666 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1908. 



vading the adjacent Spanish peninsula, as the few species of fossil 

 elephants found therein seem almost without exception to have en- 

 tered from Africa by way of Gibraltar. The gi-eat ranges of moun- 

 tains in the new world may have influenced somewhat the trend of 

 migration, but were crossed by the proboscidians at will. 



Vegetation does constitute a most effective barrier, especially in 

 the case of the tropical jungle of Central America. During the 

 Pliocene, as we shall see, after the land bridge was established, inter- 

 communication between the two Americas was very free. In the 

 Pleistocene, however, this migration of large quadrupeds gradually 

 ceased, so that in spite of the great abundance of mammoths and 

 mastodons in North America none attained a foothold south of the 

 Mexican Plateau. To-day the jungle is absolutely impenetrable for 

 all of the larger mammals except such as may be at least partially 

 arboreal in liabits. 



The migrations were forced, not voluntary, for it would appear 

 that the mighty elevations of Asia beginning in late Miocene times 

 and the conser|uent alternations of moist and arid climates, with a 

 strong tendency toward the latter, has caused these great animals to 

 disperse themselves from the rising high lands of central Asia into 

 the more stable lowlands. In these forced wanderings the land 

 bridge between Asia and Alaska was again and again discovered and 

 crossed by the migrating hordes. 



The first appearance of the proboscidians is in the Middle Eocene 

 beds of the Egyptian Fayum district. There we find in M(vrif]ie- 

 rium, the most primitive type, the forerunner of the race. Of the 

 extent of the geographical range of M cerithermm and of its successor, 

 Pala'omastodon^ we know nothing further than that they have only 

 been found within the Fayum. 



During the Oligocene the proboscidians seemingly remained in 

 Africa, though of this we have no record. Early Miocene deposits 

 of Mogara, which lies northwest of the Faj^um some five days' jour- 

 nej^, about 75 miles, give us the remains of T etrabelodon angustidens, 

 the next known type in the evolutionary series. From Tunis again 

 this species is reported, being what Professor Deperet calls the an- 

 cestral (ascending mutation) race of T. angustidens, pigmceus. This 

 race is also reported from the sands of Orleans and from the Burdi- 

 galienne of Agles (Aglie, Italy). Thus it seems as though T etrabe- 

 lodon angustidens^ the form with the maximum development of sym- 

 physis, were the one to make the exodus from Africa, not as the chil- 

 dren of Israel did, by way of the northeast, but by the land bridge 

 connecting Tunis with Sicily and the latter with Italy, and thence by 

 way of Greece to Europe and Asia. 



