No. 511] THE THEORY OF ORTHOGENESIS 409 



ideas of relationships are very liable to result from in- 

 complete knowledge of the course of evolution in the 

 group, so that it is very hazardous to select a few forms 

 of a genus and endeavor to discover the laws governing 

 their development. An illustration can be drawn from 

 this genus: hntlrri, which lies entirely within the range 

 of sirtalis, may, so far as its characters go, easily be con- 

 ceived to have been derived from the latter, and, if BO, 

 a different theory of the factors involved in its evolution 

 must be sought to explain its origin. That it has more 

 probably been developed from radix by dwarfing only 

 becomes relatively certain when the lines of evolution 

 in the genus have been worked out. This should discour- 

 age attempts to adduce as evidence for or against any 

 theory the relationships of particular forms before the 

 genus has been studied as a whole, and the general lines 

 of development determined. 



