274 THEORY OF A MIOCENE ATLANTIS. [Ch. XY. 



Upon the whole, the theory which derives the American types 

 from the east instead of the west seems by far the most natural, and 

 it seems to acquire still more claims to our favor when we study the 

 fossil shells and corals of that ancient period as well as the plants. 

 In 1850, Mr. John Carrick Moore pointed out that certain tertiary 

 shells of San Domingo exhibited affinities to the miocene shells of 

 Europe,* and that although such of the San Domingo species as 

 agreed with the living were chiefly Atlantic forms, there were some 

 so closely allied to the existing Pacific fauna as to lead him to infer 

 that there had been a channel in Miocene times through what is now 

 the Isthmus of Panama, by which the mollusca could have migrated 

 from one ocean to the other. Such an hypothesis, he observes, will 

 be the more readily accepted when we consider that the isthmus no- 

 where attains an elevation exceeding 1000 feet, which is not half the 

 height to which the marine Miocene strata of San Domingo have 

 been uplifted since their deposition. 



Similar inferences have lately been drawn by Dr. Duncan,f from 

 the corals of San Domingo, Antigua, Jamaica, Barbadoes, and other 

 West Indian islands. They are allied in a most unequivocal manner 

 to the corals of the Faluns of Vienna, Bordeaux, Dax, Saucats, 

 and Turin, while at the same time the forms are those of the Pacific 

 and not of the Caribbean Sea and Atlantic. Dr. Duncan concludes, 

 therefore, not only that there was no isthmus of Panama, but also 

 that there was no great barrier of land or Atlantic continent sepa- 

 rating the Miocene seas of Europe from the contemporaneous seas of 

 the West Indies. The bearing of these views is the more direct on 

 the theory of an Atlantis before discussed, because the affinities of 

 the marine shells and the corals belong precisely to that period (the 

 Upper Miocene), when the flora of Europe was most American. 

 There may have been, as Dr. Duncan supposes, numerous islands in 

 the Atlantic, large and small, as there are now in parts of the Pacific 

 and Indian Oceans, where corals abound, but there could not have 

 been that continuity of land which is represented in Heer's ideal map 

 of the Atlantic already cited, p. 272, which would be indispensable 

 in order to produce an affinity in so many genera and even species of 

 plants as is observed between the Recent American and the Swiss 

 Miocene flora. 



It is right, however, before concluding this subject, that I should 

 warn the reader that much of the reasoning employed by those who 

 have taken part in discussing the probable existence of a Miocene 

 Atlantis, whether as advocates or opponents of the hypothesis, has 

 proceeded on the assumption that the geographical distribution of 

 genera has been governed by laws strictly analogous to those which 

 govern the distribution of species. When Professor Heer speaks of 

 plants called by him homologous, and shows that about half of these 



* Quart. Geol. Journ., 1850, vol. iv. p. 43. f Ibid. vol. xix. p. 455. 



