45] THE NA SA L ORG A N IN A MPHIBIA —IIIGGINS 45 



related, both having retained these ancestral characters. An older larva 

 of Spelerpes, also figured by Wiedersheim (Fig. 108), has a caudal extension 

 of the antorbital process, which he calls the maxillary cartilage, whose 

 relation to the pterygoid suggests an earlier relation between the two. 



In several of the Urodeles, as shown by figures of Parker (1876), and 

 Wiedersheim (1877), a strong pterygoid reaches well forward from the 

 quadrate, absent in some of my larvae but present in the adult. In the 

 earlier larva of Amphiuma, a pterygoid process has not developed; but in 

 the later stage it has chondrified well forward, nearly to the base of the 

 antorbital. In all Urodeles, both processus antorbitalis and pterygoid are 

 developed, but the proximity of the two varies greatly throughout the 

 order, and I would hesitate to base any statement of relationships upon 

 the relative extent of development of these structures; but no where except 

 in Cryptobranchus and Ranodon is there any connection between them. 



The openness of the nasal capsule of Amphiuma, together with the 

 somewhat specialized anterior part, renders it difficult of homology with the 

 more typical Urodele, and suggests that it is far removed from any other 

 Amphibian. However, in the early method of chondrification, and in the 

 development of some of the parts, Amphiuma appears to be more closely 

 allied to Cryptobranchus. In both, the cristae trabeculorum are formed 

 early, as well as the planum verticale, the latter arising from the median 

 line of the basale and later becoming associated with the columna eth- 

 moidalis. Further relationship is evidenced in the origin of the tectale as 

 a lateral growth from the columna, giving rise in Amphiuma to the elongate 

 lamina externa, along the lateral surface of the nasal sac. Meagre as 

 these relationships may seem, Amphiuma appears to be nearer Crypto- 

 branchus than any other Urodele, the anterior parts of its capsule having 

 become secondarily acquired. 



Although Salamandra resembles Amblystoma in some ways, the many 

 differences between their capsules, such as the size of the planum verticale, 

 the persistence of the foramen ethmoidalis in the former, and the origin of 

 the ethmoidal columns, justifies their separation. In the persistence of 

 the caudal process of the planum basale in the 25 mm. Salamandra, there 

 is a resemblance to Cryptobranchus; but in the origin of the columnae 

 ethmoidales from the medial margins of the cornua, Salamandra is more 

 like Spelerpes. 



Triton and Diemictylus are far removed from the group just described, 

 but are probably related to it through Salamandra. The complete 

 development of a lamina medialis, with the deep internasal space, and the 

 anterior position of the olfactory sac in respect to the forebrain, as well as 

 the persistence throughout life of the fenestra ethmoidalis, are characters 

 common to both Triton and Diemictylus. The complete loss of a pons 

 ethmoidalis in Diemictylus converts the fenestra into a gap, only a mem- 



