390 COSMOS. 



place on widely extended fissures and submarine mountain- 

 chains, which run in directions governed by fixed laws of 

 region and grouping, and which, just as we see in the conti- 

 nental mountain chains of Central Asia, and of the Caucasus, 

 belong to different systems ; but the circumstances which 

 govern the area over which at any one particular time the 

 openings are simultaneously active, probably depend, from 

 the extremely limited number of such openings, on entirely 

 local disturbances to which the conducting fissures are sub- 

 jected. The attempt to draw lines through three now 

 simultaneously active volcanoes, whose respective distances 

 amount to between 2400 and 3000 geographical miles 

 asunder, without any intervening cases of eruption (I refer 

 to three volcanoes now in a state of ignition, Mouna Loa, 

 with Kilauea on its eastern declivity, the cone-mountain 

 of Tanna, in the new Hebrides ; and Assumption Island in 

 the North Ladrones), would afford us no information in re- 

 gard to the general formation of volcanoes in the basin of 

 the South Sea. The case is quite different if we limit our- 

 selves to single groups of islands, and look back to remote, 

 perhaps pre-historic, epochs when the numerous linearly 

 arranged, though now extinct, craters of the Ladrones 

 (Marian Islands), the New Hebrides and the Solomon's 

 Islands were active, but which certainly did not become 



positions. There is a system in their arrangement as regular as in 

 the mountain heights of a continent, and ranges of elevation are indi- 

 cated, as grand and extensive as any continent presents." Geology, 

 by J. Dana, United States' Exploring Expedition, under command of 

 Charles Wilkes, vol. x, (1849) p. 12. Dana calculates that there are 

 in the whole of the South Sea, exclusive of the small rock-islands, 

 about 350 basaltic or trachytic and 290 coral islands. He divide? 

 them into twenty-five groups, of which nineteen in the centre 

 have the direction of their axis N. 50 60 W., and the remaining 

 N. 20 30 E. It is particularly remarkable that these numerous 

 islands, with a few exceptions, such as the Sandwich Islands and New 

 Zealand, all lie between 23 23' of north and south latitude, and that 

 there is such an immense space devoid of islands eastward from the 

 Sandwich and the Nukahiva groups as far as the American shores of 

 Mexico and Peru. Dana likewise draws attention to a circumstance 

 which forms a contrast to the insignificant number of the now active 

 volcanoes, namely, that if, as is probable, the Coral Islands, when lying 

 between entirely basaltic islands, have likewise a basaltic foundation, 

 the number of submarine and subae'rial volcanic openings may be esti- 

 mated at more than a thousand (pp. 17 and 24). 



