u 



has been broken up into a number of cuboidel blocks, and long tabular masses resting on 

 these. The passages between the former are in many places broad and deep, and, where they 

 are covered by the superincumbent rocks, form dusky ca vernous hol lows, which are tenanted 

 by bats. Some of the external passages are like doorways being about 7 feet in height 

 and 6 across. At the side of one of these entrances half of the horizontal surface of one of the 

 supporting rocks is exposed. It is hollowed out into a shallow basin about 6 feet in dia- 

 meter and 6 inches deep, which is filled with vegetable debris and water. One of the hori- 

 zontal tabular masses is about 30 feet long. On its under surface, which is smooth , are 

 two semiglobular hollows. An adjacent mass is about 40 feet in length, and 8 in breadth. 

 On its under surface also I observed a cup about § feet in diameter and 1 in depth at 

 the centre. The hill is here very narrow and slopes steeply on the inner side to a man- 

 grove flat. On this side there are also several rocks. One of considerable size had a smooth 

 rounded surface unmarked by any furrows. 



The rocks along the beach , although with a few exceptions not remarkable for their size 

 and architectural features , are geologically interesting. At the farthest Chinese hut to 

 the eastward are broad flat masses stretching across the beach, and only a few feet in height. 

 Their surfaces are traversed by parallel rectilinear fissures and slight grooves, marking divi- 

 sional planes, and the direction of these, and also of the longest edges of the rocks, is almost 

 due NE. and SW. Proceeding along the beach to the westward, a fine example of con- 

 centric or parallel curved exfoliation occurs. Of what has originally been an extensive mass 

 of rock there only remain a few solid blocks, of cuboidal and rudely spherical forms, which 

 rise from the decomposed and semidecomposed bed worn down nearly to the level of the 

 beach. Embracing the rounded bases of these nuclei, and forming the bed, are the upper 

 edges of parallel curved laminae , which continue till those spreading out from an adjacent 

 nucleus meet them. Sometimes the same laminae are seen, after embracing the end of one 

 block , to bend reversely and embrace another nucleus , so that the surface or horizontal sec- 

 tion exhibits a series of narrow parallel S. shaped bands. The variation in the curves 

 according to the form of the sides of the nuclei is very great and striking. The lower 

 corner of one of the blocks is conical, and the concentric sheaths or caps have the same 

 shape. In the triangular spaces left where three systems of laminae meet, are prismatic 

 masses, solid but of a crumbling structure, and in composition similar to the laminae. 



Near one of the Chinese houses I observed that the face of a rock, freshly split by the Chi- 

 nese, was a regular curve, and on the hill the side of a large rock had a similar curve. 

 I partially re-exmined the rock mentioned ante, p. 7. The W. side is marked by deep 

 grooves, of which the axes are NE. by E. nearly. The planes in which these are formed 

 also determine the direction of the face of the rock overlooking the channel. Many of the 

 grooves on the west side are a succession of deep pear shaped cavities. I think there can be 

 no doubt that these are owing to sheathed nuclei having been gradually excavated. That 

 such nuclei are very abundantly dispersed amongst the rocks is evident. 



A little to the W. of this rock and towards the beach there is a large flattish rock. 

 One side is a curve in which parallel laminae, after retaining their continuity for some 

 thickness, part in the middle and give off two systems perpendicular to the first. If this 



