The different quarters of the world have been supposed to be characterized by the remark, that Europe has its heaths, Asia its steppes, Africa its deserts, and America its savannahs; but by this distinction, contrasts are established that are not founded either on the nature of things, or the genius of languages. The existence of a heath always supposes an association of plants of the family of ericae; the steppes of Asia are not everywhere covered with saline plants; the savannahs of Venezuela furnish not only the gramina, but with them small herbaceous mimosas, legumina, and other dicotyledonous plants. The plains of Songaria, those which extend between the Don and the Volga, and the puszta of Hungary, are real savannahs, pasturages abounding in grasses;* (* These vast steppes of Hungary are elevated only thirty or forty toises above the level of the sea, which is more than eighty leagues distant from them. See Wahlenberg's Flora Carpathianica. Baron Podmanitzky, an Hungarian nobleman, highly distinguished for his knowledge of the physical sciences, caused the level of these plains to be taken, to facilitate the formation of a canal then projected between the Danube and the Theiss. He found the line of division, or the convexity of the ground, which slopes on each side towards the beds of the two rivers, to be only thirteen toises above the height of the Danube. The widely extended pastures, which reach in every direction to the horizon, are called in the country, Puszta, and, over a distance of many leagues, are without any human habitation. Plains of this kind, intermingled with marshes and sandy tracts, are found on the western side of the Theiss, between Czegled, Csaba, Komloss, and Szarwass; and on the eastern side, between Debreczin, Karczag, and Szoboszlo. The area of these plains of the interior basin of Hungary has been estimated, by a pretty accurate calculation, to be between two thousand five hundred and three thousand square leagues (twenty to a degree). Between Czegled, Szolnok, and Ketskemet, the plain resembles a sea of sand.) while the savannahs to the east and west of the Rocky Mountains and of New Mexico produce chenopodiums containing carbonate and muriate of soda. Asia has real deserts destitute of vegetation, in Arabia, in Gobi, and in Persia. Since we have become better acquainted with the deserts in the interior of Africa, so long and so vaguely confounded together under the name of desert of Sahara (Zahra); it has been observed, that in this continent, towards the east, savannahs and pastures are found, as in Arabia, situated in the midst of naked and barren tracts. It is these deserts, covered with gravel and destitute of plants, which are almost entirely wanting in the New World. I saw them only in that part of Peru, between Amotape and Coquimbo, on the shores of the Pacific. These are called by the Spaniards, not llanos, but the desiertos of Sechura and Atacamez. This solitary tract is not broad, but it is four hundred and forty leagues long. The rock pierces everywhere through the quicksands. No drop of rain ever falls on it; and, like the desert of Sahara, north of Timbuctoo, the Peruvian desert affords, near Huaura, a rich mine of native salt. Everywhere else, in the New World, there are plains desert because not inhabited, but no real deserts.* (* We are almost tempted, however, to give the name of desert to that vast and sandy table-land of Brazil, the Campos dos Parecis, which gives birth to the rivers Tapajos, Paraguay, and Madeira, and which reaches the summit of the highest mountains. Almost destitute of vegetation, it reminds us of Gobi, in Mongolia.)
The same phenomena are repeated in the most distant regions; and, instead of designating those vast treeless plains in accordance with the nature of the plants they produce, it seems natural to class them into deserts, steppes, or savannahs; into bare lands without any appearance of vegetation, and lands covered with gramina or small plants of the dicotyledonous tribe. The savannahs of America, especially those of the temperate zone, have in many works been designated by the French term prairies; but this appears to me little applicable to pastures which are often very dry, though covered with grass of four or five feet in height. The Llanos and the Pampas of South America are really steppes. They are covered with beautiful verdure in the rainy season, but in the time of great drought they assume the aspect of a desert. The grass is then reduced to powder; the earth cracks; the alligators and the great serpents remain buried in the dried mud, till awakened from their long lethargy by the first showers of spring. These phenomena are observed on barren tracts of fifty or sixty leagues in length, wherever the savannahs are not traversed by rivers; for on the borders of rivulets, and around little pools of stagnant water, the traveller finds at certain distances, even during the period of the great droughts, thickets of mauritia, a palm, the leaves of which spread out like a fan, and preserve a brilliant verdure.
The steppes of Asia are all beyond the tropics, and form very elevated table-lands. America also has savannahs of considerable extent on the backs of the mountains of Mexico, Peru, and Quito; but its most extensive steppes, the Llanos of Cumana, Caracas, and Meta, are little raised above the level of the ocean, and all belong to the equinoctial zone. These circumstances give them a peculiar character. They have not, like the steppes of southern Asia, and the deserts of Persia, those lakes without issue, those small systems of rivers which lose themselves either in the sands, or by subterranean filtrations. The Llanos of America incline to the east and south; and their running waters are branches of the Orinoco.
The course of these rivers once led me to believe, that the plains formed table-lands, raised at least from one hundred to one hundred and fifty toises above the level of the ocean. I supposed that the deserts of interior Africa were also at a considerable height; and that they rose one above another as in tiers, from the coast to the interior of the continent. No barometer has yet been carried into the Sahara. With respect to the Llanos of America, I found by barometric heights observed at Calabozo, at the Villa del Pao, and at the mouth of the Meta, that their height is only forty or fifty toises above the level of the sea. The fall of the rivers is extremely gentle, often nearly imperceptible; and therefore the least wind, or the swelling of the Orinoco, causes a reflux in those rivers that flow into it. The Indians believe themselves to be descending during a whole day, when navigating from the mouths of these rivers to their sources. The descending waters are separated from those that flow back by a great body of stagnant water, in which, the equilibrium being disturbed, whirlpools are formed very dangerous for boats.
The chief characteristic of the savannahs or steppes of South America is the absolute want of hills and inequalities—the perfect level of every part of the soil. Accordingly the Spanish conquerors, who first penetrated from Coro to the banks of the Apure, did not call them deserts or savannahs, or meadows, but plains (llanos). Often within a distance of thirty square leagues there is not an eminence of a foot high. This resemblance to the surface of the sea strikes the imagination most powerfully where the plains are altogether destitute of palm-trees; and where the mountains of the shore and of the Orinoco are so distant that they cannot be seen, as in the Mesa de Pavones. A person would be tempted there to take the altitude of the sun with a quadrant, if the horizon of the land were not constantly misty on account of the variable effects of refraction. This equality of surface is still more perfect in the meridian of Calabozo, than towards the east, between Cari, La Villa del Pao, and Nueva Barcelona; but it extends without interruption from the mouths of the Orinoco to La Villa de Araure and to Ospinos, on a parallel of a hundred and eighty leagues in length; and from San Carlos to the savannahs of Caqueat, on a meridian of two hundred leagues. It particularly characterises the New Continent, as it does the low steppes of Asia, between the Borysthenes and the Volga, between the Irtish and the Obi. The deserts of central Africa, of Arabia, Syria, and Persia, Gobi, and Casna, present, on the contrary, many inequalities, ranges of hills, ravines without water, and rocks which pierce the sands.
The Llanos, however, notwithstanding the apparent uniformity of their surface, present two kinds of inequalities, which cannot escape the observation of the traveller. The first is known by the name of banks (bancos); they are in reality shoals in the basin of the steppes, fractured strata of sandstone, or compact limestone, standing four or five feet higher than the rest of the plain. These banks are sometimes three or four leagues in length; they are entirely smooth, with a horizontal surface; their existence is perceived only by examining their margins. The second species of inequality can be recognised only by geodesical or barometric levellings, or by the course of rivers. It is called a mesa or table, and is composed of small flats, or rather convex eminences, that rise insensibly to the height of a few toises. Such are, towards the east, in the province of Cumana, on the north of the Villa de la Merced and Candelaria, the Mesas of Amana, of Guanipa, and of Jonoro, the direction of which is south-west and north-east; and which, in spite of their inconsiderable elevation, divide the waters between the Orinoco and the northern coast of Terra Firma. The convexity of the savannah alone occasions this partition: we there find the dividing of the waters (divortia aquarum* (* "C. Manlium prope jugis [Tauri] ad divortia aquarum castra posuisse." Livy lib. 38 c. 75.)), as in Poland, where, far from the Carpathian mountains, the plain itself divides the waters between the Baltic and the Black Sea. Geographers, who suppose the existence of a chain of mountains wherever there is a line of division, have not failed to mark one in the maps, at the sources of the Rio Neveri, the Unare, the Guarapiche, and the Pao. Thus the priests of Mongol race, according to ancient and superstitious custom, erect oboes, or little mounds of stone, on every point where the rivers flow in an opposite direction.
The uniform landscape of the Llanos; the extremely small number of their inhabitants; the fatigue of travelling beneath a burning sky, and an atmosphere darkened by dust; the view of that horizon, which seems for ever to fly before us; those lonely trunks of palm-trees, which have all the same aspect, and which we despair of reaching, because they are confounded with other trunks that rise by degrees on the visual horizon; all these causes combine to make the steppes appear far more extensive than they are in reality. The planters who inhabit the southern declivity of the chain of the coast see the steppes extend towards the south, as far as the eye can reach, like an ocean of verdure. They know that from the Delta of the Orinoco to the province of Varinas, and thence, by traversing the banks of the Meta, the Guaviare, and the Caguan, they can advance three hundred and eighty leagues* (* This is the distance from Timbuctoo to the northern coast of Africa.) into the plains, first from east to west, and then from north-east to south-east beyond the Equator, to the foot of the Andes of Pasto. They know by the accounts of travellers the Pampas of Buenos Ayres, which are also Llanos covered with fine grass, destitute of trees, and filled with oxen and horses become wild. They suppose that, according to the greater part of our maps of America, this continent has only one chain of mountains, that of the Andes, which stretches from south to north; and they form a vague idea of the contiguity of all the plains from the Orinoco and the Apure to the Rio de la Plata and the Straits of Magellan.
Without stopping here to give a mineralogical description of the transverse chains which divide America from east to west, it will be sufficient to notice the general structure of a continent, the extremities of which, though situated in climates little analogous, nevertheless present several features of resemblance. In order to have an exact idea of the plains, their configuration, and their limits, we must know the chains of mountains that form their boundaries. We have already described the Cordillera of the coast, of which the highest summit is the Silla de Caraccas, and which is linked by the Paramo de las Rosas to the Nevada de Merida, and the Andes of New Grenada. We have seen that, in the tenth degree of north latitude, it stretches from Quibor and Barquesimeto as far as the point of Paria. A second chain of mountains, or rather a less elevated but much larger group, extends between the parallels of 3 and 7 degrees from the mouths of the Guaviare and the Meta to the sources of the Orinoco, the Marony, and the Essequibo, towards French and Dutch Guiana. I call this chain the Cordillera of Parime, or of the great cataracts of the Orinoco. It may be followed for a length of two hundred and fifty leagues; but it is less a chain, than a collection of granitic mountains, separated by small plains, without being everywhere disposed in lines. The group of the mountains of Parime narrows considerably between the sources of the Orinoco and the mountains of Demerara, in the Sierras of Quimiropaca and Pacaraimo, which divide the waters between the Carony and the Rio Parime, or Rio de Aguas Blancas. This is the scene of the expeditions which were undertaken in search of El Dorado, and the great city of Manoa, the Timbuctoo of the New Continent. The Cordillera of Parime does not join the Andes of New Grenada, but is separated from them by a space eighty leagues broad. If we suppose it to have been destroyed in this space by some great revolution of the globe (which is scarcely probable) we must admit that it anciently branched off from the Andes between Santa Fe de Bogota and Pamplona. This remark serves to fix more easily in the memory of the reader the geographical position of a Cordillera till now very imperfectly known. A third chain of mountains unites in 16 and 18 degrees south latitude (by Santa Cruz de la Sierra, the Serranias of Aguapehy, and the famous Campos dos Parecis) the Andes of Peru, to the mountains of Brazil. It is the Cordillera of Chiquitos which widens in the Capitania de Minas Geraes, and divides the rivers flowing into the Amazon from those of the Rio de la Plata,* (* There is only a portage or carrying-place of 5322 bracas between the Guapore (a branch of the Marmore and of the Madeira), and the Rio Aguapehy (a branch of the Jaura and of the Paraguay).) not only in the interior of the country, in the meridian of Villa Boa, but also at a few leagues from the coast, between Rio Janeiro and Bahia.* (* The Cordillera of Chiquitos and of Brazil stretches toward the south-east, in the government of the Rio Grande, beyond the latitude of 30 degrees south.)
These three transverse chains, or rather these three groups of mountains stretching from west to east, within the limits of the torrid zone, are separated by tracts entirely level, the plains of Caracas, or of the Lower Orinoco; the plains of the Amazon and the Rio Negro; and the plains of Buenos Ayres, or of La Plata. I use the term plains, because the Lower Orinoco and the Amazon, far from flowing in a valley, form but a little furrow in the midst of a vast level. The two basins, placed at the extremities of South America, are savannahs or steppes, pasturage without trees; the intermediate basin, which receives the equatorial rains during the whole year, is almost entirely one vast forest, through which no other roads are known save the rivers. The strong vegetation which conceals the soil, renders also the uniformity of its level less perceptible; and the plains of Caracas and La Plata bear no other name. The three basins we have just described are called, in the language of the colonists, the Llanos of Varinas and of Caracas, the bosques or selvas (forests) of the Amazon, and the Pampas of Buenos Ayres. The trees not only for the most part cover the plains of the Amazon, from the Cordillera de Chiquitos, as far as that of Parime; they also crown these two chains of mountains, which rarely attain the height of the Pyrenees.* (* We must except the most western part of the Cordillera of Chiquitos, between Cochabamba and Santa Cruz de la Sierra where the summits are covered with snow; but this colossal group almost belongs to the Andes de la Paz, of which it forms a promontory or spur, directed toward the east.) On this account, the vast plains of the Amazon, the Madeira, and the Rio Negro, are not so distinctly bounded as the Llanos of Caracas, and the Pampas of Buenos Ayres. As the region of forests comprises at once the plains and the mountains, it extends from 18 degrees south to 7 and 8 degrees north,* (* To the west, in consequence of the Llanos of Manso, and the Pampas de Huanacos, the forests do not extend generally beyond the parallels of 18 or 19 degrees south latitude; but to the east, in Brazil (in the capitanias of San Pablo and Rio Grande) as well as in Paraguay, on the borders of the Parana, they advance as far as 25 degrees south.) and occupies an extent of near a hundred and twenty thousand square leagues. This forest of South America, for in fact there is only one, is six times larger than France. It is known to Europeans only on the shores of a few rivers, by which it is traversed; and has its openings, the extent of which is in proportion to that of the forests. We shall soon skirt the marshy savannahs, between the Upper Orinoco, the Conorichite, and the Cassiquiare, in the latitude of 3 and 4 degrees. There are other openings, or as they are called, clear savannahs,* (* Savannas limpias, that is to say, clear of trees.) in the same parallel, between the sources of the Mao and the Rio de Aguas Blancas, south of the Sierra de Pacaraima. These last savannahs, which are inhabited by Caribs, and nomad Macusis, lie near the frontiers of Dutch and French Guiana.
Having noticed the geological constitution of South America, we shall now mark its principal features. The western coasts are bordered by an enormous wall of mountains, rich in precious metals wherever volcanic fire has not pierced through the eternal snow. This is the Cordillera of the Andes. Summits of trap-porphyry rise beyond three thousand three hundred toises, and the mean height of the chain* is one thousand eight hundred and fifty toises. (* In New Grenada, Quito, and Peru, according to measurements taken by Bouguer, La Condamine, and myself.) It stretches in the direction of a meridian, and sends into each hemisphere a lateral branch, in the latitudes of 10 degrees north, and 16 and 18 degrees south. The first of these two branches, that of the coast of Caracas, is of considerable length, and forms in fact a chain. The second branch, the Cordillera of Chiquitos and of the sources of the Guapore, is very rich in gold, and widens toward the east, in Brazil, into vast tablelands, having a mild and temperate climate. Between these two transverse chains, contiguous to the Andes, an isolated group of granitic mountains is situated, from 3 to 7 degrees north latitude; which also runs parallel to the Equator, but, not passing the meridian of 71 degrees, terminates abruptly towards the west, and is not united to the Andes of New Grenada. These three transverse chains have no active volcanoes; we know not whether the most southern, like the two others, be destitute of trachytes or trap-porphyry. None of their summits enter the limit of perpetual snow; and the mean height of the Cordillera of La Parime, and of the littoral chain of Caracas, does not reach six hundred toises, though some of its summits rise fourteen hundred toises above the level of the sea.* (* We do not reckon here, as belonging to the chain of the coast, the Nevados and Paramos of Merida and of Truxillo, which are a prolongation of the Andes of New Grenada.) The three transverse chains are separated by plains entirely closed towards the west, and open towards the east and south-east. When we reflect on their small elevation above the surface of the ocean, we are tempted to consider them as gulfs stretching in the direction of the current of rotation. If, from the effect of some peculiar attraction, the waters of the Atlantic were to rise fifty toises at the mouth of the Orinoco, and two hundred toises at the mouth of the Amazon, the flood would submerge more than the half of South America. The eastern declivity, or the foot of the Andes, now six hundred leagues distant from the coast of Brazil, would become a shore beaten by the waves. This consideration is the result of a barometric measurement, taken in the province of Jaen de Bracamoros, where the river Amazon issues from the Cordilleras. I found the mean height of this immense river only one hundred and ninety-four toises above the present level of the Atlantic. The intermediate plains, however, covered with forests, are still five times higher than the Pampas of Buenos Ayres, and the grass-covered Llanos of Caracas and the Meta.