RESEARCH IN THE EXACT SCIENCES. 189 



\Letter of G. H. Darwmi\ 



Newnham Grange, Cambridge. 



I sympathize very warmly with Professor Newcomb's plan for 

 developing the Carnegie Institution and think that it may have a 

 great future. I have been trying to picture to myself how it would 

 work out, and I see that while the gain in some subjects would be 

 great and immediate, in others it would be only collateral. 



Scientific observations may be roughly classified in two groups, 

 which, however, graduate into one another. I can best illustrate 

 my meaning by examples. 



The subject of the tides seems to belong to the group which 

 would reap immediate advantage. Observations are now published 

 in the most diverse places and are not properly coordinated. A 

 critical collection of tidal results would be a heavy task and would 

 be of much value. There is nothing in this subject which corre- 

 sponds to probable error in astronomy, for the defects depend on 

 human frailty. It would require a first-rate man to classify and 

 reject observations according to the internal evidence afforded by 

 them. When such a collection was made, generalizations would 

 follow, and the value of the conclusions would probabl}- be great. 



Meteorology and many other subjects fall into this group. The 

 distinguishing feature is that we know exactly what to observe, 

 that the mass of material is already enormous, and that it is impos- 

 sible to have too much matter, provided that it is coordinated. 



The second kind of research to which I have referred is inter- 

 mediate between observation and experiment. The subject of 

 observation is to some degree indeterminate, and it depends on the 

 investigator what he shall observe. 



I can not think of a very good example at the moment, but I may 

 perhaps illustrate my meaning b}^ supposing that we were investi- 

 gating the laws governing the drifting of sand and the formation of 

 sand dunes. It must be obvious that this is a subject of great agri- 

 cultural importance in many parts of the world. Now, it would be 

 almost useless merely to collect maps and photographs. There must 

 be a guiding mind, forming theories to be proved or disproved by 

 observation. The investigation might be expensive and troublesome, 

 but it is essentially the work of an individual. 



In this sort of case I should not look for any great gain from the 

 proposed institution, except that it would afford a fixed position, 

 with good pay, to men of ability. The exception is important, and it 



