446 MEMOIRS OF THE CARNEGIE MUSEUM 



C. propinquus and of C. j^rojyinq^ms scmborni is rather unsatisfactory, but none of 

 the known facts is opposed to our theory. With refei-ence to C. ohscurKs in Penn- 

 sylvania, I think our assumptions are well supported. Cumbanis obscurus is a Pre- 

 glacial form, helonging to the Old Mononguhela {or Spencer) River, which survived dur- 

 ing Glacial times in the Jiradiraters of this river {Lake Monongahela), and spread ant, 

 in Pustglaciid tinas, orer the irlmle of the Upper Ohio dreihuige {in ad<liti<in to the 

 Ohio a)id Monongalnla, orcr tite drainages of the Beaver and thr Alleghang Ilivers), (tnd 

 iras onlji checked in its disjx rsal in the direction, toirard tin: riionnfains hg the roughness 

 (f tlie streams. Tlie Ohio River of Postglacial times opened a way down stream, 

 but C obscuriis was unable to spread in this direction, since these parts were occu- 

 pied by another closely allied species, C. propmieptus sanborni. It has slightly 

 entered upon the territoiy of the latter (Fishing Creek), hut has not been able to 

 ci'owd it out or to coiKjuer it. Similar conditions prevail in tlie Lake Erie drainage, 

 which has been reached in consequence of stream-piracy, or else, by the help of 

 nrodern canals. Here it came into contact with C propinquus. In both cases (in 

 West Virginia and Erie County, Pennsylvania) we see that the (jtlier species show 

 indications of an inclination toward C. obscurus. I believe we have to deal here with 

 ]ii/J)ridization, but tliis will be discu.ssed later. Finally the species has crossed over 

 into the Lake Ontario di-ainage in the region of the headwaters of Genessee River, 

 presumably in consequence of stream-piracy. In the upper part of this system, in 

 Pennsylvania, it did not find any competition, and is alone represented there, while 

 in the lower part, at Rochester, N. Y., it is again fountl associated with C. propin- 

 quns. Further details with respect to these parts are not at hand. 



Comparing the distrilmtion of tlie //rr*j>/»7/n/.x-group with Adams' scheme of 

 Postglacial dispersal of the biota of North Ameri-ca, we see at a glance that the whole 

 group belongs to his northea-^tcrnbiata of the second irare (Adams, 1905, p. 5S). The 

 biotic preserve of this element, during glacial times, was not far from the southern 

 edge of the ice, in what is now the ( )hio drainage, but it was re.stricted to this 

 western part, and was not extended east of the x\lleghany Mountains. In Post- 

 glacial times this grou[) advanced northwai'd, firming part of the second wave, 

 which is most clearly seen in the ])resent distribution of (J. propinqmi^i, which largely 

 entered the coniferous forest-lielt in JMichigan, New York, and < 'anada. 'I'lie other 

 two forms (C propin(ptns .^oidxjmi and ('. olm-nrn.^) did not take nuicb jiart in the 

 migration of this wave, since they found a harrier to the noi-tli in tiie shape of the 

 continental divide, and then, after they had crossed this divide at certain points, 

 they found competition in C. [jropinquus, which had populated the whole St. 

 Lawrence drainage at an earlier date ("biocenotic barrier"). In New Y^ork state. 



