MILLIPED GENUS CHEROKIA — HOFFMAN 261 



difficult to determine whether a certain local population should be 

 considered as part of a subspecies or of the intergrading series. This 

 fact seems a common and entirely reasonable consequence of geograph- 

 ically induced speciation. 



On the other hand, C. g. georgiana and C. g. lalassa differ mainly in 

 but a single character, but in such a profound one that it separates all 

 the material examined into one group or the other. The reasons for 

 considering these two groups as conspecific have already been set 

 forth: The virtual identity in all other structural details; separate 

 distributions; the occurrence of some specimens of latassa which 

 presage the appearance of intergradient characters; and the sharing 

 by the two groups of sternal knobs on segment 7, in the region w here 

 intergradation is presumed to occur. 



The inference L draw from the relationship of latassa lo georgiana 

 is one of a division of the parent population into two parts, with the 

 barrier subsequently withdrawn before specific divergence had been 

 achieved. Judged from present distributions, the Tennessee River 

 seems entirely possible as the isolating mechanism because of the 

 change in course to the southwest through eastern Alabama. That 

 such a diversion occurred is contended by numerous zoologists, 

 although there is by no means complete agreement. Reestablishment 

 of the river in its present course curving westward through northern 

 Alabama would permit an eastward migration of the subspecifically 

 differentiated latassa and eventual contact and intergradation with 

 georgiana in the region where, today, the evidence suggests it occurs. 



The relationship of ducilla to georgiana seems more intimate and 

 suggests virtually uninterrupted contact between the two, with <hu ilia 

 representing a recent terminal race of the migration northward into 

 the mountains. So far as I know, georgiana does not cross the French 

 Broad River, whereas latassa ranges much farther north, into Ken- 

 tucky. From the great variability of georgiana in western North 

 Carolina, and its relatively narrow distribution there, I judge that the 

 species' occupation of the mountains is a fairly recent event. Pre- 

 sumably the roughness of the terrain enhances or imposes localized 

 variability; latassa varies less over a range thirty times as great. 



It therefore seems reasonable to assume that georgiana and latassa 

 diverged first, and in only one character. The race ducilla is a younger 

 offshoot of georgiana, and differs in several characters because of the 

 opportunities afforded by the invasion by its parental stock of a strong- 

 ly dissected country with different climatic conditions. The relation- 

 ship may be expressed thus: 



ducilla 

 latassa / 



\ georgiana 



parental georgiana 



