OF THE HAWAIIAN ARCHIPELAGO 273 
the New Zealand Kauri, as stranding on the windward reef of 
Funafuti. Wood-Jones (’05) gives an excellent account of 
tree-trunks and ‘“‘floating islands’’ of storm-washed vegetation 
as carriers of seeds, animals, etc., to the Cocos-Keeling group. 
He emphasizes the importance of trees with buttressed bases as 
disseminators: 
These buttresses are in ine ee of large thin wings, which taper to the trunk 
and below form like stalls in a circular stable. With- 
in a stalls much earth is held ae, by the interlacing of smaller roots, and when 
such a tree is uprooted, and set adrift to sea, it carries its earth withit. It may carry 
it for very great distances, and I have seen a buttressed tree come ashore in the atoll, 
from whose base a wheelbarrow-load of fine red earth might have been collected. 
FLOATING ROCKS AS DISSEMINATORS 
The idea of floating rocks as disseminators of littoral plants 
might be met with incredulity, were it not for the testimony of 
many reliable observers. Among the volcanic islands of the East 
Indies large blocks of pumice float for many weeks, and are carried 
many hundred miles from their points of origin. The salient 
points—prolonged flotation of the blocks; presence of numerous 
kinds of seeds in the crevices and pores of the pumice; and the 
germination of these seeds when the block is cast upon a favorable 
beach-situation—have all been corroborated by careful investi- 
gators. Ernst (’08, p. 56) states that floating blocks of pumice con- 
stitute an important dispersal agency in the Sunda-Straits region. 
Although there is very little pumice to be found on the Ha- 
waiian coasts at the present time, there is abundant evidence that 
in earlier periods in the geologic history of the islands, repeated 
volcanic explosions, resulting in pumice production, have taken 
place. There are today extensive pumice beds aroundsthe vol- 
cano Kilauea. Therefore, although pumice blocks play little or 
no part in the dispersal of plants in the Hawaiian group at present, 
it is entirely possible that they had a more important réle in 
earlier times, at least in distributing seed from island to island. 
Floating masses of dead coral may also be ranked as possible 
seed-carriers. _Wood-Jones (15) found numerous instances of 
this in the Cocos-Keeling group. The innumerable air-cavities 
in certain kinds of coral render it buoyant. The block is cast 
upon a beach at storm time; it lies there for an indefinite period; 
