HOUTMAN ABliOLHOS ISLANDS. 165 



of cemented coral fragments. At no place is the height of these islets of the 

 outer rim as great as that of Rat Island (viz. 8 feet), and, as we have already 

 pointed out, where the land does consist of cemented coral fragments the 

 limestone produced is nothing like so hard or compact as that of Rat Island. 



The eastern beaches of the islands in every case consist mainly of 

 loose coral blocks, and in places one can see how these are piled in such a 

 manner as to form very definite strata by the regular heaping of these blocks 

 on the northern shore of Wooded Island. 



In many places, perhaps in most, the surface of the coral blocks is 

 covered with sand and guano. The surface of Wooded Island appears to 

 be of quite clean coral fragments with a sparse covering of shrubs and 

 Mesemhryantheiiium. I find, however, that the guano workers shifted some 

 tons of guano several years ago from the island, and it is quite probable that 

 they also cleared this material from the general surface. 



There is a great tendency towards the formation of small or large lakes on 

 these islands. This has proceeded to such an extent that Wooded Island has 

 quite a character of its own ; a lagoon of such extent exists that the island 

 might be called a miniature atoll. Moreover, this internal lagoon attains a 

 considerable depth, for we sounded six fathoms in the centre. The water 

 of all these hollows, whether small or large, is salt, and furthermore is in 

 communication with the sea. Some idea of the mode of formation of the 

 comparatively large lagoonlet on Wooded Island can be deduced from 

 the smaller ones on the other rim islands. Thus, on the island imme- 

 diately south of Easter passage is a pit with six feet of water in it 

 at the deepest part. It is 58 feet wide across the widest part and 

 100 feet long, being roughly oval in shape. The sides of this hollow 

 overhang and are obviously being eroded at the water-level. On one side 

 a large slab of the cemented coral formation has broken off owing 

 to this undercutting. There is ample evidence that these hollows are 

 produced by the combined action of percolating sea-water and the weather, 

 and in most cases after the formation of the land. By means of a few simple 

 surveying instruments, we were able to show that the level of the sea-water 

 within the hollow was either the same as that of the sea-water outside the 

 island, or within a foot or so of it. Whilst the tide was rising, the level of 

 the water within the hollow was somewhere about one foot iielow the sea- 

 level. When the tide was falling, the level of the water within was higher 

 than sea-level by about the same amount. The communication with the sea 

 is therefore evident, and also the " lag " due to the resistance experienced by 

 the percolating water. 



Another much larger "lake" existed on the island we have just referred 

 to, but it is not so large as that on Wooded Island. In character it is similar 

 to the one already described, and it is obviously produced by the extension 

 in area of such a small one, or the union of several small ones. 



