282 THE TRANSFORMATIONS OF THE ANIMAL WORLD 



tion of a group, therefore, presents itself, whenever 

 we are able to reconstitute it with precision, in the 

 form of a broken line ; the different segments are 

 derived from sometimes very distant geographical 

 centres, and can often only be brought together 

 by the progress of geological exploration in regions 

 still imperfectly examined. It may be said that 

 the majority of attempts at phylogeny or concatena- 

 tions sketched out by palaeontologists fail especially 

 because their authors have Jiearly always sought, 

 on the spot and in the very soil of the country 

 in which they are, the different links of evolution 

 of the same group. To reconstruct a real palseonto- 

 logical history of a branch of fossil animals, one 

 must expect to have to change countries several 

 times over. 



The migrations of marine or terrestrial animals 

 are necessarily closely dependent on geographical 

 changes, such as the subsidences which open up 

 new communications between two seas hitherto 

 distant, or, on the contrary, retreats of the sea 

 which permit connections between continents origin- 

 ally separate. Thus, to mention an example taken 

 from fairly recent geological events, it was the sub- 

 sidences at the end of the Pliocene which created 

 the ^Egean and the Sea of Marmora, and opened 

 the Bosphorus and thereby enabled the existing 

 Mediterranean fauna to take possession of the 

 Black Sea region, up till then occupied by land- 

 locked seas with a very special brackish water 

 fauna of their own. 



Thus also, the closing of the Isthmus of Panama 

 at the Pliocene epoch established the very recent 



