and thirteenth chapters; and to the fact also shown here, that the very museums of land fossils are transitory, liable to be destroyed by the very agent which built them-the wash of rain. And because land animals are only found in numbers in recent land deposits, it is a mistake to argue that they did not exist during the period of the deposit of the most ancient marine strata.

The whole of the thirty-first chapter of the 'Elements' seems written to disprove Lyell's own theory of the submarine formation of valleys, and

Hills and valdistrict of cen

leys of volcanic

tral France,

contrary to

submarine and

in favour of

subaerial

to prove that they are formed by the subaerial denudation of rain and rivers. The following are theory.

extracts:

6

With regard to the Plomb du Cantal, which is a volcanic mountain several thousand feet in height,' and which has never been under the sea, and therefore should have no valleys: Valleys radiate in all directions from the natural heights of the mountain, increasing in size as they recede from those heights. Those of the Cer and Jourdanne, which are more than twenty miles in length, are of great depth, and lay open the geological structure of the mountain.'

Speaking generally, he says: "The lavas may often be traced from the crater to the nearest valley, where they usurp the channel of the river, which has frequently excavated a deep

ravine through the basalt. We have thus an opportunity of contrasting the enormous degradation which the solid and massive rock has suffered by aqueous erosion, and the integrity of the cone of sand and ashes, which has in the meantime remained uninjured on the neighbouring platform, where it was placed beyond the reach of the power of running water.'

6

With regard to the Puy de Côme, a stream of lava fills the ancient river-channel for the distance of more than a mile. The Sioule, thus dispossessed of its bed, has worked out a fresh one between the lava and the granite of its western bank; and the excavation has disclosed in one spot a wall of columnar basalt about fifty feet high. The excavation of the ravine is still in progress, every winter some columns of basalt being undermined and carried down the channel of the river, and, in the course of a few miles, rolled to sand and pebbles. Meanwhile, the cone of Côme remains stationary, its loose materials being protected by a dense vegetation, and the hill standing on a ridge not commanded by any higher ground whence floods of rainwater may descend. At another point, farther down the course of the Sioule,' 'the river has excavated a ravine through lava and subjacent gneiss to the depth of four hundred feet. On

the upper part of the precipice forming the left side of this ravine, we see a great mass of black and red scoriaceous lava; below this a thin bed of gravel, evidently an ancient river-bed, now at an elevation of fifty feet above the channel of the Sioule. The gravel again rests upon gneiss, which has been eroded to the depth of fifty feet. It it quite evident in this case, that while the basalt was gradually undermined and carried away by the force of running water, the cone whence the lava issued escaped destruction because it stood upon a platform of gneiss several hundred feet above the level of the valley in which the force of running water was exerted. The brim of the crater of the Puy de Pariou, near Clermont, is so sharp, and has been so little blunted by time, that it scarcely affords room to stand on. This and other cones, in an equally remarkable state of integrity, have stood, I conceive, uninjured, not in spite of their loose porous nature, as might at first be naturally supposed, but in consequence of it. No rills can collect where all the rain is instantly absorbed by the sand and scoriæ.'

eternity of cones, owing to their shape.

Lyell here acknowledges the denuding power Comparative of rain where it can collect in rills. But the porousness of volcanic cones is not the sole reason, nor even the main reason, why rain cannot col

lect on them. Their shape is the main reason. I endeavoured to show this in the second edition of the Tree-lifter.' I will give one more homely

illustration of the theory here.

Suppose a building five yards square, with a roof like a shed, spouted. All the rain falls into five yards of spouting. Shape the roof like a gable, the same shed of rain is dispersed over ten yards of spouting. Shape the roof like a foursided pyramid, the same shed of rain is dispersed over twenty yards of spouting instead of five.

But cones and cairns have still greater power to disperse rain rills than flat-sided pyramids. And it is to their shape mainly that volcanic cones owe their comparative eternity. Comparatively eternal also is the truncation of their shape; for this must remain till they are denuded below the bottom of the crater-in many cases many hundred feet-they will then assume the point and appearance of a simple cone of denudation. While this denudation of the crater is going on, supposing equal hardness of material, the rim should remain sharp to the last, as Lyell describes many.

How beautifully distinct at any distance is this form of the children of igneous up-throw, whose very uniformity makes variety amid forms resulting from aqueous down-pour. But to aqueous

down-pour myriads of them have yielded, and are yielding, and comparative eternity is the most perennial claim which can be made to aught existing on a globe, the very being of every department of which, organic or inorganic, is dependent on perpetual change; that is, on perpetual disintegration and reproduction.

In fact the tendency of disintegration and the wash of rain is to form cones out of single hills, and ridges out of chains of hills, with projecting spurs, each spur being itself a ridge ending in a half cone; and even these ridges are so studded with cones as to have a serrated or saw-like outline, and to have earned the modern Spanish, Portuguese, and Italian name of sierra or serra, and to have originated the Latin expression of per juga montium, the very name of hills being taken from the yoke-like appearance of contiguous cones. And in this sense, rain may be said to form hills as well as valleys. For this reason, though as Sir Humphry Davy finely remarks, no work of a mortal can be immortal, those works of man which approach nearest to immortality are cones,—the pyramid, the tumulus, and the cairn. Why do the imber edax and the fuga temporum pass with so light a touch over these? Because they begin with a form which others end in,-a form which