194 F. B. LOOMIS OKIGIN OF SOUTH AMERICAN FAUNAS 



Permian of North America and the Middle and Upper Permian of Scot- 

 land and Africa. Anomodontia are scattered through North x\merica, 

 Europe, and Africa. Edaphosaurus is North American and European. 

 Dicynoddntia are African, Scottish, and North American. There is noth- 

 ing to indicate that Africa was isolated, and the forms seemed to have 

 moved from Africa to Europe to America, or in reverse direction, easity 

 throughout the Permian. 



Eigenmann^ has used the fresh-water fishes in an extensive argument 

 for Gondwanaland. After cataloguing all the fishes of South America 

 and showing that some have come from North America, some from the 

 salt water, he lists six families, and from maps of their present distribu- 

 tion develops an argument to show that before the Tertiary, and not later 

 than the end of the Cretaceous, there must have been a land connection 

 between Africa and South America. The families are Lepidosirenidse, 

 Osteoglossidse, Siluridse, Pcecilidse, Cichilids, and Characidfe. Of Poe- 

 cilidffi he remarks : "Marine, brackish and fresh water — a connection with 

 Africa would be convenient." Of Siluridse: The Tachisurinse are ma- 

 rine, and the continental forms could have been derived from this group, 

 as could also those of Africa. This leaves four families. The Lepidosi- 

 renidse are certainly relics of an ancient group, formerly living in abun- 

 dance in both Europe and North America. The Osteoglossidse are also a 

 fairly old family, and Dapedogiossus of the Eocene Green Eiver beds 

 shows that the family was in North America at least that early. Cich- 

 ilidse also were represented by Priscacas in the Green Eiver beds and by 

 Odonteus in the Monta Bolca of Europe. These three families, then, I 

 would not feel needed an Africo- South American land connection. And 

 this leaves only the Characidse, a group of some 400 minnows and carp- 

 like forms. None of the family is known fossil. It seems to me the 

 argument from fresh-water fishes is not compelling. All that is needed 

 is to find that this last family contains or contained some form which 

 could stand salt water. The Galaxidse of Australia, South Africa, and 

 South America used also to be used for this argument, but it has been 

 shown that some of the members of the family enter salt water readily. 

 There are two fossil fish faunas in South America — an Eocene one from 

 Algoas, Brazil, which is dominantly made up of genera very close to the 

 North American Green Eiver forms, and a slightly older fauna, also from 

 Brazil, in which the genera also are closest to North American types. 



Among the arguments based on invertebrates, that on the non-marine 

 mollusca is the best, and it is summarized by Pillsbury in the report of 

 the Princeton Patagonian expeditions. An interesting map of the paths 

 of migration of land snails shoAvs one from North to South America (17 



* Princeton expeditions to Patagonia, vol. 3, 1905-11, p. 310. 



