pec 
i) 
ON THE DISPERSAL OF PLANTS. 293 
Triumfetta procumbens are also dispersed by sea-birds. Not- 
withstanding these facts, it is evident to my mind that in 
comparison with the agency of the westerly drift, birds have 
played a secondary part in stocking these islands with plants. 
The absence of the fruit-pigeon has deprived the flora of the 
Cocos-Keeling Islands of many of the conspicuous features of 
the vegetation of a coral atoll in the Western Pacific or of the 
numerous coral islands of the adjacent waters of the Indian 
Archipelago. The littoral trees so familiar to me on the 
beaches of the Solomon Islands, 4,000 miles to the eastward, 
and occurring with equal frequency in the yet more distant 
islands of the Central Pacific, are in truth nearly all repre- 
sented in the Cocos-Keeling Islands, owing their extended 
distribution from ocean to ocean mainly to the common 
agency of the currents. But here in these remote coral 
islands in the Indian Ocean, I missed the huge banyans and 
other ficoid trees, the tall Kanary, and the yellow-flowered 
Eugenia, that give height and character to the interiors of 
the numberless coral islets of the Western Pacific. Fruit- 
pigeons have not found a home here. Hence the westerly 
drift has held the sway, and these islands have become 
stocked with a monotonous and sombre flora. Rather, I 
should write in the past tense, since a few years hence but 
little of the original vegetation will remain, and some brief 
record, such as these lines record, will be all that is left to 
remind the visitor of the condition in which man first found 
these islands. 
NoTE ON THE LITTORAL PLANTS OF THE SOUTH COAST OF 
WEST JAVA. 
Since there can be no doubt that the Cocos-Keeling Islands 
have largely derived their flora from the neighbouring coasts 
of the Indian Archipelago, it will be of interest to learn what 
are the common littoral plants on the coast most adjacent to 
them, namely the south coast of West Java. Having been 
engaged for nearly six weeks in examining the geological 
structure of the southern sea-border of West Java, between 
Java Head and Cape Mandaran, promontories abont 200 miles 
apart, I am also able with some degree of confidence to 
describe the general character of the littoral vegetation of 
this coast. 
The greater portion of this extent of coast is low and 
' sandy, and for the most part destitute of coral reefs. Here 
the sea-border is formed by a low, sandy belt, usually 200 to 
