420 



THE AMERICAN NATURALIST 



[Vol. LVI 



the ear capsule by nearly its whole border. It seems to 

 me that in this case the columella is at least more free 

 than in Amhystoma, and the operculum less developed. 

 This would be in line with what is known of the habits 

 of Dicamptodon. It is a much more aquatic animal than 

 is Amhystoma. 



In the Caudate sound-transmitting apparatus, taking 

 Reed (1920) as a basis, there are the following sets of 

 conditions: 



I. Both columella and operculum present. Both free. 



Hyrwhius, Ehyacotriton. 

 II. Operculum not developed. Columella free. 



Cryptobramchus. 

 Megalobatrachus Eanodon ?, 

 Onychodactyliis ?, Dicamptodon f 



IV. Operculum developed, free. Columella fused to periotic. Stylus absent. 



Triturus, Pachytriton, P eurode:es f, 



I present. Fused together. Opcr< 



Inasmuch as II is a condition found also in larvae, there 

 is no reason to suppose that the animals in which this 

 condition occurs form a natural group. 



Condition V has been commented upon by Reed (1920), 

 and I am fully in accord with his ideas in this connection. 

 Siren and Batrachoseps are certainly not related. Both 

 are extremely specialized. Batrachoseps has certainly 

 passed through stage VII. Siren has certainly passed 

 through an ancestral period of terrestrial life, yet its 

 other peculiarities are such that it is dangerous to state 

 that its relationships are with the forms in stage IV. 



The forms which show condition VI and condition VII 



