1866.| On the Antiquity of the Volcanos of Auvergne. 211 
The annexed drawing, taken from Sir Roderick Murchison and 
Sir Charles Lyell’s paper ‘On the Excavation of the Valleys in 
Auvergne,’ will illustrate the relative position of the laya-current 
and gneiss in this locality. 
Hence it follows, that since the period at which the lava was 
ejected, a thickness of 50 feet of solid gneiss must have been 
excavated. 
Now the slowness with which the present river erodes a 
material of this description may be estimated by a fact pointed out 
by Sir Charles Lyell. in the same province, near St. Nectaire, where 
an ancient Roman bridge spans the River Couze, over a stream of 
lava, proceeding from a volcanic hill,—the Puy de Tartaret,— 
showing that a ravine, precisely like that now existing, had already 
been excavated by the river fourteen centuries ago. 
And yet the lava of the Puy de Tartaret presents all the 
appearance of a modern current, both from its having conformed to 
the sinuosities of the valley, and also from its covering a bone 
deposit at its bottom, indicatng a mammiferous Fauna, which, 
although distinct, as a whole, from that now inhabiting Auvergne, 
presents some features in common with it, as in the existence 
of the dog, deer, cat, &c., mixed with the remains of the reindeer, 
which, even so late as the time of Cesar, appears to have been 
found in the Great Hercynian Forest, and also with an animal of 
the horse tribe, differing, however, in some points from the species 
now living. 
But it is in the neighbouring province of the Vivarais that the 
most remarkable instances of the long-continued action of water 
slowly eroding to a great depth streams of lava which havé flowed 
at a comparatively recent date, are afforded. 
Before describing these, however, I must point out a cireum- 
stance which distinguishes a current of lava from one of water, 
namely, that from its viscid character it has a tendency, near its 
termination, to accumulate layer upon layer, so that its materials 
are piled up to a considerable height, instead of spreading onwards, 
as would happen to a substance of more perfect fluidity. 
Hence, when a lava stream reached the bed of a river, it 
sometimes formed a precipitous bank on one side of it, without 
appearing to have advanced to the other. 
Of this, indeed, several examples are met with in Auvergne, but 
the most remarkable cases are those to which I have alluded in the 
Vivarais. 
Tn that province, Mr. Scrope enumerates no less than six perfect 
volcanic cones, with craters on their summits still preserved in a 
State of greater or lesser integrity, from which have proceeded 
streams of lava, each of which may be traced down the sides of the 
mountain, and are seen to terminate abruptly at its foot. 
