UPOLU. 331 
denuding agencies. Back of Sangana, the unbroken declivities as 
well as the scoria in the region evince a comparatively recent action of 
the Sangana crater. Farther west, near Fasetodtai, the high unbro- 
ken cone of Tafua carries us on to a still more recent period; and 
here, as we believe, the fires of Upolu finally disappeared. This is 
the westernmost of the large craters. 
We have met with few facts that indicate, even approximately, 
the period when the volcanic action ceased. I have observed no in- 
stances of lava overlying coral, or covering deposits of coral sand such 
as now form the beaches. Along the shores westward of Apia, the 
slopes of the island pass beneath the surface of the sea, and continue 
uninterrupted for five or six miles, inclining even more gradually 
under water than above; for at this distance the depth is but seventy 
or eighty yards. From this depth, it drops off ata steep angle: within 
four hundred yards of a thirty-five fathom cast, our lead descended to 
two hundred and twenty-eight fathoms, and struck on a bottom of 
black sand ; four hundred yards farther out, we found no bottom with 
four hundred fathoms of line. The coral reef of these shores 1s a mile 
and a half wide to the line of breakers. The continuity of these 
slopes, and their length, seem to afford satisfactory evidence, that they 
belong to one and the same process of formation. 
There is evidence, however, that the coral was growing on some 
parts of the island before the fires ceased. ‘This is abundantly shown 
in the tufa craters of the islets east of Upolu, and at Tapanga Point. 
But the dilapidated condition of these craters proves that they were 
long exposed to the action of the sea, before the reefs were completed 
that now protect two of them from farther degradation. ‘The removal 
of the tufa deposits of ‘Tapanga Point, is additional evidence that the 
reefs now half a mile wide at this place, were but just begun, and in- 
sufficient to protect it from an encroaching sea, when the hill was 
formed. The point is cut through by a channel twelve feet wide 
nearly to the water level; and the amount of tufa removed just south 
of the point, was fully equal to the present extent of the point, and 
probably much larger. 
In view of these facts we conclude, that although the period of 
latest activity was subsequent to the introduction of coral to the 
shores of the island, yet it was before the reefs had become much ex- 
tended. Possibly the coral grew only in detached spots, as is now the 
case around Hawaii. 
The period of earlest eruption is still more uncertain. ‘The cha- 
