La vis — On the Structure of Rocks. 125 



igneous matter and the surface (which, however, it never is), and 

 that such a canal traversed rock strata of different permeability ; 

 then the magma enclosed in the tube would consist of a series of 

 more or less saturated aquiferous strata, superposed on each other 

 in the same order in which each part was exposed to a portion of 

 the canal wall. Now, should a sudden exit of magma occur from 

 the tube at its upper extremity, the expansion, or, in other words, 

 eruption, would take place with a violence directly proportional to 

 the amount of dissolved water, and the temperature of that por- 

 tion of the mass nearest the surface, at different periods of the 

 eruption. The eruption will therefore lull or augment, as that 

 portion, being expelled originally, occupied a more or less favour- 

 able site for the absorption of its contained volatile matter. The 

 examples given at the end of this Paper seem to indicate that this 

 departure from what is a normal type of eruption of a truly homo- 

 geneous magma is of rare occurrence. My experience in the field 

 has been chiefly drawn from the basic volcanoes of Vesuvius and 

 Roccamonfina (Leucite basalts), Mt. Yultura (Haiiyn basalt), 

 Mt. Nuovo (Phonolite), Yentotene and San Stefano of the Ponza 

 group (Andesite), and Ischia (Trachyte). Were the suppositions 

 in the above case true, and were the entire chimney or canal com- 

 pletely emptied in each eruption, then we should expect every 

 stratum of ejectamenta representing an eruption to be composed of 

 a series of components alternating with each other in direct relation 

 with the eruptive variations, and with the structure of the earth's 

 crust beneath the volcano. Besides, in any one volcano, we 

 should expect each stratum of pumice to be made up of analogous 

 components to those produced during eruptions that preceded and 

 followed it, indicating the same train of variations of activity, 

 which is not the case. Geological evidence, so far as denudation 

 has opened up to our examination the old remnants of igneous 

 dykes and chimneys, leads us to conclude that igneous canals 

 assume very irregular shapes, winding about where least resistance 

 was offered to their extension, but nearly always assuming the 

 form of a plate-like mass choking a fissure. Such fissure we know 

 may have a horizontal extension of many miles. The opening or 

 openings at the surface would be very local, and therefore the 

 exit of the igneous matter would tend to take the form of a 

 fan-shaped current with the point of orientation at the exit. Under 



