TOPOGRAPHICAL FEATURES. 421 
partly in this way that the great curve results. To illustrate :—Samoa 
is in advance of the Rarotonga line, being nearly continuous with the 
course of the Aitutaki line. The Vaitupu line is a little in advance of 
Samoa ; and, moreover, consists of two parallel series, the northeastern 
of which is in advance of the other. ‘The Southern Kingsmills con- 
stitute another line in advance of the last, and the northern are shorter 
lines each in advance of the preceding. The Ralick Islands are 
nearly continuous with the Northern Kingsmills, while the Radack 
Islands are in advance, or to the northeastward of the Ralick; the 
northern of the Radack Group being the most in advance. It is 
plain, therefore, that while part of the curve is due toa change in the 
trend of the groups as we proceed northwestward, a still larger part 
is owing to this advancing of the successive lines, each being, with a 
rare exception, a little to the northeast of the one preceding. Were 
there no curve, the range straightened out would extend from Raro- 
tonga toward the Pelews, south of the Ladrones. Were there only 
the curve depending on the change of trend in the groups (from N. 65° 
W. at Rarotonga, to N. 37° W., in the Radack Group), the Ralick 
Islands would have extended along just east of Hogoleu. ‘The differ- 
ence, therefore, between this and their present position, measures the 
amount of curvature derived from the position of the successive 
subordinate lines in the range. 
In the New Guinea chain, the successive parts are in several 
series, and although there is not a regular “advancing” or “receding” 
order, we observe the system to be so complete, that when there is 
an interval between two groups in the same line, there is, to one side, 
another range to occupy the interval. ‘Thus, opposite the interval 
between the Salomon Islands and the New Hebrides, lies the Santa 
Cruz (or Vanikoro) Group. The line of the Salomon Islands is con- 
tinued in the Admiralty Islands, and then dies out, but south, com- 
mences the large island of New Guinea. ‘There seems thus to be a 
certain relation between the several parts, which cannot be mistaken. 
The curvature in this range rises almost wholly from a change of 
trend in the separate parts. 
In the Ladrones, the curve may be due to an advancing arrange- 
ment of the parts of the group. But whether this be actually so in 
this case, and in the remaining curves alluded to, must be determined 
by future observations. 
The facts adduced are sufficient to illustrate the different ways in 
106 
