Plains of construction are such as are formed by the undisturbed extension of beds which retain the original horizontality of deposition. River deltas and alluvial flats, fens, marshes, and silted-up lakes, are instances of such plains. In this class may also be included the slightly inclined or undulating surfaces sometimes produced by the elevation of a shallow sea-bottom which special local causes have exempted from the action of marine erosion; such are the Pampas of South America, the Tundras of Siberia, and large tracts of Russia and Poland.

Plains of erosion are those which have been formed by marine erosion across the edges and outcrops of strata without reference to their inclination, flexures, or fractures. They are surfaces of planation formed by the march of the sea across the country. The limestone plains of central Ireland may be cited as an instance, and the country bordering the rivers Rhine and Moselle is another. Hills are of three kinds: 1, hills of accumulation; 2, hills of circum-detrition; 3, hills of upheaval.

Hills of accumulation are such as have been formed by the piling-up of materials upon the surface of the ground. Volcanoes are the most important instances. (See plate at the end of the volume.) Sand-hills heaped up by the action of wind on drifting sand come under the same category.

Hills of circum-detrition are such as have been left by the removal of the surrounding rocks, or are isolated by valleys of erosion.

Hills of upheaval are those formed by the elevation of a central axis faster than it can be worn down by detritive agencies.

THE

CHAPTER V.

THE ORIGIN OF LAKES.

'HE origin of the hollows or troughs in which water accumulates to form lakes has long engaged the attention of geologists, and it is now generally recognized that such hollows have been formed in many different ways. In the first place lakes may be divided into three different classes, according to their position in relation to the larger physical features of the country, as follows:I. Plain and Plateau lakes. II. Valley lakes.

III. Crater lakes.

In the first class may be placed all those lakes, whether large or small, which do not lie in the course of a definite and continuous valley. In the second, those which do lie in such a valley. These two classes include by far the larger number of lakes which exist, for the third class only includes those which lie in the craters of extinct volcanoes, and the origin of these being obvious, we need not further consider them.

I. Plain and Plateau Lakes.-There are at least three ways in which such lakes may have been formed; the hollows in which they lie may be due to: (1) original irregularity of the surface; (2) unequal elevation; (3) subsidence from solution of underlying rock.

1. Original Irregularity of Surface. We know that the sea-floors of the present day are by no means regular slopes or plains, but exhibit ridges and reefs, hollows and troughs, of various sizes and shapes. Some of these are probably ancient hollows formed when the sea-floor was a

land surface, and not yet filled up; but many are simply due to the action of currents and to the unequal deposition of material. When the sea-bottom is again raised into land some of these hollows will be occupied by fresh water and will become lakes.

Unequal surfaces of deposition are specially characteristic of deposits laid down by the agency of ice, whether land-ice or sea-ice; deposits so formed always present an irregular hummocky surface, in the hollows of which water naturally accumulates when the surface again becomes dry land. Such surfaces are common all over the northern parts of Europe and North America.

Lakes occupying hollows which have once been portions of sea-floors are not infrequent. Thus there is good reason to believe that the whole of Central Asia was once a sea comparable to the Mediterranean, the lakes both salt and fresh which now exist there being merely the deeper portions of that sea, for the creatures which live in them are chiefly marine types. Crustacea belonging to marine genera also exist in Lake Wener (Sweden), in Lake Superior, and even in Lake Titicaca (Peru). The last is 12,500 feet above the sea, but several species of a small Crustacean called Allorchestes occur in it, one of which is identical with a form still living in the Straits of Magellan. In all these cases the conclusion has been drawn that they were hollows filled originally with sea-water, which has been so gradually replaced by fresh water during the elevation of the country that some of the marine creatures were able to accommodate themselves to the change of conditions.

2. Unequal Elevation.-No part of the earth's surface seems to have been stationary for a very long period of time, speaking in a geological sense, and as movements of upheaval or subsidence affect limited areas, it often happens that elevation is proceeding in one area while a neighbouring area is stationary or sinking; consequently if the two areas form part of one continental region, the surface of this region is being slowly tilted. Such a tilt may only increase the original slopes, but on the other hand it may have a contrary effect, and may even reverse the slope of the lower and more level parts of the region.

It is obvious that in such a case the whole drainage of the country will be altered, and lakes are likely to be formed during the process.

It has recently been suggested by Mr. H. H. Howorth that such a reversal of the drainage and general slope of the country has taken place in Northern Asia.' He thinks that when the Mammoth lived in Siberia part of the Polar Sea was land, and the Siberian rivers flowed southward into the Central Asian Sea mentioned above; the present arrangement being a reversal of the drainage consequent upon the elevation of the central plateaux and the subsidence of the Arctic part of Asia. If this view is correct some of the Siberian lakes may have been formed during this gradual change of slope.

3. Unequal Weathering-It was stated in Part I., Chap. ter VI., that in the process of weathering or disintegration by atmospheric agencies some rocks were decomposed to a great depth. This is especially the case with coarse felspathic rocks, like Granite, Syenite, and Gneiss; and Pumpelly has called attention to the fact that where such rocks have been exposed to this weathering process for a long time incipient rock-basins are formed. He points out that the surface of the solid rock below the decomposed portions is very irregular, here rising into mounds and ridges, there falling into hollows and troughs, according as the rain-water has been directed toward certain places and along certain lines, and as its action has been helped by the presence or absence of vegetation, or by differences in the durability of the rock itself. It is clear, therefore, that if the loose and decomposed parts of such rock were removed by any detersive agency the underlying depressions might eventually be converted into lakes. He specially indicates certain enclosed basins in crystalline rocks in Asia as having been formed by the unequal decay of the rock and the removal of the residual materials by the wind.2

In the northern parts of both hemispheres there are large tracts of granite and gneissic rock, and before the

"Geol. Mag.," Dec. 3, vol. vii. (1890), pp. 5 and 438.

2 "Amer. Journ. Science," third series, vol. xvii. p. 139.

advent of the Glacial period or Great Ice Age, of which mention was made on p. 155, all these surfaces must have been more or less weathered in the manner described. During that period many of these tracts were swept by enormous glaciers and sheets of ice, and the higher portions of them were left in the condition of bare rocky surfaces smoothed and rounded by the passage of ice, all the loose rock and soil having been carried down to lower levels. These rocky surfaces are now covered with lakes and tarns, lying not only in the course of the valleys, but scattered on the slopes of hills and on the backs of broad ridges.

Such rock-tarns have been attributed by Sir A. Ramsay and other geologists to the direct action of ice, without due consideration of the probable previous condition of the surface, and ice has consequently been credited with a power of excavation far in excess of that which it really possesses. No one can doubt that the present surface of these districts has been swept and modelled by land-ice, but it does not follow that the hollows or rock-basins in which the lakes lie were actually formed by the ice. In the author's opinion the hollows, or most of them, were pre-existent, though filled with deposits or with decomposed rock; the ice only acted as a denudant, using that term in its proper sense of laying bare a previously formed surface: it was not the chisel which formed the surface, but only the rasper which has scraped it clean. To this point we shall recur in the sequel. We are now dealing only with those lakes which do not lie in valleys.

The extraordinary number of such lakes and tarns in certain parts of Scotland and Scandinavia ought to have suggested that some previous agency had been at work. We have Sir A. Geikie's testimony that in the north-west of Scotland "the surface of the Archæan gneiss is so thickly sprinkled with them that many tracts consist almost as much of water as of land," and in illustration he gives a map of part of the island of Lewis,' remarking also that the tarns occur not only in the lines of drainage, but are, as it were, scattered broadcast over the land.

1 "Scenery of Scotland," second edition, p. 240.