Cerf Islets. 235 



almost barren sand-cay, crcscentic in outline, about two hundred 

 yards in greatest length, and thirty yards in width. Near the 

 eastern extremity were two rude fishing-huts which seemed to 

 have been recently inhabited. They contained a turtle-spear and 

 some other fishing appliances, a hatchet, a bag of salt, a tinder- 

 bo.\-, and some other small bags which were closed up, and which 

 a delicate regard for the sacred rights of private property deterred 

 us from examining. A few pearl-shells of the species peculiar 

 to these islands lay in a heap near one of the huts. I appropriated, 

 without any scruple, some specimens of these, leaving, however, in 

 exchange, a big lump of tobacco, which I deposited in one of the 

 bags hanging from the rafters of the hut. 



Close to the concave margin of the islet was a small turtle- 

 pond, composed of stakes driven vertically into the soft sand, and 

 lashed together so as to form a circular enclosure through which 

 the shallow water flowed freely at all times of the tide. It con- 

 tained six large turtle. 



The only plants growing on the islet were a very young cocoa- 

 nut, scarcely si.x inches high, and a weed, without flowers, some- 

 what resembling a Alcscinbryanthetnnm, and evidently growing 

 wild. The latter may, I think, be considered to be the only 

 indigenous plant on the islet. In strolling over the piled-up 

 sand and broken coral, of which the surface of the islet was 

 composed, I came across three fruits of the widely-distributed 

 Barringtonia speciosa, which had evidently drifted on to the beach, 

 and had then been blown up above tide mark. 



We subsequently visited a second islet which lay about a mile 

 to the westward of the above, with which it was connected by a 

 shallow reef, probably laid bare at low tide. This second islet 

 proved to be utterly devoid of vegetation, and showed no signs 

 of having ever been inhabited. Strewn over its surface were 

 great quantities of dead shells, among which I saw examples of the 

 genera Harpa, Dolimn, Bulla, Cyprcea, Littorina, Valuta, Conns, etc. 

 From here we obtained a good view of the third islet, and could 



