Decembek 7, 1900.] 



SCIENCE. 



859 



we observe to-day. It is highly probable 

 that these conditions, which are decisive 

 not only for the fresh-water fauna but also 

 for the land gastropods, have determined 

 also the history of the mammals, which 

 may have reached Brazil only in the Plio- 

 cene. 



Although these inferences concerning the 

 different faunal elements of the neotrop- 

 ical fauna, based on the zoogeographical 

 work of the writer, seem to be quite con- 

 clusive, the matter is more difficult and 

 hypothetical if we turn to the ancient rela- 

 tions of Archiplata and Archamazonia to 

 other regions of the globe. 



The connection of Archiplata with a 

 great antarctic continent during the Cre- 

 taceous and Eocene formation seems now 

 to be generally accepted, but the historic 

 data given on the matter by Osborn are 

 very incomplete. The first to discuss the 

 question was the eminent botanist, Sir 

 William Hooker, but the work of Wallace, 

 and especiallj' his axiom of the permanence 

 of the great oceanic depths, for a long 

 time retarded further progress. Not until 

 1883 did Hutton, with reference to New Zea- 

 land, and in 1890 the writer, with reference 

 to Archiplata, turn aside to publish new 

 facts in favor of the hypothesis of Hooker, 

 which was also confirmed by Fl. Ameghino. 



In relation to the ancient connection 

 of Africa and Archamazonia I have given 

 arguments (1890) in favor of a Mesozoic 

 * archiatlantic continent,' which existed 

 during the earlier Tertiary. At first be- 

 cause of some paleontological facts noted by 

 Schlosser, I believed that this continent 

 could have transmitted Eocene mammals 

 from South Africa to Europe, an idea now 

 defended by Ameghino and Osborn ; but in 

 1893 I modified my opinion and set forth 

 the hypothesis that no Eocene placental 

 mammals had existed either in Archama- 

 zonia or in Ethiopian Africa. The ancient 

 continent uniting Archamazonia with Af- 



rica I named Archiatlantica in 1890, using 

 in 1892 the term Helenis, and in 1893 that 

 of Archhelenis, with the purpose of prevent- 

 ing confusion with the 'Atlantis ' a hypo- 

 thetical land bridge between South Europe 

 and Central America proposed by linger. 



I will not repeat here what I have said 

 elsewhere as to the intimate relation be- 

 tween the fresh water faunas of Brazil, &ui- 

 ana, and of equatorial Africa, but I shall 

 make some remarks on the geographical dis- 

 tribution of the fresh- water mussels. North 

 America agrees in its Unionidse with Eu- 

 rasia, the genera TJnio, Margaritana and 

 Anodonta being predominant. The arcbi- 

 platic element of South America is formed 

 only by the genus Unio, and by a section 

 of it which has no representatives in the 

 holarctic region, forming the subgenus 

 Nicea, which is found also in New Zea- 

 land and Australia. The numerous pre- 

 sumed genera of Unio now admitted in 

 North America all agree in the character- 

 istic sculpture of the beaks, which is quite 

 different in Nicea. I consider, therefore, 

 Niosa as a genus and the North American 

 sections of Unio only as subgenera. In the 

 archhelenic region we have representatives 

 of Unio which are more intimately allied to 

 Nicea than to Unio, no Anodontas, but very 

 numerous representatives of the Mutelidse. 

 The South American ' Anodonta ' are all 

 Glabaris, a genus of Mutelidse allied to 

 Spatha of Africa. 



Considering the geological history, we 

 find the precursors of the actual North 

 American Unionidse as far back as the 

 Jurassic period, and what we know of 

 fossil mussels of New Zealand and Archi- 

 plata are only Unios of the Niwa section. 

 On the other hand. Cretaceous deposits of 

 Bahia show us representatives of Glabaris 

 and Mycetopus. The actual conditions of 

 distribution therefore predominated even 

 in the Mesozoic time, and no explanation 

 can be given of the intimate relation be- 



