53 Till ETHNOLOGY OF TEE INDIAN ARCBIPELAOO. 



confirmed by physical, linguistic and moral evidence, little room is 

 left for hesitation in adopting the conclusions to which they lead. 

 We can only now advert to one of these indications in the briefest 

 manner. Mid nnd south Asia are strongly contrasted in their 

 physical characteristics. In historic times there have been successive 

 movements of Mid- Asian races to the W. and Si npon ihn races 

 that had preceded them there, and whom they have subdued* 

 Ethnic geography tells as that it must always have been so. 

 Middle Asia naturally nurses hardy and rode tribes. Their contests 

 cause migrations. They are pressed to the W. and 9., which are 

 more easy of access than China. The comparative softness of the 

 southern nations from the climate, and the temptation held out by 

 their wealth and refinement, provoke conquests. Independent of 

 all great aggressive movements, family migrations have always 

 taken place from M. to S. Asia ; while the difficulties of the 

 mountain barners and the rigour of the climate, have prevented 

 the stream of migration erer being from the south to the north. 

 The two regions always present^ and always will present this 

 contrast, although the progress of civilisation tends more and more 

 to modify, and may ultimately neutralise, its effects. If we now 

 view the languages of the old world in connection with this law 

 of ethnic geography, for such it seems to hare hitherto been, wo 

 might conclude, with a degree of probability, that, ra eastern S» 

 Asia, the evident connection between the Chinese and the Tartarian 

 races, arises from the former having descended from Mid- Asia, and 

 not from tho latter having ascended from the Chinese basins. 

 Proceeding westward we may, in the same way, conclude that, in 

 the Indo-Malayan basin, the Burmese and the allied languages 



E reserve some evidence of one of the southern movements of the 

 lid- Asian tribes; the Tibeto-Indian languages, of another; ami the 

 Old Indian languages, of a third, from thc 'N. W. (not N.E. as Dr 

 Prichard supposed) which connects itself more decidedly with the 

 existing Mid-Asian races, although the S. Indian are so distinct 

 from the Tartarian languages, that the period of the migration 

 must have been very ancient.* Africa again points to movements 

 long preceding the origin of the Semitic, Iranian, Old Indian and 

 even the present Tartarian races. Its languages throughout are 

 strongly allied to the Tartarian and earliest Indo-European, but 

 they have also far more Semitic tendencies than either the Tartarian 

 or its Indian allies. They therefore recede to a time when the 

 organism of the Mid-Asian languages was not fully developed, 

 but exhibited all tho chief elementary tendencies and characteristics 

 of the Turanian, Iranian, and Semitic. This evidence of the 



* Whim I speak or e»hnie moveme nts I do not mean single migrations, but 

 migrations ocm tinned during long periods, sometimes for tbousandi oi years, until 

 the migration* take a new direction, or till the censes Umt induce them, cease to 

 operate or ere opposed by new forces erWiiff in the rrffion to winch the movement* 

 were directed, or in intermediate ones. 



