BY W. G. WOOLNOUGil. 449 



considerable south-easterly extension as the Wainivobo brings 

 down abundant large granite boulders, and, as already stated, 

 the Waibowa, on the eastern side of tlie range, has granitic 

 gravels. 



The range which bounds it to the north, and of which the 

 Korobalavu is the chief point, is considerably higher than that 

 to the south. The gravels of the Waisomo prove that the 

 granites extend in this direction too. 



Rising from the south-western corner of the plain is the awe- 

 inspiring mass of Korobasabasaga, with its five towering summits 

 (Plate xiii., fig.l ). A more magnificent piece of scenery is beyond 

 the limits of imagination. It rises abruptly, and in places, sheer 

 to a height of nearly 4,000 feet above sea-level, and therefore a 

 good deal over 3,000 feet alx)ve the plain. In places there are 

 precipices of quite 3,000 feet. The mass is cleft through its 

 middle, apparently from crest to base, by a stupendous fissure, 

 whose origin I do not at present know. The mountain or rather 

 range has a general 8. W. and N.E. extension, and is composed, for 

 the most part, of very massive agglomerates of a very handsome 

 hornblende andesite described in my foraier paper (p. 529). An 

 exactly similar rock is exposed in the left bank of the Waini- 

 koroiluva below Laselase. If the scarp forming this bank is, as 

 suggested, a fault-scarp, the hornblende andesite must be older 

 than the faulting, and therefore probably older than the hyper- 

 sthene andesites of the Medrausucu Range. This is rather 

 important, as the evidence with regard to sequence of eruptions 

 is so meagre. 



In the foregoing sections I have suggested that the plain to 

 the west of the Medrausucu Range represents a " horst " of 

 granitic rock, bounded east and west by two great faults trending 

 S.S.E. and W.N.W. The existence of these faults is, I think, 

 fairly proven by geological and topographical evidence alike, and 

 their very recent date is quite probable. It is also suggested 

 that the ranges of hills bounding the Waidina and Waimanu 

 valleys may be directly due to faulting; on this latter point, 

 however, I do not wish to lay any stress. It is quite likely that 



