districts affords similar proof of the obliteration of a series of lakes by the filling up of hollows, and the cutting through of rocky barriers, a process by which running water ever labours to produce a more uniform declivity." 'There!' said Lyell, I am right either way; for my sword cuts both ways.' In Mr. Hopkins's quick reply, the words were heard, better a sword should cut only the right way.' But Mr. Martin and Humboldt interposing, adjourned the debate to that day 50,000 years.

I have said that the denudation of the Weald and that of the valley of the St. Lawrence each illustrates one of two opposite ways in which nature works the same effect, namely, the driving rivers through ridges which cross their courses. It must be confessed, however, that to level and perfect the waterslope of the line of American lakes is a more difficult job than to make that of the Weald. For in the Weald the water has always worked from above, on soft materials, and with any number of streams, while the St. Lawrence has had, single-handed, an up-hill work, through the very hardest materials. It has had to fill the lakes before it could erode the lowest part of their brims.

But though water cannot flow up hill, it can by ponding flow over hills, and so cut channels

through hills, and this, as Lyell says, is a very universal process in nature. This universal process is, however, directly opposed to the submarine theory.

Truth, if not always, is at least often paradoxical, and paradoxical as it may appear, it is nevertheless true, that rain and rivers are laying dry these, the largest bodies of fresh water in the world. And that by a double process also paradoxical. First, by filling them with alluvium. Second, by emptying them by eroding

their barriers.

With regard to the first cause, the filling them in, Lyell says, the streams which discharge their waters into Lake Superior are several hundred in number, without reckoning those of smaller size. On the northern side, which is encircled by primary mountains, the rivers sweep in many large boulders with smaller gravel and sand, chiefly composed of granite and trap rocks.' The more central deposits consist of 'finer mud' and adhesive clay.' With regard to the second cause, the emptying of the lakes by the erosion of their barriers, that this is going on, is proved by parallel terraces, marking their former levels far above their present levels. A depth of fifty feet has been thus taken from the surface of Lake Superior, which is nearly 2,000 miles in

6

This is all right. This is subaerial denudation, and very contrary to the submarine theory. Page 685, he enquires how far vegetation may retard the levelling effects of running water, which it cannot oppose, much less counterbalance. It is well known that a covering of herbage and shrubs may protect a loose soil from being carried away by rain, or even by the ordinary action of a river, and may prevent hills of loose sand from being blown away by wind; for the roots bind together the separate particles into a firm mass, and the leaves intercept the rain water, so that it dries up gradually, instead of flowing off in a mass, and with great velocity. The old Italian hydrographers make frequent mention of the increased degradation which has followed the clearing away of natural woods in several parts of Italy. A remarkable example was afforded in the Upper Val D'Arno in Tuscany, on the removal of the woods clothing the steep declivities of the hills by which that valley is bounded. When the ancient forest laws were abolished by the Grand Duke Joseph, during the last century, a considerable tract of surface in the Cassentina (the Clausentinium of the Romans) was denuded, and immediately the quantity of sand and soil washed down into the Arno increased enormously. Frisi, alluding to such

occurrences, observes, that as soon as the bushes and plants were removed the waters flowed off more rapidly, and, in the manner of floods, swept away the vegetable soil. This effect of vegetation is of high interest to the geologist, when he is considering the formation of those valleys which have been principally due to the action of rivers. The spaces intervening between valleys, whether they are flat or ridgy, when covered with vegetation, may scarcely undergo the slightest waste, as the surface may be protected by the greensward of grass; and this may be renewed, in the manner before described, from elements derived from rain-water and the atmosphere. Hence, while the river is continually bearing down matter in the alluvial plain and undermining the cliffs on each side of every valley, the height of the intervening rising grounds may remain stationary.'

it is

were widened

river banks,

would not

slope, but

must be cliffs

the

Observe the change of agents in forming If valleys valleys, in one short sentence. It begins with by erosion of rain from above, and ends with the river under- their sides mining the cliffs' from below; though evident, that if valleys were formed by undermining erosion of rivers, even in unrocky ground like our chalk, the sides would be precipitous cliffs, and the tops of intervening hills flat table land; while the enlargement of

6

with intervening table land.

Rivers may cut deep

makes wide

such valleys could only take place by a constant calamitous succession of land-slips, in the same way that cliffs yield to the erosion of the sea. Rivers have the power to cut narrow channels channels, rain or ravines; but they have very little power of widening these. Disintegration and the wash of rain widen these ravines into broad valleys. While this is going on rivers convey to the sea what rain brings to them, which would otherwise pass out of the valleys more slowly by the wash of rain.

valleys.

But rivers no more make the sediment which they carry than railroads manufacture the wares of which their traffic consists. The sediment carried by rivers is brought to them by rain, from the entire surface of their tributary ground, or waterslope. Erosion of banks is an exceptional calamity. But heavy rain never happens, but it loads the rivers with mud. Nay, it loads the along-shore waters of the sea with mud. These are matters of fact, not matters of opinion. We need no ghost' nor good eyes to ascertain them; all we need good is the umbrella and macintosh, to enable us to go out and see.

The Atlantic round Madeira, or the Mediterranean along Spain, France, and Italy, or the Egean round the Isles of Greece, forms no exception: vidimus flavum each of them; or rather