280 H. B. GUPPY, M.B., 
for many years rabbits found a home in Horsburgh Island 
and in North Keeling Island, and played their part in the 
distribution of seeds. They have now been exterminated in 
the latter island by the Cocoa-nut Crab; and in Horsburgh 
Island the crabs have nearly succeeded in destroying them 
all. 
I now come to consider the means by which this group of 
the Cocos or Keeling Islands received its original flora. My 
observations in this locality, as well as my previous experi- 
ments in the Solomon Islands, throw much light on this 
subject. First, with regard to the common littoral trees, 
Scevola Kenigit and Tournefortia argentea, which, as in the 
Pacific Islands, line the beaches on the weather coasts. It 
is often difficult to ascertain whether the seeds of these trees 
have been brought from distant regions by the currents or 
whether they have been derived from the trees already grow- 
ing on the coasts of these islands. However, there is no 
doubt that the fruits of Scewvola Kenigii will float in sea-water 
during a considerable period. Some ripe fruits, that I picked 
off a tree, continued to float buoyantly after they had been 
kept for 50 days in sea-water, losing during the early days 
of their immersion their white fleshy covering. . Notwith- 
standing this long immersion, nearly all the seeds readil 
germinated after having been sown out by Dr. Treub at 
Buitenzorg. (The series of experiments commenced by me 
in the Keeling Islands and completed by Dr. Treub at 
Buitenzorg are given further on in this paper.) Although 
doubtless usually transported directly by the currents, the 
seeds of Scawvola Kenigii are probably at times carried about 
in the crevices of floating pumice and logs. 
By a similar experiment I found that the freshly-picked 
fruits of Tournefortia argentea remained floating after 40 days 
in salt-water. During the first fortnight these small round 
hard seed-vessels split into halves and lost their outer dark 
skin, this being the condition in which they are found in the 
vegetable drift. Seven of these hemispherical seeds were 
subsequently sown out at Buitenzorg, and all readily germi- 
nated. But there is another means of transport for the 
seeds of this wide-spread tree. They occur in considerable 
numbers on the sand in the vicinity of the trees, and in this 
way often get into the crevices of stranded logs and pumice. 
I found an old log beached just above the usual high-tide 
level on the weather coast of one of the eastern islands. It 
was thoroughly honeycombed by the Teredo, long since 
dead, the empty cavities being largely filled with sand, fine 
