Fig. 86.—Vulcanello, with its three craters.
a. The most recently-formed and perfect crater, b and c. Older craters, the walls of which have been partly removed by denudation, e. Lava-currents proceeding from b. The section exposed in the cliff at d is represented in [fig. 35, p. 116].

In the case of the Lipari Islands, the fissures along which the volcanic mountains have been thrown up radiate from a common centre, and a similar arrangement can be traced in many volcanic regions, especially those in which a great central volcano has existed. In other cases, however, as in the Campi Phlegræi, the volcanic vents appear to be formed along lines which assume a parallel arrangement, and this doubtless marks the relative position of the original fissures produced in the earth's crust when these volcanoes were formed. In some other cases we find evidences of the existence of a principal fissure from the sides of which smaller cracks originated. These three kinds of arrangements of volcano-producing fissures are equally well illustrated when we study those denuded districts, in which, as we have seen, the ground-plans of volcanic structures are revealed to our view.

There is now good ground for believing that in volcanic vents, at which long-continued eruptive action takes place, the lavas of different chemical composition make their appearance in something like a definite order. It had been remarked by Scrope and other geologists at the beginning of the present century, that in many volcanic areas the acid or trachytic lavas were erupted before the basic or basaltic.

Von Richthofen, by his studies in Hungary and the volcanic districts of the Rocky Mountains, has been able to enunciate a law governing the natural order of succession of volcanic products; and although some exception to this law may be mentioned, it is found to hold good for many other districts than those in which it was first determined.

In a great number of cases it has been found that the first erupted rocks in a volcanic district are those of intermediate composition which are known as andesites. These andesites, which are especially characterised by the nature of their felspar, sometimes contain free quartz and are then known as quartz-andesites or dacites, from their abundance in Transylvania, the old Roman province of Dacia.

ORDER OF ERUPTION OF VOLCANIC PRODUCTS.

Von Richthofen suggests that another class of volcanic rocks, to which he gives the name of 'propylites,' were in every case erupted before the andesites, and in support of his views adduces the fact that in many instances propylites are found underlying andesites. But the propylites are, in chemical composition, identical with the andesites, and like them present some varieties in which quartz occurs, and others in which that mineral is absent. In their microscopic characters the propylites differ from the andesites and dacites only in the fact that the former are more perfectly crystalline in structure, being indeed in many cases quite undistinguishable from the diorites or the plutonic representatives of the andesites. The propylites also contain liquid cavities, which the andesites and dacites as a rule do not, and the former class of rocks, as Prof. Szabo well points out, are usually much altered by the passage of sulphurous and other vapours, in consequence of which they frequently contain valuable metallic ores.

The extrusion of these andesitic lavas is sometimes accompanied, and sometimes preceded or followed, by eruptions of trachytic lavas—that is, of lavas of intermediate composition which have a different kind of felspar from that prevailing in the andesites.

In the final stages of the eruptive action in most volcanic districts the lavas poured forth belong to the classes of the rhyolitic or acid, and the basaltic or basic lavas.

These facts are admirably illustrated in the case of the volcanic district of the Lipari Islands, to which we have had such frequent occasion to refer. The great central volcano of this district, which now in a ruined condition constitutes a number of small islets (see the map, [fig. 81], p. 192), is composed of andesitic lavas. The other great volcanoes thrown up along the three radiating lines of fissure are composed of andesitic and trachytic rocks. But all the more recent ejections of the volcanoes of the district have consisted either of rhyolites, as in Lipari and Vulcano, or of basalts, as in Stromboli and Vulcanello.