454 A CONTRIBUTION TO THE GEOLOGY OF VITI LEVU, FIJI, 



koroiluva just above its junction with the Navua. The main 

 Navua was in flood, so that I could not examine its gravels. 



The lower portion of the gorge of the Navua differs essentially 

 from that of the Rewa in being a perfect canon, with perpen- 

 dicular walls rising straight out of the water. The reason why 

 it has failed to perform any lateral corrasion, as the Rewa has 

 done, is that the materials it has had to excavate have been much 

 harder. The rocks of the Navua Valley are enormously massive 

 andesitic flows, with some interbedded tuffs, while those of the 

 Rewa are soft soapstones. Until it leaves the tableland and 

 begins to open out at the head of the delta, there are practically 

 no enlargements of its valley; while the Rewa and its tributaries 

 are bordered by rich river-flats and flood-|>lains which sometimes 

 show two, perhaps three, distinct terraces. 



One very striking feature of the caiion of the jSTavua is the 

 existence of numerous "hanging valleys" in its walls. Even the 

 large tributaries, such as the Wainikoroiluva, enter the main 

 stream by a series of rapids, that is, not absolutely at grade; 

 while the minor affluents often leap from the apparently unbroken 

 bank of verdure of the canon walls at heights of a couple of 

 hundred feet above the river, making the gorge one of the finest 

 pieces of scenery it has ever been my good fortune to see. 



These hanging valleys are, of course, due to the youthfulness 

 of stream-development. Neither trunk nor tributary has yet 

 reached base-level, and the differential erosion of the powerful 

 stream has sunk its bed below the level of its weaker tributaries. 

 Mr. Holmes informs me that similar structures, on an even 

 grander scale, are to be seen higher up the Navua. 



Summary of Section vi.-— Additional evidence is 

 adduced in favour of faulting having taken place along the 

 valleys of the Navua and Wainikoroiluva. A comparison is 

 drawn between the structures of the Rewa and Navua Systems. 

 Especially it is pointed out that both systems have linear eastern 

 boundaries, and draw their waters from extensive basins to the 

 west. This strongly suggests block-faulting. Additional granitic 

 areas are shown to exist west of those formerly observed. The 



