ON THE DISPERSAL OF PLANTS. 291 
from the south-east, although this is the prevailing direction 
of the trade-wind and therefore of the surface ‘current. 
For days together, the winds during the height of the trade 
season at the Cocos Islands may blow freshly from east- 
north-east to east, and in this manner by the deflection of 
the prevailing surface current coming from the south-east 
many seeds and fruits are stranded on the Cocos Islands, 
which otherwise would have drifted to the northward of the 
group. This circumstance explains how it is that vegetable 
drift together with floating pumice usually reaches these 
islands in large quantities at a time, and not as a regular and 
continuous supply. Again, the residents tell me that the 
vegetable drift arrives in greatest abundance in the months 
of December and January, when the trade-wind is less steady 
and is interrupted by the variable westerly, northerly, and 
north-easterly winds. It is just at this time that drifting 
fruits and seeds from the Indian Archipelago would be 
deflected towards the Cocos Islands, Thus it will be seen 
that our acquaintance with the winds and currents of this 
part of the Indian Ocean does not render it necessary to 
suppose that the vegetable drift from the islands of the 
Indian Archipelago has first to be transported to the vicinity 
of Western Australia before it can arrive at the Cocos or 
Keeling Islands. 
There are, in fact, many proofs that the passage is accom- 
plished direct from the Indian Archipelago. Many of the 
fruits and seeds cast up on the Cocos Islands are in a very 
fresh condition, and show no signs of having been more than 
a few weeks in the water. Others, again, however, have an 
ancient sodden appearance, and may have been floating for 
a much longer period. Yet it is a most improbable circum- 
stance that three or four living snakes which during the last 
few years have been drifted on bamboos and logs to the 
eastward coasts of these islands, could have performed the 
passage thither from the Indian Archipelago by the circuitous 
route of Western Australia, a circumstance which would 
imply a sea-voyage of some 2,000 miles in length and of 
several months’ duration. Undoubtedly they came either 
direct from the nearest coasts of the Indian Archipelago, 
700 miles away, as is most likely in the majority of cases, or 
direct from Western Australia, 1,200 miles distant, or from 
the islands intervening between these two localities. Again, 
it is highly improbable that a crocodile, that arrived at these 
islands on a large log a quarter of a century ago, could have 
performed a sea-passage of 2,000 miles. No doubt, it 
