418 PACIFIC OCEAN. 
possible irregularities in the opening of fissures from internal forces, 
for us to expect so well-defined lines, in all instances, as the Radack 
_chain and others that have been pointed out; and when the subsi- 
dence has caused a disappearance of all but the coral that is based 
upon it, there is still more reason to expect some difficulty in deter- 
mining the courses of former fissures. 
The Ladrones, as seen on the chart, constitute a line of islands ex- 
tending through a degree of latitude. ‘This line, which is nearly 
straight in its lower half and trends N.N.E., bends slightly westward 
in its northern half. Although it is evident that the bend must take 
place from a gradual veering in the line of ruptures, we cannot distin- 
guish satisfactorily their several relations. From the positions of some 
of the islands transverse to the trend of the group, and conformable to 
the west-northwest system of the Pacific, we might infer that although 
the course of the group is nearly north and south, several of the fis- 
sures were opened transversely to this course. 
The Society and Paumotu Islands indicate, by the trends of the 
several islands and lines of islands, a great number of nearly parallel 
ruptures; but they are so clustered that we do not venture to point 
out definite series. ‘They le in numerous lines, and are so related 
that they exhibit well the general principles we are endeavouring 
to illustrate. ‘The Marquesas consist of two parallel lines, a north- 
eastern and southwestern ; and Nukuhiva and Fatuhiva are two large 
islands having a transverse position. 
In the southwestern Pacific, a correspondence with the system here 
explained is obvious on a glance ata map. ‘The New Hebrides, New 
Caledonia, Salomon Islands, New Ireland and others, exhibit a series 
of parallelisms between groups and parts of groups. 
The Galapagos, as shown by Mr. Darwin, lie in three or four 
parallel lines, trending nearly northwest, while transverse lines are 
also apparent.* ‘The Canaries and Azores in the Atlantic are other 
examples. 
3. Curvatures of Ranges.—In our first chapter, the curvatures of 
some of the main ranges of the Pacific have been briefly pointed out. 
The great central chain, six thousand miles long, has a simple curva- 
ture, convex southwestward. The New Guinea chain, viewing it 
through its whole length from Southern New Caledonia, or perhaps 
* Volcanic Islands, p. 115. American Journal of Science, ii. Ser, ill, 384. 
t+ American Jour. of Science, ii. Ser. ii, 386. 
