110 H. J. JOHNSTON-LAVIS ON THE GEOLOGY 
no less than three occasions within recent historic times; and any 
one seeing the great size of it cannot but be astonished at the vast 
amount of material displaced. 
The section we have just been studying affords evidence of another 
important fact. At each new paroxysmal eruption the falling frag- 
ments were deposited in various manners; those that pitched upon 
the ridges between the valleys formed a general mantle, thickest 
towards the centre or least-inclined surfaces, thinning away as the 
slope increased, until a point would be reached where all would roll 
down into the valley beneath. Thus the ravine would receive what 
directly fell into it plus that which rolled down its sides and which 
might more or less fill it. The rain would subsequently clear them 
out, and so, choking the lower part or lagno where the inclination 
is small, would have to cut a new course, or, doing so from the begin- 
ning, change the axial position of the valley at a higher point. 
In some cases, when denudation extends to such a depth in the 
valley as to meet with some insurmountable obstacle, the eroding 
action will be deflected to the side opposite and wear away the banks 
at that point. Such a case is beautifully illustrated in the Lagno 
Cancherone, where at the 250-metre contour-line a boss of old lava 
has diverted the stream, giving its valley a serpentine form below 
that point, so that the new sections thus produced show the original 
straightness before erosion had extended down to the disturbing 
rock-mass. 
The manner in which each new mantle of pumice or lapilli covers 
the irregularities of the mountain renders the study of the strati- 
graphical portion of the subject very complicated. The radial sec- 
tions, that is those in the valleys, show each stratum to extend in 
pretty uniform thickness for a considerable distance, and there is 
apparently little unconformity between one bed and another. When, 
on the contrary, the section is tangential or transverse to the ridges 
and valleys, the variation in thickness is very great in short distances, 
and the unconformability well marked. This is well illustrated by 
the Vallone St. Patrizio, where all the deposits seem conformable 
except where a sharp double bend occurs, and the valley cuts across 
its own axis (vide Figure 2). Often where the inclination of the 
sides of a vallone are slight the latest tufa lines or coats each slope, 
which not uncommonly causes an anomaly; for if such be denuded 
away towards the upper edge, an older bed may appear to rest upon 
a younger, two good instances being the Vallone Grande, above St. 
Anastasia, and the Vallone Pallarino. A similar error might arise 
from the floor of a valley cut through lava showing a more recent 
pumice deposit which might seem to underlie the lava. It is very 
common to see deposits of the last prehistoric or pre-Plinian erup- 
tion occupying such a position; but in most cases this has been due 
to the tufa becoming consolidated and converted into a compact 
breccia by the infiltration of calciferous waters, and thus resisting 
the scouring-action of erosion. This may be well seen in the west 
branch of the Vallone Cancherone. 
A peculiar and important feature of the ridges intervening between 
