EMBRYOLOGY OF CRYPTOBRANCHUS 547 



nibranchs and the derotreraes^) as related to the land-hving 

 salamanders. 



The most prevalent view has been that the aquatic urodeles 

 are the most primitive. Parker and Haswell ('97, vol. 2, p. 291) 

 have said: ''The perennibranchiate urodeles are undoubtedly 

 the lowest of existing amphibia; they lead up, through such 

 forms as Amphiuma, with persistent gill slits but deciduous gills, 

 to the land salamanders, in which a purely terrestrial form is 

 assumed." In a group standing between fishes and the typically 

 terrestrial vertebrates, it is natural to regard the aquatic forms 

 as transitional to the terrestrial. In fishes there are usually 

 five branchial arches, and the gill slits remain open throughout 

 life. The land-living salamanders have in the adult state only 

 two branchial arches (Parker '77; Wiedersheim '77; Cope '89), 

 and the gill slits are open only to the end of larval life. In the 

 aquatic urodeles there are usually three or four branchial arches 

 (as in the larvae of the terrestrial forms), and the gill slits usually 

 remain open throughout life — conditions intermediate between 

 fishes and salamanders. In view of the occurrence of external 

 gills in all larval forms, the persistence of such gills in the peren- 

 nibranchs might, on the recapitulation theory, be regarded as 

 a primitive character. 



In the opinion of various authors, the above interpretation 

 represents a short-sighted view of the matter. Boas ('81) was 

 the first to assert that the perennibranchs are larvae that have 

 lost the ability to transform; this conclusion was reached as a 

 result of a comparative study of the circulation. Cope ('85) 

 described the retrograde metamorphosis of Siren, and concluded 

 that the present Sirens are descendants of a terrestrial type. 

 Gadow ('01, p. 66) suggested a terrestrial ancestry for the uro- 

 deles. Kingsley ('01) says: "The salamandrina form the central 

 urodele stem, and the perennibranchs and derotremes have 

 been derived from this stem by degeneration and the retention 

 of larval characters." Kingsbury ('05), on the basis of a compar- 



^ The classification of Stannius ('56) is here followed, as it seems best adapted 

 for the purposes of this discussion. 



