OF MONTE SOMMA AND VESUVIUS. 109 
delta is formed that gradually ascends the lagno until it produces too 
great an obstruction to the torrent, which is either diverted from its 
course or cuts a new passage through its earlier deposits. We see, 
therefore, that the upper third of the mountain is rendered more steep, 
since its crest 1s little affected ; but as we reach the middle third erosive 
action is rapidly gaining power and at this part reaches its maximum 
until a sudden check is put upon it at the commencement of the 
third part, or lagno, where, instead of a loss of material occurring, 
an actual gain takes place. Thus there is a tendency to produce a 
depression in the middle of the slope of the mountain, and at the 
same time to prolong the bottom part or toe. The whole group of 
phenomena may be said to render the upper parts of the mountain 
more steep and the lower parts less so, in such a manner that the 
general profile becomes more concave. This remarkable concavity I 
am inclined to attribute much more to the above cause, namely de- 
nudation, than to the sinking of the mountain by its own weight*. 
This leads us on to the inquiry as to when the valleys began to 
assume their present form; and whether they have materially altered 
their shape and position during their formation. These two im- 
portant questions open up a most interesting field of inquiry. 
here is little doubt that denuding influences had commenced 
work while the volcano was still in its period of tranquil activity, 
and also in the period of temporary extinction, prior to the change 
in its eruptive character, producing irregular shallow valleys that 
had already begun to score its sides. The exceedingly small space 
of lava exposed in each section, the fact of all these valleys being 
nearly parallel, and a few sections being at right angles to them, 
render it difficult to determine how much had already been done 
when the first ejection of the white pumice came. ‘That there were 
valleys there is no doubt; for in a circumferential section I have 
lately constructed, cutting tangentially and continuously through 
more than a quarter of all the valloni of Monte Somma, this fact is 
brought out most distinctly, though, from the partly problematical 
character that such a work must have, and from the impossibility of 
knowing the internal structure of the intervening ridges, only an 
imperfect idea can be gained of the form and depth of these ancient 
ravines. 
We see in many sections how the products of each eruption were 
partly washed away before the deposition of the next ejectamenta. 
This is well illustrated in Plate II. fig. 2 of the Vallone di Pol- 
lena, happily showing the rare occurrence of a natural transverse 
section. Here it is observed that the present valley at this point 
has shifted its axis to the right, or south-west, since the Plinian 
eruption that choked the gorge, which before that time appears 
always to have remained stationary. Lower down in the lagno a 
valley existed to the right (in ascending) of the present, now choked 
by the large quantity of black pumiceous scoria of PHasr VII., 
Period 4, which seems to have caused the lagno to take its present 
direction. Thus we see that this ravine has changed its position on 
* J. W. Judd, F.R.S., ‘ Volcanoes, 1881, p. 166. 
