444 Mr Gardiner, The Coral Reefs of [Mar. 7, 



common in Fiji, and many show clearly the sandstone formation 

 on their outer sides ; I have no doubt that many have been 

 formed by layer after layer of this rock being added to their outer 

 faces. It then a few feet back breaks down into sand, the rain 

 water perhaps, having plenty of carbonic acid gas in solution from 

 decaying organic matter, readily dissolving the finely divided 

 carbonate of lime by the deposition of which the sand was in the 

 first place consolidated. In the limestone islands in the Lau 

 group of Fiji the sandstone is much less broken down, some 

 recently dug wells in the flat showing the strata clearly; here, 

 however, the mud, caused by the wash of the sea on the limestone 

 cliffs, assists in binding the sand more closely, giving rise to a 

 much finer grained rock. 



Mr Jukes, in his narrative of H. M. S. Flys survey of the Great 

 Barrier Reef of Australia, has especially referred to this beach 

 sandstone formation. He in particular noted one island in the 

 Capricorn Group, where it was being washed away, probably 

 owing to a change of current from some reef growing up. After 

 dismissing the idea of elevation as a "difficult and gratuitous 

 hypothesis," he suggests that " the present structure must have 

 been produced in the interior of a mass of loose sand by the 

 infiltration of sea or rain water, or some other cause of which Ave 

 are ignorant." " After the interior of such a mass of sand had 

 been consolidated, the loose exterior may have been washed away 

 and the solid rock exposed." In fact Jukes considered that this 

 rock is formed in the interior of a mass of sand, the formation 

 gradually proceeding outwards, while it seems to me patent that 

 it is rather formed on the outside and broken down within. 



At Oinafa the sand flat has a height over most of its surface of 

 a few inches above high tide level, but 100 yards from the coast 

 opposite the anchorage there is a brackish swamp 1 — 3 feet below 

 this level about 800 yards long by 300 yards in greatest breadth 

 with in places pools 6 feet or more deep. On it more or less 

 buried in the sand are small blocks of coral strewn about, and a 

 crowbar cannot be forced down far without meeting the coral-rock, 

 so that the flat here appears to have been formed on the inner 

 edge of the reef. 



From a consideration of the foregoing facts and the evidence, 

 collected at Rotuma, it appears to me probable that the reef has 

 been formed entirely at the present level. A gradual outgrowth 

 from the shore, due principally to nullipore growth, took place. 

 When it had reached a certain breadth, masses began to be added 

 to the rim, as at Funafuti, causing it to extend more rapidly. 

 A boat channel was then formed, partially by solution and partially 

 by the scour of the tide. As the whole broadened, this deepened 

 and sandstone flats were formed on its inner parts. 



