324 R. D. Oldtiam — On jjrohahle Changes in the Geography [No. 4, 



and wandered over higher ground. In the text there is nothing to shew 

 that General Cunningham appreciated this difficulty, but in the map a 

 compromise seems to be attempted which, the usual fate of such attempts, 

 can hardly be called satisfactory. I say seems, for in this map — as in all 

 the maps illustrating the work, but more conspicuously in this — an 

 attempt has been made to represent, without any distinguishing mark, 

 both the present and the ancient courses of the rivers. On the map in 

 question (No. IX), the " Narra B. (ancient course of the Indus) " leaves 

 the existing course of the Indus about thirty miles north of Aror and 

 flows nearly due south to Jakrao, whence a course is marked running S. W". 

 by Brahmanabad to Patala. From Jakrao, another course diverges to the 

 S. E., and, after reaching the latitude of Amarkot, turns S. S. W. and 

 flows into the Han — or perhaps into a lake, for it is by no means clear 

 whether General Cunningham supposed the Ran to have existed in 

 Alexander's times — shortly after joining a branch of the Indus which 

 flows S. S. E. from Patala, but whether this eastern line is supposed to 

 mark an ancient course of the Indus or to represent the dry bed of the 

 Narra is not clearly shewn, but either supposition would be equally im- 

 possible. The accounts of the Arab historians and geographers shew 

 that from the 8th century the Indus flowed past Mansura, until, in the 

 13th century, it abandoned this course for one further to the west, which 

 it has since maintained, and the supposition that the Eastern Narra 

 marks the ancient course of the Indus lands us on one of the horns of 

 a dilemma, for, if the Indus flowed down the Narra as far as Jakrao, 

 and the present continuation was then in existoDce, it is inconceiv- 

 able that the river should have left this lowland to wander up hill, 

 through the higher land to the west ; nor, if this line is meant to re- 

 present the present channel of the Eastern Narra, which did not exist 

 in Alexander's time, is it possible satisfactorily to exjDlain the excavation 

 of this channel. I have not written the above in any spirit of captious 

 criticism, but merely to shew the difficulty that attaches to the elucida- 

 tion of the ancient geography of Sind if we accept the prevalent idea, 

 inconsistent as it is with the known principles of physical geography, 

 that the Eastern Narra represents an ancient course of the Indus. 



§ 2. The Indus in its course through Sind flows between banks that 

 are raised above the general level of the country, which slopes away on 

 either side. This is a feature common to all rivers which are raising 

 the level of their alluvial plains by the deposit of silt, but, at Bukkur, 

 the Indus exhibits a feature which is exceedingly rare, if not without a 

 parallel, in the case of any other river, for here it flows at the higher 

 level through a gap in a low range of hills surrounded on either side by 



