November 30, 1917] 



SCIENCE 



525 



biological problems, tben pbysical prob- 

 lems, then chemical problems and so on, so 

 reconstiiicting a chart similar to Chart 1 

 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 ca- 

 pable 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 laboratory for which ac- 

 commodation, 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 which the funds 

 available will allow for each part of the 

 work. 



Specialized laboratories may originate in 

 various ways, but it seems clear that with 

 an increasing total amount of research and 

 with an increasing realization of the im- 

 portance of research more laboratories will 

 be developed and no doubt laboratories 

 which originally were of the divergent type 

 will with their growth tend to split into a 

 linked group of convergent laboratories. 

 Consider, for instance, a very large indus- 

 trial research laboratory covering a wide 

 field of research and dealing with many 

 different types of problems. There are two 

 types of organization possible to such a 

 laboratory. It might be divided according 

 to the branches of science in which the 

 workers were proficient. It might have, for 

 instance, chemical divisions, physical divi- 

 sions, and so on, but if the groups of prob- 

 lems dealt with were reasonably permanent 

 in their character it would more probably 

 develop into a group of convergent labora- 

 tories 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 working on the same general 

 problem. Any national laboratory which 

 is developed for industrial research, for 

 instance, should almost certainly be organ- 

 ized as a group of convergent laboratories 

 rather than as a group of separate physical, 

 chemical, engineering, etc., laboratories. 



We may expect then that the general or- 

 ganization of scientific research will tend 

 towards the production of numbers of spe- 

 cialized laboratories, each of which will be 

 working on an inter-related group of prob- 

 lems and attacking it from vai-ious stand- 

 points. 



Some of the questions relating to the in- 

 ternal organization suitable for these con- 

 vergent laboratories have already been dis- 

 cussed in a former paper- and I need only 

 add here that the "conference" system de- 

 scribed there as a method of actually carry- 

 ing on the scientific work of the research 

 laboratory has continued to prove quite 

 satisfactory. 



2. THE CLASSIFICATION OP SCIENTIFIC 

 KNOWLEDGE 



The work of the research laboratories is 

 published by various methods in the form 

 of scientific papers, and with the increas- 

 ing 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 ade- 

 quately with the current literature 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 publish journals giving abstracts 

 in a conveniently indexed form of all the 

 important papers published, and these ab- 

 stract journals are of great value in search- 

 ing for information on special subjects. 



2 ' ' The Organization of Industrial Scientific Ee- 

 searcli," Science, 1916, p. 763. 



