452 MEMOIRS OF THE CARNEGIE MUSEUM 



h. Origin of the distribution of C. carol irius. 



Generally, conforming to tlie subgenus Bartonius, the center of radiation of this 

 species is to be sought in the southern part of the Appalacliian system. It has fol- 

 lowed in its migration the strike of the mountains, keeping to the higher parts of 

 the latter. Thus it has entered soutliern Pennsylvania, being restricted here to the 

 highest portions of the Alleghen}^ Plateau. 



The lowest elevation at which I found it is at Ohiopyle, Fayette County, 1,250 

 feet, and at Dunbar, P'ayette ('ounly, 1,260 feet. (At the latter place a few strag- 

 glers — two specimens — were taken as low as 1,070 feet, associated with (J. diogenes, 

 but here they had apparently come down from the top of the mountain, where this 

 species was abundant at 1,260 feet.) All other localities in Pennsylvania were 

 highei', generally between 1,500 and 2,000 feet. 



The northern boundary of this species in our state is formed by two difi'erent, 

 opposite features in the physical geography. Between Chestnut and I^aurel Hill 

 Ridges it is a cross divide of the longitudinal valley ; between Laurel Hill and the 

 Alleghany Front tlie deep erosion of the original longitudinal valley by the head- 

 waters of the Conemaugh River forms the boundary. We do not know much of the 

 geological history of this I'egion, but it seems to me that the floors of these high 

 valleys with their extensive clay deposits form a part of a former base-level, namely, 

 that of the Old Tertiary peneplain identified with the Harrisburg peneplain by 

 Campbell (1903, p. 293). In northern Somerset and southern Cambria Counties 

 this has been eroded by the ( 'onemaugh system, thus removing a good deal of the 

 clay bottoms, which seem to l)e an essential condition for this species, and conse- 

 quently the lack of this feature, or its interruption l)y the < 'onemaugh system at the 

 northern end of Somerset County, has formed here the Ijarrier to the dispersal of C. 

 carol inus. 



To all appearances C. carolinus is a Postglacial immigrant into this state. The 

 northern boundaries in both of the longitudinal valleys are rather insignificant, and 

 we should expect that C. carolinus, being a chimney-builder and able to leave the 

 water for a considerable time, should be able, like'V'. hartoni, to cross boundaries of 

 this character. We should even expect that it would be better fitted to do so than 

 C. bartoni. In fact 0. carolinus must have done so repeatedly on its way from the 

 South, being found in the upper drainages of rivers running in difierent directions, 

 for instance, the upper Youghiogheny in Maiwland, the u])per Potomac in ^laiy- 

 land and West V^irginia, upper Decker's Creek (tributary of the Monongahela), and 

 upper Cheat River in West Virginia.'' That it has been checked in I'ennsylvania 



^' As to stream adjustments and migration of divides in Garrett County, Maryland, See Abbe, 1002, p. 47, .53. 



