378 



SCIENCE 



[N. S. Vol. XXVII. No. ( 



focus this combined effort upon specific 

 problems. Speaking in general terms, it is 

 evident that the bulk of the investiga- 

 tion going on at present is not being con- 

 ducted on this principle. The productive 

 investigators in the various laboratories of 

 the world are working independently. The 

 problems that engage their attention are 

 determined by personal interests or acci- 

 dents of equipment or opportunity, and 

 their researches are not correlated except 

 in so far as certain problems come to the 

 front from time to time, and by the general 

 interest which they excite attract for a 

 period numerous workers to a common line 

 of study. The fortunate investigator who 

 unearths a new idea, or devises a new 

 method of importance, is sure to have many 

 followers, and there results for a while a 

 certain kind of cooperation, which is lack- 

 ing, however, in the element of intelligent 

 coordination; so that oftentimes there is 

 an apparent waste of energy and material, 

 due to the fact that the individual in- 

 vestigator is unable or unwilling to make 

 full use of the results obtained by his 

 coworkers. The method of independent 

 investigations needs no apology or defense, 

 and we should be careful not to minimize 

 its importance. The competition that it 

 implies encourages originality and carries 

 with it all the benefits that accrue from 

 differences in point of view. It is to this 

 kind of investigation that we must look for 

 our epoch-making discoveries, so at least 

 we may, infer from the past history of 

 science. No one surely has any wish other 

 than to see this kind of research grow in 

 volume and importance in this country. 

 Nevertheless we may ask whether it is not 

 possible that in some ways better and 

 quicker results would be obtained by 

 directed cooperation. In the accumulation 

 of reliable data, for example, by the con- 

 centrated application of approved methods 



of work. It needs no argument, I am sure, 

 to convince any experienced worker in sci-' 

 ence that eventually such accumulated 

 knowledge will cause of itself the destruc- 

 tion of false theories and the development 

 of newer and truer points of view. In my 

 own subject, at least, it is undoubtedly a 

 fact that brilliant discoveries have come, 

 as a rule, not as a bolt from the blue, but 

 from a slow accumulation of diverse facts 

 and theories which, eventually, in the mind 

 of some one gifted worker, when the time 

 was ripe, have burst forth as a new con- 

 ception. Our individual workers of genius 

 must be supplied with raw material in the 

 way of facts and theories in order that 

 their talent may be productive of real good, 

 and it is in the accumulation of this raw 

 material that most of us make our contribu- 

 tions to the advancement of science. It is 

 in this direction also, as well as in the 

 utilitarian application of scientific knowl- 

 edge, that cooperative work, as defined' 

 above, might be depended upon to greatly 

 accelerate the rate of progress. From the 

 point of view here adopted the success of 

 cooperation in scientific investigation must 

 depend chiefly upon the possibility of devis- 

 ing an efficient organization for carrying it 

 on, and obviously two essential require- 

 ments of such an organization are, first, 

 that it shall possess sufficient dignity and 

 authority to make its direction respected, 

 and second, that it shall have at its dis- 

 posal sufficient funds to pay for the ex- 

 penses of the work. Several possible ways 

 may be suggested for developing such a 

 mechanism for cooperative research. In 

 the first place it is quite possible that any 

 body of scientific men may cooperate by a 

 series of conferences and some sort of a 

 voluntary compact. A notable example of 

 an important effort of this kind is found in 

 the International Union for Cooperation in 

 Solar Research, already referred to. Quite 



