MILLIPED GENUS CHEROKIA — HOFFMAN 251 



both in body form and color pattern with typical georgiana from 

 farther south. 



Intermediate specimens from the Nantahala Range, at Wayah Bald, 

 for instance, are the narrowest outside the range of ducilla, the average 

 w/l being 24.3 percent. Millipeda from Smokemont, at a low eleva- 

 tion (about 2,000 feet) at the very base of the Smokies and less than 

 10 miles from Newfound Gap, are closer in every respect to the Nanta- 

 hala population than to ducilla. Specimens from the type locality of 

 due ilia are the broadest of the subspecies, five measurable males aver- 

 aging 23.4 percent. These topotypes agree with narrower specimens 

 from the high Smokies in color pattern and gonopod structure, however 

 even though certainly not structurally typical of the population here 

 called ducilla. Soco Gap is located in the Plott Balsam Mountains, 

 a southeasterly trending satellite of the Great Smoky massif, which 

 forms the boundary between Jackson and Haywood Counties, North 

 Carolina, for about 30 miles. At its southeastern end, perhaps 20 

 miles below Soco Gap, the local Cherokia population bears little close 

 resemblence to typical ducilla. 



The local picture with reference to color pattern is somewhat more 

 complicated. C. g. georgiana of the lowlands is either transversely 

 banded or trimaculate with very broad median spots (these absent, 

 however, from a single local population in southeastern Alabama) . In 

 the Blue Ridge foothills of north Georgia the median spots tend to 

 become smaller, and this reduction produces the trimaculate Cherokia, 

 which is characteristic of the intermediate population as shown by 

 hollow circles on the map. To the northeast, along the South Carolina 

 border and into Pisgah Ridge, however, the median spots tend to 

 become even smaller and create a local variant (discussed on page 234), 

 in which they are obviously being lost. 



The basic color pattern in Cherokia (and many other xystodesmid 

 genera) appears to be trimaculation. This inference is drawn both 

 from its prevalence and from the fact that the penult instars of cross- 

 banded adults are trimaculate. In some genera, as well, support is 

 drawn from the geographic location of bimaculate or banded popula- 

 tions on the periphery, or at extremes, of the range of vicarious tri- 

 maculate populations. With so little known of the main georgiana 

 lowland population, it would be premature to speculate on the inde- 

 pendent recurrence of bands and spots in the range, but it is certainly 

 noteworthy to disclose microcvolutionary developments within the 

 intergrade population. 



Aside from the tendency toward reduction of the median spots, there 

 are two local phenomena. First, in a very localized region shown by 

 fine stipple on the map, many of the specimens (perhaps a third or 



