NO. 2211, GENERA DESMOONATHUS AND LEUROONATHUS—DUNN. 425 



Morphologically, as I have shown, carolinensis is derived from 

 fusca. Ecologically, while fusca and monticola have apparently simi- 

 lar habitats, actually the chraate of the range of monticola is more 

 humid that that of the range oi fusca. Consequently, in degree of 

 aquatic habitat, fusca is between monticola and carolinensis. 



With the retreat of the ice carolinensis followed the coniferous 

 forest to the southern mountains, where they both remain. Northern 

 representatives of the form have developed into ocTirophaea. 



Monticola also retreated to the mountains because of the greater 

 humidity. It now occupies the deciduous forest in the transition 

 zone in the southern moun tarns. 



Postglacially also occurred the overlapping of the ranges oi fusca 

 and monticola. It matters not whether we assign the differentiation 

 of these two closely related, ecologically similar, forms to the period 

 preceding the last glaciation or to that of the ice itself. At any rate, 

 we are compelled to hypothesize for them ranges originally adjacent 

 but distinct, differing in that the range of monticola was more humid. 

 Perhaps it will best suit the facts of differentiation to suppose that 

 prior to the last glaciation monticola occupied the mountains and 

 fusca the piedmont and coastal plam. 



It is highly probable that the entire extension of the genus west of 

 the Appalachian Valley is postglacial, as neither monticola nor guadra- 

 maculata are known from the Cumberlands. 



There are indications, such as the apparent isolation of brimleyorum 

 in Arkansas and the rarity and disconnected range oi fusca in Indiana 

 and Illinois, that in early postglacial times forms of Desmognaihus 

 extended farther west than now. In this connection the remarks of 

 Gaines (1895) on the recent extinction of fusca near Vincennes, 

 Indiana; of McAtee (1907) on its former presence in Monroe County, 

 Indiana; and Hahn's (1908) failure to find it in Lawrence County, 

 Indiana, are of interest. It should be noted that failure to find 

 fusca is of much greater significance than failure to find most other 

 species of reptiles and amphibians, for where it is present it is much 

 the commonest salamander. 



For the differentiation of monticola from quadramaculata we do 

 not need to assume a distinct geographic range, as they are ecolog- 

 ically different. But we are compelled to place the time of differ- 

 entiation at a more remote period than that of any of the other forms, 

 because quadramaculata is the most prunitive as well as the most 

 distinct form in the genus. 



The above is only a hypothesis, and only in so far as it accomits 

 for the knoAvn facts of distribution and relationship is it worthy of 

 consideration. If it incites discussion and criticism the author will " 

 be well pleased. 



Besides these arguments drawn from the species of Desmognatliu^, 

 it is significant that in the southern Alleghcnies occurs the only 



