From what has been said, it will be seen that, not only do the volcanoes of the globe usually assume a linear arrangement, but nearly the whole of them can be shown to be thrown up along three well-marked bands and the branches proceeding from them. The first and most important of these bands is nearly 10,000 miles in length, and with its branches contains more than 150 active volcanoes; the second is 8,000 miles in length, and includes about 100 active volcanoes; the third is much more broken and interrupted, extends to a length of nearly 1,000 miles, and contains about 50 active vents. The volcanoes of the eastern coast of Africa, with Mauritius, Bourbon, Rodriguez, and the vents along the line of the Red Sea, may be regarded as forming a fourth and subordinate band.
Thus we see that the surface of the globe is covered by a network of volcanic bands, all of which traverse it in sinuous lines with a general north-and-south direction, giving off branches which often run for hundreds of miles, and sometimes appear to form a connection between the great bands.
These four bands of volcanic vents, running in a general north-and-south direction, separate four unequal areas within which the exhibitions of volcanic activity are feeble or quite unknown. The two grandest of the bands of volcanic activity, with their branches, form an almost complete series encircling the largest of the oceans.
To this rule of the linear arrangement of the volcanic vents of the globe and their accumulation along certain well-marked bands, there are two very striking exceptions, which we must now proceed to notice.
In the very centre of the continent formed by Europe and Asia, the largest unbroken land-mass of the globe, there rises from the great central plateau the remarkable volcanoes of the Thian Shan Range. The existence of these volcanoes, of which only obscure traditional accounts had reached Europe before the year 1858, appears to be completely established by the researches of the Russian traveller Semenof. Three volcanic vents appear to exist in this region: the active volcanoes of Boschan and Turfan or Hot-schen, and the solfatara of Urumtsi. At a point situated about half-way between these three volcanoes and the sea, another active vent, that of Ujung-Holdongi, is said to exist. Other volcanic phenomena have been stated to occur in the great plateau of Central Asia, but the existence of some at least of these appears to rest on very doubtful evidence. The only accounts which we have of the eruptions of these Thian Shan volcanoes are contained in Chinese histories and treatises on geography; and a great service would be rendered to science could they be visited by some competent explorer.
EXCEPTIONALLY-SITUATED VOLCANOES.
The second exceptionally-situated volcanic group is that of the Sandwich Islands. While the Thian Shan volcanoes rise in the centre of the largest unbroken land-mass, and stand on the edge of the loftiest and greatest plateau in the world, the volcanoes of the Sandwich Islands rise almost in the centre of the largest ocean and from almost the greatest depths in that ocean. All round the Sandwich Islands the sea has a depth of from 2,000 to 3,000 fathoms, and the island-group culminates in several volcanic cones which rise to the height of nearly 14,000 feet above the sea-level. The volcanoes of the Sandwich Islands are unsurpassed in height and bulk by those of any other part of the globe.
With the exception of the two isolated groups of the Thian Shan and the Sandwich Islands, nearly all the active volcanoes of the globe are situated near the limits which separate the great land- and water-masses of the globe—that is to say, they occur either on the parts of continents not far removed from their coast-lines, or on islands in the ocean not very distant from the shores.
The fact of the general proximity of volcanoes to the sea, is one which has frequently been pointed out by geographers, and may now be regarded as being thoroughly established. Even the apparently anomalous case of the Thian Shan volcanoes is susceptible of explanation if we remember the fact, now well ascertained by geological researches, that as late certainly as Pliocene times, a great inland sea spread over the districts where the Caspian, the Sea of Aral, and many other isolated lakes are now found. Upon the southern shore of this sea rose the volcanoes of the Thian Shan, some of which have not yet fallen into a state of complete extinction.
But although the facts concerning the general proximity of volcanoes to the ocean may be admitted to be thoroughly established, yet inferences are sometimes hastily drawn from these facts which the latter, if fairly considered, will not be found to warrant. It is frequently assumed that we may refer all the remarkable phenomena of volcanic action to the penetration of sea-water to a mass of incandescent lava in the earth's crust, and to the chemical or mechanical action which would result from this meeting of sea-water and molten rock. And this conclusion is supposed to find support in the circumstance that many of the gases and volatile substances emitted from volcanic vents are such as would be produced by the decomposition of the various salts contained in sea-water.