16 PROFESSOR W. J. SOLLAS, 



The lower part of the pinnacles is in many cases obviously composed of 

 fragmentary corals, forming a breccia, continuous with that of the ocean platform ; 

 the upper part has at first sight much more the appearance of coral that has 

 grown in place, though it is hard to say how much of its branching irregular form is 

 merely the result of sub-aerial weathering. As a good deal depends on this point 

 I naturally paid considerable attention to it, and examined very carefully one 

 pinnacle in jjarticular. It presented many branching forms of coral rising upwards 

 in the direction of growth, and some turbinate forms still standing on a naiTow 

 base, though a few were evidently overturned. I was anxious to obtain inde- 

 pendent testimony in this matter, and consequently invited my companions 

 Messrs. Gardiner and Hedley to examine the pinnacle along with me ; but on 

 doing so they arrived at contradictory conclusions, one holding that the corals 

 had grown in place, and the other, that they had been washed into position by 

 the waves. Subsequently I examined many similar pinnacles, and though frequently 

 on the point of deciding in favovu- of their having grown where they are found, I 

 was never able to come to a quite definite conclusion. It is very easy to be deceived 

 in the matter ; on one occasion when I thought I had found an indubitable instance of 

 a coral still retaining its original position of growth, I discovered on breaking it 

 open with a hammer a Tridacna shell enclosed within, which was obviously 

 inverted. 



The fact, however, that their basal part consists of breccia, and not of once growing 

 reef has some significance, for this must have been formed when the ocean was 

 washing fragments across the reef into the lagoon ; and if the iq^per part was 

 formed by coral growth in place this would involve a subsequent subsidence, since 

 followed by elevation. 



The question raised by the existence of these pinnacles is rendered all the more 

 difficult by our almost total ignorance of the conditions under which the consolidation 

 of loose superficial debris is effected. Evidence of disintegration and erosion met us 

 on every hand, and wherever the solid platform was exposed to the fury of tlie 

 breakers it was in process of being broken up and driven inland ; indeed much of the 

 material of the outer ridge has been contributed by the destruction of the hard 

 platform ; but nowhere could I discover any evidence of consolidation in progress. 

 Where nullipores are in active growth they serve to cement together whatever 

 material is associated with them, but much of the hard rock of the superficial strata 

 consists of calcareous sand and rounded pebbles cemented together by carbonate of 

 lime, deposited from solution without the intervention of organic agency. In making 

 preparations for our second boring, we dug several trenches, as much as 10 to 12 feet 

 deep, in the loose material of Funafuti islet without discovering any hard deposit till 

 we sank a pit near the seaward margin, when we encountered a loosely cemented bed of 

 foraminiferal sand at a depth of 1 1 feet. This locality, however, may have been an 

 exceptional one ; it occurs where the land is so low that the natives use it as a portage 



