106 ORIGIN, ETC., OF THE FAUNA. 
Cyprinoidea (Characins, Electric Eels, Carps, Loaches) and Siluroidea (Catfishes). 
Of the Cyprinoids the Cypriniformes are absent from South America; the Characi- 
formes (Characins) are represented by five families—four endemic and the other, the 
most generalized, found in Africa also; whilst the Gymnotiformes (Electric Fels) are 
peculiar to the Neotropical Region. The Siluroids include the archaic Diplomystes 
of Chile and eight endemic families, of which the most generalized, Pimelodide, is 
related to the African and Indian Bagride. The only other Neotropical fishes of much 
importance geographically are the Cichlid Perches, also found in Africa, for the 
Cyprinodonts are partly marine and the sea has aided in their dispersal. 
Origin of N eotropical Fishes.—It has been suggested by various authors that 
the ancestors of the characteristic neotropical faunal groups may have reached South 
America in late mesozcic or very early tertiary times either from Australia wid 
Antarctica, from North America, or from Africa vié an Atlantic continent. Recently * 
I have attempted to show that the evidence for the supposed connection with Australia 
does not warrant the conclusions that have been reached; so far as the true fresh-water 
fishes are concerned, there is only one family common to Australia and South America, 
the Osteoglosside, a generalized, ancient, and widely distributed group represented at 
the present day by a few remnants. As to North America, it has not, at the present 
day, a single family in common with South America, and the fossils show that it has 
been just as distinct throughout the tertiary. Cat-fish (Rhineastes), an Osteoglossid 
(Dapedoglossus), and a Perch (Priscacara) from the Green River Shales (Lower Eocene) 
in Wyoming have been cited as showing neotropical affinities +. In my opinion, 
Rhineastes may belong to the cosmopolitan and marine family Ariide, Dapedoglossus 
seems to be nearer to the Indo-Australian Scleropages than to the South American 
Osteoglossum, and Priscacara is not a Cichlid. All the Cichlid, whether African or 
American, have the palate toothless, and the caudal fin formed of 16 principal rays, 
14 branched. Priscacara has teeth on the vomer and 17 principal caudal rays, 
15 of them branched. After examination of the specimens in the British Museum, 
I conclude that Priscacara is a member of the endemic Nearctic family Centrar- 
chide, and is closely related to the modern Eupomotis, in which genus the 
enlargement and coalescence of the lower pharyngeals may also be seen. If the 
ancestors of the present Neotropical fishes came from North America, no trace of 
them has yet been discovered. 
The hypothesis of a former land-connection between South America and Africa 
receives strong support from the Fishes; although no genera of the Characide and 
Cichlide are common to the two continents, the close relationship of Brycon and 
* Brit. Antarctic ‘Terra Nova’ Exped., Fish. 1914. 
T Osborn, ‘ The Age of Mammals,’ p. 136. 
