Yol. 66.] 
VOLCANO OF MATAVANTJ IN SAY All. 
637 
level of the lava vary frequently; and the places where the lava 
appeared to flow out a few days previously can then be seen to 
be solid, and the caves into which it seemed to flow to have solid 
ends. The motion must be due at any rate to convection currents 
like those in a boiling pot or kettle, but it is still a question whether 
this is all. 
The movement in Matavanu looks similar but much more active, 
the waves of liquid basalt are much larger, the splashes remain 
longer visible when they strike the walls, and the fountains are 
much larger and generally more active. The molten lava also can 
be seen to pour like a cataract into the tunnel, and visibly falls 
into the sea at the other end. Matavanu is therefore certainly a 
river. Kilauea may be a river, hut is more probably a boiling pot. 
Does the lava of Kilauea at the times when the lake empties 
itself periodically, say once in five years, also discharge itself into 
the sea? This has been often supposed, but absolute proof has been 
wanting. The analogy of Matavanu now supports this hypothesis ; 
and the appearances which I saw, along with Governor Drear of 
Hawaii, in some enormous cracks and subsidences between Kilauea 
and the sea give support to the idea. 1 
In conclusion, I wish to tender my thanks to Dr. Solf, Governor 
of German Samoa, for countenance and introductions; to Vice- 
Admiral von Coerper of the Imperial German Navy, and to the 
Captain and Officers of S.M.S. Leipzig , especially Capt.-Lieut. von 
Luck and Ober-Lieut. von Sastrow for hospitable conveyance on 
their warship; and above all to Mr. Kichard Williams, Amtmann 
of Savaii, wdiom I have so often quoted above, for much hospitality 
and practical assistance. Thanks are also due to Messrs. Ludwig 
Schroeder of Safune and Herbert Edwin Eae of Apia and to 
Serjeant Suisala (himself a native chief) for practical assistance of 
various kinds. 
EXPLANATION OF PLATES XLY-LII. 
Plate XLY. 
Fig. 1. Matavanu from the sea.—A background of old volcanic mountains; 
then the new cone of Matavanu, surmounted and partly concealed by 
a cumulus and a lower stratum of condensed vapour. In front of 
this are seen the lava-fields covered with a little fleecy vapour, and 
themselves partly surrounding two old cones. Still nearer is the 
tropical jungle, and then the sea. 
2. Matavanu: the cone and upper lava-field seen from the 
south.—In the distance is the cone. The lip of the crater extends 
from a point near the right of the plate to nearly the same distance 
from the left edge. The field in front is formed of slaggy lava 
(pahoelioe), and appears to have 6tood originally at a higher level, up 
to the ridge at the foot of the cone. After the formation of a crust, 
the lava beneath it has found a vent and the surface has subsided 
considerably, especially to the right and centre of the foreground. 
1 A discussion of these subsidences would, however, be premature until all 
the photographs are developed and can be examined. 
