Chap, xii.] COSMOPOLITAN PLANTS. 25 1 



of coral-reef rock, without, as far as is known, any blown-sand 

 formation. The sand on the beaches is scanty. The presence 

 of blown sand-rock on coral islands must depend on the freedom 

 of some part of the coast from breakwaters of coral, in order 

 that a heavy surf may form sand in abundance. In Bermuda 

 the sand is derived from the unsheltered side of the island. 



In some rock, about 30 feet above sea level, I saw, as Dana 

 describes, some Brain Corals imbedded in the position in which 

 they had grown. About the reefs are to be seen curious cylin- 

 drical blocks of coral standing on end, and often hollowed out 

 at the top. These arise from the growing of a mass of or- 

 dinarily rounded coral until the top reaches the surface of the 

 water or an insufficient depth to allow of further growth. The 

 top of the mass then dies, whilst growth goes on at the sides, 

 and the dead core is hollowed out by decay. 



The surface of the rock in Tonga is covered with a reddish 

 soil, like that of Bermuda. It is so hidden with soil and vege- 

 tation that it is very difficult to observe the rock structure. 

 The wells, round holes sunk to a depth of four or five feet close 

 to the shore, show a mere continuation of the reef-structure of 

 the shore covered by about a foot of soil. 



I was interested to recognise amongst the littoral plants of 

 Tonga, many forms which I had gathered on the shores of the 

 far-distant Bermuda. They were cosmopolitan tropical plants, 

 and became familiar objects on nearly all the tropical shores 

 visited subsequently. One plant grows in Tonga which is 

 almost identical with one occurring in Kerguelen's Land, but 

 it again is a cosmopolitan, a water weed, Nitella flexilis. To 

 remind one of Australia, there are Casuarina trees in Tonga, 

 but they are nowhere abundant. 



In every direction in Tonga are large tracts of land which 

 have been under cultivation, but are now overrun with a wild 

 growth, affording plain evidence of the reduction of the popu- 

 lation. These tracts are overrun with a dense low tangle of 

 several species of convolvulus and a trailing bean. The posi- 

 tion of the more recent clearings is marked in the distance by 

 the projection from the main mass of dark foliage of the dead 

 branches of trees that have had their bark ringed. These, with 

 a species of Acacia (?), which at the time of our visit in winter 

 had a yellow tint upon its foliage, formed a marked feature in a 

 general view of the vegetation from a distance. 



There are naturally no indigenous mammals in Tonga except 



bats. A large Fruit-bat, probably Pteropits keraiidreuii, which 



occurs in Fiji and Samoa and also in the Caroline Islands,* is 



* "Journal des Museum Godefifroy, Heft IP 1873." "Die Carolinen 



Insel Yap oder Guap." 



