FISHES OF SOUTH AND MIDDLE AMERICA 525 



These are all marine families, some of which have also developed 

 fresh-water forms in Europe and North America as well as in South 

 America. The fresh-water forms of South America and Africa are 

 local adaptations of marine families that require no change in the 

 present condition to account for their origin. 



The second group comprises the Lepidosirenida?, Osteoglossidge, 

 Siluridge, Characinida?, Pceciliidas and Cichlidas. Of these the 

 Lepidosirenida? are relicts of a former widely distributed group, and 

 it requires no land connection to satisfactorily account for their pres- 

 ence in Africa and South America. The Pceciliidge live indifferently 

 in marine, brackish water or fresh water. They reach their maximum 

 development in the fresh waters of Mexico and Central America. 

 The marine species are found along the shores, not at sea, and there 

 is, therefore, at present, no known means of getting them from the 

 American to the African shore. Nevertheless, Fundulus is found on 

 both sides of the Atlantic, and there must have been an intermigration 

 much more recent than the youngest possible land connection between 

 Africa and South America, or else there has been a very long persistence 

 of this type. A land connection, while not absolutely required for 

 this family, would be very convenient. 



The Siluridge are in part marine. All of the South American 

 forms of Siluridge can be derived from the marine Tachisuringe, and 

 the same is probably true of the American members of the family. 

 Furthermore, the catfishes are found in North America, Europe and 

 Asia and ha,ve been recorded in North America from the Tertiary. 

 A land connection between Africa and South America is, therefore, 

 not absolutely required to account for their presence in both continents, 

 though, as in the case of the Pceciliidse, such a connection would be 

 very convenient. 



The Cichlidse and Characinidae are abundant in tropical America 

 and in Africa, a few species of Cichlidse being also found in India. 

 There is no known means by which these two forms could have crossed 

 the existing gap between Africa and South America. There has been 

 no exchange of species in recent times, for there is no species or genus 

 common to the two continents. The South American and African 

 elements of these two families must have been derived from some 

 intermediate land mass or must have gone from one continent to the 

 other over a land bridge. That this connection, whatever it was, must 

 have been obliterated before the tertiary, is evidenced by the facts that 

 the tertiary deposits of Taubate and Parana show existing genera and 

 that there are many South American types, as the Gymnotidas, Electro- 

 phoridse, Bunocephalidae, Loricariida?, Argiidse, Pygidiidae, Callichthy- 

 ida3, Hypophthalmidre and others not found in Africa that have all 



