406 J. STANLEY GARDINER. 



seldom more than 30 yards across. In the lagoon shoals also a similar nan-ow breadth of the 

 reefs round the velu, where no land of any sort exists, was also observed. The solution where 

 the reef is protected, eats up almost to the growing coral layer, and, if any accident befall 

 these animals so that they die, it is easy to perceive how that the water of the velu and 

 of the atoll-lagoon will quickly eat a passage through the reef, and join one another. In 

 this connection the south-west reef of South Nilandu is most noticeable, as it appears to 

 have had a long linear velu, which has now become joined to the lagoon of the atoll, 

 leaving, nevertheless, a series of isolated I'eef masses, parts of the original great breakwater. 



The land of both atolls is nearly all built up of sand, and there are many fresh banks 

 of the same formed on reefs in the lagoons, though it is doubtful whether many — or indeed 

 any — of these will ever form islands, covered with tiinber. The changes in the land are mainly 

 local, depending on the passages, etc. in the immediate vicinity of each separate island. No 

 signs of pinnacles of elevated rock were observed, and it is possible that the elevation, such 

 as it is in the Maldive region, had no part in the formation of the land that exists in either 

 of these atolls^ 



Within the lagoons no changes in depths were found. The bottom in most places is 

 of hard rock or sand, and in South Nilandu its fauna was richer than on any other bank 

 in the group. Indeed six dredgings in this atoll yielded as great a variety of animals as 

 any consecutive eighteen elsewhere, and the number of specimens of some of the species of 

 Crustacea, Echinoderma and Mollusca was extraordinary. Of the passages into the lagoons 

 only those we traversed in going into and out of each atoll were examined. That of Maimbudu 

 was found to have a shoal of 7 fathoms in its centre within a few yards of the 40 fathom 

 line, the steep of the reef commencing from the shallowest water of the channel. In Mawaftiri 

 passage three shoals, almost reaching the surface, were found. 



IX. KoLUMADULU AND Haddumati' (Figs. 106 and 107). 



These atolls resemble one another closely in that they have single and relatively perfect 

 encircling reefs, enclosing large lagoons with 40 — 4.5 fathoms of water. The lagoon shoals 

 of each are comparatively small, although fairly numerous, with the regular, almost perpen- 

 dicular cliffs usually found around them. None have land, unless situated close to large 

 passages into the atoll lagoons, and none are faro. Both have for the greater part of their 

 circumference single rather narrow reefs, without pools of water in the centre. Both are 

 exposed on the east and also on the west to the full force of the sea, and on the other 

 sides are afforded little or no protection by the neighbouring banks. Lastly the land in each 

 is mainly situated to the south and east, in some positions consisting of lines of islands 

 almost joining one another, while to the north and west the reefs are very bare. 



The islands are for the most part formed of sand alone, and, as a general rule, are to 

 seaward bounded by a well-defined sand flat, boulder zone, and reef flat. A few, however, 

 send out heads as far as the boulder zone, and these are generally washing away at their 

 ends. Behind, their necks are usually being cut through so as to separate the rocky outer 



1 That this is really the case is, however, distinctly - My knowledge of these atolls is confined to two hurried 



unlikely. Our time was too short to enable us to stop and visits in the s.s. Ileafaee in going to and from Suvadiva. la 



examine the individual islands, and our attention was mainly addition Mr Forster Cooper spent 10 days iu Kolumadulu 



taken up with the almost universal faro formation of the and 8 in Haddumati, dredging and sounding the deeper 



shoals. waters of their lagoons. 



