428 PERCY SLADEN TRUST EXPEDITION 



and therefore the flora and fauna were quite neglected. Birds seemed scarce, but I noticed 

 a dnnyris and a rail {Dryolimnus abbotti ?) which seemed to be the Assumption species. 

 No remains of land tortoises were obtained, but 1 shall not be surprised to hear of their 

 being found when the guano is dug *. 



Insects were very numerous and it was a matter of great regret that few could 

 be taken ; butterflies were more common than on any other island and comprised one 

 species, the beautiful Precis rhadama, not noted elsewhere. The bush in places resembles 

 Aldabra Pemphis scrub and in others the open country of Assumption : it was very bare 

 and dead during my visit and hardly a score of plants were obtained in flower, though 

 during the dry season it is supposed to be the most luxuriant of all the Aldabra series. 

 I had intended to revisit Astove to complete my investigation and take photographs on 

 the homeward voyage ; unfortunately during a dead calm the strong current carried the 

 schooner altogether out of her course and it was impossible to put back. 



VII. COSMOLEDO. 



Cosmoledo is a large atoll about 9 miles long by 7 miles broad, situated in 

 lat. 9° 41' 20" S. and long. 47° 32' 12" E. Contrasted with the other two atolls visited 

 it is remarkable in having less than 10 miles of its 24 miles of circumference capped with 

 land. The latter is divided into eight main islands, of which Menai and Wizard Islands 

 are the most important. 



The islands in all cases are fundamentally of coral rock, though sand has been 

 blown on from all sides and in some cases has completely hidden the rock under large 

 sand dunes and ridges. The rock is elevated coral rock, which in most places has been 

 much metamorphosed, and which always contains varying amounts of phosphate. "Platin" 

 rock was not found, and corals in position of growth were only occasionally observed. It 

 was obvious that the islands were once larger and there is evidence in the form of small 

 islets and rocks to show that the land-rim was formerly almost perfect. Erosion is still 

 progressing rapidly in many parts of the atoll, an actual measurement being obtained 

 at a point on the lagoon side of Wizard Island, which showed the loss of 15 yards in 

 fifteen years (1893—1908). 



The nature of the islands visited is as follows : the North-East Islands and Goelette 

 Island are almost entirely rocky t ; the shores to seaward and towards the lagoon are 

 formed by overhanging cliffs as at Aldabra. Extensive phosphate deposits have existed 

 on all, and on West North-East Island is a certain amount of a phosphate rock, which 

 appears to be composed of sand and rubble cemented together by phosphoric acid. It was 

 interesting to find eggs of a giant land tortoise embedded in this rock, thus proving that 

 once these reptiles lived on Cosmoledo. 



Menai Island is rocky, but is also considerably overlaid with sand, which has been 

 blown up from the lagoon. The shore is formed partly by cliffs and partly by sand : 



* A letter from Capt. Parcou just received informs me of the find of large bones cemented into a rocky 

 pinnacle. I have every confidence that these will be found to be those of giant tortoises, 

 t There is a small dune on the lagoon side of Great North-East Island. 



