NO. 2211. GENERA DESMOGNATHUS AND LEV ROGNATHVS— DUNN. 423 



Specimens examined from localities as f oUows : 



North Carolina. — Stream between Linville and Blowing Rock, 

 about 10 miles from the latter (type locality), 3; Blowing Rock, 1; 

 Roan Mountain, 1; Pineola, 1. 



DISCUSSION OF THE ORIGIN OF THE VARIOUS FORMS. 



Adams (1902) gives 10 criteria for determining the center of dispersal 

 of a related group of animals. These are : 



1. Location of greatest differentiation of a type, 



2. Location of dominance or great abundance of individuals. 



3. Location of synthetic or closely related forms. 



4. Location of maximum size of individuals. 



5. Location of greatest productiveness and its relative stability. 



6. Continuity and convergence of lines of dispersal. 



7. Location of least dependence upon a restricted habitat. 



8. Continuity and directness of individual variations or modifica- 

 tions radiating from the center of origin along the highways of dis- 

 persal. 



9. Direction indicated by biogeographical affinities. 



10. Direction indicated by the annual migration routes in birds. 

 The last is, of course, inapplicable in this instance. 



In regard to the first criterion it has been shown that the three 

 groups of Desmognathus occur together only in the southern Alle- 

 ghenies. 



In regard to abundance, Moore's already quoted remarks on quadra- 

 maculata apply, and Brimley (1912) says of carolinensis, "abundant 

 throughout the regions collected in, particularly at the high alti- 

 tudes" and of quadramaculata (with which he included monticola), 

 "abundant, the most characteristic salamander of the rocky, moun- 

 tain streams." 



As to the third criterion, monticola and carolinensis are closer to 

 quadramaculata than are any other members of their groups. Quad- 

 ramaculata is the largest species in the genus, and monticola and 

 carolinensis are the largest in their groups. 



It has been shown that productiveness, judged by the number of 

 eggs, is, in fusca, dependent on size. The largest fasca seen are from 

 North Carolina, and in the mountains of that State occurs monticola, 

 a close relative oi fusca, and stiU larger. 



The sixth and ninth criteria are closely similar and may be treated 

 together. The dispersal of the ochropliaea group seems to have been 

 northward along the mountains. The fusca group took the plains 

 route radiating from the mountains as fusca and southward along 

 the coast as auriculata. An isolated trans-Mississippi form of the 

 latter has developed into hrimleyorum. In Virginia and northward 

 fy^ca has invaded the mountains into the range of monticola. It 



