406 



THE AMERICAN NATURALIST [Vol. XLTII 



sions on this subject without the controlling evidence of 

 experimentation, certain general relations seem to be 

 apparent. 



I have stated that the variations in each form fluctuate 

 about a mean, but that this mean varies, approaching in 

 the intermediate region that which characterizes the next 

 form on its line of descent. That the different forms 

 also originated in this way is shown in the numerous 

 instances where they actually intergrade. This is essen- 

 tially the idea of mutation (in the sense of Waagen and 

 Scott) or ' ' phylogenetic variation", which is not to be 

 distinguished from individual variation by any character 

 of quality or quantity, but by the fact that it pursues a 

 determinate direction by the gradual shifting of the nor- 

 mal type. Conn 5 states that : 



It should be noticed that these considerations in regard to variations 

 along definite lines have less significance in connection with such char- 

 acters as can be supposed to advance by general averages. Some organs 

 have been advancing in definite directions for long generations, but if 

 the advance consists of an increase or decrease in size of the organs there 



If it be an advantage to have an organ increase in size, and if variations 

 in this organ occur around an average type, then without any necessity 

 of supposing a special law directing variations, we can understand how 

 natural selection will continue to increase the size of the organ in 



All this is very true, for it is only orthoselection work- 

 ing on the average type, but as selection seems to be de- 

 barred in this case, the characters being non-utilitarian, 

 we apparently have in the garter-snakes a case of evolu- 

 tion along fixed lines as the result of definite variation. 

 Even if we could admit the selection of fluctuating varia- 

 tions as the dominant factor in the evolution of these 

 forms, we should encounter the additional difficulty that 

 it (selection) is apparently unable to create a new species, 

 the form slipping back to the original condition when 

 the selection ceases. The nature of the variations also 

 seems to me to debar de Vries 's theory of mutations as a 



s Conn, H. W., "The Method of Evolution," p. 146. 



