434 MEMOIKS OF THE CARNEGIE MUSEUM 



wliat toward C. obscurm, and one individual lias l)een found wiiich represents 

 typical C. obscurus. In Middle Island ('reek near St. Mary's, Pleasants ('ounty. 

 West Virginia, which is al)out twenty-five miles further down the Ohio, the few 

 specimens collected seem to be typical C. propinquus sanJmrtii. 



Thus it appears that C obscurus goes down the Ohio River to alxjut Mounds- 

 ville. West Virginia. All the tributaries of the Ohio in the Panhandle possess this 

 species, and very probably it will Ije found also in Ohicj on the opp<jsite side of the 

 river. But crossing over the divide between this part of the Ohio and the Muskin- 

 gum-Tuscarawas River in Ohio, we again find ('. projiiii<jnii.'< ><anJ>iinn in the drain- 

 age of the latter. The western boundary of (J. obscurus consequently is formed Ijy 

 the divide just mentioned, but this line crosses the Ohio River between Mounds- 

 ville and New Martinsville, West Virginia (PL XLII, Fig. 2, and PI. XLIII). 



Further to the south in West Virginia in the drainage of the u})per Mononga- 

 hela this species has not been traced. It surely goes up the Monongahela bey(jnd 

 the southern l)Oundary line of Pennsylvania, Init how far has not been ascertained. 



The fact that C. obscurns is found also in the Potomac drainage, in Wills Creek, 

 between Hyndman, Bedford ( 'uunty, Pennsylvania, and EUerslie, Alleghany County, 

 Maryland, deserves special menticjn, and will l)e eommenteil upon elsewhere. 



b. Orujhi of the (li.sfriJnitlnv of ('. propivquus, propi nqiius sanhornu niid C (iJ)scnr(is. 



In order to get a fair understanding of the distribution of these forms, we must 

 take notice of the Preglaeial physiography of the region in which they are found, 

 for, as we shall see below, we are led to believe that these forms ai'e of Preglaeial 

 age, and survived during the Glacial Period in the southern parts of the drainage 

 systems, which now constitute that of the ( Jhio. 



First of all, we should bear in mind that at the end of the Tertiary Pei'iod before 

 the ice pressed down from the north, the Ohio River in its present form did not 

 exist. In the whole region, drained now by the middle and upper (jhio, the drain- 

 age was at that time not to the west, but to the north, and it was collected by a 

 river running in a northeasterly direction toward the present (iulf of St. Lawrence, 

 (the Erigan River or Ancient Cirand River). ^' 



Disregarding some smaller streams, for instance the Old Middle and ( >ld U])pcr 

 Alleghany, which do not concern us here, three main rivers, tributaiy to the Erigan 

 River, have been traced with uKjre or less accuracy, and the evidence for their ex- 

 istence, although fragmentary, leaves no doubt as to the general correctness of the 



■"This is tlie opinion of Spencer (1881 and 1894). Others believe that tliis river drained toward the Southwest, 

 into the Mississippi ; see (Jrabau, 1901, maps, p. 44 and 45 (Dundas River). 



