1830,] 
On the Excavation of V allies. 
243 
might groove out the bed of the Garry ; but the same current could not excavate 
the° bed of the Tilt, at right angles to the other. We believe, indeed, that on this 
subject geologists, with few exceptions, are now disposed to adopt the opinions so 
beautifully developed by Playfair, in his illustrations of the Huttonian Theory, and 
which derive no small confirmation from the researches of the two writers now be- 
fore us. 
It is perhaps unknown to some of our readers, that there are about three hun- 
dred volcanoes in the south of France, not such volcanoes as Arthur’s Seat and 
North Berwick Law presented to the eye of the fanciful Faujas St. Fond, but real 
igneous cones, in general with open craters in their summits, surrounded by mass- 
es of slags, pumice, and ashes, and with long currents of perfect lava on their sides, 
— volcanoes, in short, differing in no essential circumstance from Etna or Vesu- 
vias, when in a state of quiescence. History is silent as to their eruptions ; but 
this only shows that the period of their activity must have been very remote ; for 
that they were once active, and that a long period elapsed between the first and last 
of their eruptions, is attested by facts which cannot be mistaken. Having premis- 
ed these remarks, we shall endeavour to make the reader acquainted with the phe- 
nomena which Messrs. Lyell and Murchison describe in detail, and make the 
groundwork of their conclusions. 
Some of the streams of ancient lava in Auvergne have been traced over an extent 
of thirty miles or more. As the lava in its progress, like any other fluid body, seeks 
the lowest levels, it necessarily flowed into the beds of the rivers. In this way it some- 
times formed barriers across their courses, or filled up their channels for several 
miles ; a lake was consequently formed, whose waters rose till they sought out some 
lateral channel, or flowed over the mound of lava which had arrested their progress. 
There are rivers in Auvergne whose courses have been entirely changed in this 
way. But it is a curious fact, which is confirmed by what is observed in active 
volcanoes, that currents of lava do not flow with a level surface like streams of 
water, but swell up in the middle, and have their sides often sloping at an angle of 
30 or 40 degrees. Let us now suppose that the line ABC 
t 
\ 
\ 
\ 
\ 
^ 
B 
represents a transverse section of a cavity cut in the primitive rock (gneiss) by a 
stream of water, which- flows over a bed of gravel at the bottom B. A “ coulee” 
or current of lava flows into this cavity from the right, and fills it up to D E, which 
represents the curved surface of the lava. The water stopped by the irruption will 
rise till it escape through the triangular opening at E, formed by the sloping side of 
the current of lava M E, and the original bank of primitive rock E C. So long as 
the old part of the river’s bed above the intruding lava forms a lake, all the sand, 
gravel, and stones, which are the great engines of attrition, are detained by it, and 
the limpid water which escapes, makes little or no impression on the rock. But when 
these travelled materials, having filled up the lake, begin to be rolled over the new 
barrier, the erosive action commences ; and as it goes on, the lava D E is worn 
away on the one side, and the gneiss E C on the other, till a new channel M L S X 
is worked out exactly like the old, with a stratum of water-worn pebbles at the 
bottom. We may now conceive a second irruption of lava to take place from the 
same side ; the waters again collect ; a lake is formed, and afterwards filled up ; 
and a third channel is scooped out, higher and farther to the left than M L S X. 
Auvergne abounds in examples of such changes as we have now described ; and 
we are thus enabled to understand how it often happens there, that one side of a 
