THE PHYSICAL GEOGRAPHY AND GEOLOGY OF THE AMAZON VALLEY.

The basin of the Amazon surpasses in dimensions that of any other river in the world. It is entirely situated in the Tropics, on both sides of the Equator, and receives over its whole extent the most abundant rains. The body of fresh water emptied by it into the ocean, is therefore far greater than that of any other river; not only absolutely, but probably also relatively to its area, for as it is almost entirely covered by dense virgin forests, the heavy rains which penetrate them do not suffer so much evaporation as when they fall on the scorched Llanos of the Orinooko or the treeless Pampas of La Plata. For richness of vegetable productions and universal fertility of soil it is unequalled on the globe, and offers to our notice a natural region capable of supporting a greater population, and supplying it more completely with the necessaries and luxuries of life, than any other of equal extent. Of this wonderful district we will now describe the principal physical peculiarities.

From about 4° north latitude to 20° south, every stream that flows down the eastern slope of the Andes is a tributary of the Amazon. This is as if every river, from St. Petersburg to Madrid, united their waters into one mighty flood.

The MaraÑon, which is generally considered the main stream of the Amazon, deserves that title on several accounts. It rises to the westward of all the other great tributaries, and it receives all the waters which flow nearest to the Pacific, and most remote in a direct line from the mouth of the river. It flows for a considerable distance in the most westerly valley of the Andes, separated by one range only from the Pacific, and at the point where it breaks through the eastern chain of the Andes, in 78° west longitude, is already a large river, on a meridian where all the other streams which can lay a claim to be considered the head-waters of the Amazon have as yet no existence. On going up the Amazon from its mouth, it is that branch on which you can keep longest in the general east and west direction of the river; and if the actual length of its course is considered, it still keeps its place, for I find that there is not more than ten or twenty miles' difference between it and the Uaycali, reckoning to the most distant source of the latter; and its course is at present so uncertain, that future surveys may increase or diminish it considerably.

These considerations, I think, decide the question as to the propriety of considering the MaraÑon as the true source of the Amazon. We find that from its origin in the Lake Lauricocha, to its mouth in longitude 50° west, in length, following the main curves, but disregarding the minuter windings, is 2,740 English miles.

Its extent, in a straight line from east to west, is about 2,050 miles; and from north to south, its tributary streams cover a space of 1,720 miles.

The whole area of its basin, not including that of the Tocantins, which I consider a distinct river, is 2,330,000 English square miles, or 1,760,000 nautical square miles. This is more than a third of all South America, and equal to two-thirds of all Europe. All western Europe could be placed in it without touching its boundaries, and it would even contain our whole Indian empire.

The numerous tributary streams of the Amazon, many of them equal to the largest rivers of Europe, differ remarkably in the colour of their waters, the character of the vegetation on their banks, and the animals that inhabit them. They may be divided into three groups,—the white-water rivers, the blue-water rivers, and the black-water rivers.

The main stream of the Amazon itself is a white-water river, this name being applied to those waters which are of a pale yellowish olive-colour. This colour does not seem to depend entirely on free earthy matter, but rather on some colouring material held in solution; for in lakes and inlets, where the waters are undisturbed and can deposit all their sediment, they still retain the colour.

The waters of the Amazon continue of the same colour up to the mouth of the Uaycali, when they become blue or transparent, and the white waters are extended up that branch.

This has been taken as an evidence of the UaycÁli being the main stream of the Amazon; but I cannot consider that it has anything to do with the question. It is evident that if equal quantities of clear and muddy water are mixed together, the result will differ very little from the latter in colour, and if the clear water is considerably more in quantity the resulting mixture will still be muddy. But the difference of colour between the white-and blue-water rivers, is evidently owing to the nature of the country they flow through: a rocky and sandy district will always have clear-water rivers; an alluvial or clayey one, will have yellow or olive-coloured streams. A river may therefore rise in a rocky district, and after some time flow through an alluvial basin, where the water will of course change its colour, quite independently of any tributaries which may enter it near the junctions of the two formations.

The IÇa and JapurÁ have waters very similar in colour to the Amazon. The Rio Branco, a branch of the Rio Negro from the north, is remarkable for its peculiar colour: till I saw it, I had not believed it so well deserved its name. The Indians and traders had always told me that it was really white, much more so than the Amazon; and on descending the Rio Negro in 1852 I passed its mouth, and found that its waters were of a milky colour mixed with olive. It seemed as if it had a quantity of chalk in solution, and I have little doubt of there being on its banks considerable beds of the pure white clay which occurs in many parts of the Amazon, and which helps to give the waters their peculiar whiteness. The Madeira and PurÚs have also white waters in the wet season, when their powerful currents bring down the alluvial soil from their banks; but in the dry season they are a dark transparent brownish-olive.

All the rivers that rise in the mountains of Brazil have blue or clear water. The Tocantins, the XingÚ, and the TapajÓz, are the chief of this class. The Tocantins runs over volcanic and crystalline rocks in the lower parts of its course, and its waters are beautifully transparent; the tide, however, enters for some miles, and renders it turbid, as also the XingÚ. The TapajÓz, which enters the Amazon about five hundred miles above ParÁ, is clear to its mouth, and forms a striking contrast to the yellow flood of that river.

It is above the Madeira that we first meet with the curious phenomenon of great rivers of black water. The Rio Negro is the largest and most celebrated of these. It rises in about 2° 30´ N. lat., and its waters are there much blacker than in the lower part of its course. All its upper tributaries, the smaller ones especially, are very dark, and, where they run over white sand, give it the appearance of gold, from the rich colour of the water, which, when deep, appears inky black. The small streams which rise in the same district, and flow into the Orinooko, are of the same dark colour. The Cassiquiare first pours in some white or olive-coloured water. Lower down, the CababurÍs, MaravihÁ, and some smaller white-water streams help to dilute it, and then the Rio Branco adds its flood of milky water. Notwithstanding all this, the Rio Negro at its mouth still appears as black as ink; only in shallow water it is seen to be paler than it is up above, and the sands are not dyed of that pure golden tint so remarkable there.

On the south of the Amazon there are also some black-water streams—the Coary, the Teffe, the JuruÁ, and some others. The inhabitants have taken advantage of these, to escape from the plague of the mosquitoes, and the towns of Coary and Ega are places of refuge for the traveller on the Upper Amazon, those annoying insects being scarcely ever found on the black waters. The causes of the peculiar colour of these rivers are not, I think, very obscure; it appears to me to be produced by the solution of decaying leaves, roots, and other vegetable matter. In the virgin forests, in which most of these streams have their source, the little brooks and rivulets are half choked up with dead leaves and rotten branches, giving various brown tinges to the water. When these rivulets meet together and accumulate into a river, they of course have a deep brown hue, very similar to that of our bog or peat water, if there are no other circumstances to modify it. But if the streams flow through a district of soft alluvial clay, the colour will of course be modified, and the brown completely overpowered; and I think this will account for the anomalies observed, of streams in the same districts being of different colours. Those whose sources are pretty well known are seen to agree with this view. The Rio Negro, the Atabapo, the Isanna, and several other smaller rivers, have their sources and their whole course in the deep forest; they flow generally over clean granite rocks and beds of sand, and their streams are gentle, so as not to wear away the soft parts of their banks.

The IÇa, JapurÁ, and Upper Amazon, on the contrary, flow through a long extent of alluvial country, and, having their sources on the slopes of the Andes, are much more liable to sudden floods, and by their greater velocity bring down a quantity of sediment. In fact, it seems clear, that a thorough knowledge of the course of each river would enable us to trace the colour of its waters to the various peculiarities of the country through which it flows.

With the exception of the streams rising in the Andes, the boundaries of the Amazon basin, or the most distant sources of its tributaries on the north and south, are comparatively little elevated above the level of the sea. The whole basin, with the exception of a very small portion, is one great plain of the most perfect and regular character.

The true altitude of the source in the Lake Lauricocha has not been ascertained. At Tomependa Humboldt states it to be 1,320 feet above the sea: this is as near as possible 2,000 miles in a straight line from the mouth; so that the average rise is only eight inches in a mile. But if we take the height at Tabatinga, on the boundary of Brazil, which, according to Spix and Martius is 670 feet, we shall find, the distance being about 1,400 miles, that the rise is only five and a half inches per mile. If we had the height of Barra do Rio Negro accurately, we should no doubt find the rise to that point not more than two or three inches in a mile. The distance is, in a straight line, about 700 miles, and we may therefore probably estimate the height at less than 200, and perhaps not more than 150 feet.

This height I am inclined to believe quite great enough, from some observations I made with an accurate thermometer, reading to tenths of a degree, on the temperature of boiling water. This instrument I received from England, after leaving ParÁ. The mean of five observations at Barra, some with river and some with rain-water, gave 212.5° as the temperature of boiling water; a remarkable result, showing that the barometer must stand there at more than thirty inches, and that unless it is, in the months of May and August, considerably more than that at the sea-level, Barra can be but very little elevated above the sea.

For the height of the country about the sources of the Rio Negro, Humboldt is our only authority. He gives 812 feet as the height of SÃo Carlos; he, however, states that the determination is uncertain, owing to an accident happening to the barometer; I may, therefore, though with great diffidence, venture to doubt the result. The distance, in a straight line, from the mouth of the Rio Negro to SÃo Carlos, is rather less than from the same point to Tabatinga, whose height is 670 feet. The current, however, from Tabatinga is much more rapid than down the Rio Negro, the lower part of which has so little fall, that in the month of January, when the Amazon begins to rise, the water enters the mouth of the Rio Negro, and renders that river stagnant for several hundred miles up. The falls of the Rio Negro I cannot consider to add more than fifty feet to the elevation, as above and below them the river is not very rapid. Thus, from this circumstance alone, we should be disposed to place SÃo Carlos at a rather less elevation than Tabatinga, or at about 600 feet. My observations up the Rio Negro gave consistent results. At Castanheiro, about five hundred miles up, the temperature of boiling water was 212·4°, at the mouth of the UaupÉs 212·2°, and at a point just below SÃo Carlos, 212·0°. This would not give more than 250 feet for the height of SÃo Carlos above Barra; and, as we have estimated this at 200 feet above the sea, the height of SÃo Carlos will become 450 feet, which I think will not be found far from the truth.

The velocity of the current varies with the width of the stream and the time of the year; we have little accurate information on this subject. In a Brazilian work on the Province of ParÁ, the Madeira is stated to flow 2,970 braÇas, or about three and a half miles, an hour in the wet season. At Obidos I made an observation in the month of November, when the Amazon is at the lowest level, and found it four miles an hour; but this by no means represents the current in the rainy season. On descending to ParÁ, in the month of June, 1852, I found that we often floated down about five miles an hour, and as the wind was strong directly up the river, it probably retarded us, rather than helped us on, our vessel not being rigged in the best manner. Martius calculates that 500,000 cubic feet of water per second pass Obidos. This agrees pretty well with my own calculations of the quantity in the dry season; when the river is full, it is probably much greater. If we suppose, on a moderate calculation, that seventy-two inches, or six feet, of rain fall annually over the whole Amazon valley, it will amount to 1,500,000 cubic feet per second, the whole of which must either evaporate, or flow out of the mouth of the Amazon; so that if we increase the amount given by Martius by one half, to take in the lower part of the Amazon and to allow for the whole year, we shall have the evaporation as one half of the rain falling annually.

It is a fact which has been frequently stated, and which seems fully established, that the Amazon carries its fresh waters out into the ocean, which it discolours for a distance of a hundred and fifty miles from its mouth. It is also generally stated that the tide flows up the river as far as Obidos, five hundred miles from the mouth. These two statements appear irreconcilable, for it is not easy to understand how the tides can flow up to such a great distance, and yet no salt water enter the river. But the fact appears to be, that the tide never does flow up the river at all. The water of the Amazon rises, but during the flood as well as the ebb the current is running rapidly down. This takes place even at the very mouth of the river, for at the island of Mexiana, exposed to the open sea, the water is always quite fresh, and is used for drinking all the year round. But as salt water is heavier than fresh, it might flow up at the bottom, while the river continued to pour down above it; though it is difficult to conceive how this could take place to any extent without some salt water appearing at the margins.

The rising of the water so far up the river can easily be explained, and goes to prove also that the slope of the river up to where the tide has any influence cannot be great; for as the waters of the ocean rose, the river would of course be banked up, the velocity of its current still forcing its waters onward; but it is not easy to see how the stream could be thus elevated to a higher level than the waters of the ocean which caused the rise, and we should therefore suppose that at Obidos, where the tidal rise ceases to be felt, the river is just higher than the surface of the ocean at the highest spring-tides. A somewhat similar phenomenon is seen at the mouth of the TapajÓz. Here, at the end of the dry season, there is but a small body of water, and the current is very sluggish. The Amazon, however, rises considerably with the tides, and its waters then become higher than those of the TapajÓz, and they therefore enter into that river and force it back; we then see the Amazon flowing rapidly down, at the same time that the TapajÓz is flowing up.

It seems to be still a disputed question among geographers, whether the ParÁ river is or is not a branch of the Amazon. From my own observation, I am decidedly of opinion that it is not: it appears to me to be merely the outlet of the TocantÍns and of numerous other small streams. The canal or channel of TagipurÚ, which connects it with the Amazon, and by which all the trade between ParÁ and the interior is carried on, is one of a complete network of channels, along which the tide ebbs and flows, so as in a great measure to disguise the true direction and velocity of its current. It seems probable that not a drop of Amazon water finds its way by this channel into the ParÁ river, and I ground my opinion upon the following facts.

It is well known, that in a tidal river the ebb-tide will continue longer than the flood, because the stream of the river requires to be overcome, and thus delays the commencement of the flood, while it facilitates that of the ebb. This is very remarkable in all the smaller rivers about ParÁ. Taking this as our guide, we shall be able to ascertain which way the current in the TagipurÚ sets, independently of the tide.

On my journey from ParÁ to the Amazon, our canoe could only proceed with the tide, having to wait moored to the bank while it was against us, so that we were of course anxious to find the time of our tedious stoppages diminished. Up to a certain point, we always had to wait more time than we were moving, showing that the current set against us and towards ParÁ; but after passing that point, where there was a bend, and several streams met, we had but a short time to wait, and a long ebb in our favour, showing that the current was with us or towards the Amazon, whereas it would evidently have been different had there been any permanent current flowing from the Amazon through the TagipurÚ towards ParÁ.

I therefore look upon the TagipurÚ as a channel formed by the small streams between the TocantÍns and XingÚ, meeting together about MelgÁco, and flowing through a low swampy country in two directions, towards the Amazon, and towards the ParÁ river.

At high tides the water becomes brackish, even up to the city of ParÁ, and a few miles down is quite salt. The tide flows very rapidly past ParÁ, up all the adjacent streams, and as far as the middle of the TagipurÚ channel; another proof that a very small portion, if any, of the Amazon water is there to oppose it.

The curious phenomenon of the bore, or "pirorÓco," in the rivers GuamÁ and MojÚ, I have described and endeavoured to explain in my Journal, and need not now repeat the account of it. (See page 89.)

Our knowledge of the courses of most of the tributaries of the Amazon is very imperfect. The main stream is tolerably well laid down in the maps as far as regards its general course and the most important bends; the details, however, are very incorrect. The numerous islands and parallel channels,—the great lakes and offsets,—the deep bays,—and the varying widths of the stream, are quite unknown. Even the French survey from ParÁ to Obidos, the only one which can lay claim to detailed accuracy, gives no idea of the river, because only one channel is laid down. I obtained at Santarem a manuscript map of the lower part of the river, much more correct than any other I have seen. It was, with most of my other papers, lost on my voyage home; but I hope to be able to obtain another copy from the same party. The Madeira and the Rio Negro are the only other branches of the Amazon whose courses are at all accurately known, and the maps of them are very deficient in anything like detail. The other great rivers, the Xingu, the TapajÓz, the PurÚs, CoarÍ, Teffe, JuruÁ, JutaÍ, JabarÍ, IÇÁ, JapurÁ, etc., though all inserted in our maps, are put in quite by guess, or from the vaguest information of the general direction of their course. Between the TocantÍns and the Madeira, and between the Madeira and the Uaycali, there are two tracts of country of five hundred thousand square miles each, or each twice as large as France, and as completely unexplored as the interior of Africa.

The Rio Negro is one of the most unknown in its characteristic features; although, as before stated, its general course is laid down with tolerable accuracy. I have narrated in my Journal how I was prevented from descending on the north side of it, and thus completing my survey of its course.

The most remarkable feature is the enormous width to which it spreads,—first, between Barra and the mouth of the Rio Branco, and from thence to near St. Isabel. In some places, I am convinced, it is between twenty and thirty miles wide, and, for a very great distance, fifteen to twenty. The sources of the rivers UaupÉs, Isanna, XiÉ, Rio Negro, and Guaviare, are very incorrectly laid down. The Serra Tunuhy is generally represented as a chain of hills cutting off these rivers; it is, however, a group of isolated granite peaks, about two thousand feet high, situated on the north bank of the river Isanna, in about 1° north latitude and 70° west longitude. The river rises considerably beyond them, in a flat forest-country, and further west than the Rio Negro, for there is a path across to the InirizÁ, a branch of the Guaviare which does not traverse any stream, so that the Rio Negro does not there exist.

My own journey up the UaupÉs extended to near 72° west longitude. Five days further in a small canoe, or about a hundred miles, is the Jurupari caxoeira, the last fall on the river. Above that, traders have been twelve days' journey on a still, almost currentless river, which, by the colour of its water, and the aspect of its vegetation, resembles the Upper Amazon. For all this distance, which must reach very nearly to the base of the Andes, the river flows through virgin forest. But the Indians in the upper part say there are campos, or plains, and cattle, further up; and they possess Spanish knives and other articles, showing that they have communications with the civilised inhabitants of the country to the east of BogotÁ.

I am therefore strongly inclined to believe that the rivers AriarÍ and others, rising about a hundred miles south of BogotÁ, are not, as shown in all our maps, the sources of the Guaviare, but of the UaupÉs, and that the basin of the Amazon must therefore be here extended to within sixty miles of the city of BogotÁ. This opinion is strengthened by information obtained from the Indians of JavÍta, who annually ascend the Guaviare to fish in the dry season, and who state that the river is very small, and in its upper part, where some hills occur and the forest ends, it is not more than a hundred yards wide; whereas the UaupÉs, at the furthest point the traders have reached, is still a large river, from a quarter of a mile to a mile in width.

The Amazon and all its branches are subject, like most tropical rivers, to an annual rise and fall of great regularity. In the main stream, and in all the branches which flow from the Andes, the waters begin to rise in December or January, when the rains generally commence, and continue rising till June, when the fine weather has just set in. The time when the waters begin to fall is about the 21st of June,—seldom deviating more than a few days from this date. In branches which have their sources in a different direction, such as the Rio Negro, the time of rising does not coincide. On that river the rains do not commence steadily till February or March, when the river rises with very great rapidity, and generally is quite filled by June, and then begins to fall with the Amazon. It thus happens that in the months of January and February, when the Amazon is rising rapidly, the Rio Negro is still falling in its upper part; the waters of the Amazon therefore flow into the mouth of the Rio Negro, causing that river to remain stagnant like a lake, or even occasionally to flow back towards its source. The total rise of the Amazon between high and low water mark has not been accurately ascertained, as it cannot be properly determined without a spirit-level; it is, however, certainly not less than forty, and probably often fifty feet. If therefore we consider the enormous water surface raised fifty feet annually, we shall gain from another point of view an idea of the immense quantity of water falling annually in the Amazon valley. We cannot take the length of the Amazon with its main tributaries at less than ten thousand miles, and their average width about two miles; so that there will be a surface of twenty thousand square miles of water, raised fifty feet every year. But it is not only this surface that is raised, for a great extent of land on the banks of all the rivers is flooded to a great depth at every time of high water. These flooded lands are called, in the language of the country, "gapÓ" and are one of the most singular features of the Amazon. Sometimes on one side, sometimes on both, to a distance of twenty or thirty miles from the main river, these gapÓs extend on the Amazon, and on portions of all its great branches. They are all covered with a dense virgin forest of lofty trees, whose stems are every year, during six months, from ten to forty feet under water. In this flooded forest the Indians have paths for their canoes, cutting across from one river to another, which are much used, to avoid the strong current of the main stream. From the mouth of the river Tapajoz to Coary, on the SolimÕes, a canoe can pass, without once entering the Amazon: the path lies across lakes, and among narrow inland channels, and through miles of dense flooded forest, crossing the Madeira, the Purus, and a hundred other smaller streams. All along, from the mouth of the Rio Negro to the mouth of the IÇa, is an immense extent of gapÓ, and it reaches also far up into the interior; for even near the sources of the Rio Negro, and on the upper waters of the UaupÉs, are extensive tracts of land which are annually overflowed.

In the whole country around the mouth of the Amazon, round the great island of MarajÓ, and about the mouths of the TocantÍns and XingÚ, the diurnal and semi-monthly tides are most felt, the annual rise and fall being almost lost. Here the low lands are overflowed at all the spring-tides, or every fortnight, subjecting all vegetation to another peculiar set of circumstances. Considerable tracts of land, still covered with vegetation, are so low, that they are flooded at every high water, and again vary the conditions of vegetable growth.

GEOLOGY.

Fully to elucidate the Geology of the Amazon valley, requires much more time and research than I was able to devote to it. The area is so vast, and the whole country being covered with forests renders natural sections so comparatively scarce, that the few distant observations one person can make will lead to no definite conclusions.

It is remarkable that I was never able to find any fossil remains whatever,—not even a shell, or a fragment of fossil wood, or anything that could lead to a conjecture as to the state in which the valley existed at any former period. We are thus unable to assign the geological age to which any of the various beds of rock belong.[4]

My notes, and a fine series of specimens of the rocks of the Rio Negro, were lost, and I have therefore very few materials to go upon.

Granite seems to be, in South America, more extensively developed than in any other part of the world. Darwin and Gardner found it in every part of the interior of Brazil, in La Plata, and Chile. Up the XingÚ Prince Adalbert met with it. Over the whole of Venezuela and New Granada, it was found by Humboldt. It seems to form all the mountains in the interior of Guiana, and it was met with by myself over the whole of the upper part of the Rio Negro, and far up the UaupÉs towards the Andes.

From what I could see of the granitic formation of the Upper Rio Negro, it appeared to be spread out in immense undulating areas, the hollows of which, being filled up with alluvial deposits, form those beds of earth and clay which occur, of various dimensions, everywhere in the midst of the granite formation. In these places grow the lofty virgin forests, while on the scantily covered granite rocks, and where beds of sand occur, are the more open catinga forests, so different in their aspect and peculiar in their vegetation. What strikes one most in this great formation, is its almost perfect flatness. There are no ranges of mountains, or even slightly elevated plateaus; all is level, except the abrupt peaks that rise suddenly from the plain, to a height of from one hundred to three thousand feet. In the Upper Rio Negro these peaks are very numerous. The first is the Serra de JacamÍ, a little above St. Isabel; it rises immediately from the bank of the river, on the south side, to a height of about six hundred feet. Several others are scattered about, but the Serras de CuricuriarÍ are the most lofty. They consist of a group of three or four mountains, rising abruptly to the height of near three thousand feet; towards their summits are immense precipices and jagged peaks. Higher up, on the same river, is another group of rather less height. On the UaupÉs are numerous hills, some conical, others dome-shaped, but all keeping the same character of abrupt elevations, quite independent of the general profile of the country. About the falls of the river UaupÉs there are small hills of granite, broken about in the greatest confusion. Great chasms or bowls occur, and slender pillars of rock rise above the surrounding forest like dead trunks of giant trees. Up the river Isanna, the Tunuhy mountains are a similar isolated group. The Cocof is a quadrangular or cubical mass, about a thousand feet in elevation, which forms the boundary between Brazil and Venezuela; and behind it are the PirapocÓ, and the serras of the CababurÍs, which seem rather more extensive, and form something more like a connected range of hills.

But the great peculiarity of them all is, that the country, does not perceptibly rise to their bases; they spring up abruptly, as if elevated by some local isolated force. I ascended one of the smaller of these serras as far as practicable, and have recorded my impressions of it in my Journal. (See page 153.)

The isolation and abrupt protrusion of these mountains is not, however, altogether without parallel in the Andes itself. This mighty range, from all the information I can obtain, rises with almost equal abruptness from an apparently level plain. The Andes of Quito, and southward to the Amazon, is like a huge rocky rampart, bounding the great plain which extends in one unbroken imperceptible slope from the Atlantic Ocean to its base. It is one of the grandest physical features of the earth,—this vast unbroken plain,—that mighty and precipitous mountain-range.

The granitic rocks of the Rio Negro in general contain very little mica; in some places, however, that mineral is abundant, and exists in large plates. Veins of pure quartz are common, some of very great size; and numerous veins or dykes of granite, of a different colour or texture. The direction of these is generally nearer east and west, than north and south.

Just below the falls of the Rio Negro are some coarse sandstone rocks, apparently protruding through the granite, dipping at an angle of 60° or 70° south-south-west. (Plate II. c.) Near the same place a large slab of granite rock exhibits quantities of curiously twisted or folded quartz veins (Plate II. b.), which vary in size from a line to some inches in diameter, and are folded in a most minute and regular manner.

On an island in the river, near this place, are finely stratified crystalline rocks, dipping south from 70° to vertical, and sometimes waved and twisted.

The granite often exhibits a concentric arrangement of laminÆ, particularly in the large dome-shaped masses in the bed of the river (Plate III. a, c), or in portions protruding from the ground (Plate III. b). Near SÃo Gabriel, and in the UaupÉs, large masses of pure quartz rock occur, and the shining white precipices of the serras are owing, I have no doubt, to the same cause. At Pimichin, near the source of the Rio Negro, the granite contains numerous fragments of stratified sandstone rock imbedded in it (Plate II. a); I did not notice this so distinctly at any other locality.

High up the river UaupÉs there is a very curious formation. All along the river-banks there are irregular fragments of rocks, with their interstices filled up with a substance that looks exactly like pitch. On examination, it is found to be a conglomerate of sand, clay, and scoriÆ, sometimes very hard, but often rotten and easily breaking to pieces; its position immediately suggests the idea of its having been liquid, for the fragments of rocks appear to have sunk in it.

Coarse volcanic scoriÆ, with a vitreous surface, are found over a very wide area. They occur at CaripÉ, near ParÁ,—above BaiÃo, in the TocantÍns,—at the mouth of the TapajÓz,—at Villa Nova, on the Amazon,—above Barra, on the Rio Negro, and again up the UaupÉs. A small conical hill behind the town of Santarem, at the mouth of the TapajÓz, has all the appearance of being a volcanic cone.

The neighbourhood of ParÁ consists entirely of a coarse iron sandstone, which is probably a continuation of the rocks observed by Mr. Gardner at Maranham and in the Province of Piauhy, and which he considered to belong to the chalk formation. Up the TocantÍns we found fine crystalline stratified rocks, coarse volcanic conglomerates, and fine-grained slates. At the falls were metamorphic slates and other hard crystalline rocks; many of these split into flat slabs, well adapted for building, or even for paving, instead of the stones now imported from Portugal into ParÁ. In the serras of Montealegre, on the north bank of the Amazon, are a great variety of rocks,—coarse quartz conglomerates, fine crystalline sandstones, soft beds of yellow and red sandstones, and indurated clay rocks. These beds are all nearly horizontal, but are much cleft and shattered vertically; they are alternately hard and soft, and by their unequal decay have formed the hanging stones and curious cave described in my Journal.

The general impression produced by the examination of the country is, that here we see the last stage of a process that has been going on, during the whole period of the elevation of the Andes and the mountains of Brazil and Guiana, from the ocean. At the commencement of this period, the greater portion of the valleys of the Amazon, Orinooko, and La Plata must have formed a part of the ocean, separating the groups of islands (which those elevated lands formed on their first appearance) from each other. The sediment carried down into this sea by the rapid streams, running down the sides of these mountains, would tend to fill up and level the deeper and more irregular depressions, forming those large tracts of alluvial deposits we now find in the midst of the granite districts. At the same time volcanic forces were in operation, as shown by the isolated granite peaks which in many places rise out of the flat forest district, like islands from a sea of verdure, because their lower slopes, and the valleys between them, have been covered and filled up by the sedimentary deposits. This simultaneous action of the aqueous and volcanic forces, of submarine earthquakes and marine currents, shaking up, as it were, and levelling the mass of sedimentary matter brought down from the now increasing surface of dry land, is what has produced that marvellous regularity of surface, that gradual and imperceptible slope, which exists over such an immense area.[5]

At the point where the mountains of Guiana approach nearest to the chain of the Andes, the volcanic action appears to have been continued in the interval between them, throwing up the serras of CuricuriarÍ, Tunuhy, and the numerous smaller granite mountains of the UaupÉs; and it is here probably that dry land first appeared, connecting Guiana and New Granada, and forming that slightly elevated ridge which is now the watershed between the basins of the Amazon and Orinooko. The same thing occurs in the southern part of the continent, for it is where the mountains of Brazil, and the eastern range of the Bolivian Andes, stretch out to meet each other, that the sedimentary deposits in that part appear to have been first raised above the water, and thus to have determined the limits of the basin of the Amazon on the south. The Amazon valley would then have formed a great inland gulf or sea, about two thousand miles long and seven or eight hundred wide.

The rivers and mountain-torrents pouring into it on every side, would gradually fill up this great basin; and the volcanic action still visible in the scoriÆ of the Tocantins and Tapajoz, and the shattered rocks of Montealegre, would all tend to the levelling of the vast area, and to determining the channels of the future rivers. This process, continuing for ages, would at length narrow this inland sea, almost within the limits of what is now gapÓ, or flooded land. Ridges, gradually elevated a few feet above the waters, would separate the tributary streams; and then the eddies and currents would throw up sandbanks as they do now, and gradually define the limits of the river, as we now see it. And changes are yet going on. New islands are yearly forming in the stream, large tracts of flooded land are being perceptibly raised by the deposits upon them, and the numerous great lakes are becoming choked with aquatic plants, and filled up with sediment.

The large extent of flat land on the banks of the river will still continue to be flooded, till some renewed earthquakes raise it gradually above the waters; during which time the stream will work for itself a wider and deeper bed, capable of containing its accumulated flood. In the course of ages perhaps this might be produced by the action of the river itself, for at every annual inundation a deposit of sediment is formed, and these lands must therefore be rising, and would in time become permanently elevated above the highest rise of the river. This, however, would take a very long time, for as the banks rose, the river, unable to spread its waters over the adjoining country, would swell higher, and flow more rapidly than before, and so overflow a country elevated above the level of its former inundations.

The complete history of these changes,—the periods of elevation and of repose, the time when the dividing ridges first rose above the waters, and the comparative antiquity of the tributary streams,—cannot be ascertained till the country has been more thoroughly explored, and the organic remains, which must doubtless exist, be brought forward, to give us more accurate information respecting the birth and growth of the Amazon.