492 K. MALLET ON THE MECHANISM OF 



of little -coherent material lying between them. These lava beds 

 probably seldom if ever traverse the entire circumference of the 

 crater, but merely snch portion of it as represented the width of the 

 overflow or lava stream at its brim. Were such plates or shelves 

 of intercalated lava free from joints and unbroken, they would, by 

 their cohesion, offer great resistance to the opening of any fissure 

 through them ; but from the effects of their cooling after outflow, 

 and from the immense and unequal pressures to which they have 

 been subjected by the superincumbent mass of the cone, they are all, 

 we must suppose (as, indeed, observation indicates), greatly broken. 

 Still, however, such intercalated plates of lava must greatly increase 

 the resistance of that portion of the crater-walls in which they 

 occur to becoming fissured in the way described. 



Prom all these causes, and the concurrence of many other cir- 

 cumstances which will readily occur to those familiar with the 

 features of volcanic mountains, it will be obvious that fissures com- 

 menced at the interior and propagated into the mass of volcanic 

 cones can seldom be uniformly distributed round the crater, being 

 chiefly formed at the weakest places, and can be but rarely pro- 

 duced in perfectly regular vertical planes having a truly radial 

 direction from the crater ; and the disparity from such normal form 

 or direction, which must always exist more or less, will become 

 greater with the increased length of the fissure (that is, with the 

 distance from the crater), having at every point of its course been 

 forced through heterogeneous material and such as may have its 

 directions of easiest division in very different directions at different 

 points of the path. Thus, for example, the position assigned by 

 Lyell in the Yal del Bove to the ancient crater of Trifoglietto is in 

 a right line more than a mile from the nearest points of the escarp- 

 ment in which the orientation of the dykes is observable, and by 

 the convergence of which the position of Trifoglietto was fixed. 

 Again, assuming, as is commonly done, that the axis of the ancient 

 Somma was the same in position as that of the present Vesuvius, 

 the distance of that point from which the dykes of Somma are 

 supposed to have emanated is more than a mile. 



In traversing such a distance through material highly irregular 

 as regards hardness, cohesion, and direction of easiest fracture, and 

 remembering that all these may vary in any one dyke at every few 

 fathoms of its depth, we are forced to conclude that, even supposing 

 the fissures to have commenced perfectly normally at the respective 

 craters of Somma and Etna, they can scarcely but have diverged 

 largely in many cases, both in dip and strike, and in different 

 degrees at different parts of their height, from the directions in. which 

 they were first started. 



A fissure, for example, commencing vertically and, let us suppose, 

 in a direction due north in the interior of the ancient crater of 

 Somma, and being caused by the diversity as to material through 

 which it is opened to diverge gradually, so as to take a curved 

 direction towards the east or west, must have its orientation changed 

 to an extent proportionate to the distance it has traversed outwards ; 



