1917.] NATURAL SCIENCES OF PHILADELPHIA. 13 



premaxillary fontanel and that "in the only skeleton I have examined 

 the prefrontal is present one side and wanting on the other" would 

 bring it near Leurognathus. 



The next group, Desmognathiis and Leurognathus, agree in the 

 absence of prefrontals, poorly developed vomerine teeth, and in a 

 peculiar collar on the atlas where the temporal muscles originate. 

 But the absence of a premaxillary fontanel in Leurognathus and the 

 presence of well-developed vomerine teeth in some male and all 

 female Desmognathiis prevent us dividing one genus from the other. 

 All the preceding genera have typically aquatic larvae. Desmogna- 

 thus seems to tend towards terrestrialism in that one species lays 

 eggs on land and the aquatic larva of another is unknown and may 

 not exist. Typhlotriton and Autodax stand between this group and 

 the Plethodon group, but there is some uncertainty as to the actual 

 relations. Both have prefrontals and one premaxillary with a 

 fontanel. Autodax is a Pacific-coast form with absolutely terres- 

 trial existence. Typlotriton is a blind cave form from Missouri 

 with an aquatic larval stage. In both the vertebrae become 

 opisthocoelous with age. Both, according to Cope, have the tem- 

 porals originating on the atlas. Typhlotriton, whose cranial structure 

 Ave recently examined, has not the transverse dorsal ridge on the 

 atlas which characterizes Desmognathiis and Leurognathus. Whether 

 Autodax has it is not known, but the tendency of the temporal 

 muscles to originate on the atlas is accentuated in these forms, as 

 Cope pointed out. Geographical probabilities and life histories 

 suggest Typhlotriton as next of kin, while the hooked jaw of the 

 most terrestrial species of Desmognathus (ochrophea) would indicate 

 Autodax. 



The next group is Plethodon and its two derivatives, Hemi- 

 dactylium in the east and Batrachoseps in the west, is, so far as known, 

 entirely terrestrial. The vertebrae appear to remain amphicoelous 

 throughout life. Plethodon with two premaxillaries stands at the 

 ])ase. Hemidactylium differs in the loss of a toe, Batrachoseps in 

 the loss of prefrontals, the consolidation of the premaxillaries and 

 the loss of a toe. Heptoglossa known only from one specimen, has 

 the tongue attached, the toes webbed and opisthocfielous vertebrae. 

 Beyond this it is unknown. Probably it should stand between- 

 Plethodon and Autodax. Typhlomolge is a permanent larva of some 

 plethodont. More cannot be said of its relation?^ other than its 

 affinities are more likely at the aquatic Spelerpes end of the series 

 rather than at the terrestrial Plethodon end. 



