546 DISCUSSION ox RESULTS RECORDED 



Dr. A. Smith Woodward, F.R.S., Y.P.Z.S., remarked that 

 nearly all the vertebrates in South America which seemed to 

 suggest a direct land-connection with the Old World through 

 Africa, were either late-Tertiaiy immigrants from North America 

 or senile members of pre-Tertiary cosmopolitan groups. Most of 

 the resemblances in the faunas of the two countries usually noted 

 Avere in animals of which the ancestry was entirely unknown. 

 The only resemblances already explained by palaeontology were 

 due to the survival in the two southern continents of remnants 

 or refugees of formerly widely-spread faunas, which had become 

 extinct in the more progressiv^e northern hemisphere. Pakeon- 

 tologists began to distinguish between the characters of animals 

 Avhich were real marks of affinity and others which were the 

 inevitable and oft-repeated concomitants of maturity and senility 

 in a race. It must be possible to distinguish these characters in 

 a group of animals before the latter can be used in discussing 

 questions of geographical distribution. 



Mr. 0. Tate Regan, M.A., F.Z.S., said : — South America has 

 a very rich and vai^ied freshwater fish-fauna ; with the excep- 

 tion of the Osteoglossidse, a generalized and ancient group 

 represented at the present day by a few remnants, it has not 

 a single family in common with either North AmeHca or with 

 Australia. On the othei" hand, three South-American families, 

 Lepidosirenidfe, Characidse, and Cichlidge occur also in Africa, 

 and the South-American Catfishes of the family Pimelodidas ai-e 

 closely Isolated to the Ba^Tid£e of Africa and India. 



If South America and Africa were one continent in Cretaceous 

 times, and the connection between them persisted until the 

 beginning of the Eocene, these facts would be satisfactorily 

 explained. Alternative hypotheses are that the fa.milies common 

 to South America and Africa were formerly marine and entered 

 their rivers from the sea, or that they were formerly northei'ii 

 and migrated southwards, becoming extinct in the north. Against 

 the former it may be urged that the Lepidosirenidfe are obviously 

 adapted for life in fresh water and unfitted for life in the sea, 

 that the Characidfe are Cyprinoids, a strictly freshwater group, 

 and that if the Cichlida? were formerly tropical shore-fishes, 

 entering rivers, it is curious that they did not establish them- 

 selves in the southern rivers of North Ainerica. The second 

 hypothesis is nnsa,tisfactory, for when the slownaess of dispersal of 

 freshwater fishes is taken into account the improbability is great 

 that several groups should have made these extended journeys, 

 with the final result that closely related genera arrived in Africa 

 and South America. Hydrographical changes, such as the union 

 of rivers formerly distinct or the capture by one river of the 

 tributaries of another, are the means by' which the dispersal 

 of freshwater fishes is accomplished ; for such fishes migration 

 appears to be difficult, survival relatively easy. No known 

 northern fossils can be referred to these African and South 



