362 CORALS AND CORAL ISLANDS. 
it at 1,150 feet—possibly 1,750. These islands, therefore, 
although toward the limits of the subsiding area, were still far 
within it. The valley-bays of the islets of the lagoon are of 
great depth, and afford additional evidence of the subsidence. 
Tahitian Islands.—The Tahitian Islands, along with Sa- 
moa and the Feejees, are near the southern limits of the area 
pointed out. Twenty-five miles to the north of Tahiti, within 
sight from its peaks, lies the coral island Tetuaroa, a register 
of subsidence. Tahiti itself, by its barrier reefs, gives evidence 
of the same kind of change; amounting, however, as we have 
estimated, to a depression of but two hundred and fifty or 
three hundred feet. The northwestern islands of the group 
lie more within the coral area, and correspondingly, they have 
wider reefs and channels, and deep bays, indicating a greater 
amount of subsidence. 
Samoan or Navigator Growp—tThe island of Upolu has 
extensive reefs, which, in many parts are three-fourths of a 
mile wide, but no inner channel. The subsidence is estimated 
on page 158, at one or two hundred feet. The volcanic land 
west of Apia, declines with an unbroken gradual slope of 
one to three degrees beneath the sea. The absence of a low 
cliff is probable evidence of a depression, as has been else- 
where shown. The island of Tutuila has abrupt shores, deep 
bays and little coral. It appears probable, therefore, that it 
has experienced a greater subsidence than Upolu. Yet the 
central part of Upolu has very similar bays on the north, 
which would afford apparently the same evidence; and it is 
quite possible that the facts indicate a sinking which either 
preceded the ejections that now cover the eastern and west- 
ern extremities of Upolu, or accompanied this change of level. 
The large island of Savaii, west of Upolu, has small reefs, 
small because, probably, of volcanic action; for it bears, every- 
