[ 22 ] 
of fuch groups, which have given way through time. 
This I have more particularly obferved in the vulcanic 
,-and columnar hills of Ifenchaux in Velay ; in which the 
groups of columns are often fo united with the body of 
lava^ that they form, in a manner, but one folid, though 
figured mafs. Monte RolTo is precifely of this charadter; 
nor did I obferve a fingle column, or fragment even, 
loofe; thofe in the drawing being merely introduced to 
fliow the forms of the columns. Neither are there here, 
or in the bafaltine hills near Ifenchaux, any figns of a 
crater: on the contrary, thefe hills, as well as Monte 
Roffo, are moftly terminated by regular convex fummits, 
that form a folid mafs. And that fufed mafles fliould 
frequently concrete in fuch a form we need not wonder, 
if we refle^l on the effervefcent and expanfive property 
of fire. The phaenomenon of horizontal vulcanic hills 
is accountable upon another principle, and feems chiefly 
to depend on the Rate of thofe hills before their ignition; 
as I fliall endeavour to prove in my account of thofe of 
Auvergne ; and of which the vulcanic hills of the Vero- 
nefe and Vicentine diftri6ts afford alfo fingular inftances, 
which I fliall now confider. It is difficult to fay in 
what ftate vulcanic hills of a particular and regular ftruc- 
ture, like the bafaltine hills, for inftance, may have pre- 
exifted, before their alterations by fire, fince they afford 
evident proofs, not only of a liquefadtion, but of an in- 
tire new organization ; by which means all marks of their 
former characters are totally effaced. Notwithftanding 
which, fince thefe organizations are generated locally, 
fome 
