PART II.

A MORE EXTENDED INVESTIGATION REGARDING

THE EARTH'S SURFACE.

It is the province of Hydrography to deal with the oceanic world; Geography proper concerns itself simply with solid forms. The Hydrography of the globe we must pass over, however. Apart from the fact that it would lead us into studies of the most extended nature, it forms strictly one department of nautical science. sides, there is the less occasion to speak of it here at length, that works of great excellence have been published relating to this branch. We turn therefore to the land, and shall study the world of waters only so far as it exerts influence on the land.

Be

By land we mean the islands as well as the continents, for, as remarked before, the difference between them is merely relative. To the land division of the globe, however, belong all rivers and the internal fresh-water lakes, however large. The basis of difference does not lie in the fact that one part of the globe is water, the other part land, but in the fact that one is a tract of uniform evenness, the other of constantly-varying surface-the internal rivers and lakes being only frills, so to speak, to the elevated region, and not sharing the sea-level of the great

oceanic mass. Uniformity of surface is, then, the chief characteristic of the sea; a want of it, of the land. A mathematical level is a thing unknown on extended districts, and an approximation to it is very rare. Even the

The

basins of former seas do not display a perfectly level bed. The plains of North Germany are characterised by their billowy rolling. The flats along the Danube in Hungary, and along the Po in North Italy, have really important deviations from a true level, though the eye is not able to discern them. Milan is four hundred feet above the Adriatic; but the eye does not discern that it is not at the centre of a plain as perfect as the surface of the sea itself, and yet that plain does shelve gradually away till the Adriatic checks and defines it. Pesth is two hundred and fifteen feet above the ocean-level, yet the gradual decline to the Black Sea is undiscernible to the eye. immense plains along the Amazon-even the celebrated llanos on the Orinoco, which Alexander von Humboldt likens to inland seas of verdure-have a not insignificant slope from west to east. The middle point of these llanos, near the city of Calabozo, about 100 German geographical miles from the sea, he found to be 180 feet above the sealevel-far lower, indeed, than Milan or Pesth, relatively, yet at a perceptible elevation. All of these plains were once the bottom of the sea; the Adriatic laved the base of the Apennines and the Cottian Alps, and the Atlantic swept westward over the llanos of the Orinoco and the Essequibo, having the Sierra de Venezuela on the north and the Sierra Parima on the south, till it was checked by the Cordilleras of Merida and Pamplona.

Depression and elevation, then, are the characteristics of the land. They are both measured from the level of the sea; their absolute altitude is reckoned from the

imaginary sea-level, extended over the whole globe. Their mutual relations to each other are determined from their relative heights. The absolute elevation above the level of the ocean can be determined in a number of ways. If the heights to be measured are in the immediate vicinity of the sea, a simple system of triangulation will effect it. If they are removed from the sea, the difficulties are greater, and increase according to the distance from the sea. The heights of great inland mountains are determined by complicated operations with the spirit-level, protracted trigonometrical calculations, the unwearied and skilful use of the barometer, and constant appeal to the boiling-point of water. The description of these methods falls within the province of Physics.

As the determination of the heights of the loftiest mountains could not be made before the appointments of scientific explorers had attained to a certain degree of accuracy and delicacy, the knowledge of them in former times was almost wholly relative. The inquiries of La Condamine, Saussure, and de Luc, in the Andes and the Swiss Alps, are almost the only ones to be trusted among those of the older observers. All unscientific travellers, without accurate instruments, confounded absolute heights with relative, and innumerable errors crept therefore into the earlier text-books. It is only within the most recent times that Hypsometry has attained to the dignity of a science.

To meet and counteract the errors alluded to above, and current in the loose language of popular speech, we shall use a new and indeed arbitrary terminology-arbitrary because the data which mensuration will sometimes furnish are now, in part, wanting. We will divide the earth, not relatively but absolutely, into highlands and

lowlands. The great districts often met, whose elevations are very moderate, we call Lowlands; they are, for the most part, immense plains, varying but little above the level of the sea. The great districts which enclose mountain-ranges we call Highlands, and sometimes plateaus. True highlands can often embrace very extended and elevated plains, and these plains again may include hills and mountains. This does not affect their character as highlands, which lies in the fact of elevation rather than in more or less modified variation of surface. There may be vast variety in the physical manifestations of a great plateau district, entirely independent of the relative effects. produced by the distribution of its surface into plains, rolling land, hills, and mountains.

In the lowlands there may exist hills to some extent, and these may even be combined in ranges, provided only that they do not violate the uniform characteristics of the district in which they are found.

The highlands are generally met with in the interior of the continents; the lowlands at the coasts. Yet there are exceptions to this.

In the transitions from lowlands to highlands there is great diversity. We can speak of three distinct bases of discrimination: a sudden and abrupt ascent, a rise in elevation so gradual as scarcely to be perceived, and a terrace formation. Yet in these there is a blending of one variety with another; there is no line sharply defined, where we can say that one form ends and an other begins. There are constantly found modifications of these three transitional phases. The plains along the Indus and the Ganges rise sharply to the plateaus of Tibet. The flat Pacific coast of South America is exchanged with equal abruptness for the highlands of Peru.

The transition is a gradual one from the lowlands of North Germany, along the Baltic and the North Sea, through Saxony and Bohemia to the Bavarian highlands, north of the Alps. The Spanish highlands form a series of terraces, increasing in height from south to north. The immense plateaus of central Asia are also terrace formations, of diminishing elevation as they advance to Siberia; so, too, are the eastern plateaus of Peru, falling off in altitude toward the plains of the Amazon.

Just as varied are the heights taken from the sea-level of the leading plateaus. Yet they never rise to a point of elevation comparable with those of isolated mountain peaks or ranges. These attain, in no insignificant numbers, the height of 24,000 feet, while some ascend thousands of feet beyond that. In Mount Everest, of the Himalaya chain, the loftiest summit yet measured (29,000 feet) is found, although it may be that future investigations more to the south will disclose yet greater heights.

HIGHLANDS.

Continuous highlands or plateaus seldom attain an elevation greater than a half or a third of the loftiest mountains; the most elevated range in altitude from 8000 to 12,000 feet above the sea-level. On an average, they lie about from 4000 to 5000 feet above the sea. We take the last height as a convenient point of demarcation between the two classes of highlands-those of the first and those of the second magnitude. It is an arbitrary point, of course, and the division there must remain without a natural base to rest upon, till more results in Hypsometry shall have determined the real point of average between the