AFRO-INDIAN REGIONS. 



235 



Urtica ; bis, (compare No. 1 Clermont Tonnerre to the high coral-island of Metia), but 



not seen on Taheiti. Low and herbaceous. 

 Pandanus ; bis (No. 1 Clermont Tonnerre coral-island to Taheiti). Abundant. 

 Lepturus ; bis, (compare No. 1 Clermont Tonnerre to Taiara coral-island, and perhaps 



Taheiti). 



Polypodium ; bis (No. 1) Disappointment coral-islands (to Metia, and pei'haps Taheiti). 



c. Rose Coral-island. 



Continuing Westward, the Vincennes after six days sail, arrived in 

 sight of what was at first mistaken for a rock in mid-ocean. Captain 

 Vanderford, in passing by "nearly twenty years" previously, during 

 '•a strong breeze while the surf was beating heavily," had in like 

 manner been deceived. 



14. Rose coral-island. As the Vincennes drew near, the supposed 

 rock proved to be a dense clump of trees upon a small coral-island ; 

 "in S. Lat. 14° 18' and W. Long. 168° 05', and about eighty miles" 

 from the nearest rocky island of the Samoan Group. Rose coral- 

 island, though hardly " two miles" in diameter, was found to contain 

 a lagoon ; having a large outlet deep enough to admit a ship, and " six 

 fathoms" inside, except at the centre of the lagoon, where a column 

 of coral-rock rises to within " three feet" of the surface. The tide rising 

 "about five feet," most of the island is submerged at highwater; but 

 the outline continues distinguishable by the line of surf, with here 

 and there a projecting massive coral-block ; at all times, we found 

 much water beating over from the windward, and running in streams 

 into the lagoon, and afterwards in a strong current through the outlet. 



For the first time on a coral-island, the mineral kingdom was re- 

 presented ; several blocks of vesicular lava being met with by our 

 party; in all instances resting upon the coral-shelf, not imbedded. 

 Two or three of these blocks were seen by myself, the largest weighing 

 perhaps twenty pounds. From the mineral composition, they had 

 evidently been derived from some volcanic island ; and there seemed 

 no means of transportation, unless entangled in driftwood. This 

 actually takes place at the Tarawan coral-islands ; where Mr. Hale 

 found a native name for " basaltic stones in the roots of trees drifted" 

 to those shores. 



Two banks on the rim were never overflowed by the tide ; one of 

 them devoid of vegetable growth ; the other, containing the above- 



