206 BULLETIN 17 5, UNITliD STATES NATIONAL MUSEUM 



northwest obviously converge in the southwest in the range of sayi 

 affinis. The series of forms in each of these lines is not linear, how- 

 ever, and in each case except that of vertebralis a subsidiary center 

 of dispersal for the group may be determined. Thus, in the deppei 

 group, d. deppei is clearly ancestral to d. jani and lineaticollis, in the 

 eastern line s. sayi leads to m. ruthveni, from wliich the other three 

 subspecies of melanoleucus obviously have been derived independently 

 of one another, and c. catenifer and c. annectens undoubtedly have 

 arisen separately from c. deserticola. 



2. Continuity and directness of individual variations or modifications 

 radiating from the center of origin along the highways of dispersal. — As 

 has been stated above, the characters showing any pronounced geo- 

 graphic variations vary in general along the lines of radiation from 

 the center of origin for the form, if affinis is accepted as the central 

 form of the genus, to the periphery of its range. A complete con- 

 tinuity of variation from form to form is impossible in a genus in which 

 the evolutionary series are not strictly linear, and where the lines of 

 dispersal fork on the periphery of the range into two or more diverging 

 branches. In such a case, if the variational trend is continuous into 

 one of the diverging lines of each branching, as in each of the evolu- 

 tionary series of Pituophis, the requirement may be considered to be 

 fulfilled. 



3. Direction indicated by biogeographical affinities. — The geographic 

 probability that affinis is the form nearest the evolutionary center of 

 the genus is clearly indicated by a consideration of the distribution 

 of the included forms. Thus affinis is near the geographic center of 

 the genus, and has a range contiguous with or overlapping that of the 

 form of each radiating evolutionary line which is nearest to it in scale 

 and pattern characters, and thus obviously most closely related. 



It is evident that in accordance with these three criteria affinis must 

 be accepted as located at the probable center of origin for the genus. 



The main principle of dispersal propounded by Matthew (1915, 

 p; 180) is as follows: 



Whatever agencies may be assigned as the principal cause of evolution of a race, 

 it should be at first most progressive at its point of original dispersal, and it will 

 continue this progress at that point in response to whatever stimulus origmally 

 caused it, and spread out in successive waves of migration, each wave a stage 

 higher than the previous one. At any one time, therefore, the most advanced 

 stages should be nearest the center of dispersal, the most conservative farthest 

 from it. 



Since every animal is dependent for its existence upon a favorable 

 environment, this factor must be accepted as of the utmost importance 

 in its relation to evolutionary change. Whether a changed environ- 

 ment affects the animal directly, or only indirectly as the agent of 

 natural selection, is immaterial in this connection. In either case, it 



