NON-MATHEMATICAL SCIENCES 183 



substances we might, after a certain number of 

 cultures, suddenly find the other appearing as 

 a mutation, and this would then breed true. If 

 there were only two isomers we should find occa- 

 sional changes back and forth from one to the 

 other. But suppose that it is a complex mole- 

 cule with a still finite but practically countless 

 number of isomers, then the chance would be 

 negligible that any such mutations would go 

 back to a form that had previously existed. We 

 should see a process of evolution, each molecule 

 reproducing itself exactly, until an accidental 

 rearrangement would set a new molecule to 

 propagating itself. Would not this be repro- 

 duction with the transmission of acquired char- 

 acteristics "^ 



You may object to my using terms drawn 

 from biology ; but suppose that this hypotheti- 

 cal experiment could be realized, which seems 

 not unlikely, and suppose we could discover a 

 whole chain of phenomena, leading by imper- 

 ceptible gradations from the simplest chemical 

 molecule to the most highly developed organism. 

 Would we then say that my preparation of this 

 volume is only a chemical reaction, or, con- 

 versely, that a crystal is thinking about the con- 

 cepts of science ? Nothing could be more absurd. 



