526 VOLCANIC ROCKS OF [Ch. XXX. 



of altered sedimentary rocks ejected during eruptions. We may easily 

 conceive that the first explosions would act with the greatest violence; 

 rending and shattering whatever solid masses obstructed the escape of 

 lava and the accompanying gases, so that great heaps of ejected pieces 

 of rock would naturally occur in the tufaceous breccias formed by the 

 earliest eruptions. But when a passage had once been opened, and an 

 habitual vent established, the materials thrown out would consist of 

 liquid lava, which would take the form of sand and scorige, or of angu- 

 lar fragments of such solid lavas as may have choked up the vent. 



Among the fragments which abound in the tufaceous breccias of 

 Somma, none are more common than a saccharoid dolomite, supposed 

 to have been derived from an ordinary limestone altered by heat and 

 volcanic vapours. 



Carbonate of lime enters into the composition of so many of the 

 simple minerals found in Somma, that M. Mitscherlich, with much pro- 

 bability, ascribes their great variety to the action of the volcanic heat 

 on subjacent masses of limestone. 



Dikes of Somma. — The dikes seen in the great escarpment which 

 Somma presents towards the modern cone of Vesuvius are very nume- 

 rous. They are for the most part vertical, and traverse at right angles 

 the beds of lava, scoriae, volcanic breccia, and sand, of which the ancient 

 cone is composed. They project in relief several inches, or sometimes 

 feet, from the face of the cliff, being extremely compact, and less de- 

 structible than the intersected tuffs and porous lavas. In vertical extent 

 they vary from a few yards to 500 feet, and in breadth from 1 to 12 feet. 

 Many of them cut all the inclined beds in the escarpment of Somma 

 from top to bottom, others stop short before they ascend above half way, 

 and a few terminate at both ends, either in a point or abruptly. In 

 mineral composition they scarcely differ from the lavas of Somma, the 

 rock consisting of a base of leucite and augite, through which large 

 crystals of augite and some of leucite are scattered.* Examples are not 

 rare of one dike cutting through another, and in one instance a shift or 

 fault is seen at the point of intersection. 



In some cases, however, the rents seem to have been filled laterally, 

 when the walls of the' crater had been broken by star-shaped cracks, as 

 seen in the accompanying wood-cut (fig. 663). But the shape of these 

 rents is an exception to the general rule; for nothing is more remarka- 

 ble than the usual parallelism of the opposite sides of the dikes, which 

 correspond almost as regularly as the two opposite faces of a wall of 

 masonry. This character appears at first the more inexplicable, when 

 we consider how jagged and uneven are the rents caused b}^ earthquakes 

 in masses of heterogeneous composition, like those composing the cone 

 of Somma. In explanation of this phenomenon, M. Necker refers us 

 to Sir W. Hamilton's account of an eruption of Vesuvius in the year 



*. L. A. Necker, Mem. de la Soc. de Pliys. et d'Hist. Nat. de Geneve, torn. ii. 

 part i. Nov. 1822. 



