46 BULLETIN 31, UNITElJ STATES NATIONAL MUSEUM. 



T]io characters wliich restrict the family are as follows: 



No etliuioid. Paiatiue boues not prolonged over parasphcuoids, bear- 

 ing teeth oil their posterior margins. Orbitosphenoid separated from 

 prootic by membranous walls. Internal wall of vestibule osseous. Car- 

 pus and tarsus ossified. Vertebrte amphicoelous. Prefrontals and 

 pterygoids present. Premaxillaries fully developed. Parasphenoid 

 without dentigerous plates. An otoglossal cartilage; onl}^ one, the first 

 epibranchial ; second basibranchial isolateil. 



We may here observe the significance of the features defining this 

 family. Two of the characters assigned are what I have termed morphic; 

 that is, one has not been assumed after possession of the other, nor is 

 it identical with the immature stage of the same. Such are the short- 

 ened form of the palatine bones, as compared with the posteriorly pro- 

 duced laminae of the SalamandridfB, and the absence of deutigerous 

 plates on the parasphenoid in the Plethodontidse is a character of the 

 same kind. Under such circumstances we infer that the famihes ex- 

 hibit an outogeuy modified by ccenogeny. 



The biconcave vertebrae constitute a persistence of a larval feature. 



Tlie presence of pterygoids has the same significance with reference 

 to other families. 



The ossification of the carpus and tarsus are characters in which this 

 group develops beyond the larval condition which is permanent in the 

 family Plethodontidse. 



Thus of eight characters two are morphic and six developmental; 

 of the six, two are of advanced development and four of repressed 

 development, as comi)ared with other families. 



The writer characterized this family nearly as above in the Journal of 

 the Philadelphia Academy, 1866, 105. Dr. Hallowell proposed it in the 

 same work, 1858, 337, but on insufficient characters. Many of the char- 

 acters of the principal genus Amblystoma had been already pointed 

 out by Professor Baird. The genera included by Hallowell were Am- 

 blystoma, Xiplwnura Tsch., and Onychodactylus Tschudi. Gray had 

 previously embraced the same genera with Heterotriton Gray, in his 

 lirst section of the Plethodontidai, which corresponds with this family. 

 The writer in 1859 embraced Onychodactylus, Amblystoma, Camarataxis 

 Cope, and Megalobatrachus Tschudi. In the above cited essay of 1866 

 the genera are limited to the two first mentioned with EnsaUna Gray. 



The ravestigation of the subject which I gave in mj'' monograph of this 

 family, published in 1869, resulted in the following disposition of these 

 supposed genera, Baird having already shown the identity of Xiplio- 

 nura with Amblystoma. " Heterotriton is identical with Amblystoma. 

 Megalobatrachus, the great salamander of Japan, I have determined 

 to pertain to the Cryi^tobranchidte. The genus Camarataxis, as will 

 appear further on, was established on a larval character, permanent 

 in some individuals it is true, but not permanent in any species. On 

 the other hand, there is some probability that one or both of the species 

 of Ilynobius Tschudi from Japan enter the family, but this I am not 



