78 Aiinnal Address. [Feb. 



aaid this, in its turn, in a corrupted form, has given its name to their 

 dialect of Ahirvali.*^ I may give another instance. The last census 

 gives only 4,500 Koches in Bengal. The Koches are a strong Tibeto- 

 Buiman race, which certainly once occupied a large portion of Bengal 

 proper. Now the linguistic survey lias discovered 217,500 more of 

 these Koches who live in the North-Central Districts of Rajshahi, 

 Purnea and Malda. This illustrates how important it is to go on with 

 the survey, and not to stop it in its present half-finished condition. 



I have already remarked that we may fully expect the results of 

 the Linguistic Survey to lead to great advances of our knowledge of the 

 history, the inter-relation, and distribution of the languages of Northern 

 India. In order to show what we may expect in this direction, I cannot 

 do better than communicate to you the substance of a note which Dr. 

 Grierson has been good enough to place in my hands. 



The extensive studies which I made of the North Indian vernaculars, 

 when I was preparing my Comparative Grammar of the Gaudian 

 Languages, had led me to the conclusion set out in the Introduction to 

 that Grammar, that there must have been two consecutive Aryan inva- 

 sions of India, and that the second set of invaders entered the domains 

 of the first " like a wedge." Dr. Grierson informs me that all his 

 studies, subsequent to that publication, have confirmed, in a most strik- 

 ing way, my theory, which even then was not an altogether new 

 suggestion. He is of opinion tliat it will ultimately be shown that there 

 are much plainer signs of this double invasion in ancient Indian Litera- 

 ture, than has hitherto been supposed. Thus he believes it can be 

 shown that the war between Vi9vamitra and Va^istha was a war between 

 these two tribes, in which Vacistha represents the first comers, and 

 Vicvamitra represents their new-come rivals, who had settled on the 

 Sarasvati, and had alreadv driven the older tribe, partly to the East to 

 beyond the Gandak and into Magadha, partly South into the Pancala 

 country, and partly West to the banks of the Indus, where Sudas, 

 Va9istha's master, lived. He further believes that the Kuru-Pancala 

 war of the Mahabharata was in its essence a struggle between these two 

 tribes, the Kurus representing the new-comers and the Paiicalas the 

 old ones ; and that if this theory is borne in mind in reading the 

 V"a9istha-Vicvamitra hymns of the Rg-veda and the Mahabharata, and 

 if a proper study is made of the geography of the period and of the 

 tribes mentioned and the sides they took, it will receive remarkable 

 confirmation. 



41 Mr. V. A. SrnitVi, iu the Journal of the Royal Asiatic Society for 1897, p. 891, 

 following Sir A. Cunningham, places the Ahira farther south, between Jhansi and 

 Bhilsa in the Gwaliyar State. 



