364 CORALS AND CORAL ISLANDS. 
sion that was derived over fifty years since from the size and 
distribution of the atolls; for the line, A A, lies in the 3,000- 
4,000 area of the central Pacific, and, moreover, passes over 
the deepest point in it yet observed, — the sounding giving 
a depth of 3,448 fathoms. The deep central area of the 
Pacific apparently extends, in the direction of this axial line, 
to the still deeper waters east of Japan. 
The southern boundary line of the coral area, as we have 
laid it down, hes within the area of subsidence, although near 
its limits. This area has been prolonged southeastward in 
some places beyond the boundary line. One of the regions 
of this prolongation lies between the Samoan or Navigator 
Group and the Feejees and Tonga Group; another is east of — 
Samoa, along by the Hervey Group. Each of these exten- 
sions trends parallel with the groups of islands. Tt would 
seem, therefore, that the Society and Samoan Islands were 
regions of less change of level than the deep seas either side 
of them; that, therefore, instead of a uniform subsidence over — 
the subsiding area, shading off toward the borders, there were 
troughs of greater subsidence, whose courses were parallel to 
the ranges of islands; that, im other words, there were in the 
ocean's bottom, a few broad synclinal and anticlinal flexures, | 
having a common direction nearly parallel to the axial line 
of the Pacific. The Marquesas and Fanning Groups lie in a 
common line, and thus may mark the course of a great cen- 
tral anticlinal in the oceanic basin. 
The Hawaiian range has probably experienced its greatest 
subsidence to the northwest, where the islands are all atolls, 
and show some evidences of recent sinking; and this north- 
western extremity of the range is nearer to the axis of the area 
of subsidence, above laid down, than is the southwestern. 
What is the extent of the subsidence indicated by the coral 
