working places in proximity) because they were work- 

 ing on the same general problem. Any national labora- 

 tory which is developed for industrial research, for 

 instance, should almost certainly be 

 organised as a group of convergent 

 laboratories rather than as a group 

 of separate physical, chemical, en- 

 gineering, etc., laboratories. 



We may expect, then, that the 

 general organisation of scientific re- 

 search will tend towards the produc- 

 tion of numbers of specialised 

 laboratories, each of which will be 

 working on an interrelated group of 

 problems, and attacking it from 

 various points of view. 

 Some of the questions relating to the internal organ- 

 isation suitable for these convergent laboraitories have 

 already been discussed in a former paper, ^ and I need 

 only add here that the "conference" system described 

 there as a method of actually carrying on the scientific 

 work of the research laboratory has continued to prove 

 quite satisfactory. 



(2) The Classification of Scientific Knowledge. 

 The work of the research laboratories is published 

 by various methods in the form of scientific papers, 

 and with the increasing amount of research done the 

 number of technical journals is increasing steadily, so 

 that the workers in most branches of science find it 

 difficult to keep up adequately with the current litera- 

 ture, and especially those who become interested in the 

 light thrown upon their own problem by other branches 

 of science find it a task of great magnitude to acquaint 

 themselves adequately with the literature. In order to 

 meet this difficulty the various scientific societies pub- 

 lish journals giving abstracts in a conveniently indexed 

 form of all the important papers published, and these 

 abstract journals are of great value in searching for 

 information on special subjects. 



In spite of these abstract journals the task of obtain- 

 ing all the references to the literature on a given 

 subject is still a formidable one, and might be very 

 much simplified by the adoption of some radical 

 changes in the organisation of the abstraction and 

 classification of scientific knowledge. In the first 



2 "The Organisation of Industrial Scientific Research," Science, 1916 

 p. 763 Nature, 1916 pp. 411 and 431. 



