DESCRIPTION OF THE EXPEDITION. 155 
faces, but growing as fast to seaward. Poule still exists as a patch of sand with two 
conspicuous bushes. Our last visit was to Ressource, where we found the greater part 
of its southern spit gone, while two great accumulations of sand to the north-west had 
increased its size considerably. 
In the lagoon we took a series of soundings, finding a sandy bottom. There were a 
few coral-patches, noticeably one in the centre. ‘They were not conspicuous, being 
mostly round masses of Porites or Goniastrea with branching Pocillopora and Stylophora. 
The heads were healthy on the surface, but underneath were rotten with boring 
organisms. Many had fallen owing to their own weight, and were being broken wp into 
little fragments, which would be triturated up into fine mud by the very abundant 
Holothurians and other sand-feeders. This then would be removed by the tides, so that 
the lagoon would be in nowise filled in by their means. 
Off Darros we took two successful dredgings in 35 to 40 fathoms, and then proceeded 
south so as to run a line of soundings due east from the Amirante Bank to the entrance 
of Desroches, which is really a submerged atoll save for the reef and island of the same 
name on its southern rim. The distance between its 100-fathom line and that of the 
Amirante Bank proved to be less than 8 miles. We got soundings of 24, 24, and 23 
fathoms, and then in less than a mile dropped off to 588 fathoms, getting 874 fathoms 
in the middle of the passage and apparently almost as steep a rise against Desroches. 
We entered through the west passage, passing across the lagoon, which is singularly 
free from shoals, and anchored off the island, where we spent two clear days. 
The island of Desroches is about 3} miles long by 500 to 800 yards broad, lying 
almost east and west. It is hollowed out into small bays with outstanding points to the 
south, where the shore is largely formed of beach sandstone and there is a wide reef. 
At the east end there has been some loss, hut it is not so extensive as to the west, where 
the beach ends in a marked cliff on either side of a long point, which continues the 
seaward beach of the island on to the reef. Off it beach sandstone forms a ridge, 
gradually decreasing in height for 300 yards further westwards on the reef-flat. The 
loss at present visible at this end is about 600 yards and may have been much more. 
How far the land ever extended over the existing reef it is difficult to estimate, but the 
appearance of the whole rather suggests that it was once almost coterminous with the 
edge of the broad weed-covered reef-flat to the south, east and west. This reef-flat bends 
round the ends of the island and is growing along the north side, but still leaves an 
open coast for about a mile in the centre, where only patches of coral are found. Here 
the shore is continually growing further and further out, any loss to the south being 
more than counterbalanced by this gain. 
The settlement was situated in the centre of the north side of the island, and consisted 
of about 30 men employed in preparing coconut-oil. Everything was of the most 
primitive condition and in decay, a pleasing island, zoologically speaking. Behind the 
houses is a great clump of Hernandea and Barringtonia, and from them radiate avenues 
of Casuarina towards each end of the island. These were planted, we were told, about 
70 years ago, after which the island was abandoned. Sowing themselves the casuarinas 
soon covered almost the whole island, until a fresh settlement was made about 1880, 
