AMPHIBIA. 305 



to been generally considered to constitute an abnormal type of the 

 MoUusca are in reality the true progenitors of the fishes. The facts 

 in the case, however, require further substantiation. 



The phylogenetic relationship existing between the selachians 

 and ganoids is equally obscure as that which exists between fishes 

 generally and the other classes of animals. Both types are known 

 to us in their oldest forms from very nearly the same horizon, and 

 consequently give no indication as to priority of birth. Much 

 more positive indication in this direction is afforded with respect 

 to the lung-fishes and teleosts, the former of which appear to be 

 clearly related to, and to have been derived from, the dipteroid 

 ganoids (of the type of Bipterus), and the latter to have held a 

 similar relationship to the rhomb-scaled ganoids, possibly of the 

 type of Leptolepis. Prom the former, apparently, have descended 

 the amphibians, while the latter have continued to develop as the 

 dominant fish-fauna of existing waters. That the fresh-water forms 

 are modified descendants of types originally inhabiting the seas 

 there can be no reasonable doubt. It is impossible to state when 

 the earliest differentiation of marine and fresh-water forms was 

 effected, but there is every reason for supposing that it dates back 

 far into the Paleozoic era, and that some, if not many, of the 

 Devonian fishes were of a strictly fresh- water habit. 



AMPHIBIA. 



The most salient facts that present themselves in connection 

 with the geographical distribution of the Amphibia are, first, their 

 almost complete absence from oceanic islands — the Seychelles, New 

 Caledonia, and the Feejee and Solomon Islands forming island 

 groups exceptional to a general rule — and, secondly, the very nearly 

 universal limitation of the tailed forms, sirens, newts, salamanders, 

 &c., to the Northern Hemisphere. The nature of the first condition 

 has already been discussed in treating of the dispersal of animals 

 generally. The total number of known forms are comprised, ac- 

 cording to the latest researches of Boulenger, in somewhat more 

 than one hundred and forty genera and nine hundred species, of 

 which only twenty-seven genera and seventy species belong to the 

 urodelous or tailed division, eleven genera and thirty-one species 

 are coecilians (Apoda), and the remainder, one hundred and five 

 genera and eight hundred species, frogs and toads (Anura). 

 21 



