Ch. X.] 
PARALLEL SIDES OF DIKES. 
123 
the mountain, and, whilst in a state of perfect fusion, continued 
their course in those channels, which were sometimes full to 
the brim, and at other times more or less so according to the 
quantity of matter in motion. 
* These channels, upon examination after an eruption, I 
have found to be in general from two to five or six feet wide, 
and seven or eight feet deep. They were often hid from the 
sight by a quantity of scoriae that had formed a crust over 
them, and the lava, having been conveyed in a covered way for 
some yards, came out fresh again into an open channel- After 
an eruption I have walked in some of those subterraneous or 
covered galleries, which were exceedingly curious, the sides, 
top, and bottom, being worn perfectly smooth and even in most 
parts, by the violence of the currents of the red-hot lavas, 
which they had conveyed for many weeks successively.' 
In another place, in the same memoir, he describes the liquid 
and red-hot matter as being received c into a regular channel, 
raised upon a sort of wall of scorise and cinders, almost perpen- 
dicularly, of about the height of eight or ten feet, resembling 
much an ancient aqueduct *.' 
Now, if the lava in these instances had not run out from the 
covered channel, in consequence of the declivity whereon it was 
placed — if, instead of the space being left empty, the lava had 
been retained within until it cooled and consolidated, it would 
then have constituted a small dike with parallel sides. But the 
walls of a vertical fissure through which lava has ascended in 
its way to a volcanic vent, must have been exposed to the same 
erosion as the four sides of the channels before adverted to. 
The prolonged and uniform friction of the heavy fluid as it flows 
upwards cannot fail to wear and smooth down the surfaces on 
which it rubs, and the intense heat must melt all such masses 
as project and obstruct the passage of the incandescent fluid. 
We do not mean to assert that the sides of fissures caused 
by earthquakes are never smooth and parallel, but they are 
usually uneven, and are often seen to have been so where volcanic 
* Phil. Trans., vol. lxx. 1780. 
