372 • H. D. REED 



The Plethodontidae and Desmognathidae may be considered 

 together, since the sound-transmitting apparatus in these fam- 

 ihes is identical. Here the columella is reduced to that extreme 

 that it takes no part in the formation of the fenestral plate, 

 merely becoming joined to it as the stylus columellae. In these 

 families the identity of the columella is not lost, even after its 

 attachment to the side of the plate. The differences between the 

 sound-transmitting apparatus of the two families are slight, cer- 

 tainly of not more than generic value, which argues for the inclu- 

 sion of the two groups in the same family, as was suggested by 

 Moore ('00) in his study of the vertebrae and later adopted by 

 Dunn ('17). Moore's point is well taken. The vertebrae of 

 urodeles are acentrous. In such a shell of bone formed by the 

 descent of the dorsal elements, a cavity necessarily results and 

 its later partly or completely filled state, which resembles an opis- 

 thocoelian vertebra, is without significance in this connection. 



As stated above, it seems clear that the tjrpe of sound-trans- 

 mitting apparatus which obtains in the plethodontid legion could 

 have come about originally only under the influence of or in 

 adaptation to a terrestrial abode. For the existence of the isth- 

 mus fenestralis no reason suggests itself, unless it be that dis- 

 cussed in connection with Spelerpes in the first part of this paper 

 or the persistence of the earlier cutting-out process of the oper- 

 culum, which is here chiefly that of growth of the fenestral lips 

 into the membrane. 



The plethodontid legion, as a whole, has descended from ter- 

 restrial forms, the Necturidae, Amphiumidae, and Typhlomolg- 

 idae having returned to an aquatic abode. The extreme spe- 

 cialization of the sound-transmitting apparatus in these second- 

 arily aquatic species leads to the belief that the legion has lived 

 in a terrestrial environment from a very remote period. The 

 unusually long larval period of Spelerpes and Gyrinophilus 

 should be interpreted as a tendency of these modern species to 

 return to an aquatic life as Necturus has done. The short larval 

 period of Desmognathus, therefore, represents the normal pe- 

 riod, or one in which the tendency is toward a more strictly ter- 



