272 



River, from the Hospital and Hartley and Riley's bars or beaches ; these two 

 places, in fact, may be likened to two ripples placed by nature in her sluice 

 box, the Molyneux River, and there are, without doubt, many equally rich 

 places. At the Dunstan township, the road enters the gorge, and continues 

 by the river side to Cromwell ^ at the former place the banks are about 35 feet 

 high, and at the latter 75 feet. The gorge is filled with immense water and 

 ice-worn rocks and boulders, and at places on either bank a terrace formation 

 is observable. 



Our road to the Wakatipu, or Queenstown Basin, lies up the Kawaraii 

 River, which, at Cromwell, joins the River Molyneux in its course from the 

 Wanaka Lake. After crossing the plain, and temporarily losing sight of the 

 river, we again fall in with it on entering the Kawarau Gorge, through which 

 we pass in company with it till we reach the Arrow District, at the lower end 

 of the Queenstown Basin. The Arrow River, which takes its rise under 

 Mount Hyde, flows into the Queenstown Basin, and merges into the Kawarau 

 not far from the Arrow Blufli', and has here cut through the slate rock perpen- 

 dicularly to a depth of about 200 feet. The Kawarau River is also fed in the 

 Wakatipu Basin by the Shotover River, which takes its rise between Centaur 

 Creek and Treble Cone. The Kawarau Gorge is filled with the same evidence 

 of water and ice action as the Dunstan Gorge. It will be remembered, that 

 till we reach the level of the Wakatipu Lake we have been steadily ascending, 

 and in both the Cromwell and this basin, as we pi'ogress, we advance against 

 walls of terraces that continually rise on the horizon to our view, and these on 

 our return journey entirely disappeai", or rather no such appearance is 

 observable, from which circumstance I presume, that the weight of water 

 pressing downwards has swept all before it, and pi-evented the formation of 

 any such accumulations to form this terrace appearance, these terraces being 

 composed entirely of drift. 



In the Dunstan Basin the same thing is observable. Starting from the 

 Manuherikia township towards the upper watershed of the valley at Dunstan 

 Creek, terraces are seen under the Dunstan Ranges, but I believe no corres- 

 ponding formation on the other or lower side, bounded by Blackstone Hill and 

 the Raggedy Ranges. 



I am informed that the features of the Arrow and Shotover Rivers are 

 precisely similar. From QueenstoAvn the road to Skipper's Creek does not 

 continue directly up the gorge through which the Shotover River passes, but 

 traverses a high saddle, and joins that river near its junction with the Moke 

 Creek. On either side of this river are what are styled upper terraces, flat 

 areas of land, having a scarcely perceptible inclination towards the river, so 

 that where the gorge or valley widens, some considerable breadth of such land 

 occurs, so much as neaidy a quarter of a mile in depth at places. I have been 

 as far up this river as Skipper's Creek, which, but for the small quantity of 



