No. 554] ADAPTATION THROUGH SELECTION 



67 



many qualities, for the most part minor qualities, which 

 so far as we can judge, are non-adaptive. I would sug- 

 gest for your consideration the propositions that the 

 broader adaptive features are due to natural selection, 

 and that the non-adaptive minor characters are the result 

 partly of factors within the organism, and partly of ex- 

 ternal factors other than natural selection. This paper 

 will not refer further to these other external factors, but 

 will discuss one of the factors within the organism. . 



Evidence accumulated during the last decade seems to 

 indicate that fluctuating variations are not significant, 

 and that only mutations are stable and can serve as a 

 foundation for evolution. Perhaps we have accepted this 

 statement a little too hastily. The actual evidence in its 

 favor is not all that could be desired, but it seems to 

 be the general opinion to-day. If this opinion is well 

 founded, it is through the study of the origin and nature 

 of mutation that we may hope to gain most light upon 

 the problem of the origin of adaptation. 



When one conies to think of it, we really know very 

 little of mutation. Few instances have been carefully 

 observed and recorded. But in the midst of this scant 

 knowledge of mutation one fact stands out clearly, that 

 mutation apparently is not indeterminate, occurring in 

 all directions equally, and changing from generation to 

 generation, as would be the case were it purely fortuitous. 

 Our best studied instances of mutation are still seen in 

 (Enothera lamarckiana, and in all the observations upon 

 the mutation of this species nothing is more patent than 

 that there are certain definite tendencies for particular 

 mutants of very complex and very clearly recognizable 

 types to appear generation after generation, in numbers 

 that are very much greater than could be accounted for 

 by any chance aggregation of unit characters to make 

 these few well-marked types. 



In 1905 I wrote as follows: 



