geognostic description of south america, north of the river amazon, and east of the meridian of the sierra nevada de merida.

the object of this memoir is to concentrate the geological observations which i collected during my journeys among the mountains of new andalusia and venezuela, on the banks of the orinoco and in the llanos of barcelona, calabozo and the apure; consequently, from the coast of the caribbean sea to the valley of the amazon, between 2 and 10 1/2° north latitude.

the extent of country which i traversed in different directions was more than 15,400 square leagues. it has already formed the subject of a geological sketch, traced hastily on the spot, after my return from the orinoco, and published in 1801. at that period the direction of the cordillera on the coast of venezuela and the existence of the cordillera of parime were unknown in europe. no measure of altitude had been attempted beyond the province of quito; no rock of south america had been named; there existed no description of the superposition of rocks in any region of the tropics. under these circumstances an essay tending to prove the identity of the formations of the two hemispheres could not fail to excite interest. the study of the collections which i brought back with me, and four years of journeying in the andes, have enabled me to rectify my first views, and to extend an investigation which, by reason of its novelty, had been favourably received. that the most remarkable geological relations may be the more easily seized, i shall treat aphoristically, in different sections, the configuration of the soil, the general division of the land, the direction and inclination of the beds and the nature of the primitive, intermediary, secondary and tertiary rocks.

section 1.

configuration of the country. inequalities of the soil. chains and groups of mountains. divisionary ridges. plains or llanos.

south america is one of those great triangular masses which form the three continental parts of the southern hemisphere of the globe. in its exterior configuration it resembles africa more than australia. the southern extremities of the three continents are so placed that, in sailing from the cape of good hope (latitude 33° 55′) to cape horn (latitude 55° 58′), and doubling the southern point of van diemen’s land (latitude 43° 38′), we see those lands stretching out towards the south pole in proportion as we advance eastward. a fourth part of the 571,000 square sea leagues* which south america comprises is covered with mountains distributed in chains or gathered together in groups. the other parts are plains forming long uninterrupted bands covered with forests or gramina, flatter than in europe, and rising progressively, at the distance of 300 leagues from the coast, between 30 and 170 toises above the level of the sea. the most considerable mountainous chain in south america extends from south to north according to the greatest dimension of the continent; it is not central like the european chains, nor far removed from the sea-shore, like the himalaya and the hindoo–koosh; but it is thrown towards the western extremity of the continent, almost on the coast of the pacific ocean. referring to the profile which i have given* of the configuration of south america, in the latitude of chimborazo and grand para, across the plains of the amazon, we find the land low towards the east, in an inclined plane, at an angle of less than 25 seconds on a length of 600 leagues; and if, in the ancient state of our planet, the atlantic ocean, by some extraordinary cause, ever rose to 1100 feet above its present level (a height one-third less than the table-lands of spain and bavaria), the waves must, in the province of jaen de bracamoros, have broken upon the rocks that bound the eastern declivity of the cordilleras of the andes. the rising of this ridge is so inconsiderable compared to the whole continent that its breadth in the parallel of cape saint roche is 1400 times greater than the average height of the andes.

[* almost double the extent of europe.]

[* map of columbia according to the astronomical observations of humboldt by a.h. brue 1823.]

we distinguish in the mountainous part of south america a chain and three groups of mountains, namely, the cordillera of the andes, which the geologist may trace without interruption from cape pilares, in the western part of the straits of magellan, to the promontory of paria opposite the island of trinidad; the insulated group of the sierra nevada de santa marta; the group of the mountains of the orinoco, or of la parime; and that of the mountains of brazil. the sierra de santa marta being nearly in the meridian of the cordilleras of peru and new grenada, the snowy summits descried by navigators in passing the mouth of the rio magdalena are commonly mistaken for the northern extremity of the andes. i shall soon prove that the colossal group of the sierra de santa marta is almost entirely separate from the mountains of ocana and pamplona which belong to the eastern cordillera of new grenada. the hot plains through which runs the rio cesar, and which extend towards the valley of upar, separate the sierra nevada from the paramo de cacota, south of pamplona. the ridge which divides the waters between the gulf of maracaibo and the rio magdalena is in the plain on the east of the laguna zapatoza. if, on the one hand, the sierra de santa marta has been erroneously considered (on account of its eternal snow, and its longitude) to be a continuation of the cordillera of the andes, on the other hand, the connexion of that same cordillera with the coast mountains of the provinces of cumana and caracas has not been recognized. the littoral chain of venezuela, of which the different ranges form the montana de paria, the isthmus of araya, the silla of caracas and the gneiss-granite mountains north and south of the lake of valencia, is joined between porto cabello, san felipe and tocuyo to the paramos de las rosas and niquitao, which form the north-east extremity of the sierra de merida, and the eastern cordillera of the andes of new grenada. it is sufficient here to mention this connexion, so important in a geological point of view; for the denominations of andes and cordilleras being altogether in disuse as applied to the chains of mountains extending from the eastern gulf of maracaibo to the promontory of paria, we shall continue to designate those chains (stretching from west to east) by the names of littoral chain, or coast-chain of venezuela.

of the three insulated groups of mountains, that is to say, those which are not branches of the cordillera of the andes and its continuation towards the shore of venezuela, one is on the north, and the other two on the west of the andes: that on the north is the sierra nevada de santa marta; the two others are the sierra de la parime, between 4 and 8° of north latitude, and the mountains of brazil, between 15 and 28° south latitude. this singular distribution of great inequalities of soil produces three plains or basins, comprising a surface of 420,600 square leagues, or four-fifths of all south america, east of the andes. between the coast-chain of venezuela and the group of the parime, the plains of the apure and the lower orinoco extend; between the group of parime and the brazil mountains are the plains of the amazon, of the rio negro and the madeira, and between the groups of brazil and the southern extremity of the continent are the plains of rio de la plata and of patagonia. as the group of the parime in spanish guiana, and of the brazil mountains (or of minas geraes and goyaz), do not join the cordillera of the andes of new grenada and upper peru towards the west, the three plains of the lower orinoco, the amazon, and the rio de la plata, are connected by land-straits of considerable breadth. these straits are also plains stretching from north to south, and traversed by ridges imperceptible to the eye but forming divortia aquarum. these ridges (and this remarkable phenomenon has hitherto escaped the attention of geologists) are situated between 2 and 3° north latitude, and 16 and 18° south latitude. the first ridge forms the partition of the waters which fall into the lower orinoco on the north-east, and into the rio negro and the amazon on the south and south-east; the second ridge divides the tributary streams of the right bank of the amazon and the rio de la plata. these ridges, of which the existence is only manifested, as in volhynia, by the course of the waters, are parallel with the coast-chain of venezuela; they present, as it were, two systems of counter-slopes partially developed, in the direction from west to east, between the guaviare and the caqueta, and between the mamori and the pilcomayo. it is also worthy of remark that in the southern hemisphere the cordillera of the andes sends an immense counterpoise eastward in the promontory of the sierra nevada de cochabamba, whence begins the ridge stretching between the tributary streams of the madeira and the paraguay to the lofty group of the mountains of brazil or minas geraes. three transversal chains (the coast-mountains of venezuela, of the orinoco or parime, and the brazil mountains) tend to join the longitudinal chain (the andes) either by an intermediary group (between the lake of valencia and tocuyo), or by ridges formed by the intersection of counter-slopes in the plains. the two extremities of the three llanos which communicate by land-straits, the llanos of the lower orinoco, the amazon, and the rio de la plata or of buenos ayres, are steppes covered with gramina, while the intermediary llano (that of the amazon) is a thick forest. with respect to the two land-straits forming bands directed from north to south (from the apure to caqueta across the provincia de los llanos, and the sources of the mamori to rio pilcomayo, across the province of mocos and chiquitos) they are bare and grassy steppes like the plains of caracas and buenos ayres.

in the immense extent of land east of the andes, comprehending more than 480,000 square sea leagues, of which 92,000 are a mountainous tract of country, no group rises to the region of perpetual snow; none even attains the height of 1400 toises. this lowering of the mountains in the eastern region of the new continent extends as far as 60° north latitude; while in the western part, on the prolongation of the cordillera of the andes, the highest summits rise in mexico (latitude 18° 59′) to 2770 toises, and in the rocky mountains (latitude 37 to 40°) to 1900 toises. the insulated group of the alleghenies, corresponding in its eastern position and direction with the brazil group, does not exceed 1040 toises.* the lofty summits, therefore, thrice exceeding the height of mont blanc, belong only to the longitudinal chain which bounds the basin of the pacific ocean, from 55° south to 68° north latitude, that is to say, the cordillera of the andes. the only insulated group that can be compared with the snowy summits of the equinoctial andes, and which attains the height of nearly 3000 toises, is the sierra de santa marta; it is not situated on the east of the cordilleras, but between the prolongation of two of their branches, those of merida and veragua. the cordilleras, where they bound the caribbean sea, in that part which we designate by the name of coast chain of venezuela, do not attain the extraordinary height (2500 toises) which they reach in their prolongation towards chita and merida. considering separately the groups of the east, those of the shore of venezuela, of the parime, and brazil, we see their height diminish from north to south. the highest summits of each group are the silla de caracas (1350 toises), the peak of duida (1300 toises), the itacolumi and the itambe* (900 toises). but, as i have elsewhere observed, it would be erroneous to judge the height of a chain of mountains solely from that of the most lofty summits. the peak of the himalayas, accurately measured, is 676 toises higher than chimborazo; chimborazo is 900 toises higher than mont blanc; and mont blanc 653 toises higher than the peak of nethou.* these differences do not furnish the relative average heights of the himalayas, the andes, the alps and the pyrenees, that is, the height of the back of the mountains, on which arise the peaks, needles, pyramids, or rounded domes. it is that part of the back where passes are made, which furnishes a precise measure of the minimum of the height of the great chains. in comparing the whole of my measures with those of moorcroft, webb, hodgson, saussure and ramond, i estimate the average height of the top of the himalayas, between the meridians of 75 and 77°, at 2450 toises; the andes* (at peru, quito and new grenada), at 1850 toises; the summit of the alps and pyrenees at 1150 toises. the difference of the mean height of the cordilleras (between 5° north and 2° south latitude) and the swiss alps, is consequently 200 toises less than the difference of their loftiest summits; and in comparing the passes of the alps, we see that their average height is nearly the same, although peak nethou is 600 toises lower than mont blanc and mont rosa. between the himalaya* and the andes, on the contrary, (considering those chains in the limits which i have just indicated), the difference between the mean height of the ridges and that of the loftiest summits presents nearly the same proportions.

[* the culminant point of the alleghenies is mount washington in new hampshire, latitude 44 1/4°. according to captain partridge its height is 6634 english feet.]

[* according to the measure of mm. spix and martius the itambe de villa de principe is 5590 feet high.]

[* the peak iewahir, latitude 30° 22′ 19′; longitude 77° 35′ 7′ east of paris, height 4026 toises, according to mm. hodgson and herbert.]

[* this peak, called also peak of anethou or malahita, or eastern peak of maladetta, is the highest summit of the pyrenees. it rises 1787 toises and consequently exceeds mont perdu by 40 toises.]

[* in the passage of quindiu, between the valley of the magdalena and that of the rio cauca, i found the culminant point (la garita del parama) to be 1798 toises; it is however, regarded as one of the least elevated. the passages of the andes of guanacas, guamani and micuipampa, are respectively 2300, 1713, and 1817 toises above sea-level. even in 33° south latitude the road across the andes between mendoza and valparaiso is 1987 toises high. i do not mention the col de l’assuay, where i passed, near la ladera de cadlud, on a ridge 2428 toises high, because it is a passage on a transverse ridge joining two parallel chains.]

[* the passes of the himalaya that lead from chinese tartary into hindostan (nitee–ghaut, bamsaru, etc.) are from 2400 to 2700 toises high.]

taking an analogous view of the groups of mountains at the east of the andes, we find the average height of the coast-chain of venezuela to be 750 toises; of the sierra parime, 500 toises; of the brazilian group, 400 toises; whence it follows that the mountains of the eastern region of south america between the tropics are, when compared to the medium elevation of the andes, in the relation of one to three.

the following is the result of some numerical statements, the comparison of which affords more precise ideas on the structure of mountains in general.*

[* the cols or passes indicate the minimum of the height to which the ridge of the mountains lowers in a particular country. now, looking at the principal passes of the alps of switzerland (col terret, 1191 toises, mont cenis, 1060 toises; great saint bernard, 1246 toises; simplon, 1029 toises; and on the neck of the pyrenees, benasque, 1231 toises; pinede, 1291 toises; gavarnic, 1197 toises; cavarere, 1151 toises; it would be difficult to affirm that the pyrenees are lower than the average height of the swiss alps.]

names of the chains of mountains. the highest summits in toises. mean height of the ridge in toises. proportion of the mean height of the ridges to that of the highest summits.

himalayas (between north latitude 30° 18′ and 31° 53′, and longitude 75° 23′ and 77° 38′) 4026 2450 1: 1.6.

cordillera of the andes (between latitude 5 and 2° south) 3350 1850 1: 1.8.

alps of switzerland 2450 1150 1: 2.1.

pyrenees 1787 1150 1: 1.5.

littoral chain of venezuela 1350 750 1: 1.8.

group of the mountains of the parime 1300 500 1: 2.6.

group of the mountains of brazil 900 500 1: 2.3.

if we distinguish among the mountains those which rise sporadically, and form small insulated systems,* and those that make part of a continued chain,* we find that, notwithstanding the immense height* of the summits of some insulated systems, the culminant points of the whole globe belong to continuous chains — to the cordilleras of central asia and south america.

[* as the groups of the canaries, the azores, the sandwich islands, the monts–dores, and the euganean mountains.]

[* the himalayas, the alps, and the andes.]

[* among the insulated systems, or sporadic mountains, mowna–roa is generally regarded as the most elevated summit of the sandwich islands. its height is computed at 2500 toises, and yet at some seasons it is entirely free from snow. an exact measure of this summit, situated in very frequented latitudes, has for 25 years been desired in vain by naturalists and geologists.]

in that part of the andes with which i am best acquainted, between 8° south latitude and 21° north latitude, all the colossal summits are of trachyte. it may almost be admitted as a general rule that whenever the mass of mountains rises in that region of the tropics much above the limit of perpetual snow (2300 to 2470 toises), the rocks commonly called primitive (for instance, gneiss-granite or mica-slate) disappear, and the summits are of trachyte or trappean-porphyry. i know only a few rare exceptions to this law, and they occur in the cordilleras of quito where the nevados of conderasto and cuvillan, situated opposite to the trachytic chimborazo, are composed of mica-slate and contain veins of sulphuret of silver. thus in the groups of detached mountains which rise abruptly from the plains the loftiest summits, such as mowna–roa, the peak of teneriffe, etna and the peak of the azores, present only recent volcanic rocks. it would, however, be an error to extend that law to every other continent, and to admit, as a general rule, that, in every zone, the greatest elevations have produced trachytic domes: gneiss-granite and mica-slate constitute the summits of the ridge, in the almost insulated group of the sierra nevada of grenada and the peak of malhacen,* as they also do in the continuous chain of the alps, the pyrenees and probably the himalayas.* these phenomena, discordant in appearance, are possibly all effects of the same cause: granite, gneiss, and all the so-styled primitive neptunian mountains, may possibly owe their origin to volcanic forces, as well as the trachytes; but to forces of which the action resembles less the still-burning volcanoes of our days, ejecting lava, which at the moment of its eruption comes immediately into contact with the atmospheric air; but it is not here my purpose to discuss this great theoretic question.

[* this peak, according to the survey of m. clemente roxas, is 1826 toises above the level of the sea, consequently 39 toises higher than the loftiest summit of the pyrenees (the granitic peak of nethou) and 83 toises lower than the trachytic peak of teneriffe. the sierra nevada of grenada forms a system of mountains of mica-slate, passing to gneiss and clay-slate, and containing shelves of euphotide and greenstone.]

[* if we may judge from the specimens of rocks collected in the gorges and passes of the himalayas or rolled down by the torrents.]

after having examined the general structure of south america according to considerations of comparative geology, i shall proceed to notice separately the different systems of mountains and plains, the mutual connection of which has so powerful an influence on the state of industry and commerce in the nations of the new continent. i shall give only a general view of the systems situated beyond the limits of the region which forms the special object of this memoir. geology being essentially founded on the study of the relations of juxtaposition and place, i could not treat of the littoral chain and the chain of the parime separately, without touching on the other systems south and west of venezuela.

a. systems of mountains.

a.1. cordilleras of the andes.

this is the most continuous, the longest, the most uniform in its direction from south to north and north-north-west, of any chain of the globe. it approaches the north and south poles at unequal distances of from 22 to 33°. its development is from 2800 to 3000 leagues (20 to a degree), a length equal to the distance from cape finisterre in galicia to the north-east cape (tschuktschoi–noss) of asia. somewhat less than one half of this chain belongs to south america, and runs along its western shores. north of the isthmus of cupica and of panama, after an immense lowering, it assumes the appearance of a nearly central ridge, forming a rocky dyke that joins the great continent of north america to the southern continent. the low lands on the east of the andes of guatimala and new spain appear to have been overwhelmed by the ocean and now form the bottom of the caribbean sea. as the continent beyond the parallel of florida again widens towards the east, the cordilleras of durango and new mexico, as well as the rocky mountains, merely a continuation of those cordilleras, appear to be thrown still further westward, that is, towards the coast of the pacific ocean; but they still remain eight or ten times more remote from it than in the southern hemisphere. we may consider as the two extremities of the andes, the rock or granitic island of diego ramirez, south of cape horn, and the mountains lying at the mouth of mackenzie river (latitude 69°, longitude 130 1/2°), more than twelve degrees west of the greenstone mountains, known by the name of the copper mountains, visited by captain franklin. the colossal peak of saint elias and that of mount fairweather, in new norfolk, do not, properly speaking, belong to the northern prolongation of the cordilleras of the andes, but to a parallel chain (the maritime alps of the north-west coast), stretching towards the peninsula of california, and connected by transversal ridges with a mountainous land, between 45 and 53° of latitude, with the andes of new mexico (rocky mountains). in south america the mean breadth of the cordillera of the andes is from 18 to 22 leagues.* it is only in the knots of the mountains, that is where the cordillera is swelled by side-groups or divided into several chains nearly parallel, and reuniting at intervals, for instance, on the south of the lake of titicaca, that it is more than 100 to 120 leagues broad, in a direction perpendicular to its axis. the andes of south america bound the plains of the orinoco, the amazon, and the rio de la plata, on the west, like a rocky wall raised across a crevice 1300 leagues long, and stretching from south to north. this upheaved part (if i may be permitted to use an expression founded on a geological hypothesis) comprises a surface of 58,900 square leagues, between the parallel of cape pilesar and the northern choco. to form an idea of the variety of rocks which this space may furnish for the observation of the traveller, we must recollect that the pyrenees, according to the observations of m. charpentier, occupy only 768 square sea leagues.

[* the breadth of this immense chain is a phenomenon well worthy of attention. the swiss alps extend, in the grisons and in the tyrol, to a breadth of 36 and 40 leagues, both in the meridians of the lake at como, the canton of appenzell, and in the meridian of bassano and tegernsee.]

the name of andes in the quichua language (which wants the consonants d, f, and g) antis, or ante, appears to me to be derived from the peruvian word anta, signifying copper or metal in general. anta chacra signifies mine of copper; antacuri, copper mixed with gold; and puca anta, copper, or red metal. as the group of the altai mountains* takes its name from the turkish word altor or altyn, in the same manner the cordilleras may have been termed “copper-country,” or anti-suyu, on account of the abundance of that metal, which the peruvians employed for their tools. the inca garcilasso, who was the son of a peruvian princess, and who wrote the history of his native country in the first years of the conquest, gives no etymology of the name of the andes. he only opposes anti-suyu, or the region of summits covered with eternal snow (ritiseca), to the plains or yuncas, that is, to the lower region of peru. the etymology of the name of the largest mountain chain of the globe cannot be devoid of interest to the mineralogic geographer.

[* klaproth. asia polyglotta page 211. it appears to me less probable that the tribe of the antis gave its name to the mountains of peru.]

the structure of the cordillera of the andes, that is, its division into several chains nearly parallel, which are again joined by knots of mountains, is very remarkable. on our maps this structure is indicated but imperfectly; and what la condamine and bouguer merely guessed, during their long visit to the table-land of quito, has been generalized and ill-interpreted by those who have described the whole chain according to the type of the equatorial andes. the following is the most accurate information i could collect by my own researches and an active correspondence of twenty years with the inhabitants of spanish america. the group of islands called tierra del fuego, in which the chain of the andes begins, is a plain extending from cape espiritu santo as far as the canal of san sebastian. the country on the west of this canal, between cape san valentino and cape pilares, is bristled with granitic mountains covered (from the morro de san agueda to cabo redondo) with calcareous shells. navigators have greatly exaggerated the height of the mountains of tierra del fuego, among which there appears to be a volcano still burning. m. de churruca found the height of the western peak of cape pilares (latitude 52° 45′ south) only 218 toises; even cape horn is probably not more than 500 toises* high. the plain extends on the northern shore of the straits of magellan, from the virgin’s cape to cabo negro; at the latter the cordilleras rise abruptly, and fill the whole space as far as cape victoria (latitude 52° 22′). the region between cape horn and the southern extremity of the continent somewhat resembles the origin of the pyrenees between cape creux (near the gulf of rosas) and the col des perdus. the height of the patagonian chain is not known; it appears, however, that no summit south of the parallel of 48° attains the elevation of the canigou (1430 toises) which is near the eastern extremity of the pyrenees. in that southern country, where the summers are so cold and short, the limit of eternal snow must lower at least as much as in the northern hemisphere, in norway, in latitude 63 and 64°; consequently below 800 toises. the great breadth, therefore, of the band of snow that envelopes these patagonian summits, does not justify the idea which travellers form of their height in 40° south latitude. as we advance towards the island of chiloe, the cordilleras draw near the coast; and the archipelago of chonos or huaytecas appears like the vestiges of an immense group of mountains overwhelmed by water. narrow estuaries fill the lower valleys of the andes, and remind us of the fjords of norway and greenland. we there find, running from south to north, the nevados de maca (latitude 45° 19′), of cuptano (latitude 44° 58′), of yanteles (latitude 43° 52′), of corcovado, chayapirca (latitude 42° 52′) and of llebean (latitude 41° 49′). the peak of cuptana rises like the peak of teneriffe, from the bosom of the sea; but being scarcely visible at thirty-six or forty leagues distance, it cannot be more than 1500 toises high. corcovado, situated on the coast of the continent, opposite the southern point of the island of chiloe, appears to be more than 1950 toises high; it is perhaps the loftiest summit of the whole globe, south of the parallel of 42° south latitude. on the north of san carlos de chiloe, in the whole length of chile to the desert of atacama, the low western regions not having been overwhelmed by floods, the andes there appear farther from the coast. the abbe molina affirms that the cordilleras of chile form three parallel chains, of which the intermediary is the most elevated; but to prove that this division is far from general, it suffices to recollect the barometric survey made by mm. bauza and espinosa, in 1794, between mendoza and santiago de chile. the road leading from one of those towns to the other, rises gradually from 700 to 1987 toises; and after passing the col des andes (la cumbre, between the houses of refuge called las calaveras and las cuevas), it descends continually as far as the temperate valley of santiago de chile, of which the bottom is only 409 toises above the level of the sea. the same survey has made known the minimum of height at chile of the lower limit of snow, in 33° south latitude. the limit does not lower in summer to 2000 toises.* i think we may conclude according to the analogy of the snowy mountains of mexico and southern europe, and considering the difference of the summer temperature of the two hemispheres, that the real nevadas at chile, in the parallel of valdivia (latitude 40°), cannot be below 1300 toises; in valparaiso (latitude 33°) not lower than 2000 toises, and in that of copiapo (latitude 27°) not below 2200 toises of height. these are the limit-numbers, the minimum of elevation, which the ridge of the andes of chile must attain in different degrees of latitude, to enable their summits to rise above the line of perpetual snow. the numerical results which i have just marked and which are founded on the laws of distribution of heat, have still the same importance which they possessed at the time of my travels in america; for there does not exist in the immense extent of the andes, from 8° south latitude to the straits of magellan, one nevada of which the height above the sea-level has been determined, either by a simple geometric measure, or by the combined means of barometric and geodesic measurements.

[* it is very distinctly seen at the distance of 60 miles, which, without calculating the effects of terrestrial refraction, would give it a height of 498 toises.]

[* on the southern declivity of the himalayas snow begins (3° nearer the equator) at 1970 toises.]

between 33 and 18° south latitude, between the parallels of valparaiso and arica, the andes present towards the east three remarkable spurs, the sierra de cordova, the sierra de salta, and the nevados de cochabamba. travellers partly cross and partly go along the side of the sierra de cordova (between 33 and 31° of latitude) in their way from buenos ayres to mendoza; it may be said to be the most southern promontory which advances, in the pampas, towards the meridian of 65°; it gives birth to the great river known by the name of desaguadero de mendoza and extends from san juan de la frontera and san juan de la punta to the town of cordova. the second spur, called the sierra de salta and the jujui, of which the greatest breadth is 25° of latitude, widens from the valley of catamarca and san miguel del tucuman, in the direction of the rio vermejo (longitude 64°). finally, the third and most majestic spur, the sierra nevada de cochabamba and santa cruz (from 22 to 17 1/2° of latitude), is linked with the knot of the mountains of porco. it forms the points of partition (divortia aquarum, between the basin of the amazon and that of the rio de la plata. the cachimayo and the pilcomayo, which rise between potosi, talavera de la puna, and la plata or chuquisaca, run in the direction of south-east, while the parapiti and the guapey (guapaiz, or rio de mizque) pour their waters into the mamori, to north-east. the ridge of partition being near chayanta, south of mizque, tomina and pomabamba, nearly on the southern declivity of the sierra de cochabamba in latitude 19 and 20°, the rio guapey flows round the whole group, before it reaches the plains of the amazon, as in europe the poprad, a tributary of the vistula, makes a circuit in its course from the southern part of the carpathians to the plains of poland. i have already observed above, that where the mountains cease (west* of the meridian of 66 1/2°) the partition ridge of cochabamba goes up towards the north-east, to 16° of latitude, forming, by the intersection of two slightly inclined planes, only one ridge amidst the savannahs, and separating the waters of the guapore, a tributary of the madeira, from those of the aguapehy and jauru, tributaries of the rio paraguay. this vast country between santa cruz de la sierra, villabella, and matogrosso, is one of the least known parts of south america. the two spurs of cordova and salta present only a mountainous territory of small elevation, and linked to the foot of the andes of chile. cochabamba, on the contrary, attains the limit of perpetual snow (2300 toises) and forms in some sort a lateral branch of the cordilleras, diverging even from their tops between la paz and oruro. the mountains composing this branch (the cordillera de chiriguanaes, de los sauces and yuracarees) extend regularly from west to east; their eastern declivity* is very rapid, and their loftiest summits are not in the centre, but in the northern part of the group.

[* i agree with captain basil hall, in fixing the port of valparaiso in 71° 31′ west of greenwich, and i place cordova 8° 40′, and santa cruz de la sierra 7° 4′ east of valparaiso. the longitudes mentioned in the text refer always to the meridian of the observatory of paris.]

[* for much information concerning the sierra de cochabamba i am indebted to the manuscripts of my countryman, the celebrated botanist taddeus haenke, which a monk of the congregation of the escurial, father cisneros, kindly communicated to me at lima. mr. haenke, after having followed the expedition of alexander malaspina, settled at cochabamba in 1798. a part of the immense herbal of this botanist is now at prague.]

the principal cordillera of chile and upper peru is, for the first time, ramified very distinctly into two branches, in the group of porco and potosi, between latitude 19 and 20°. these two branches comprehend the table-land extending from carangas to lamba (latitude 19 3/4 to 15°) and in which is situated the small mountain lake of paria, the desaguadero, and the great laguna of titicaca or chucuito, of which the western part bears the name of vinamarca. to afford an idea of the colossal dimensions of the andes, i may here observe that the surface of the lake of titicaca alone (448 square sea leagues) is twenty times greater than that of the lake of geneva, and twice the average extent of a department of france. on the banks of this lake, near tiahuanacu, and in the high plains of callao, ruins are found which bear evidence of a state of civilization anterior to that which the peruvians assign to the reign of the inca manco capac. the eastern cordillera, that of la paz, palca, ancuma, and pelechuco, join, north-west of apolobamba, the western cordillera, which is the most extensive of the whole chain of the andes, between the parallels 14 and 15°. the imperial city of cuzco is situated near the eastern extremity of this knot, which comprehends, in an area of 3000 square leagues, the mountains of vilcanota, carabaya, abancai, huando, parinacochas, and andahuaylas. though here, as in general, in every considerable widening of the cordillera, the grouped summits do not follow the principal axis in uniform and parallel directions, a phenomenon observable in the general disposition of the chain of the andes, from latitude 18°, is well worthy the attention of geologists. the whole mass of the cordilleras of chile and upper peru, from the straits of magellan to the parallel of the port of arica (18° 28′ 35′), runs from south to north, in the direction of a meridian at most 5° north-east; but from the parallel of arica, the coast and the two cordilleras east and west of the alpine lake of titicaca, abruptly change their direction and incline to north-west. the cordilleras of ancuma and moquehua, and the longitudinal valley, or rather the basin of titicaca, which they inclose, take a direction north 42° west. further on, the two branches again unite in the group of the mountains of cuzco, and thence their direction is north 80° west. this group of which the table-land inclines to the north-east, forms a curve, nearly from east to west, so that the part of the andes north of castrovireyna is thrown back more than 242,000 toises westward. this singular geological phenomenon resembles the variation of dip of the veins, and especially of the two parts of the chain of the pyrenees, parallel to each other, and linked by an almost rectangular elbow, 16,000 toises long, near the source of the garonne;* but in the andes, the axes of the chain, south and north of the curve, do not preserve parallelism. on the north of castrovireyna and andahuaylas (latitude 14°), the direction is north 22° west, while south of 15°, it is north 42° west. the inflexions of the coast follow these changes. the shore separated from the cordillera by a plain 15 leagues in breadth, stretches from camapo to arica, between 27 1/2 and 18 1/2° latitude north 5° east; from arica to pisco, between 18 1/2 and 14° latitude at first north 42° west, afterwards north 65° west; and from pisco to truxillo, between 14 and 8° of latitude north 27° west. the parallelism between the coast and the cordillera of the andes is a phenomenon the more worthy of attention, as it occurs in several parts of the globe where the mountains do not in the same manner form the shore.

[* between the mountain of tentenade and the port d’espot.]

after the great knot of mountains of cuzco and parinacochas, in 14° south latitude, the andes present a second bifurcation, on the east and west of the rio jauja, which throws itself into the mantaro, a tributary stream of the apurimac. the eastern chain stretches on the east of huanta, the convent of ocopa and tarma; the western chain, on the west of castrovireyna, huancavelica, huarocheri, and yauli. the basin, or rather the lofty table-land which is inclosed by these chains, is nearly half the length of the basin of chucuito or titicaca. two mountains covered with eternal snow, seen from the town of lima, and which the inhabitants name toldo de la nieve, belong to the western chain, that of huarocheri.

north-west of the valleys of salcabamba, in the parallel of the ports of huaura and guarmey, between 11 and 10° latitude, the two chains unite in the knot of the huanuco and the pasco, celebrated for the mines of yauricocha or santa rosa. there rise two peaks of colossal height, the nevados of sasaguanca and of la viuda. the table-land of this knot of mountains appears in the pambas de bombon to be more than 1800 toises above the level of the ocean. from this point, on the north of the parallel of huanuco (latitude 11°), the andes are divided into three chains: the first, and most eastern, rises between pozuzu and muna, between the rio huallaga, and the rio pachitea, a tributary of the ucayali; the second, or central, is between the huallaga, and the upper maranon; the third, or western, between the upper maranon and the coast of truxillo and payta. the eastern chain is a small lateral branch which lowers into a range of hills: its direction is first north-north-east, bordering the pampas del sacramento, afterwards it turns west-north-west, where it is broken by the rio huallaga, in the pongo, above the confluence of chipurana, and then it loses itself in latitude 6 1/4°, on the north-west of lamas. a transversal ridge seems to connect it with the central chain, south of paramo de piscoguanuna (or piscuaguna), west of chachapoyas. the intermediary or central chain stretches from the knot of pasco and huanuco, towards north-north-west, between xican and chicoplaya, huacurachuco and the sources of the rio monzan, between pataz and pajatan, caxamarquilla and moyobamba. it widens greatly in the parallel of chachapoyas, and forms a mountainous territory, traversed by deep and extremely hot valleys. on the north of the paramo de piscoguanuna (latitude 6°) the central chain throws two branches in the direction of la vellaca and san borja. we shall soon see that this latter branch forms, below the rio neva a tributary stream of the amazon, the rocks that border the famous pongo de manseriche. in this zone, where north peru approximates to the confines of new grenada in latitude 10 and 5°, no summit of the eastern and central chains rises as high as the region of perpetual snow; the only snowy summits are in the western chain. the central chain, that of the paramos de callacalla, and piscoguanuna, scarcely attains 1800 toises, and lowers gently to 800 toises; so that the mountainous and temperate tract of country which extends on the north of chachapoyas towards pomacocha, la vellaca and the source of the rio nieva is rich in fine cinchona trees. after having passed the rio huallaga and the pachitea, which with the beni forms the ucayali, we find, in advancing towards the east, only ranges of hills. the western chain of the andes, which is the most elevated and nearest to the coast, runs almost parallel with the shore north 22° west, between caxatambo and huary, conchucos and guamachuco, by caxamarca, the paramo de yanaguanga, and montan, towards the rio de guancabamba. it comprises (between 9 and 7 1/2°) the three nevados de pelagatos, moyopata and huaylillas. this last snowy summit, situated near guamachuco (in 7° 55′ latitude), is the more remarkable, since from thence on the north, as far as chimborazo, on a length of 140 leagues, there is not one mountain that enters the region of perpetual snow. this depression, or absence of snow, extends in the same interval, over all the lateral chains; while, on the south of the nevado de huaylillas, it always happens that when one chain is very low, the summits of the other exceed the height of 2460 toises. it was on the south of micuipampa (latitude 7° 1 minute) that i found the magnetic equator.

the amazon, or as it is customary to say in those regions, the upper maranon, flows through the western part of the longitudinal valley lying between the cordilleras of chachapayas and caxamarca. comprehending in one point of view, this valley, and that of the rio jauja, bounded by the cordilleras of tarma and huarocheri, we are inclined to consider them as one immense basin 180 leagues long, and crossed in the first third of its length, by a dyke, or ridge 18,000 toises broad. in fact, the two alpine lakes of lauricocha and chinchaycocha, where the river amazon and the rio de jauja take their rise, are situated south and north of this rocky dyke, which is a prolongation of the knot of huanuco and pasco. the amazon, on issuing from the longitudinal valley which bounds the chains of caxamarca and chachacocha, breaks the latter chain; and the point where the great river penetrates the mountains, is very remarkable. entering the amazon by the rio chamaya or guancabamba, i found opposite the confluence, the picturesque mountain of patachuana; but the rocks on both banks of the amazon begin only between tambillo and tomependa (latitude 5° 31′, longitude 80° 56′). from thence to the pongo de rentema, a long succession of rocks follow, of which the last is the pongo de tayouchouc, between the strait of manseriche and the village of san borja. the course of the amazon, which is first directed north, then east, changes near puyaya, three leagues north-east of tomependa. throughout the whole distance between tambillo and san borja, the waters force a way, more or less narrow, across the sandstones of the cordillera of chachapoyas. the mountains are lofty near the embarcadero, at the confluence of the imasa, where large trees of cinchona, which might be easily transplanted to cayenne, or the canaries, approach the amazon. the rocks in the famous strait of manseriche are scarcely 40 toises high; and further eastward the last hills rise near xeberos, towards the mouth of the rio huallaga.

i have not yet noticed the extraordinary widening of the andes near the apolobamba. the sources of the rio beni being found in the spur which stretches northward beyond the confluence of that river with the apurimac, i shall give to the whole group the name of “the spur of beni.” the following is the most certain information i have obtained respecting those countries, from persons who had long inhabited apolobamba, the real das minas of pasco, and the convent of ocopa. along the whole eastern chain of titicaca, from la paz to the knot of huanuco (latitude 17 1/2 to 10 1/2°) a very wide mountainous land is situated eastward, at the back of the declivity of the andes. it is not a widening of the eastern chain itself, but rather of the small heights that surround the foot of the andes like a penumbra, filling the whole space between the beni and the pachitca. a chain of hills bounds the eastern bank of the beni to latitude 8°; for the rivers coanache and magua, tributaries of the ucayali (flowing in latitude 6 and 7°) come from a mountainous tract between the ucayali and the javari. the existence of this tract in so eastern a longitude (probably longitude 74°), is the more remarkable, as we find at four degrees of latitude further north, neither a rock nor a hill on the east of xeberos, or the mouth of the huallaga (longitude 77° 56′).

we have just seen that the spur of beni, a sort of lateral branch, loses itself about latitude 8°; the chain between the ucayali and the huallaga terminates at the parallel of 7°, in joining, on the west of lamas, the chain of chachapayas, stretching between the huallaga and the amazon. finally, the latter chain, to which i have given the designation of central, after forming the rapids and cataracts of the amazon, between tomependa and san borja, turns to north-north-west, and joins the western chain, that of caxamarca, or the nevados of pelagatos and huaylillas, and forms the great knot of the mountains of loxa. the mean height of this knot is only from 1000 to 1200 toises: its mild climate renders it peculiarly favourable to the growth of the cinchona trees, the finest kinds of which are found in the celebrated forest of caxanuma and uritusinga, between the rio zamora and the cachiyacu, and between tavacona and guancabamba. before the cinchona of popayan and santa fe de bogota (north latitude 2 1/2 to 5°), of huacarachuco, huamalies and huanuco (south latitude 9 to 11°) became known, the group of the mountains of loxa had for ages been regarded as the sole region whence the febrifuge bark of cinchona could be obtained. this group occupies the vast territory between guancabamba, avayaca, ona and the ruined towns of zamora and loyola, between latitude 5 1/2 and 3 1/4°. some of the summits (the paramos of alpachaca, saraguru, savanilla, gueringa, chulucanas, guamani, and yamoca, which i measured) rise from 1580 to 1720 toises, but are not even sporadically covered with snow, which in this latitude falls only above 1860 to 1900 toises of absolute height. eastward, in the direction of the rio santiago and the rio de chamaya, two tributary streams of the amazon, the mountains lower rapidly: between san felipe, matara, and jaen de bracamoros, they are not more than 500 or 300 toises.

as we advance from the mica-slate mountain of loxa towards the north, between the paramos of alpachaca and sara (in latitude 3° 15′) the knot of mountains ramifies into two branches which comprehend the longitudinal valley of cuenca. this separation continues for a length of only 12 leagues; for in latitude 2° 27′ the two cordilleras again re-unite in the knot of assuy, a trachytic group, of which the table-land near cadlud (2428 toises high) nearly enters the region of perpetual snow.

the group of the mountains of assuy, which affords a very frequented pass of the andes between cuenca and quito (latitude 2 1/2 to 0° 40′ south) is succeeded by another division of the cordilleras, celebrated by the labours of bouguer and la condamine, who placed their signals sometimes on one, sometimes on the other of the two chains. the eastern chain is that of chimborazo (3350 toises) and carguairazo; the western is the chain of the volcano sangay, the collanes, and of llanganate. the latter is broken by the rio pastaza. the bottom of the longitudinal basin that bounds those two chains, from alausi to llactacunga, is somewhat higher than the bottom of the basin of cuenca. north of llactacanga, 0° 40′ latitude, between the tops of yliniza (2717 toises) and cotopaxi (2950 toises), of which the former belongs to the chain of chimborazo, and the latter to that of sangay, is situated the knot of chisinche; a kind of narrow dyke that closes the basin, and divides the waters between the atlantic and the pacific. the alto de chisinche is only 80 toises above the surrounding table-lands. the waters of its northern declivity form the rio de san pedro, which, joining the rio pita, throws itself into the gualabamba, or rio de las esmeraldas. the waters of the southern declivity, called cerro de tiopullo, run into the rio san felipe and the pastaza, a tributary stream of the amazon.

the bipartition of the cordilleras re-commences and continues from 0° 40′ latitude south to 0° 20′ latitude north; that is, as far as the volcano of imbabura near the villa of ibarra. the eastern cordillera presents the snowy summits of antisana (2992 toises), of guamani, cayambe (3070 toises) and of imbabura; the western cordillera, those of corazon, atacazo, pichinca (2491 toises) and catocache (2570 toises). between these two chains, which may be regarded as the classic soil of the astronomy of the 18th century, is a valley, part of which is again divided longitudinally by the hills of ichimbio and poignasi. the table-lands of puembo and chillo are situated eastward of those hills; and those of quito, inaquito and turubamba lie westward. the equator crosses the summit of the nevado de cayambe and the valley of quito, in the village of san antonio de lulumbamba. when we consider the small mass of the knot of assuy, and above all, of that of chisinche, we are inclined to regard the three basins of cuenca, hambato and quito as one valley (from the paramo de sarar to the villa de ibarra) 73 sea leagues long, from 4 to 5 leagues broad, having a general direction north 8° east, and divided by two transverse dykes one between alausi and cuenca (2° 27′ south latitude), and the other between machache and tambilbo (0° 40′). nowhere in the cordillera of the andes are there more colossal mountains heaped together than on the east and west of this vast basin of the province of quito, one degree and a half south, and a quarter of a degree north of the equator. this basin which, next to the basin of titicaca, is the centre of the most ancient native civilization, touches, southward, the knot of the mountains of loxa, and northward the tableland of the province of los pastos.

in this province, a little beyond the villa of ibarra, between the snowy summits of cotocache and imbabura, the two cordilleras of quito unite, and form one mass, extending to meneses and voisaco, from 0° 21′ north latitude to 1 degree 13 minutes. i call this mass, on which are situated the volcanoes of cumbal and chiles, the knot of the mountains of los pastos, from the name of the province that forms the centre. the volcano of pasto, the last eruption of which took place in the year 1727, is on the south of yenoi, near the northern limit of this group, of which the inhabited table-lands are more than 1600 toises above sea-level. it is the thibet of the equinoctial regions of the new world.

on the north of the town of pasto (latitude 1 degree 13 minutes north; longitude 79° 41′) the andes again divide into two branches and surround the table-land of mamendoy and almaguer. the eastern cordillera contains the sienega of sebondoy (an alpine lake which gives birth to the putumayo), the sources of the jupura or caqueta, and the paramos of aponte and iscanse. the western cordillera, that of mamacondy, called in the country cordillera de la costa, on account of its proximity to the shore of the pacific, is broken by the great rio de patias, which receives the guativa, the guachicon and the quilquase. the table-land or intermediary basin has great inequalities; it is partly filled by the paramos of pitatumba and paraguay, and the separation of the two chains appeared to me indistinct as far as the parallel of almaguer (latitude 1 degree 54 minutes; longitude 79° 15′). the general direction of the andes, from the extremity of the basin of the province of quito to the vicinity of popayan, changes from north 8° east to north 36° east; and follows the direction of the coast of esmeralda and barbacoas.

on the parallel of almaguer, or rather a little north-east of that town, the geological structure of the ground displays very remarkable changes. the cordillera, to which we have given the name of eastern, that of the lake of sebondoy, widens considerably between pansitara and ceja. the knot of the paramo de las papas and of socoboni gives birth to the great rivers of cauca and magdalena, and is divided into two chains, latitude 2° 5′ east and west of la plata, vieja and timana. these two chains continue nearly parallel as far as 5° of latitude, and they bound the longitudinal valley through which winds the rio magdalena. we shall give the name of the eastern cordillera of new grenada to that chain which stretches towards santa fe de bogota, and the sierra nevada de merida, east of magdalena; the chain which lies between the magdalena and the cauca, in the direction of mariquita, we will call the central cordillera of new grenada; and the chain which continues the cordillera de la costa from the basin of almaguer, and separates the bed of the rio cauca from the platiniferous territory of choco, we will designate the western cordillera of new grenada. for additional clearness, we may also name the chain, that of suma paz, after the colossal group of mountains on the south of santa fe de bogota, which empties the waters of its eastern declivity into the rio meta. the second chain may bear the name of the chain of guanacas or quindiu, after the two celebrated passages of the andes, on the road from santa fe de bogota to popayan. the third chain may be called the chain of choco, or of the shore. some leagues south of popayan (latitude 2° 21′ north), west of paramo de palitara and the volcano of purace, a ridge of mica-slate runs from the knot of the mountains of sacoboni to north-west, and divides the waters between the pacific and the caribbean sea; they flow from the northern declivity into the rio cauca, and from the southern declivity, into the rio de patias.

the tripartition of the andes (north latitude 1 3/4 to 2 1/4°) resembles that which takes place at the source of the amazon in the knot of the mountains of huanuco and pasco (latitude 11° south); but the most western of the three chains that bound the basins of the amazon and the huallaga, is the loftiest; while that of choco, or the shore, is the least elevated of the three chains of new grenada. ignorance of this tripartition of the andes in that part of south america near the rio atrato and the isthmus of panama, has led to many erroneous opinions respecting the possibility of a canal that should connect the two seas.

the eastern chain of the andes of new grenada* preserves its parallelism during some time with the two other chains, those of quindiu and choco; but beyond tunja (latitude 5 1/2°) it inclines more towards the north-east, passing somewhat abruptly from the direction north 25° east to that of north 45° east. it is like a vein that changes its direction; and it rejoins the coast after being greatly enlarged by the grouping of the snowy mountains of merida. the tripartition of the cordilleras, and above all, the spreading of their branches, have a vast influence on the prosperity of the nations of new grenada. the diversity of the superposed table-lands and climates varies the agricultural productions as well as the character of the inhabitants. it gives activity to the exchange of productions, and renews over a vast surface, north of the equator, the picture of the sultry valleys and cool and temperate plains of peru. it is also worthy of remark that, by the separation of one of the branches of the cordilleras of cundinamarca and by the deviation of the chain of bogota towards the north-east, the colossal group of the mountains of merida is enclosed in the territory of the ancient capitania-general of venezuela, and that the continuity of the same mountainous land from pamplona to barquisimeto and nirgua may be said to have facilitated the political union of the columbian territory. as long as the central chain (that of quindiu) presents its snowy summits, no peak of the eastern chain (that of la suma paz) rises, in the same parallels, to the limit of perpetual snow. between latitude 2 and 5 1/2° neither the paramos situated on the east of gigante and neiva, nor the tops of la suma paz, chingasa, guachaneque, and zoraca, exceed the height of 1900 to 2000 toises; while on the north of the parallel of paramo d’erve (latitude 5° 5′), the last of the nevados of the central cordillera, we discover in the eastern chain the snowy summits of chita (latitude 5° 50′), and of mucuchies (latitude 8° 12′). hence it results that from latitude 5° the only mountains covered with snow during the whole year are the cordilleras of the east; and although the sierra nevada of santa marta is not, properly speaking, a continuation of the nevados of chita and mucuchies (west of patute and east of merida), it is at least very near their meridian.

[* i employ a systematic denomination, for the name of the andes is unknown in the countries situated north of the equator.]

having now arrived at the northern extremity of the cordilleras, comprehended between cape horn and the isthmus of panama, we shall proceed to notice the loftiest summits of the three chains which separate in the knot of the mountains of socoboni, and the ridge of roble (latitude 1 degree 50 minutes to 2° 20′). i begin with the most eastern chain, that of timana and suma paz, which divides the tributary streams of the magdalena and the meta: it runs by the paramos de chingasu, guachaneque, zoraca, toquillo (near labranza grande), chita, almorsadero, laura, cacota, zumbador and porqueras, in the direction of the sierra nevada de merida. these paramos indicate ten partial risings of the back of the cordilleras. the declivity of the eastern chain is extremely rapid on the eastern side, where it bounds the basin of the meta and the orinoco; it is widened on the west by the spurs on which are situated the towns of santa fe de bogota, tunja, sogamoso and leiva. they are like tablelands fixed to the western declivity, and are from 1300 to 1400 toises high; that of bogota (the bottom of an ancient lake) contains fossil bones of the mastodon, in the plain called (from them) the campo de gigantes, near suacha.

the intermediary, or central chain, runs east of popayan, by the high plains of mabasa, the paramos of guanacas, huila, savelillo, iraca, baraguan, tolima, ruiz and herveo, towards the province of antioquia. in 5° 15′ of latitude this chain, the only one that shows traces of recent volcanic fire, in the summits of sotara and purace, widens considerably towards the west, and joins the western chain, which we have called the chain of choco, because the platiniferous land of that province lies on the slope opposite the pacific ocean. by the union of the two chains, the basin of the province of popayan is close on the north of cartago viejo; and the river of cauca, issuing from the plain of buga, is forced, from the salto de san antonio, to la boca del espiritu santo, to open its way across the mountains, along a course of from 40 to 50 leagues. the difference of the level is very remarkable in the bottom of the two parallel basins of cauca and magdalena. the former, between cali and cantago, is from 500 to 404 toises; the latter, from neiva to ambalema, is from 265 to 150 toises high. according to different geological hypotheses, it may be said either that the secondary formations have not accumulated to the same thickness between the eastern and central, as between the central and western chains; or, that the deposits have been made on the base of primitive rocks, unequally upheaved on the east and west of the andes of quindiu. the average difference of the thickness of these formations is 300 toises. the rocky ridge of the angostura of carare branches from the south-east, from the spur of muzo, through which winds the rio negro. by this spur, and by those that come from the west, the eastern and central chains approach between nares, honda, and mendales. in fact, the bed of the rio magdalena is narrowed in 5 and 5° 18′, on the east by the mountains of sergento, and on the west by the spurs that are linked with the granitic mountains of maraquito and santa ana. this narrowing of the bed of the river is in the same parallel with that of the cauca, near the salto de san antonio; but, in the knot of the mountains of antioquia the central and western chains join each other, while between honda and mendales, the tops of the central and eastern chains are so far removed that it is only the spurs of each system that draw near and are confounded together. it is also worthy of remark that the central cordillera of new grenada displays the loftiest summit of the andes in the northern hemisphere. the peak of tolima (latitude 4° 46′) which is almost unknown even by name in europe, and which i measured in 1801, is at least 2865 toises high. it consequently surpasses imbabura and cotocache in the province of quito, the chiles of the table-lands of los pastos, the two volcanoes of popayan and even the nevados of mexico and mount saint elias of russian america. the peak of tolima, which in form resembles cotapaxi, is perhaps inferior in height only to the ridge of the sierra nevada de santa marta, which may be considered as an insulated system of mountains.

the eastern chain, also called the chain of choco and the east coast (of the pacific), separates the provinces of popayan and antioquia from those of barbacoas, raposo and choco. it is in general but little elevated, compared to the height of the central and eastern chains; it however presents great obstacles to the communications between the valley of cauca and the shore. on its western slope lies the famous auriferous and platiniferous land,* which has during ages yielded more than 13,000 marks of gold annually. this alluvial zone is from ten to twelve leagues broad; its maximum of productiveness lies between the parallels of 2 and 6° latitude; it sensibly impoverishes towards the north and south, and almost entirely disappears between 1 1/4 degree north latitude and the equator. the auriferous soil fills the basin of cauca, as well as the ravines and plains west of the cordillera of choco; it rises sometimes nearly 600 toises above the level of the sea, and descends at least 40 toises.* platinum (and this fact is worthy of attention) has hitherto been found only on the west of the cordillera of choco, and not on the east, notwithstanding the analogy of the fragments of rocks of greenstone, phonolite, trachyte, and ferruginous quartz, of which the soil of the two slopes is composed. from the ridge of los robles, which separates the table-land of almaguer from the basin of cauca, the western chain forms, first, in the cerros de carpinteria, east of the rio san juan de micay, the continuation of the cordillera of sindagua, broken by the rio patias; then, lowering northward, between cali and las juntas de dagua, and at the elevation of 800 to 900 toises, it sends out considerable spurs (latitude 4 1/4 to 5°) towards the source of the calima, the tamana and the andagueda. the two former of these auriferous rivers are tributary streams of the rio san juan del choco; the second empties its waters into the atrato. this widening of the western chain forms the mountainous part of choco: here, between the tado and zitara, called also francisco de quibdo, lies the isthmus of raspadura, across which a monk traced a navigable line of communication between the two oceans. the culminant point of this system of mountains appears to be the peak of torra, situated south-east of novita.

[* choco, barbacoas and brazil are the only countries in which the existence of grains of platinum and palladium has hitherto been fully ascertained. the small town of barbacoas is situated on the left bank of the rio telembi (a tributary of patias or the rio del castigo) a little above the confluence of telembi and the guagi or guaxi, nearly in latitude 1 degree 48 minutes. the ancient provincia, or rather the partido del raposo, comprehends the insalubrious land extending from the rio dagua, or san buenaventura, to the rio iscuande, the southern limit of choco.]

[* m. caldas assigns to the upper limit of the zone of gold-washings, only the height of 350 toises. semanario tome 1 page 18; but i found the seraderos[?] of quilichao, on the north of popayan, to be 565 toises high.]

the northern extremity of this enlargement of the cordillera of choco, which i have just described, corresponds with the junction formed on the east, between the same cordillera and the central chain, that of quindiu. the mountains of antioquia, on which we have the excellent observations of mr. restrepo, may be called a knot of mountains, and on the northern limit of the plains of buga, or the basin of cauca, they join the central and western chains. the ridge of the eastern cordillera is at the distance of thirty-five leagues from this knot, so that the contraction of the bed of the rio magdalena, between honda and ambalema, is caused only by the approximation of the spurs of mariquita and guaduas. there is not, therefore, properly speaking, a group of mountains between latitude 5 and 5 1/4°, uniting the three chains at once. in the group of the province of antioquia, which forms the junction of the central and western cordilleras, we may distinguish two great masses; one between the magdalena and the cauca, and the other between the cauca and the atrato. the first of these masses, which is linked most immediately to the snowy summits of herveo, gives birth on the east to the rio de la miel and the nare; and on the north to porce and nechi; its average height is only from 1200 to 1350 toises. the culminant point appears to be near santa rosa, south-west of the celebrated valley of bears (valle de osos). the towns of rio negro and marinilla are built on table-lands 1060 toises high. the western mass of the knot of the mountains of antioquia, between the cauca and the atrato, gives rise, on its western descent, to the rio san juan, bevara, and murri. it attains its greatest height in the alto del viento, north of urrao, known to the first conquistadores by the name of the cordilleras of abide or dabeida. this height (latitude 7° 15′) does not, however, exceed 1500 toises. following the western slope of this system of mountains of antioquia, we find that the point of partition of the waters that flow towards the pacific and the caribbean sea (latitude 5 1/2 and 6° ) nearly corresponds with the parallel of the isthmus of raspadura, between the rio san juan and the atrato. it is remarkable that in this group, more than 30 leagues broad, without sharp summits, between latitude 5 1/4 and 7°, the highest masses rise towards the west; while, further south, before the union of the two chains of quindiu and choco, we saw them on the east of cauca.

the ramifications of the knot of antioquia, on the north of the parallel 7°, are very imperfectly known; it is observed only that their lowering is in general more rapid and complete towards the north-west, in the direction of the ancient province of biruquete and darien, than towards the north and north-east, on the side of zaragoza and simiti. from the northern bank of the rio nare, near its confluence with the samana, a spur stretches out, known by the name of la simitarra, and the mountains of san lucar. we may call it the first branch of the group of antioquia. i saw it, in going up the rio magdalena, on the west, from the regidor and the mouth of the rio simiti, as far as san bartolome (on the south of the mouth of the rio sogamozo); while, eastward, in latitude 7 3/4 and 8 1/4°, the spur of the mountains of ocana appear in the distance; they are inhabited by some tribes of molitone indians. the second branch of the group of antioquia (west of samitarra) commences at the mountains of santa rosa, stretches out between zaragoza and caceres, and terminates abruptly at the confluence of the rio nechi (latitude 8° 33′): at least if the hills, often conical, between the mouth of the rio sinu and the small town of tolu, or even the calcareous heights of turbaco and popa, near carthagena, may not be regarded as the most northern prolongation of this second branch. a third advances towards the gulf of uraba or darien, between the rio san jorge and the atrato. it is linked southward with the alto del viento, or sierra de abide, and is rapidly lost, advancing as far as the parallel of 8°. finally, the fourth branch of the andes of antioquia, situated westward of zitara and the rio atrato, undergoes, long before it enters the isthmus of panama, such a depression, that between the gulf of cupica and the embarcadero of the rio naipipi, we find only a plain across which m. gogueneche has projected a canal for the junction of the two seas. it would be interesting to know the configuration of the strata between cape garachine, or the gulf of st. miguel, and cape tiburon, especially towards the source of the rio tuyra and chucunaque or chucunque, so as to determine with precision where the mountains of the isthmus of panama begin to rise; mountains whose elevation does not appear to be more than 100 toises. the interior of darfur is not more unknown to geographers than the humid, insalubrious forest-land which extends on the north-west of betoi and the confluence of the bevara with the atrato, towards the isthmus of panama. all that we positively know of it hitherto is that between cupica and the left bank of the atrato there is either a land-strait, or a total absence of the cordillera. the mountains of the isthmus of panama, by their direction and their geographical position, may be considered as a continuation of the mountains of antioquia and choco; but on the west of bas–atrato, there is scarcely a ridge in the plain. we do not find in this country a group of interposed mountains like that which links (between barquisimeto, nirgua and valencia) the eastern chain of new grenada (that of suma paz and the sierra nevada de merida) to the cordillera of the shore of venezuela.

the cordillera of the andes, considered in its whole extent, from the rocky wall of the island of diego ramirez to the isthmus of panama, is sometimes ramified into chains more or less parallel, and sometimes articulated by immense knots of mountains. we distinguish nine of those knots, and consequently an equal number of branching-points and ramifications. the latter are generally bifurcations. the andes are twice only divided into three chains; in the knot of huanuco, near the source of the amazon, and the huallaga (latitude 10 to 11°) and in the knot of the paramo de las papas (latitude 2°), near the source of the magdalena and the cauca. basins, almost shut in at their extremities, parallel with the axis of the cordillera and bounded by two knots and two lateral chains, are characteristic features of the structure of the andes. among these knots of mountains some, for instance those of cuzco, loxa and los pastos, comprise 3300, 1500 and 1130 square leagues, while others no less important in the eye of the geologist are confined to ridges or transversal dykes. to the latter belong the altos de chisinche (latitude 0° 40′ south) and the los robles (latitude 2° 20′ north), on the south of quito and popayan. the knot of cuzco, so celebrated in the annals of peruvian civilization, presents an average height of from 1200 to 1400 toises, and a surface nearly three times greater than the whole of switzerland. the ridge of chisinche, which separates the basins of tacunga and quito, is 1580 toises high, but scarcely a mile broad. the knots or groups which unite several partial chains have not the highest summits, either in the andes or, for the most part, in the great mountain ranges of the old continent; it is not even certain that there is always in those knots a widening of the chain. the greatness of the mass, and the height so long attributed to points whence several considerable branches issue, was founded either on theoretic ideas or on false measures. the cordilleras were compared to rivers that swell as they receive a number of tributary streams.

among the basins which the andes present, and which form probably as many lakes or small inland seas, those of titicaca, rio jauja and the upper maranon, comprise respectively 3500, 1300, and 2400 square leagues of surface.* the first is so encompassed that no drop of water can escape except by evaporation; it is like the enclosed valley of mexico,* and of those numerous circular basins which have been discerned in the moon, and which are surrounded by lofty mountains. an immense alpine lake characterizes the basin of tiahuanaco or titicaca; this phenomenon is the more worthy of attention, as in south america there are scarcely any of those reservoirs of fresh water which are found at the foot of the european alps, on the northern and southern slopes, and which are permanent during the season of drought. the other basins of the andes, for instance, those of jauja, the upper maranon and cauca, pour their waters into natural canals, which may be considered as so many crevices situated either at one of the extremities of the basin, or on its banks, nearly in the middle of the lateral chain. i dwell on this articulated form of the andes, on those knots or transverse ridges, because, in the continuation of the andes called the cordilleras of the shore of venezuela, we shall find the same transverse dykes, and the same phenomena.

[* i here subjoin some measures interesting to geologists. area of the andes, from tierra del fuego to the paramo de las rosas (latitude 9 1/4° north), where the mountainous land of tocuyo and barquesimeto begins, part of the cordillera of the shore of venezuela, 58,900 square leagues, (20 to a degree) the four spurs of cordova, salta, cochabamba and beni alone, occupy 23,300 square leagues of this surface, and the three basins contained between latitude 6 and 20° south measure 7200 square leagues. deducting 33,200 square leagues for the whole of the enclosed basins and spurs, we find, in latitude 65°, the area of the cordilleras elevated in the form of walls, to be 25,700 square leagues, whence results (comprehending the knots, and allowing for the inflexion of the chains) an average breadth of the andes of 18 to 20 leagues. the valleys of huallaga and the rio magdalena are not comprehended in these 58,900 square leagues, on account of the diverging direction of the chain, east of cipoplaya and santa fe de bogota.]

[* we consider it in its primitive state, without respect to the gap or cleft of the mountains, known by the name of desaghue de huehuetoca.]

the ramification of the andes and of all the great masses of mountains into several chains merits particular consideration in reference to the height more or less considerable of the bottom of the enclosed basins, or longitudinal valleys. geologists have hitherto directed more attention to the successive narrowing of these basins, their depth compared with the walls of rock that surround them, and the correspondence between the re-entering and the salient angles, than to the level of the bottom of the valleys. no precise measure has yet fixed the absolute height of the three basins of titicaca, jauja and the upper maranon;* but i was fortunate enough to be able to determine the six other basins, or longitudinal valleys, which succeed each other, as if by steps, towards the north. the bottom of the valley of cuenca, between the knots of loxa and assuay, is 1350 toises; the valley of allansi and of hambato, between the knot of the assuay and the ridge of chisinche, 1320 toises; the valley of quito in the eastern part, 1340 toises, and in the western part, 1490 toises; the basin of almaguer, 1160 toises; the basin of the rio cauca, between the lofty plains of cali, buga, and cartago, 500 toises; the valley of magdalena, first between neiva and honda, 200 toises; and further on, between honda and mompox, 100 toises of average height above the level of the sea.* in this region, which has been carefully measured, the different basins lower very sensibly from the equator northward. the elevation of the bottom of enclosed basins merits great attention in connection with the causes of the formation of the valleys. i do not deny that the depressions in the plains may be sometimes the effect of ancient pelagic currents, or slow erosions. i am inclined to believe that the transversal valleys, resembling crevices, have been widened by running waters; but these hypotheses of successive erosions cannot well be applied to the completely enclosed basins of titicaca and mexico. these basins, as well as those of jauja, cuenca and almaguer, which lose their waters only by a lateral and narrow issue, owe their origin to a cause more instantaneous, more closely linked with the upheaving of the whole chain. it may be said that the phenomenon of the narrow declivities of the sarenthal and of the valley of eysack in the tyrol, is repeated at every step, and on a grander scale, in the cordilleras of equinoctial america. we seem to recognize in the cordilleras those longitudinal sinkings, those rocky vaults, which, to use the expression of a great geologist,* “are broken when extended over a great space, and leave deep and almost perpendicular rents.”

[* i am inclined to believe that the southern part of the basin of the upper maranon, between huary and huacarachuco, exceeds 350 toises.]

[* in the region of the andes comprehended between 4° of south latitude and 2° of north, the longitudinal valleys or basins inclosed by parallel chains are regularly between 1200 and 1500 toises high; while the transversal valleys are remarkable for their depression, or rather the rapid lowering of their bottom. the valley of patias, for instance, running from north-east to south-west is only 350 toises of absolute height, even above the junction of the rio guachion with the quilquasi, according to the barometric measures of m. caldas; and yet it is surrounded by the highest summits, the paramos de puntaurcu and mamacondy. going from the plains of lombardy, and penetrating into the alps of the tyrol, by a line perpendicular to the axis of the chain, we advance more than 20 marine leagues towards the north, yet we find the bottom of the valley of the adige and of eysack near botzen, to be only 182 toises of absolute height, an elevation which exceeds but 117 toises that of milan. from botzen however, to the ridge of brenner (culminant point 746 toises) is only 11 leagues. the valais is a longitudinal valley; and in a barometric measurement which i made very recently from paris to naples and berlin, i was surprised to find that from sion to brigg, the bottom of the valley rises only to from 225 to 350 toises of absolute height; nearly the level of the plains of switzerland, which, between the alps and the jura, are only from 274 to 300 toises.]

[* von buch, tableau du tyrol meridional page 8 1823.]

if, to complete the sketch of the structure of the andes from tierra del fuego to the northern polar sea, we pass the boundaries of south america, we find that the western cordillera of new grenada, after a great depression between the mouth of the atrato and the gulf of cupica, again rises in the isthmus of panama to 80 or 100 toises high, augmenting towards the west, in the cordilleras of veragua and salamanca,* and extending by guatimala as far as the confines of mexico. within this space it extends along the coast of the pacific where, from the gulf of nicoya to soconusco (latitude 9 1/2 to 16°) is found a long series of volcanoes,* most frequently insulated, and sometimes linked to spurs or lateral branches. passing the isthmus of tehuantepecor huasacualco, on the mexican territory, the cordillera of central america extends on toward the intendancia of oaxaca, at an equal distance from the two oceans; then from 18 1/2 to 21° latitude, from misteca to the mines of zimapan, it approximates to the eastern coast. nearly in the parallel of the city of mexico, between toluca, xalapa and cordoba, it attains its maximum height; several colossal summits rising to 2400 and 2770 toises. farther north the chain called sierra madre runs north 40° west towards san miguel el grande and guanaxuato. near the latter town (latitude 21° 0′ 15′) where the richest silver mines of the known world are situated, it widens in an extraordinary degree and separates into three branches. the most eastern branch advances towards charcas and the real de catorce, and lowers progressively (turning to north-east) in the ancient kingdom of leon, in the province of cohahuila and texas. that branch is prolonged from the rio colorado de texas, crossing the arkansas near the confluence of the mississippi and the missouri (latitude 38° 51′). in those countries it bears the name of the mountains of ozark,* and attains 300 toises of height. it has been supposed that on the east of the mississippi (latitude 44 to 46°) the wisconsin hills, which stretch out to north-north-east in the direction of lake superior, may be a continuation of the mountains of ozark. their metallic wealth seems to denote that they are a prolongation of the eastern cordillera of mexico. the western branch or cordillera occupies a part of the province of guadalajara and stretches by culiacan, aripe and the auriferous lands of the pimeria alta and la sonora, as far as the banks of the rio gila (latitude 33 to 34°), one of the most ancient dwellings of the aztek nations. we shall soon see that this western chain appears to be linked by the spurs that advance to the west, with the maritime alps of california. finally, the central cordillera of anahuac, which is the most elevated, runs first from south-east to north-west, by zacatecas towards durango, and afterwards from south to north, by chihuahua, towards new mexico. it takes successively the names of sierra de acha, sierra de los mimbres, sierra verde, and sierra de las grullas, and about the 29 and 39° of latitude, it is connected by spurs with two lateral chains, those of the texas and la sonora, which renders the separation of the chains more imperfect than the trifurcations of the andes in south america.

[* if it be true, as some navigators affirm, that the mountains at the north-western extremity of the republic of columbia, known by the names of silla de veragua, and castillo del choco, be visible at 36 leagues distance, the elevation of their summits must be nearly 1400 toises, little lower than the silla of caracas.]

[* see the list of twenty-one volcanoes of guatimala, partly extinct and partly still burning, given by arago and myself, in the annuaire du bureau des longitudes pour 1824 page 175. no mountain of guatimala having been hitherto measured, it is the more important to fix approximately the height of the volcan de agua, or the volcano of pacaya, and the volcan de fuego, called also volcano of guatimala. mr. juarros expressly says that this volcano which, by torrents of water and stones, destroyed, on the 11th september, 1541, the ciudad vieja, or almolonga (the ancient capital of the country, which must not be confounded with the ancient guatimala), is covered with snow, during several months of the year. this phenomenon would seem to indicate a height of more than 1750 toises.]

[* ozark is at once the ancient name of arkansas and of the tribe of quawpaw indians who inhabit the banks of that great river. the culminant point of the mountains of ozark is in latitude 37 1/2°, between the sources of the white and osage rivers.]

that part of the cordilleras of mexico which is richest in silver beds and veins, is comprehended between the parallels of oaxaca and cosiquiriachi (latitude 16 1/2 to 29°); the alluvial soil that contains disseminated gold extends some degrees still further northwards. it is a very striking phenomenon that the gold-washing of cinaloa and sonora, like that of barbacoas and choco on the south and north of the isthmus of panama, is uniformly situated on the west of the central chain, on the descent opposite the pacific. the traces of a still-burning volcanic fire which was no longer seen, on a length of 200 leagues, from pasto and popayan to the gulf of nicoya (latitude 1 1/4 to 9 1/2°), become very frequent on the western coast of guatimala (latitude 9 1/2 to 16°); these traces of fire again cease in the gneiss-granite mountains of oaxaca, and re-appear, perhaps for the last time, towards the north, in the central cordillera of anahuac, between latitude 18 1/4 and 19 1/2°, where the volcanoes of taxtla, orizaba, popocatepetl, toluca, jorullo and colima appear to be situated in a crevice* extending from east-south-east to west-north-west, from one ocean to the other. this line of summits, several of which enter the limit of perpetual snow, and which are the loftiest of the cordilleras from the peak of tolima (latitude 40° 46′ north), is almost perpendicular to the great axis of the chain of guatimala and anahuac, advancing to the 27th parallel, uniformly north 42° east. a characteristic feature of every knot, or widening of the cordilleras, is that the grouping of the summits is independent of the general direction of the axis. the backs of the mountains in new spain form very elevated plains, along which carriages can roll for an extent of 400 leagues, from the capital to santa–fe and taos, near the sources of rio del norte. this immense table-land, in 19 and 24 1/2°, is constantly at the height of from 950 to 1200 toises, that is, at the elevation of the passes of the great saint bernard and the splugen. we find on the back of the cordilleras of anahuac, which lower progressively from the city of mexico towards taos, a succession of basins: they are separated by hills little striking to the eye of the traveller because they rise only from 250 to 400 toises above the surrounding plains. the basins are sometimes closed, like the valley of tenochtitlan, where lie the great alpine lakes, and sometimes they exhibit traces of ancient ejections, destitute of water.

[* on this zone of volcanoes is the parallel of the greatest heights of new spain. if the survey of captain basil hall afford results alike certain in latitude and in longitude, the volcano of colima is north of the parallel of puerto de navidad in latitude 19° 36′; and, like the volcano of tuxtla, if not beyond the zone, at least beyond the average parallel of the volcanic fire of mexico, which parallel seems to be between 18° 59′ and 19° 12′.]

between latitude 33 and 38°, the rio del norte forms, in its upper course, a great longitudinal valley; and the central chain seems here to be divided into several parallel ranges. this distribution continues northward, in the rocky mountains,* where, between the parallels of 37 and 41°, several summits covered with eternal snow (spanish peak, james peak and big horn) are from 1600 to 1870 toises of absolute height. towards latitude 40° south of the sources of the paduca, a tributary of the rio de la plata, a branch known by the name of the black hills, detaches itself towards the north-east from the central chain. the rocky mountains at first seem to lower considerably in 46 and 48°; and then rise to 48 and 49°, where their tops are from 1200 to 1300 toises, and their ridge near 950 toises. between the sources of the missouri and the river lewis, one of the tributaries of the oregon or columbia, the cordilleras form in widening, an elbow resembling the knot of cuzco. there, also, on the eastern declivity of the rocky mountains, is the partition of water between the caribbean sea and the polar sea. this point corresponds with those in the andes of south america, at the spur of cochabamba, on the east, latitude 19° 20′ south; and in the alto de los robles (latitude 2° 20′ north), on the west. the ridge that separates the rocky mountains extends from west to east, towards lake superior, between the basins of the missouri and those of lake winnipeg and the slave lake. the central cordillera of mexico and the rocky mountains follow the direction north 10° west, from latitude 25 to 38°; the chain from that point to the polar sea prolongs in the direction north 24° west, and ends in the parallel 69°, at the mouth of the mackenzie river.*

[* the rocky mountains have been at different periods designated by the names of chypewyan, missouri, columbian, caous, stony, shining and sandy mountains.]

[* the eastern boundary of the rocky mountains lies:—

in 38° latitude: 107° 20′ longitude.

in 40° latitude: 108° 30′ longitude.

in 63° latitude: 124° 40′ longitude.

in 68° latitude: 130° 30′ longitude.]

in thus developing the structure of the cordilleras of the andes from 56° south to beyond the arctic circle, we see that its northern extremity (longitude 130° 30′) is nearly 61° of longitude west of its southern extremity (longitude 60° 40′); this is the effect of the long-continued direction from south-east to north-west north of the isthmus of panama. by the extraordinary breadth of the new continent, in the 30 and 60° north latitude, the cordillera of the andes, continually approaching nearer to the western coast in the southern hemisphere, is removed 400 leagues on the north from the source of the rio de la paz. the andes of chile may be considered as maritime alps,* while, in their most northern continuation, the rocky mountains are a chain in the interior of a continent. there is, no doubt, between latitude 23 and 60° from cape saint lucas in california, to alaska on the western coast of the sea of kamschatka, a real littoral cordillera; but it forms a system of mountains almost entirely distinct from the andes of mexico and canada. this system, which we shall call the cordillera of california, or of new albion, is linked between latitude 33 and 34° with the pimeria alta, and the western branch of the cordilleras of anahuac; and between latitude 45 and 53°, with the rocky mountains, by transversal ridges and spurs that widen towards the east. travellers who may at some future time pass over the unknown land between cape mendocino and the source of the rio colorado, may perhaps inform us whether the connexion of the maritime alps of california or new albion, with the western branch of the cordilleras of mexico, resembles that which, notwithstanding the depression, or rather total interruption observed on the west of the rio atrato, is admitted by geographers to exist between the mountains of the isthmus of panama and the western branch of the andes of new grenada. the maritime alps, in the peninsula of old california, rise progressively towards the north in the sierra of santa lucia (latitude 34 1/2°), in the sierra of san marcos (latitude 37 to 38°) and in the snowy mountains near cape mendocino (latitude 39° 41′); the last seem to attain at least the height of 1500 toises. from cape mendocino the chain follows the coast of the pacific, but at the distance of from twenty to twenty-five leagues. between the lofty summits of mount hood and mount saint helen, in latitude 45 3/4°, the chain is broken by the river columbia. in new hanover, new cornwall and new norfolk these rents of a rocky coast are repeated, these geologic phenomena of the fjords that characterize western patagonia and norway. at the point where the cordillera turns towards the west (latitude 58 3/4°, longitude 139° 40′) there are two volcanic peaks, one of which (mount saint elias) perhaps equals cotopaxi in height; the other (fair–weather mountain) equals the height of mount rosa. the elevation of the former exceeds all the summits of the cordilleras of mexico and the rocky mountains, north of the parallel 19 1/4°; it is even the culminant point in the northern hemisphere, of the whole known world north of 50° of latitude. north-west of the peaks of saint elias and fair–weather the chain of california widens considerably in the interior of russian america. volcanoes multiply in number as we advance westward, in the peninsula of alaska and the fox islands, where the volcano ajagedan rises to the height of 1175 toises above the level of the sea. thus the chain of the maritime alps of california appears to be undermined by subterraneous fires at its two extremities; on the north in 60° of latitude, and on the south, in 28°, in the volcanoes of the virgins.* if it were certain that the mountains of california belong to the western branch of the andes of anahuac, it might be said that the volcanic fire, still burning, abandons the central cordillera when it recedes from the coast, that is, from the volcano of colima; and that the fire is borne on the north-west by the peninsula of old california, mount saint elias, and the peninsula of alaska, towards the aleutian islands and kamschatka.

[* geognostically speaking, a littoral chain is not a range of mountains forming of itself the coast; this name is extended to a chain separated from the coast by a narrow plain.]

[* volcanes de las virgenes. the highest summit of old california, the cerro de la giganta (700 toises), appears to be also an extinguished volcano.]

i shall terminate this sketch of the structure of the andes by recapitulating the principal features that characterize the cordilleras, north-west of darien.

latitude 8 to 11°. mountains of the isthmus of panama, veragua and costa rica, slightly linked to the western chain of new grenada, which is that of choco.

latitude 11 to 16°. mountains of nicaragua and guatimala; line of volcanoes north 50° west, for the most part still burning, from the gulf of nicoya to the volcano of soconusco.

latitude 16 to 18°. mountains of gneiss-granite in the province of oaxaca.

latitude 18 1/2 to 19 1/2°. trachytic knot of anahuac, parallel with the nevados and the burning volcanoes of mexico.

latitude 19 1/2 to 20°. knot of the metaliferous mountains of guanaxuato and zacatecas.

latitude 21 3/4 to 22°. division of the andes of anahuac into three chains:

eastern chain (that of potosi and texas), continued by the ozark and wisconsin mountains, as far as lake superior.

central chain (of durango, new mexico and the rocky mountains), sending on the north of the source of the river platte (latitude 42°) a branch (the black hills) to north-east, widening greatly between the parallels 46 and 50°, and lowering progressively as it approaches the mouth of mackenzie river (latitude 68°).

western chain (of cinaloa and sonora). linked by spurs to the maritime alps, or mountains of california.

we have yet no means of judging with precision the elevation of the andes south of the knot of the mountains of loxa (south latitude 3° 5), but we know that on the north of that knot the cordilleras rise five times higher than the majestic elevation of 2600 toises:

in the group of quito, 0 to 2° south latitude (chimborazo, antisano, cayambe, cotopaxi, collanes, yliniza, sangay, tungurahua.)

in the group of cundinamarca, latitude 4 3/4° north (peak of tolima, north of the andes of quindiu).

in the group of anahuac, from latitude 18° 59′ to 19° 12′ (popocatepetl or the great volcano of mexico, and peak of orizaba). if we consider the maritime alps or mountains of california and new norfolk, either as a continuation of the western chain of mexico, that of sonora, or as being linked by spurs to the central chain, that of the rocky mountains, we may add to the three preceding groups:

the group of russian america, from latitude 60 to 70° (mount saint elias). over an extent of 63° of latitude, i know only twelve summits of the andes which reach the height of 2600 toises, and consequently exceed by 140 toises, the height of mont blanc. only three of these twelve summits are situated north of the isthmus of panama.

2. insulated group of the snowy mountains of santa marta.

in the enumeration of the different systems of mountains, i place this group before the littoral chain of venezuela, though the latter, being a northern prolongation of the cordillera of cundinamarca, is immediately linked with the chain of the andes. the sierra nevado of santa marta is encompassed within two divergent branches of the andes, that of bogota, and that of the isthmus of panama. it rises abruptly like a fortified castle, amidst the plains extending from the gulf of darien, by the mouth of the magdalena, to the lake of maracaybo. the old geographers erroneously considered this insulated group of mountains covered with eternal snow, as the extremity of the high cordilleras of chita and pamplona. the loftiest ridge of the sierra nevada de santa marta is only three or four leagues in length from east to west; it is bounded (at nine leagues distance from the coast) by the meridians of the capes of san diego and san augustin. the culminant points, called el picacho and horqueta, are near the western border of the group; they are entirely separated from the peak of san lorenzo, also covered with eternal snow, but only four leagues distant from the port of santa marta, towards the south-east. i saw this latter peak from the heights that surrounded the village of turbaco, south of carthagena. no precise measurement has hitherto given us the height of the sierra nevada, which dampier affirms to be one of the highest mountains of the northern hemisphere. calculations founded on the maximum of distance at which the group is discerned at sea, give a height of more than 3004 toises. that the group of the mountains of santa marta is insulated is proved by the hot climate of the lands (tierras calientes) that surround it. low ridges and a succession of hills indicate, perhaps, an ancient connection between the sierra nevada de santa marta on one side, by the alto de las minas, with the phonolitic and granitic rocks of the penon and banca, and on the other, by the sierra de perija, with the mountains of chiliguana and ocana, which are the spurs of the eastern chain of the andes of new grenada. in this latter chain, the febrifuge species of cinchona (corollis hirsutis, staminibus inclusis) are found in the sierra nevada de merida; but the real cinchona, the most northern of south america, is found in the temperate region of the sierra nevada de santa marta.

3. littoral chain of venezuela.

this is the system of mountains the configuration and direction of which have excited so powerful an influence on the cultivation and commerce of the ancient capitania general of venezuela. it bears different names, as the mountains of coro, of caracas, of the bergantin, of barcelona, of cumana, and of paria; but all these names belong to the same chain, of which the northern part runs along the coast of the caribbean sea. this system of mountains, which is 160 leagues long,* is a prolongation of the eastern cordillera of the andes of cundinamarca. there is an immediate connection of the littoral chain with the andes, like that of the pyrenees with the mountains of asturia and galicia; it is not the effect of transversal ridges, like the connection of the pyrenees with the swiss alps, by the black mountain and the cevennes. the points of junction are between truxillo and the lake of valencia.

[* it is more than double the length of the pyrenees, from cape creux to the point of figuera.]

the eastern chain of new grenada stretches north-east by the sierra nevada de merida, as well as by the four paramos of timotes, niquitao, bocono and las rosas, of which the absolute height cannot be less than from 1400 to 1600 toises. after the paramo of las rosas, which is more elevated than the two preceding, there is a great depression, and we no longer see a distinct chain or ridge, but merely hills, and high table-lands surrounding the towns of tocuyo and barquisimeto. we know not the height even of cerro del altar, between tocuyo and caranacatu; but we know by recent measures that the most inhabited spots are from 300 to 350 toises above sea-level. the limits of the mountainous land between tocuyo and the valleys of aragua are, the plains of san carlos on the south, and the rio tocuyo on the north; the rio siquisique flows into that river. from the cerro del altar on the north-east towards guigue and valencia, succeed, as culminant points, the mountains of santa maria (between buria and nirgua); then the picacho de nirgua, supposed to be 600 toises high; and finally las palomeras and el torito (between valencia and nirgua). the line of water-partition runs from west to east, from quibor to the lofty savannahs of london, near santa rosa. the waters flow on the north, towards the golfo triste of the caribbean sea; and on the south, towards the basins of the apure and the orinoco. the whole of this mountainous country, by which the littoral chain of caracas is linked to the cordilleras of cundinamarca, was celebrated in europe in the middle of the nineteenth century; for that part of the territory formed of gneiss-granite, and lying between the rio tocuyo and the rio yaracui, contains the auriferous veins of buria, and the copper-mine of aroa which is worked at the present day. if, across the knot of the mountains of barquisimeto, we trace the meridians of aroa, nirgua and san carlos, we find that on the north-west that knot is linked with the sierra de coro, and on the north-east with the mountains of capadare, porto cabello and the villa de cura. it may be said to form the eastern wall of that vast circular depression of which the lake of maracaybo is the centre and which is bounded on the south and west by the mountains of merida, ocana, perija and santa marta.

the littoral chain of venezuela presents towards the centre and the east the same phenomena of structure as those observed in the andes of peru and new grenada; namely, the division into several parallel ranges and the frequency of longitudinal basins or valleys. but the irruptions of the caribbean sea having apparently overwhelmed, at a very remote period, a part of the mountains of the shore, the ranges or partial chains are interrupted and some basins have become oceanic gulfs. to comprehend the cordillera of venezuela in mass we must carefully study the direction and windings of the coast from punta tucacas (west of porto cabello) as far as punta de la galera of the island of trinidad. that island, those of los testigos, marguerita and tortuga constitute, with the mica-slates of the peninsula of araya, one and the same system of mountains. the granitic rocks which appear between buria, duaca and aroa cross the valley of the rio yaracui and draw near the shore, whence they extend, like a continuous wall, from porto cabello to cape codera. this prolongation forms the northern chain of the cordillera of venezuela and is traversed in going from south to north, either from valencia and the valleys of aragua, to burburata and turiamo, or from caracas to la guayra. hot springs* issue from those mountains, those of las trincheras (90.4°) on its southern slope and those of onoto and mariara on its southern slope. the former issue from a granite with large grains, very regularly stratified; the latter from a rock of gneiss. what especially characterizes the northern chain is a summit which is not only the loftiest of the system of the mountains of venezuela, but of all south america, on the east of the andes. the eastern summit of the silla of caracas, according to my barometric measurement made in 1800, is 1350 toises high,* and notwithstanding the commotion which took place on the silla during the great earthquake of caracas, that mountain did not sink 50 or 60 toises, as some north american journals asserted. four or five leagues south of the northern chain (that of mariara, la silla and cape codera) the mountains of guiripa, ocumare and panaquire form the southern chain of the coast, which stretches in a parallel direction from guigue to the mouth of the rio tuy, by the guesta of yusma and the guacimo. the latitudes of the villa de cura and san juan, so erroneously marked on our maps, enabled me to ascertain the mean breadth of the whole cordillera of venezuela. ten or twelve leagues may be reckoned as the distance from the descent of the northern chain which bounds the caribbean sea, to the descent of the southern chain bounding the immense basin of the llanos. this latter chain, which also bears the name of the inland mountains, is much lower than the northern chain; and i can hardly believe that the sierra de guayraima attains the height of 1200 toises.

[* the other hot springs of the cordillera of the shore are those of san juan, provisor, brigantin, the gulf of cariaco, cumucatar and irapa. mm. rivero and boussingault, who visited the thermal waters of mariara in february, 1823, during their journey from caracas to santa fe de bogota, found their maximum to be 64° centigrade. i found it at the same season only 59.2°. has the great earthquake of the 26th march, 1812, had an influence on the temperature of these springs? the able chemists above mentioned were, like myself, struck with the extreme purity of the hot waters that issue from the primitive rocks of the basin of aragua. those of onoto, which flow at the height of 360 toises above the level of the sea, have no smell of sulphuretted hydrogen; they are without taste, and cannot be precipitated, either by nitrate of silver or any other re-agent. when evaporated they have an inappreciable residue which consists of a little silica and a trace of alkali; their temperature is only 44.5°, and the bubbles of air which are disengaged at intervals are at onoto, as well as in the thermal waters of mariara, pure nitrogen. the waters of mariara (244 toises) have a faint smell of sulphuretted hydrogen; they leave, by evaporation, a slight residuum, that yields carbonic acid, sulphuric acid, soda, magnesia and lime. the quantities are so small that the water is altogether without taste. in the course of my journey i found only the springs of cumangillas hotter than the thermal waters of las trincheras: they are situated on the south of porto cabello. the waters of comangillas are at the height of 1040 toises and are alike remarkable for their purity and their temperature of 96.3° centigrade.]

[* the silla of caracas is only 80 toises lower than the canigou in the pyrenees.]

the two partial chains, that of the interior, and that which runs along the coast, are linked by a ridge or knot of mountains known by the names of altos de las cocuyzas (845 toises) and the higuerote (835 toises between los teques and la victoria) in longitude 69° 30′ and 69° 50′. on the west of this ridge lies the enclosed basin* of the lake of valencia or the valles de aragua; and on the east the basin of caracas and of the rio tuy. the bottom of the first-mentioned basins is between 220 and 250 toises high; the bottom of the latter is 460 toises above the level of the caribbean sea. it follows from these measures that the most western of the two longitudinal valleys enclosed by the littoral cordillera is the deepest; while in the plains near the apure and the orinoco the declivity is from west to east; but we must not forget that the peculiar disposition of the bottom of the two basins, which are bounded by two parallel chains, is a local phenomenon altogether separate from the causes on which the general structure of the country depends. the eastern basin of the cordillera of venezuela is not shut up like the basin of valencia. it is in the knot of the mountains of las cocuyzas, and of higuerote, that the serrania de los teques and oripoto, stretching eastward, form two valleys, those of the rio guayre and rio tuy; the former contains the town of caracas and both unite below the caurimare. the rio tuy runs through the rest of the basin, from west to east, as far as its mouth which is situated on the north of the mountains of panaquire.

[* this basin contains a small system of inland rivers which do not communicate with the ocean. the southern chain of the litteral cordillera of venezuela is so depressed on the south-west that the rio pao is separated from the tributary streams of the lake of tacarigua or valencia. towards the east the rio tuy, which takes its rise on the western declivity of the knot of mountains of las cocuyzas, appears at first to empty itself into the valleys of aragua; but hills of calcareous tufa, forming a ridge between consejo and victoria, force it to take its course south-east.]

cape codera seems to terminate the northern range of the littoral mountains of venezuela but this termination is only apparent. the coast forms a vast nook, thirty-five sea leagues in length, at the bottom of which is the mouth of the rio unare and the road of nueva barcelona. stretching first from west to east, in the parallel of 10° 37′, this coast recedes at the parallel 10° 6′, and resumes its original direction (10° 37′ to 10° 44′) from the western extremity of the peninsula of araya to the eastern extremities of montana de paria and the island of trinidad. from this dissection of the coast it follows that the range of mountains bordering the shore of the provinces of caracas and barcelona, between the meridian 66° 32′ and 68° 29′ (which i saw on the south of the bay of higuerote and on the north of the llanos of pao and cachipo), must be considered as the continuation of the southern chain of venezuela and as being linked on the west with the sierras de panaquire and ocumare. it may therefore be said that between cape codera and cariaco the inland chain itself forms the coast. this range of very low mountains, often interrupted from the mouth of the rio tuy to that of the rio neveri, rises abruptly on the east of nueva barcelona, first in the rocky island of chimanas, and then in the cerro del bergantin, elevated probably more than 800 toises, but of which the astronomical position and the precise height are yet alike unknown. on the meridian of cumana the northern chain (that of cape codera and the silla of caracas) again appears. the micaceous slate of the peninsula of araya and maniquarez joins by the ridge or knot of mountains of meapire the southern chain, that of panaquire the bergantin, turimiquiri, caripe and guacharo. this ridge, not more than 200 toises of absolute height, has, in the ancient revolutions of our planet, prevented the irruption of the ocean, and the union of the gulfs of paria and cariaco. on the west of cape codera the northern chain, composed of primitive granitic rocks, presents the loftiest summits of the whole cordillera of venezuela; but the culminant points east of that cape are composed in the southern chain of secondary calcareous rocks. we have seen above that the peak of turimiquiri, at the back of the cocollar, is 1050 toises, while the bottoms of the high valleys of the convent of caripe and of guardia de san augustin are 412 and 533 toises of absolute height. on the east of the ridge of meapire the southern chain sinks abruptly towards the rio arco and the guarapiche; but, on quitting the main land, we again see it rising on the southern coast of the island of trinidad which is but a detached portion of the continent, and of which the northern side unquestionably presents the vestiges of the northern chain of venezuela, that is, of the montana de paria (the paradise of christopher columbus), the peninsula of araya and the silla of caracas. the observations of latitude i made at the villa de cura (10° 2′ 47′), the farm of cocollar (10° 9′ 37′) and the convent of caripe (10° 10′ 14′), compared with the more anciently known position of the south coast of trinidad (latitude 10° 6′), prove that the southern chain, south of the basins of valencia and of tuy* and of the gulfs of cariaco and paria, is still more uniform in the direction from west to east than the northern chain from porto cabello to punta galera. it is highly important to know the southern limit of the littoral cordillera of venezuela because it determines the parallel at which the llanos or the savannahs of caracas, barcelona and cumana begin. on some well-known maps we find erroneously marked between the meridians of caracas and cumana two cordilleras stretching from north to south, as far as latitude 8 3/4°, under the names of cerros de alta gracia and del bergantin, thus describing as mountainous a territory of 25 leagues broad, where we should seek in vain a hillock of a few feet in height.

[* the bottom of the first of these four basins bounded by parallel chains is from 230 to 460 toises above, and that of the two latter from 30 to 40 toises below the present sea-level. hot springs gush from the bottom of the gulf of the basin of cariaco, as from the bottom of the basin of valencia on the continent.]

turning to the island of marguerita, composed, like the peninsula of araya, of micaceous slate, and anciently linked with that peninsula by the morro de chacopata and the islands of coche and cubagua, we seem to recognize in the two mountainous groups of macanao and la vega de san juan traces of a third coast-chain of the cordillera of venezuela. do these two groups of marguerita, of which the most westerly is above 600 toises high, belong to a submarine chain stretching by the isle of tortuga, towards the sierra de santa lucia de coro, on the parallel of 11°? must we admit that in latitude 11 1/4 and 12 1/2° a fourth chain, the most northerly of all, formerly stretched out in the direction of the island of hermanos, by blanquilla, los roques, orchila, aves, buen ayre, curacao and oruba, towards cape chichivacoa? these important problems can only be solved when the chain of islands parallel with the coast has been properly examined. it must not be forgotten that a great irruption of the ocean appears to have taken place between trinidad and grenada,* and that no where else in the long series of the lesser antilles are two neighbouring islands so far removed from each other. we observe the effect of the rotatory current in the direction of the coast of trinidad, as in the coasts of the provinces of cumana and caracas, between cape paria and punta araya and between cape codera and porto cabello. if a part of the continent has been overwhelmed by the ocean on the north of the peninsula of araya it is probable that the enormous shoal which surrounds cubagua, coche the island of marguerita, los frailes, la sola and the testigos marks the extent and outline of the submerged land. this shoal or placer, which is of the extent of 200 square leagues, is well known only to the tribe of the guayqueries; it is frequented by these indians on account of its abundant fishery in calm weather. the gran placer is believed to be separated only by some canals or deep furrows of the bank of grenada from the sand-bank that extends like a narrow dyke from tobago to grenada, and which is known by the lowering of the temperature of the water and from the sand-banks of los roques and aves. the guayquerie indians and, generally speaking, all the inhabitants of the coast of cumana and barcelona, are imbued with an idea that the water of the shoals of marguerita and the testigos diminishes from year to year; they believe that, in the lapse of ages, the morro do chacopata on the peninsula of araya will be joined by a neck of land to the islands of lobos and coche. the partial retreat of the waters on the coast of cumana is undeniable and the bottom of the sea has been upheaved at various times by earthquakes; but these local phenomena, which it is so difficult to account for by the action of volcanic force, the changes in the direction of currents, and the consequent swelling of the waters, are very different from the effects manifested at once over the space of several hundred square leagues.

[* it is affirmed that the island of trinidad is traversed in the northern part by a chain of primitive slate, and that grenada furnishes basalt. it would be important to examine of what rock the island of tobago is composed; it appeared to me of dazzling whiteness; and on what point, in going from trinidad northward, the trachytic and trappean system of the lesser antilles begins.]

4. group of the mountains of parime.

it is essential to mineralogical geography to designate by one name all the mountains that form one system. to attain this end, a denomination belonging to a partial group only may be extended over the whole chain; or a name may be employed which, by reason of its novelty, is not likely to give rise to homogenic mistakes. mountaineers designate every group by a special denomination; and a chain is generally considered as forming a whole only when it is seen from afar bounding the horizon of the plains. we find the name of snowy mountains (himalaya, imaus) repeated in every zone, white (alpes, alb), black and blue. the greater part of the sierra parime is, as it were, edged round by the orinoco. i have, however, avoided a denomination having reference to this circumstance, because the group of mountains to which i am about to direct attention extends far beyond the banks of the orinoco. it stretches south-east, towards the banks of the rio negro and the rio branco, to the parallel of 1 1/2° north latitude. the geographical name of parime has the advantage of reviving recollections of the fable of el dorado, and the lofty mountains which, in the sixteenth century, were supposed to surround the lake rupunuwini, or the laguna de parime. the missionaries of the orinoco still give the name of parime to the whole of the vast mountainous country comprehended between the sources of the erevato, the orinoco, the caroni, the rio parime* (a tributary of the rio branco) and the rupunuri or rupunuwini, a tributary of the rio essequibo. this country is one of the least known parts of south america and is covered with thick forests and savannahs; it is inhabited by independent indians and is intersected by rivers of dangerous navigation, owing to the frequency of shoals and cataracts.

[* the rio parime, after receiving the waters of the uraricuera, joins the tacutu, and forms, near the fort of san joacquim, the rio branco, one of the tributary streams of the rio negro.]

the system of the mountains of parime separates the plains of the lower orinoco from those of the rio negro and the amazon; it occupies a territory of trapezoidal form, comprehended between the parallels of 3 and 8°, and the meridians of 61 and 70 1/2°. i here indicate only the elements of the loftiest group, for we shall soon see that towards south-east the mountainous country, in lowering, draws near the equator, as well as to french and portuguese guiana. the sierra parime extends most in the direction north 85° west and the partial chains into which it separates on the westward generally follow the same direction. it is less a cordillera or a continuous chain in the sense given to those denominations when applied to the andes and caucasus than an irregular grouping of mountains separated the one from the other by plains and savannahs. i visited the northern, western and southern parts of the sierra parime, which is remarkable by its position and its extent of more than 25,000 square leagues. from the confluence of the apure, as far as the delta of the orinoco, it is uniformly three or four leagues removed from the right bank of the great river; only some rocks of gneiss-granite, amphibolic slate and greenstone advance as far as the bed of the orinoco and create the rapids of torno and of la boca del infierno.* i shall name successively, from north-north-east to south-south-west, the different chains seen by m. bonpland and myself as we approached the equator and the river amazon. first. the most northern chain of the whole system of the mountains of parime appeared to us to be that which stretches (latitude 7° 50′) from the rio arui, in the meridian of the rapids of camiseta, at the back of the town of angostura, towards the great cataracts of the rio carony and the sources of the imataca. in the missions of the catalonian capuchins this chain, which is not 300 toises high, separates the tributary streams of the orinoco and those of the rio cuyuni, between the town of upata, cupapui and santa marta. westward of the meridian of the rapids of camiseta (longitude 67° 10′) the high mountains in the basin of the rio caura only commence at 7° 20′ of latitude, on the south of the mission of san luis guaraguaraico, where they occasion the rapids of mura. this chain stretches westward by the sources of the rio cuchivero, the cerros del mato, the cerbatana and maniapure, as far as tepupano, a group of strangely-formed granitic rocks surrounding the encaramada. the culminant points of this chain (latitude 7° 10′ to 7° 28′) are, according to the information i gathered from the indians, situated near the sources of cano de la tortuga. in the chain of the encaramada there are some traces of gold. this chain is also celebrated in the mythology of the tamanacs; for the painted rocks it contains are associated with ancient local traditions. the orinoco changes its direction at the confluence of the apure, breaking a part of the chain of the encaramada. the latter mountains and scattered rocks in the plain of the capuchino and on the north of cabruta may be considered either as the vestiges of a destroyed spur or (on the hypothesis of the igneous origin of granite) as partial eruptions and upheavings. i shall not here discuss the question whether the most northerly chain, that of angostura and of the great fall of carony, be a continuation of the chain of encaramada. third. in navigating the orinoco from north to south we observe, alternately, on the east, small plains and chains of mountains of which we cannot distinguish the profiles, that is, the sections perpendicular to their longitudinal axes. from the mission of the encaramada to the mouth of the rio qama i counted seven recurrences of this alternation of savannahs and high mountains. first, on the south of the isle cucuruparu rises the chain of chaviripe (latitude 7° 10′); it stretches, inclining towards the south (latitude 6° 20′ to 6° 40′), by the cerros del corozal, the amoco, and the murcielago, as far as the erevato, a tributary of the caura. it there forms the rapids of paru and is linked with the summits of matacuna. fourth. the chain of chaviripe is succeeded by that of the baraguan (latitude 6° 50′ to 7° 5′), celebrated for the strait of the orinoco, to which it gives its name. the saraguaca, or mountain of uruana, composed of detached blocks of granite, may be regarded as a northern spur of the chain of the baraguan, stretching south-west towards siamacu and the mountains (latitude 5° 50′) that separate the sources of the erevato and the caura from those of the ventuari. fifth. the chain of carichana and of paruaci (latitude 6° 25′), wild in aspect, but surrounded by charming meadows. piles of granite crowned with trees and insulated rocks of prismatic form (the mogote of cocuyza and the marimaruta or castillito of the jesuits) belong to this chain. sixth. on the western bank of the orinoco, which is low and flat, the peak of uniana rises abruptly more than 3000 feet high. the spurs (latitude 5° 35′ to 5° 40′) which this peak sends eastward are crossed by the orinoco in the first great cataract (that of mapura or the atures); further on they unite together and, rising in a chain, stretch towards the sources of the cataniapo, the rapids of ventuari, situated on the north of the confluence of the asisi (latitude 5° 10′) and the cerro cunevo. seventh. five leagues south of the atures is the chain of quittuna, or of maypures (latitude 15° 13′), which forms the bar of the second great cataract. none of those lofty summits are situated on the west of the orinoco; on the east of that river rises the cunavami, the truncated peak of calitamini and the jujamari, to which father gili attributes an extraordinary height. eighth. the last chain of the south-west part of the sierra parime is separated by woody plains from the chain of maypures; it is the chain of the cerros de sipapo (latitude 4° 50′); an enormous wall behind which the powerful chief of the guaypunabi indians intrenched himself during the expedition of solano. the chain of sipapo may be considered as the beginning of the range of lofty mountains which bound, at the distance of some leagues, the right bank of the orinoco, where that river runs from south-east to north-west, between the mouth of the ventuari, the jao and the padamo (latitude 3° 15′). in ascending the orinoco, above the cataract of maypures, we find, long before we reach the point where it turns, near san fernando del atabapo, the mountains disappearing from the bed of the river, and from the mouth of the zama there are only insulated rocks in the plains. the chain of sipapo forms the south-west limit of the system of mountains of parime, between 70 1/2 and 68° of longitude. modem geologists have observed that the culminant points of a group are less frequently found at its centre than towards one of its extremities, preceding, and announcing in some sort, a great depression* of the chain. this phenomenon is again observed in the group of the parime, the loftiest summits of which, the duida and the maraguaca, are in the most southerly range of mountains, where the plains of the cassiquiare and the rio negro begin.

[* to this series of advanced rocks also belong those which pierce the soil between the rio aquire and the rio barima; the granitic and amphibolic rocks of the vieja guayana and of the town of angostura; the cerro de mono on the south-east of muitaco or real corono; the cerro of taramuto near the alta gracia, etc.]

[* as seen in mont blanc and chimborazo.]

these plains or savannahs which are covered with forests only in the vicinity of the rivers do not, however, exhibit the same uniform continuity as the llanos of the lower orinoco, of the meta and of buenos ayres. they are interrupted by groups of hills (cerros de daribapa) and by insulated rocks of grotesque form which pierce the soil and from a distance fix the attention of the traveller. these granitic and often stratified masses resemble the ruins of pillars or edifices. the same force which upheaved the whole group of the sierra parime has acted here and there in the plains as far as beyond the equator. the existence of these steeps and sporadic hills renders it difficult to determine the precise limits of a system in which the mountains are not longitudinally ranged as in a vein. as we advance towards the frontier of the portuguese province of the rio negro the high rocks become more rare and we no longer find the shelves or dykes of gneiss-granite which cause rapids and cataracts in the rivers.

such is the surface of the soil between 68 1/2 and 70 1/2° of longitude, between the meridian of the bifurcation of the orinoco and that of san fernando de atabapo; further on, westward of the upper rio negro, towards the source of that river, and its tributary streams the xie and the uaupes (latitude 1 to 2 1/4°, longitude 72 to 74°) lies a small mountainous tableland, in which indian traditions place a laguna de oro, that is, a lake surrounded with beds of auriferous earth.* at maroa, the most westerly mission of the rio negro, the indians assured me that that river as well as the inirida (a tributary of the guavare) rises at the distance of five days’ march, in a country bristled with hills and rocks. the natives of san marcellino speak of a sierra tunuhy, nearly thirty leagues west of their village, between the xie and the icanna. la condamine learned also from the indians of the amazon that the quiquiari comes from a country of mountains and mines. now, the iquiari is placed by the french astronomer between the equator and the mouth of the xie (ijie), which identifies it with the iguiare that falls into the icanna. we cannot advance in the geologic knowledge of america without having continually recourse to the researches of comparative geography. the small system of mountains, which we may provisionally call that of the sources of the rio negro and the uaupes, and the culminant points of which are not probably more than 100 or 120 toises high, appears to extend southward to the basin of rio yupura, where rocky ridges form the cataracts of the rio de los enganos and the salto grande de yupura (south latitude 0° 40′ to north latitude 0° 28′), and the basin of the upper guaviare towards the west. we find in the course of this river, from 60 to 70 leagues west of san fernando del atabapo, two walls of rocks bounding the strait (nearly 3° 10′ north latitude and 73 3/4° longitude) where father maiella terminated his excursion. that missionary told me that, in going up the guaviare, he perceived near the strait (angostura) a chain of mountains bounding the horizon on the south. it is not known whether those mountains traverse the guaviare more to the west, and join the spurs which advance from the eastern cordillera of new grenada, between the rio umadea and the rio ariari, in the direction of the savannahs of san juan de los llanos. i doubt the existence of this junction. if it really existed, the plains of the lower orinoco would communicate with those of the amazon only by a very narrow land-strait, on the east of the mountainous country which surrounds the source of the rio negro: but it is more probable that this mountainous country (a small system of mountains, geognostically dependent on the sierra parime) forms as it were an island in the llanos of guaviare and yupura. father pugnet, principal of the franciscan convent at popayan, assured me, that when he went from the missions settled on the rio caguan to aramo, a village situated on the rio guayavero, he found only treeless savannahs, extending as far as the eye could reach. the chain of mountains placed by several modern geographers, between the meta and the vichada, and which appears to link the andes of new grenada with the sierra parime, is altogether imaginary.

[* according to the journals of acunha and fritz the manao indians (manoas) obtained from the banks of the yquiari (iguiare or iguare) gold of which they made thin plates. the manuscript notes of don apollinario also mention the gold of the rio uaupes. la condamine, voyage a l’amazone. we must not confound the laguna de oro, which is said to be found in going up the uaupes (north latitude 0° 40′) with another gold lake (south latitude 1 degree 10 minutes) which la condamine calls marahi or morachi (water), and which is merely a tract often inundated between the sources of the jurubech (urubaxi) and the rio marahi, a tributary stream of the caqueta.]

we have now examined the prolongation of the sierra parime on the west, towards the source of the rio negro: it remains for us to follow the same group in its eastern direction. the mountains of the upper orinoco, eastward of the raudal of the guaharibos (north latitude 1 degree 15 minutes longitude 67° 38′), join the chain of pacaraina, which divides the waters of the carony and the rio branco, and of which the micaceous schist, resplendent with silvery lustre, figures so conspicuously in raleigh’s el dorado. the part of that chain containing the sources of the orinoco has not yet been explored; but its prolongation more to the east, between the meridian of the military post of guirior and the rupunuri, a tributary of the essequibo, is known to me through the travels of the spaniards antonio santos and nicolas rodriguez, and also by the geodesic labours of two portuguese, pontes and almeida. two portages but little frequented* are situated between the rio branco and the rio essequibo, south of the chain of pacaraina; they shorten the land-road leading from the villa del rio negro to dutch guiana. on the contrary, the portage between the basin of the rio branco and that of the carony crosses the summit of the chain of pacaraina. on the northern slope of this chain rises the anocapra, a tributary of the paraguamusi or paravamusi; and on the southern slope, the araicuque, which, with the uraricapara, forms the famous valley of inundations, above the destroyed mission of santa rosa (latitude 3° 46′, longitude 65° 10′). the principal cordillera, which appears of little breadth, stretches on a length of 80 leagues, from the portage of anocapra (longitude 65° 35′) to the left bank of the rupunuri (longitude 61° 50′), following the parallels of 4° 4′ and 4° 12′. we there distinguish from west to east the mountains of pacaraina, tipique, tauyana, among which rises the rio parime (a tributary of the uraricuera), tubachi, christaux (latitude 3° 56′, longitude 62° 52′) and canopiri. the spanish traveller, rodriguez, marks the eastern part of the chain by the name of quimiropaca; but preferring to adopt general names, i continue to give the name of pacaraina to the whole of this cordillera which links the mountains of the orinoco to the interior of dutch and french guiana, and which raleigh and keymis made known in europe at the end of the 16th century. this chain is broken by the rupunuri and the essequibo, so that one of their tributary streams, the tavaricuru, takes its rise on the southern declivity, and the other, the sibarona, on the northern. on approaching the essequibo, the mountains are more developed towards the south-east, and extend beyond 2 1/2° north latitude. from this eastern branch of the chain of pacaraina the rio rupunuri rises near the cerro uassari. on the right bank of the rio branco, in a still more southern latitude (between 1 and 2° north) is a mountainous territory in which the caritamini, the padaviri, the cababuri (cavaburis) and the pacimoni take their source, from east to west. this western branch of the mountains of pacaraina separates the basin of rio branco from that of the upper orinoco, the sources of which are probably not found east of the meridian of 66 15 minutes: it is linked with the mountains of unturan and yumariquin, situated south-east of the mission of esmeralda. thence it results that, while on the west of the cassiquiare, between that river, the atabapo, and the rio negro, we find only vast plains, in which rise some little hills and insulated rocks; real spurs stretch eastward of the cassiquiare, from north-west to south-east, and form a continued mountainous territory as far as 2° north latitude. the basin only, or rather the transversal valley of the rio branco, forms a kind of gulf, a succession of plains and savannahs (campos) several of which penetrate from south to north, into the mountainous land between the eastern and western branches of the chain of pacaraina, to the distance of eight leagues north of the parallel of san joaquin.

[* the portages of sarauru and the lake amucu.]

we have just examined the southern part of the vast system of the mountains of parime, between 2 and 4° of latitude, and between the meridians of the sources of the orinoco and the essequibo. the development of this system of mountains northward between the chain of pacaraina and rio cuyuni, and between the meridians 66 and 61 3/4°, is still less known. the only road frequented by white men is that of the river paragua, which receives the paraguamusi, near the guirior. we find indeed, in the journal of nicolas rodriguez, that he was constantly obliged to have his canoe carried by men (arrastrando) past the cataracts which intercept the navigation; but we must not forget a circumstance of which my own experience furnished me with frequent proofs — that the cataracts in this part of south america are often caused only by ridges of rocks which do not form mountains. rodriguez names but two between barceloneta and the mission of san jose; while the missionaries place more to the east, in 6° latitude, between the rio caroni and the cuyuni, the serranias of usupama and rinocote. the latter crosses the mazaruni, and forms thirty-nine cataracts in the essequibo, from the military post of arinda (latitude 5° 30′) to the mouth of rupunuri.

with respect to the continuation of the system of the mountains of parime, south-east of the meridian of the essequibo, the materials are entirely wanting for tracing it with precision. the whole interior of dutch, french and portuguese guiana is a terra incognita; and the astronomical geography of those countries has scarcely made any progress during the space of thirty years. if the american limits recently fixed between france and portugal should one day cease to be mere diplomatic illusions and acquire reality in being traced on the territory by means of astronomical observations (as was projected in 1817), this undertaking would lead geographical engineers to that unknown region which, at 3 1/2° west of cayenne, divides the waters between the coast of guiana and the amazon. till that period, which the political state of brazil seems to retard, the geognostic table of the group of parime can only be completed by scattered notions collected in the portuguese and dutch colonies. in going from the uassari mountains (latitude 2° 25′, longitude 61° 50′) which form a part of the eastern branch of the cordillera of pacaraina, we find towards the east a chain of mountains, called by the missionaries acaray and tumucuraque. those two names are found on our maps between 1/2 and 3° north latitude. raleigh first made known, in 1596, the system of the mountains of parime, between the sources of the rio carony and the essequibo, by the name of wacarima (pacarima), and the jesuits acunha and artedia furnished, in 1639, the first precise notions of that part of this system which extends from the meridian of essequibo to that of oyapoc. there they place the mountains of yguaracuru and paraguaxo, the former of which gives birth to a gold river (rio de oro), a tributary of the curupatuba;* and according to the assertion of the natives, subterraneous noises are sometimes heard from the latter. the ridge of this chain of mountains, which runs in a direction south 85° east from the peak of duida near the esmeralda (latitude 3° 19′), to the rapids of the rio manaye near cape nord (latitude 1 degree 50 minutes), divides, in the parallel of 2°, the northern sources of the essequibo, the maroni and the oyapoc, from the southern sources of the rio trombetas, curupatuba and paru. the most southern spurs of this chain approach nearer to the amazon, at the distance of fifteen leagues. these are the first heights which we perceived after having left xeberos and the mouth of the huallaga. they are constantly seen in navigating from the mouth of the rio topayo towards that of paru, from the town of santarem to almeirim. the peak tripoupou is nearly in the meridian of the former of those towns and is celebrated among the indians of upper maroni. it is said that farther eastward, at melgaco, the serras do velho and do paru are still distinguished in the horizon. the real boundaries of this series of sources of the rio trombetas are better known southward than northward, where a mountainous country appears to advance in dutch and french guiana, as far as within twenty to twenty-five leagues of the coast. the numerous cataracts of the rivers of surinam, maroni and oyapoc, prove the extent and the prolongation of rocky ridges; but in those regions nothing indicates the existence of continued plains or table-lands some hundred toises high, fitted for the cultivation of the plants of the temperate zone.

[* when we know that in tamanac gold is called caricuri; in carib, caricura: in peruvian, cori (curi), we easily recognize in the names of the mountains and rivers (yguara-curu, cura-patuba) which we have just marked, the indication of auriferous soil. such is the analogy of the imported roots in the american tongues, which otherwise differ altogether from each other, that 300 leagues west of the mountain ygaracuru, on the banks of the caqueta, pedro de ursua heard of the province of caricuri, rich in gold washings. the curupatuba falls into the amazon near the villa of monte alegre, north-east of the mouth of the rio topayos.]

the system of the mountains of parime surpasses in extent nineteen times that of the whole of switzerland. even considering the mountainous group of the sources of the rio negro and the xie as independent or insulated amidst the plains, we still find the sierra parime (between maypures and the sources of the oyapoc) to be 340 leagues in length; its greatest breadth (the rocks of imataca, near the delta of the orinoco, at the sources of the rio paru) is 140 leagues. in the group of the parime, as well as in the group of the mountains of central asia, between the himalaya and the altai, the partial chains are often interrupted and have no uniform parallelism. towards the south-west, however (between the strait of baraguan, the mouth of the rio zama and the esmeralda), the line of the mountains is generally in the direction of north 70° west. such is also the position of a distant coast, that of portuguese, french, dutch and english guiana, from cape north to the mouth of the orinoco; such is the mean direction of the course of the rio negro and yupura. it is desirable to fix our attention on the angles formed by the partial chains, in different regions of america, with the meridians; because on less extended surfaces, for instance in germany, we find also this singular co-existence of groups of neighbouring mountains following laws of direction altogether different, though every separate group exhibits the greatest uniformity in the line of chains.

the soil on which the mountains of parime rise, is slightly convex. by barometric measures i found that, between 3 and 4° north latitude, the plains are elevated from 160 to 180 toises above sea-level. this height will appear considerable if we reflect that at the foot of the andes of peru, at tomependa, 900 leagues from the coast of the atlantic ocean, the llanos or plains of the amazon rise only to the height of 194 toises. the distinctive characteristics of the group of the mountains of parime are the rocks of granite and gneiss-granite, the total absence of calcareous secondary formations, and the shelves of bare rock (the tsy of the chinese deserts), which occupy immense spaces in the savannahs.

5. group of the brazil mountains.

this group has hitherto been marked on the maps in a very erroneous way. the temperate table-lands and real chains of 300 to 500 toises high have been confounded with countries of exceedingly hot temperature, and of which the undulating surface presents only ranges of hills variously grouped. but the observations of scientific travellers have recently thrown great light on the orography of portuguese america. the mountainous region of brazil, of which the mean height rises at least to 400 toises, is comprehended within very narrow limits, nearly between 18 and 28° south latitude; it does not appear to extend, between the provinces of goyaz and matogrosso, beyond longitude 53° west of the meridian of paris.

when we regard in one view the eastern configuration of north and south america, we perceive that the coast of brazil and guiana, from cape saint roque to the mouth of the orinoco (stretching from south-east to north-west), corresponds with that of labrador, as the coast from cape saint roque to the rio de la plata corresponds with that of the united states (stretching from south-west to north-east). the chain of the alleghenies is opposite to the latter coast, as the principal cordilleras of brazil are nearly parallel to the shore of the provinces of porto seguro, rio janeiro and rio grande. the alleghenies, generally composed of grauwacke and transition rocks, are somewhat loftier than the almost primitive mountains (of granite, gneiss and mica-slate) of the brazilian group; they are also of a far more simple structure, their chains lying nearer to each other and preserving, as in the jura, a more uniform parallelism.

if, instead of comparing those parts of the new continent situated north and south of the equator, we confine ourselves to south america, we find on the western and northern coasts in their whole length, a continued chain near the shore (the andes and the cordillera of venezuela), while the eastern coast presents masses of more or less lofty mountains only between the 12 and 30° south latitude. in this space, 360 leagues in length, the system of the brazil mountains corresponds geologically in form and position with the andes of chile and peru. its most considerable portion lies between the parallels 15 and 22°, opposite the andes of potosi and la paz, but its mean height is five toises less, and cannot even be compared with that of the mountains of parime, jura and auvergne. the principal direction of the brazilian chains, where they attain the height of from four to five hundred toises, is from south to north, and from south-south-west to north-north-east; but, between 13 and 19° the chains are considerably enlarged, and at the same time lowered towards the west. ridges and ranges of hills seem to advance beyond the land-straits which separate the sources of the rio araguay, parana, topayos, paraguay, guapore and aguapehy, in 63° longitude. as the western widening of the brazilian group, or rather the undulations of the soil in the campos parecis, correspond with the spurs of santa cruz de la sierra, and beni, which the andes send out eastward, it was formerly concluded that the system of the mountains of brazil was linked with that of the andes of upper peru. i myself laboured under this error in my first geologic studies.

a coast chain (serra do mar) runs nearly parallel with the coast, north-east of rio janeiro, lowering considerably towards rio doce, and losing itself almost entirely near bahia (latitude 12° 58′). according to m. eschwege* some small ridges reach cape saint roque (latitude 5° 12′). south-east of rio janeiro the serra do mar follows the coast behind the island of saint catherine as far as torres (latitude 29° 20′); it there turns westward and forms an elbow stretching by the campos of vacaria towards the banks of the jacuy.

[* geognostiches gemulde von brasilien, 1822. the limestone of bahia abounds in fossil wood.]

another chain is situated westward of the shore-chain of brazil. this is the most lofty and considerable of all and is called the chain of villarica. mr. eschwege distinguishes it by the name of serra do espinhaco and considers it as the principal part of the whole structure of the mountains of brazil. this cordillera loses itself northward,* between minas novas and the southern extremity of the capitania of bahia, in 16° latitude. it is there more than 60 leagues removed from the coast of porto seguro; but southward, between the parallels of rio janeiro and saint paul (latitude 22 to 23°), in the knot of the mountains of serra da mantiquiera, it draws so near to the cordillera of the shore (serra do mar), that they are almost confounded together. in the same manner the serra do espinhaco follows constantly the direction of a meridian, towards the north; while towards the south it runs south-east, and terminates about 25° latitude. the chain reaches its highest elevation between 18 and 21°; and there the spurs and table-lands at its back are of sufficient extent to furnish lands for cultivation where, at successive heights, there are temperate climates comparable to the delicious climates of xalapa, guaduas, caracas and caripe. this advantage, which depends at once on the widening of the mass of the chain and of its spurs, is nowhere found in the same degree east of the andes, not even in chains of more considerable absolute height, as those of venezuela and the orinoco. the culminant points of the serra do espinhaco, in the capitania of minas geraes, are the itambe (932 toises), the serra da piedade, near sabara (910 toises), the itacolumi, properly itacunumi (900 toises), the pico of itabira (816 toises), the serras of caraca, ibitipoca and papagayo. saint hilaire felt piercing cold in the month of november (therefore in summer) in the whole cordillera of lapa, from the villa do principe to the morro de gaspar suares.

[* the rocky ridges that form the cataract of paulo affonso, in the rio san francisco, are supposed to belong to the northern prolongation of the serra do espinhaco, as a series of heights in the province of seara (fetid calcareous rocks containing a quantity of petrified fish) belong to the serra dos vertentes.]

we have just noticed two chains of mountains nearly parallel but of which the most extensive (the littoral chain) is the least lofty. the capital of brazil is situated at the point where the two chains draw nearest together and are linked together on the east of the serra de mantiqueira, if not by a transversal ridge, at least by a mountainous territory. old systematic ideas respecting the rising of mountains in proportion as we advance into a country, would have warranted the belief that there existed, in the capitania of mato grosso, a central cordillera much loftier than that of villarica or do espinhaco; but we now know (and this is confirmed by climateric circumstances) that there exists no continued chain, properly speaking, westward of rio san francisco, on the frontiers of minas geraes and goyaz. we find only a group of mountains, of which the culminant points are the serras da canastra (south-west of paracatu) and da marcella (latitude 18 1/2 and 19.10°), and, further north, the pyrenees stretching from east to west (latitude 16° 10′) between villaboa and mejaponte). m. eschwege has named the group of mountains of goyaz the serra dos vertentes, because it divides the waters between the southern tributary streams of the rio grande or parana, and the northern tributary streams of rio tucantines. it runs southward beyond the rio grande (parana), and approaches the chain of espinpapo in 23° latitude, by the serra do franca. it attains only the height of 300 or 400 toises, with the exception of some summits north-west of paracatu, and is consequently much lower than the chain of villarica.

further on, west of the meridian of villaboa, there are only ridges and a series of low hills which, on a length of 12°, form the division of water (latitude 13 to 17°) between the araguay and the paranaiba (a tributary of the parana), between the rio topayos and the paraguay, between the guapore and the aguapehy. the serra of san marta (longitude 15 1/2°) is somewhat lofty, but maps have vastly exaggerated the height of the serras or campos parecis north of the towns of cuyaba and villabella (latitude 13 to 14°, longitude 58 to 62°). these campos, which take their name from that of a tribe of wild indians, are vast, barren table-lands, entirely destitute of vegetation; and in them the sources of the tributary streams of three great rivers, the topayos, the madeira and the paraguay, take their rise.

according to the measures and geologic observations of m. eschwege, the high summits of the serra do mar (the coast-chain) scarcely attain 660 toises; those of the serra do espinhaco (chain of villarica), 950 toises; those of serra de los vertentes (group of canastra and the brazilian pyrenees), 450 toises. further west the surface of the soil seems to present but slight undulations; but no measure of height has been made beyond the meridian of villaboa. considering the system of the mountains of brazil in their real limits, we find, except some conglomerates, the same absence of secondary formations as in the system of the mountains of the orinoco (group of parime). these secondary formations, which rise to considerable heights in the cordillera of venezuela and cumana, belong only to the low regions of brazil.

b. plains (llanos) or basins.

in that part of south america situated on the east of the andes we have successively examined three systems of mountains, those of the shore of venezuela, of the parime and brazil: we have seen that this mountainous region, which equals the cordillera of the andes, not in mass, but in area and horizontal section of surface, is three times less elevated, much less rich in precious metals adhering to the rock, destitute of recent traces of volcanic fire and, with the exception of the coast of venezuela, little exposed to the violence of earthquakes. the average height of the three systems diminishes from north to south, from 750 to 400 toises; those of the culminant points (maxima of the height of each group) from 1350 to 1000 or 900 toises. hence it results that the loftiest chain, with the exception of the small insulated system of the sierra nevada of santa marta, is the cordillera of the shore of venezuela, which is itself but a continuation of the andes. directing our attention northward, we find in central america (latitude 12 to 30°) and north america (latitude 30 to 70°), on the east of the andes of guatimala, mexico and upper louisiana, the same regular lowering which struck us towards the south. in this vast extent of land, from the cordillera of venezuela to the polar circle, eastern america presents two distinct systems, the group of the mountains of the west indies (which in its eastern part is volcanic) and the chain of the alleghenies. the former of these systems, partly covered by the ocean, may be compared, with respect to its relative position and form, to the sierra parime; the latter, to the brazil chains, running also from south-west to north-east. the culminant points of those two systems rise to 1138 and 1040 toises. such are the elements of this curve, of which the convex summit is in the littoral chain of venezuela:

america, east of the andes.

column 1: systems of mountains.

column 2: maxima of heights in toises.

brazil group: itacolumi 900 (south latitude 20 1/2°).

parime group: duida 1300 (north latitude 3 1/4°).

littoral chain of venezuela: silla of caracas 1350 (north latitude 10 1/2°).

group of the west indies: blue mountains 1138 (north latitude 18 1/5°).

chain of the alleghenies: mount washington 1040 (north latitude 44 1/4°).

i have preferred indicating in this table the culminant points of each system to the mean height of the line of elevation; the culminant points are the results of direct measures, while the mean height is an abstract idea somewhat vague, particularly when there is only one group of mountains, as in brazil, parime and the west indies, and not a continued chain. although it cannot be doubted that, among the five systems of mountains on the east of the andes, of which one only belongs to the southern hemisphere, the littoral chain of venezuela is the most elevated (having a culminant point of 1350 toises, and a mean height from the line of elevation of 750), we yet recognise with surprise that the mountains of eastern america (whether continental or insular) differ very inconsiderably in their height above the level of the sea. the five groups are all nearly of an average height of from 500 to 700 toises; and the culminant points (maxima of the lines of elevation) from 1000 to 1300 toises. that uniformity of structure, in an extent twice as great as europe, appears to me a very remarkable phenomenon. no summit east of the andes of peru, mexico and upper louisiana rises beyond the limit of perpetual snow.* it may be added that, with the exception of the alleghenies, no snow falls sporadically in any of the eastern systems which we have just examined. from these considerations it results, and above all, from the comparison of the new continent with those parts of the old world which we know best, with europe and asia, that america, thrown into the aquatic hemisphere* of our planet, is still more remarkable for the continuity and extent of the depressions of its surface, than for the height and continuity of its longitudinal ridge. beyond and within the isthmus of panama, but eastward of the cordillera of the andes, the mountains scarcely attain, over an extent of 600,000 square leagues, the height of the scandinavian alps, the carpathians, the monts–dores (in auvergne) and the jura. one system only, that of the andes, comprises in america, over a long and narrow zone of 3000 leagues, all the summits exceeding 1400 toises high. in europe, on the contrary, even considering the alps and the pyrenees as one sole line of elevation, we still find summits far from this line or principal ridge, in the sierra nevada of grenada, sicily, greece, the apennines, perhaps also in portugal, from 1500 to 1800 toises high.* the contrast between america and europe, with respect to distribution of the culminant points, which attain from 1300 to 1500 toises, is the more striking, as the low eastern mountains of south america, of which the maximum of elevation is only from 1300 to 1400 toises, are situated beside a cordillera of which the mean height exceeds 1800 toises, while the secondary system of the mountains of europe rises to maxima of elevation of 1500 to 1800 toises, near a principal chain of at least 1200 toises of average height.

[* not even the white mountains of the state of new hampshire, to which mount washington belongs. long before the accurate measurement of captain partridge i had proved (in 1804), by the laws of the decrement of heat, that no summit of the white mountains could attain the height assigned to them by mr. cutler, of 1600 toises.]

[* the southern hemisphere, owing to the unequal distribution of seas and continents, has long been marked as eminently aquatic; but the same inequality is found when we consider the globe as divided not according to the equator but by meridians. the great masses of land are stinted between the meridian of 10° west, and 150° east of paris, while the hemisphere eminently aquatic begins westward of the meridian of the coast of greenland, and ends on the east of the meridian of the eastern coast of new holland and the kurile isles. this unequal distribution of land and water has the greatest influence on the distribution of heat over the surface of the globe, on the inflexions of the isothermal lines, and the climateric phenomena in general. for the inhabitants of the central parts of europe the aquatic hemisphere may be called western, and the land hemisphere eastern; because in going to the west we reach the former sooner than the latter. it is the division according to the meridians, which is intended in the text. till the end of the 15th century the western hemisphere was as much unknown to the nations of the eastern hemisphere, as one half of the lunar globe is to us at present, and will probably always remain.]

[* culminant points; malhacen of grenada, 1826 toises; etna, according to captain william henry smith, 1700 toises; monte corno of the apennines, 1489 toises. if mount tomoros in greece and the serra gaviarra of portugal enter, as is alleged, into the limit of perpetual snow, those summits, according to their position in latitude, should attain from 1400 to 1600 toises. yet on the loftiest mountains of greece, tomoros, olympus in thessaly, polyanos in dolope and mount parnassus, m. pouqueville saw, in the month of august, snow lying only in patches, and in cavities sheltered from the rays of the sun.]

maxima of the line of elevation in the same parallels.

andes of chile, upper peru. knots of the mountains of porco and cuzco, 2500 toises.: group of the brazil mountains; a little lower than the cevennes 900 to 1000 toises.

andes of popayan and cundinamarca. chain of guacas, quindiu, and antioquia. more than 2800 toises.: group of parime mountains; little lower than the carpathians; 1300 toises.

insulated group of the snowy mountains of santa marta. it is believed to be 3000 toises high.: littoral chain of venezuela; 80 toises lower than the scandinavian alps; 1350 toises.

volcanic andes of guatimala, and primitive andes of oaxaca, from 1700 to 1800 toises.: group of the west indies, 170 toises higher than the mountains of auvergne, 1140 toises.

andes of new mexico and upper louisiana (rocky mountains) and further west. the maritime alps of new albion, 1600 to 1900 toises.: chain of the alleghenies; 160 toises higher than the chains of jura and the gates of malabar; 1040 toises.

this table contains the whole system of mountains of the new continent; namely: the andes, the maritime alps of california or new albion and the five groups of the east.

i may subjoin to the facts i have just stated an observation equally striking; in europe the maxima of secondary systems, which exceed 1500 toises, are found solely on the south of the alps and pyrenees, that is, on the south of the principal continental ridge. they are situated on the side where that ridge approaches nearest the shore, and where the mediterranean has not overwhelmed the land. on the north of the alps and pyrenees, on the contrary, the most elevated secondary systems, the carpathian and the scandinavian mountains* do not attain the height of 1300 toises. the depression of the line of elevation of the second order is consequently found in europe as well as in america, where the principal ridge is farthest removed from the shore. if we did not fear to subject great phenomena to too small a scale, we might compare the difference of the height of the alps and the mountains of eastern america, with the difference of height observable between the alps or the pyrenees, and the monts dores, the jura, the vosges or the black forest.

[* the lomnitzer spitz of the carpathians is, according to m. wahlenberg, 1245 toises; sneehattan, in the chain of dovrefjeld in norway (the highest summit of the old continent, north of the parallel of 55°), is 1270.]

we have just seen that the causes which upheaved the oxidated crust of the globe in ridges, or in groups of mountains, have not acted very powerfully in the vast extent of country stretching from the eastern part of the andes towards the old world; that depression and that continuity of plains are geologic facts, the more remarkable, as they extend nowhere else in other latitudes. the five mountain systems of eastern america, of which we have stated the limits, divide that part of the continent into an equal number of basins of which only that of the caribbean sea remains submerged. from north to south, from the polar circle to the straits of magellan, we see in succession:

1. the basin of the mississippi and of canada.

an able geologist, mr. edwin james, has recently shown that this basin is comprehended between the andes of new mexico, or upper louisiana, and the chains of the alleghenies which stretch northward in crossing the rapids of quebec. it being quite as open northward as southward, it may be designated by the collective name of the basin of the mississippi, the missouri, the river st. lawrence, the great lakes of canada, the mackenzie river, the saskatchewan and the coast of hudson’s bay. the tributary streams of the lakes and those of the mississippi are not separated by a chain of mountains running from east to west, as traced on several maps; the line of partition of the waters is marked by a slight ridge, a rising of two counter-slopes in the plain. there is no chain between the sources of the missouri and the assineboine, which is a branch of the red river and of hudson’s bay. the surface of these plains, almost all savannah, between the polar sea and the gulf of mexico, is more than 270,000 square sea leagues, nearly equal to the area of the whole of europe. on the north of the parallel of 42° the general slope of the land runs eastward; on the south of that parallel it inclines southward. to form a precise idea how little abrupt are these slopes we must recollect that the level of lake superior is 100 toises; that of lake erie, 88 toises, and that of lake ontario, 36 toises above the level of the sea. the plains around cincinnati (latitude 39° 6′) are scarcely, according to mr. drake, 80 toises of absolute height. towards the west, between the ozark mountains and the foot of the andes of upper louisiana (rocky mountains, latitude 35 to 38°), the basin of the mississippi is considerably elevated in the vast desert described by mr. nuttal. it presents a series of small table-lands, gradually rising one above another, and of which the most westerly (that nearest the rocky mountains, between the arkansas and the padouca), is more than 450 toises high. major long measured a base to determine the position and height of james peak. in the great basin of the mississippi the line that separates the forests and the savannahs runs, not, as may be supposed, in the manner of a parallel, but like the atlantic coast, and the allegheny mountains themselves, from north-east to south-west, from pittsburg towards saint louis, and the red river of nachitoches, so that the northern part only of the state of illinois is covered with gramina. this line of demarcation is not only interesting for the geography of plants, but exerts, as we have said above, great influence in retarding culture and population north-west of the lower mississippi. in the united states the prairie countries are more slowly colonized; and even the tribes of independent indians are forced by the rigour of the climate to pass the winter on the banks of rivers, where poplars and willows are found. the basins of the mississippi, of the lakes of canada and the st. lawrence, are the largest in america; and though the total population does not rise at present beyond three millions, it may be considered as that in which, between latitude 29 and 45° (longitude 74 to 94°), civilization has made the greatest progress. it may even be said that in the other basins (of the orinoco, the amazon and buenos ayres) agricultural life scarcely exists; it begins, on a small number of points only, to supersede pastoral life, and that of fishing and hunting nations. the plains between the alleghenies and the andes of upper louisiana are of such vast extent that, like the pampas of choco and buenos ayres, bamboos (ludolfia miega) and palm-trees grow at one extremity, while the other, during a great part of the year, is covered with ice and snow.

2. the basin of the gulf of mexico, and of the caribbean sea.

this is a continuation of the basin of the mississippi, louisiana and hudson’s bay. it may be said that all the low lands on the coast of venezuela situated north of the littoral chain and of the sierra nevada de merida belong to the submerged part of this basin. if i treat here separately of the basin of the caribbean sea, it is to avoid confounding what, in the present state of the globe, is partly above and partly below the ocean. the recent coincidence of the periods of earthquakes observed at caracas and on the banks of the mississippi, the arkansas and the ohio, justifies the geologic theories which regard as one basin the plains bounded on the south, by the littoral cordillera of venezuela; on the east, by the alleghenies and the series of the volcanoes of the west indies; and on the west, by the rocky mountains (mexican andes) and by the series of the volcanoes of guatimala. the basin of the west indies forms, as we have already observed, a mediterranean with several issues, the influence of which on the political destinies of the new continent depends at once on its central position and the great fertility of its islands. the outlets of the basin, of which the four largest* are 75 miles broad, are all on the eastern side, open towards europe, and agitated by the current of the tropics. in the same manner as we recognize, in our mediterranean, the vestiges of three ancient basins by the proximity of rhodes, scarpanto, candia, and cerigo, as well as by that of cape sorello of sicily, the island of pantelaria and cape bon, in africa; so the basin of the west india islands, which exceeds the mediterranean in extent, seems to present the remains of ancient dykes which join* cape catoche of yucatan to cape san atonio of the island of cuba; and that island to cape tiburon of st. domingo; jamaica, the bank of la vibora and the rock of serranilla to cape gracias a dios on the mosquito shore. from this situation of the most prominent islands and capes of the continent, there results a division into three partial basins. the most northerly has long been distinguished by a particular denomination, that of the gulf of mexico; the intermediary or central basin may be called the sea of honduras, on account of the gulf of that name which makes a part of it; and the southern basin, comprehended between the caribbean islands and the coast of venezuela, the isthmus of panama, and the country of the mosquito indians, would form the caribbean sea. the modern volcanic rocks distributed on the two opposite banks of the basin of the west indies on the east and west, but not on the north and south, is also a phenomenon worthy of attention. in the caribbean islands, a group of volcanoes, partly extinct and partly burning, stretches from 12 to 18°; and in the cordilleras of guatimala and mexico from latitude 9 to 19 1/2°. i noticed on the north-west extremity of the basin of the west indies that the secondary formations dip towards south-east; along the coast of venezuela rocks of gneiss and primitive mica-slate dip to north-west. the basalts, amygdaloids, and trachytes, which are often surmounted by tertiary limestones, appear only towards the eastern and western banks.

[* between tobago and grenada; saint martin and the virgin isles; porto rico and saint domingo; and between the little bank of bahama and cape canaveral of florida.]

[* i do not pretend that this hypothesis of the rupture and the ancient continuity of lands can be extended to the eastern foot of the basin of the west indies, that is, to the series of the volcanic islands in a line from trinidad to porto rico.]

3. the basin of the lower orinoco, or the plains of venezuela.

this basin, like the plains of lombardy, is open to the east. its limits are the littoral chain of venezuela on the north, the eastern cordillera of new grenada on the west, and the sierra parime on the south; but as the latter group extends on the west only to the meridian of the cataracts of maypures (longitude 70° 37′), there remains an opening or land-strait, running from north to south, by which the llanos of venezuela communicate with the basin of the amazon and the rio negro. we must distinguish between the basin of the lower orinoco, properly so called (north of that river and the rio apure), and the plains of meta and guaviare. the latter occupy the space between the mountains of parime and new grenada. the two parts of this basin have an opposite direction; but being alike covered with gramina, they are usually comprehended in the country under the same denomination. those llanos extend, in the form of an arch, from the mouth of the orinoco, by san fernando de apure, to the confluence of the rio caguan with the jupura, consequently along a length of more than 360 leagues.

(3a.) part of the basin of venezuela running from east to west.

the general slope is eastward, and the mean height from 40 to 50 toises. the western bank of that great sea of verdure (mar de yerbas) is formed by a group of mountains, several of which equal or exceed in height the peak of teneriffe and mont blanc. of this number are the paramos del almorzadero, cacota, laura, porquera, mucuchies, timotes, and las rosas. the height of the northern and southern banks is generally less than 500 or 600 toises. it is somewhat extraordinary that the maximum of the depression of the basin is not in its centre, but on its southern limit, at the sierra parime. it is only between the meridians of cape codera and cumana, where a great part of the littoral cordillera of venezuela has been destroyed, that the waters of the llanos (the rio unare and the rio neveri) reach the northern coast. the partition ridge of this basin is formed by small table-lands, known by the names of mesas de amana, guanipa and jonoro. in the eastern part, between the meridians 63 and 66°, the plains or savannahs run southward beyond the bed of the orinoco and the imataca, and form (as they approach the cujuni and the essequibo) a kind of gulf along the sierra pacaraina.

(3b.) part of the basin of venezuela running from south to north.

the great breadth of this zone of savannahs (from 100 to 120 leagues) renders the denomination of land-strait somewhat improper, at least if it be not geognostically applied to every communication of basins bounded by high cordilleras. perhaps this denomination more properly belongs to that part in which is situated the group of almost unknown mountains that surround the sources of the rio negro. in the basin comprehended between the eastern declivity of the andes of new grenada and the western part of the sierra parime, the savannahs, as we have observed above, stretch far beyond the equator; but their extent does not determine the southern limits of the basin here under consideration. these limits are marked by a ridge which divides the waters between the orinoco and the rio negro, a tributary stream of the amazon. the rising of a counter-slope almost imperceptible to the eye, forms a ridge that seems to join the eastern cordillera of the andes to the group of the parime. this ridge runs from ceja (latitude 1 degree 45 minutes), or the eastern slope of the andes of timana, between the sources of the guayavero and the rio caguan, towards the isthmus that separates the tuamini from pimichin. in the llanos, consequently, it follows the parallels of 20° 30′ and 2° 45′. it is remarkable that we find the divortia aquarum further westward on the back of the andes, in the knot of mountains containing the sources of the magdalena, at a height of 900 toises above the level of the llanos, between the caribbean sea and the pacific ocean, and almost in the same latitude (1 degree 45 minutes to 2° 20′). from the isthmus of javita towards the east, the line of the partition of waters is formed by the mountains of the parime group; it first rises a little on the north-east towards the sources of the orinoco (latitude 3° 45′?) and the chain of pacaraina (latitude 4° 4′ to 4° 12′); then, during a course of 80 leagues, between the portage of the anocapra and the banks of the rupunuri, it runs very regularly from west to east; and finally, beyond the meridian 61° 50′, it again deviates towards lower latitudes, passing between the northern sources of the rio suriname, the maroni, the oyapoc and the southern sources of rio trombetas, curupatuba, and paru (latitude 2° to 1 degree 50 minutes). these facts suffice to prove that this first line of partition of the waters of south america (that of the northern hemisphere) traverses the whole continent between the parallels of 2 and 4°. the cassiquiare alone has cut its way across the ridge just described. the hydraulic system of the orinoco displays the singular phenomenon of a bifurcation where the limit of two basins (those of the orinoco and the rio negro) crosses the bed of the principal recipient. in that part of the basin of the orinoco which runs in the direction of from south to north, as well as in that running from west to east, the maxima of depression are found at the foot of the sierra parime, we may even say, on its outline.

4. the basin of the rio negro and the amazon.

this is the central and largest basin of south america. it is exposed to frequent equatorial rains, and the hot and humid climate develops a force of vegetation to which nothing in the two continents can be compared. the central basin, bounded on the north by the parime group, and on the south by the mountains of brazil, is entirely covered by thick forests, while the two basins at the extremities of the continent (the llanos of venezuela and the lower orinoco, and the pampas of buenos ayres or the rio de la plata) are savannahs or prairies, plains without trees and covered with gramina. this symmetric distribution of savannahs bounded by impenetrable forests, must be connected with physical revolutions which have operated simultaneously over great surfaces.

(4a.) part of the basin of the amazon, running from east to west, between 2 degrees north and 12 degrees south; 880 leagues in length.

the western shore of this basin is formed by the chain of the andes, from the knot of the mountains of huanuco to the sources of the magdalena. it is enlarged by the spurs of the rio beni,* rich in gem-salt, and composed of several ranges of hills (latitude 8° 11′ south) which advance into the plains on the eastern bank of the paro. these hills are transformed on our maps into upper cordilleras and andes of cuchao. towards the north the basin of the amazon, of which the area (244,000 square leagues) is only one-sixth less than the area of all europe, rises in a gentle slope towards the sierra parime. at 68° of west longitude the elevated part of this sierra terminates at 3 1/2° north latitude. the group of little mountains surrounding the source of the rio negro, the inirida and the xie (latitude 2°) the scattered rocks between the atabapo and the cassiquiare, appear like groups of islands and rocks in the middle of the plain. some of those rocks are covered with signs or symbolical sculpture. nations, very different from those who now inhabit the banks of the cassiquiare, penetrated into the savannahs; and the zone of painted rocks, extending more than 150 leagues in breadth, bears traces of ancient civilization. on the east of the sporadic groups of rocks (between the meridian of the bifurcation of the orinoco and that of the confluence of the essequibo with the rupunuri) the lofty mountains of the parime commence only in 3° north latitude; where the plains of the amazon terminate.

[* the real name of this great river, respecting the course of which geographers have been so long divided, is uchaparu, probably water (para) of ucha; peni also signifies river or water; for the language of the maypures has very many analogies with that of the moxos; and veni (oueni) signifies water in maypure, as una in moxo. perhaps the river retained the name of maypure, after the indians who spoke that language had emigrated northward in the direction of the banks of the orinoco.]

the limits of the plains of the amazon are still less known towards the south than towards the north. the mountains that exceed 400 toises of absolute height do not appear to extend in brazil northward of the parallels 14 or 15° of south latitude, and west of the meridian of 52°; but it is not known how far the mountainous country extends, if we may call by that name a territory bristled with hills of one hundred or two hundred toises high. between the rio dos vertentes and the rio de tres barras (tributary streams of the araguay and the topayos) several ridges of the monts parecis run northward. on the right bank of the topayos a series of little hills advance as far as the parallel of 5° south latitude, to the fall (cachoeira) of maracana; while further west, in the rio madeira, the course of which is nearly parallel with that of the topayos, the rapids and cataracts indicate no rocky ridges beyond the parallel of 8°. the principal depression of the basin of which we have just examined the outline, is not near one of its banks, as in the basin of the lower orinoco, but at the centre, where the great recipient of the amazon forms a longitudinal furrow inclining from west to east, under an angle of at least 25°. the barometric measurements which i made at javita on the banks of the tuamini, at vasivia on the banks of the cassiquiare and at the cataract of rentema, in the upper maranon, seem to prove that the rising of the llanos of the amazon northward (at the foot of the sierra parime) is 150 toises, and westward (at the foot of the cordillera of the andes of loxa), 190 toises above the sea-level.

(4b.) part of the basin of the amazon stretching from south to north.

this is the zone or land-strait by which, between 12 and 20° of south latitude, the plains of the amazon communicate with the pampas of buenos ayres. the western bank of this zone is formed by the andes, between the knot of porco and potosi, and that of huanuco and pasco. part of the spurs of the rio beni, which is but a widening of the cordilleras of apolobamba and cuzco and the whole promontory of cochabamba, advance eastward into the plains of the amazon. the prolongation of this promontory has given rise to the idea that the andes are linked with a series of hills which the serras dos parecis, the serra melgueira, and the supposed cordillera of san fernando, throw out towards the west. this almost unknown part of the frontiers of brazil and upper peru merits the attention of travellers. it is understood that the ancient mission of san jose de chiquitos (nearly latitude 17°, longitude 67° 10′, supposing santa cruz de la sierra, in latitude 17° 25′, longitude 66° 47′) is situated in the plains, and that the mountains of the spur of cochabamba terminate between the guapaix (rio de mizque) and the parapiti, which lower down takes the names of rio san miguel and rio sara. the savannahs of the province of chiquitos communicate on the north with those of moxos, and on the south with those of chaco; but a ridge or line of partition of the waters is formed by the intersection of two gently sloping plains. this ridge takes its origin on the north of la plata (chuquisaca) between the sources of the guapaix and the cachimayo, and it ascends from the parallel of 20° to that of 15 1/2° south latitude, consequently on the north-east, towards the isthmus of villabella. from this point, one of the most important of the whole hydrography of america, we may follow the line of the partition of the water to the cordillera of the shore (serra do mar). it is seen winding (latitude 17 to 20°) between the northern sources of the araguay, the maranhao or tocantines, the rio san francisco and the southern sources of the parana. this second line of partition which enters the group of the brazil mountains on the frontier of capitania of goyaz separates the flowings of the basin of the amazon from those of the rio de la plata, and corresponds, south of the equator, with the line we have indicated in the northern hemisphere (latitude 2 to 4°), on the limits of the basins of the amazon and the lower orinoco.

if the plains of the amazon (taking that denomination in the geognostic sense we have given it) are in general distinguished from the llanos of venezuela and the pampas of buenos ayres, by the extent and thickness of their forests, we are the more struck by the continuity of the savannahs in that part running from south to north. it would seem as though this sea of verdure stretched forth an arm from the basin of buenos ayres, by the llanos of tucuman, manso, chuco, the chiquitos, and the moxos, to the pampas del sacramento and the savannahs of napo, guaviare, meta and apure. this arm crosses, between 7 and 3° south latitude, the basin of the forests of the amazon; and the absence of trees on so great an extent of territory, together with the preponderance which the small monocotyledonous plants have acquired, is a phenomenon of the geography of plants which belongs perhaps to the action of ancient pelagic currents or other partial revolutions of our planet.

5. plains of the rio de la plata, and of patagonia, from the south-western slope of the group of the brazil mountains to the strait of magellan; from 20 to 53 degrees of latitude.

these plains correspond with those of the mississippi and of canada in the northern hemisphere. if one of their extremities approaches less nearly to the polar regions, the other enters much further into the region of palm-trees. that part of this vast basin extending from the eastern coast towards the rio paraguay does not present a surface so perfectly smooth as the part situated on the west and the south-east of the rio de la plata, and which has been known for ages by the name of pampas, derived from the peruvian or quichua language.* geognostically speaking these two regions of east and west form only one basin, bounded on the east by the sierra de villarica or do espinhaco, which loses itself in the capitania of san paul, near the parallel of 24°; issuing on the north-east by little hills, from the serra da canastra and the campos parecis towards the province of paraguay; on the west by the andes of upper peru and chile; and on the north-west by the ridge of the partition of the waters which runs from the spur of santa cruz de la sierra, across the plains of the chiquitos, towards the serras of albuquerque (latitude 19° 2′) and san fernando. that part only of this basin lying on the west of the rio paraguay, and which is entirely covered with gramina, is 70,000 square leagues. this surface of the pampas or llanos of manse, tucuman, buenos ayres and eastern patagonia is consequently four times greater than the surface of the whole of france. the andes of chile narrow the pampas by the two spurs of salta and cordova; the latter promontory forms so projecting a point that there remains (latitude 31 to 32°) a plain only 45 leagues broad between the eastern extremity of the sierra de cordova and the right bank of the river paraguay, stretching in the direction of a meridian, from the town of nueva coimbra to rosario, below santa fe. far beyond the southern frontiers of the old viceroyalty of buenos ayres, between the rio colorado and the rio negro (latitude 38 to 39°) groups of mountains seem to rise in the form of islands in the middle of a muriatiferous plain. a tribe of indians of the south (tehuellet) have there long borne the characteristic name of men of the mountains (callilehet) or serranos. from the parallel of the mouth of the rio negro to that of cabo blanco (latitude 41 to 47°) scattered mountains on the eastern patagonian coast denote more considerable inequalities inland. all that part, however, of the straits of magellan, from the virgins’ cape to the north cape, on the breadth of more than 30 leagues, is surrounded by savannahs or pampas; and the andes of western patagonia only begin to rise near the latter cape, exercising a marked influence on the direction of that part of the strait nearest the pacific, proceeding from south-east to north-west.

[* hatan pampa signifies in that language, a great plain. we find the word pampa also in riobamba and guallabamba; the spaniards, in order to soften the geographical names, changing the p into b.]

if we have given the plains or great basins of south america the names of the rivers that flow in their longitudinal furrows, we have not meant by so-doing to compare them to mere valleys. in the plains of the lower orinoco and the amazon all the lines of the declivity doubtless reach a principal recipient, and the tributaries of tributary streams, that is the basins of different orders, penetrate far into the group of the mountains. the upper parts or high valleys of the tributary streams must be considered in a geological table as belonging to the mountainous region of the country, and beyond the plains of the lower orinoco and the amazon. the views of the geologist are not identical with those of the hydrographer. in the basin of the rio de la plata and patagonia the waters that follow the lines of the greatest declivities have many issues. the same basin contains several valleys of rivers; and when we examine nearly the polyedric surface of the pampas and the portion of their waters which, like the waters of the steppes of asia, do not go to the sea, we conceive that these plains are divided by small ridges or lines of elevation, and have alternate slopes, inclined, with reference to the horizon, in opposite directions. in order to point out more clearly the difference between geological and hydrographic views, and to prove that in the former, abstracting the course of the waters which meet in one recipient, we obtain a far more general point of view, i shall here again recur to the hydrographic basin of the orinoco. that immense river rises on the southern slope of the sierra parime. it is bounded by plains on the left bank, from the cassiquiare to the mouth of the atabapo, and flows in a basin which, geologically speaking, according to one great division of the surface of south america into three basins, we have called the basin of the rio negro and the amazon. the low regions, which are bounded by the southern and northern declivities of the parime and brazil mountains, and which the geologist ought to mark by one name, contain, according to the no less precise language of hydrography, two basins of rivers, those of the upper orinoco and the amazon, separated by a ridge that runs from javita towards esmeralda. from these considerations it results that a geological basin (sit venia verbo) may have several recipients and several emissaries, divided by small ridges almost imperceptible; it may at the same time contain waters that flow to the sea by different furrows independent of each other, and the systems of inland rivers flowing into lakes more or less charged with saline matter. a basin of a river, or hydrographic basin, has but one recipient, one emissary; if, by a bifurcation, it gives a part of its waters to another hydrographic basin, it is because the bed of the river, or the principal recipient, approaches so near the banks of the basin or the ridge of partition that the ridge partly crosses it.

the distribution of the inequalities of the surface of the globe does not present any strongly marked limits between the mountainous country and the low regions, or geologic basins. even where real chains of mountains rise like rocky dykes issuing from a crevice, spurs more or less considerable, seem to indicate a lateral upheaving. while i admit the difficulty of properly defining the groups of mountains and the basins or continuous plains, i have attempted to calculate their surfaces according to the statements contained in the preceding sheets.

geographical location. area in square marine leagues.

1. mountainous part:

andes 58,900

littoral chain of venezuela 1,900

sierra nevada de merida 200

group of the parime 25,800

system of the brazil mountains 27,600.

total 114,400.

2. plains:

llanos of the lower orinoco, the meta, and the guaviare 29,000

plains of the amazon 260,400

pampas of rio de la plata and patagonia 135,200

plains between the eastern chain of the andes of cundinamarca and the chain of choco 12,300

plains of the shore on the west of the andes 20,000.

>total 456,900.

the whole surface of south america contains 571,300 square leagues (20 to a degree), and the proportion of the mountainous country to the region of the plains is as 1 to 3.9. the latter region, on the east of the andes, comprises more than 424,600 square leagues, half of which consists of savannahs; that is to say, it is covered with gramina.

section 2.

general partition of ground. direction and inclination of the strata. relative height of the formations above the level of the ocean.

in the preceding section we have examined the inequalities of the surface of the soil, that is to say, the general structure of the mountains and the form of the basins rising between those variously grouped mountains. these mountains are sometimes longitudinal, running in narrow bands or chains, similar to the veins that preserve their directions at great distances, as the andes, the littoral chain of venezuela, the serra do mar of brazil, and the alleghenies of the united states. sometimes they are in masses with irregular forms, in which upheavings seem to have taken place as on a labyrinth of crevices or a heap of veins, as for example in the sierra parime and the serra dos vertentes. these modes of formation are linked with a geognostic hypothesis, which has at least the recommendation of being founded on facts observed in remote times, and which strongly characterize the chains and groups of mountains. considerations on the aspect of a country are independent of those which indicate the nature of the soil, the heterogeneity of matter, the superposition of rocks and the direction and inclination of strata.

in taking a general view of the geological constitution of a chain of mountains, we may distinguish five elements of direction too often confounded in works of geognosy and physical geography. these elements are:—

1. the longitudinal axis of the whole chain.

2. the line that divides the waters (divortia aquarum).

3. the line of ridges or elevation passing along the maxima of height.

4. the line that separates two contiguous formations into horizontal

sections.

5. the line that follows the fissures of stratification.

this distinction is the more necessary, there existing probably no chain on the globe that furnishes a perfect parallelism of all these directing lines. in the pyrenees, for instance, 1, 2, 3, do not coincide, but 4 and 5 (that is, the different formations which come to light successively, and the direction of the strata) are obviously parallel to 1, or to the direction of the whole chain. we find so often in the most distant parts of the globe, a perfect parallelism between 1 and 5, that it may be supposed that the causes which determine the direction of the axis (the angle under which that axis cuts the meridian) are generally linked with causes that determine the direction and inclination of the strata. this direction of the strata is independent of the line of the formations, or their visible limits at the surface of the soil; the lines 4 and 5 sometimes cross each other, even when one of them coincides with 1, or with the direction of the longitudinal axis of the whole chain. the relief of a country cannot be precisely explained on a map, nor can the most erroneous opinions on the locality and superposition of the strata be avoided, if we do not apprehend with clearness the relation of the directing lines just mentioned.

in that part of south america to which this memoir principally relates, and which is bounded by the amazon on the south, and on the west by the meridian of the snowy mountains (sierra nevada) of merida, the different bands or zones of formations (4) are sensibly parallel with the longitudinal axis (1) of the chains of mountains, basins or interposed plains. it may be said in general that the granitic zone (including under that denomination the rocks of granite, gneiss and mica-slate) follows the direction of the cordillera of the shore of venezuela, and belongs exclusively to that cordillera and the group of the parime mountains; since it nowhere pierces the secondary and tertiary strata in the llanos or basin of the lower orinoco. thence it results that the same formations do not constitute the region of plains and that of mountains.

if we may be allowed to judge of the structure of the whole sierra parime, from the part which i examined in 6° of longitude, and 4° of latitude, we may believe it to be entirely composed of gneiss-granite; i saw some beds of greenstone and amphibolic slate, but neither mica-slate, clay-slate, nor banks of green limestone, although many phenomena render the presence of mica-slate probable on the east of the maypures and in the chain of pacaraina. the geological formation of the parime group is consequently still more simple than that of the brazilian group, in which granites, gneiss and mica-slate are covered with thonschiefer, chloritic quartz (itacolumite), grauwacke and transition-limestone; but those two groups exhibit in common the absence of a real system of secondary rocks; we find in both only some fragments of sandstone or silicious conglomerate. in the littoral cordillera of venezuela the granitic formations predominate; but they are wanting towards the east, and especially in the southern chain, where we observe (in the missions of caripe and around the gulf of cariaco) a great accumulation of secondary and tertiary calcareous rocks. from the point where the littoral cordillera is linked with the andes of new grenada (longitude 71 1/2°) we observe first the granitic mountains of aroa and san felipe, between the rivers yaracui and tocuyo; these granitic formations extend on the east of the two coasts of the basin of the valleys of aragua, in the northern chain, as far as cape codera; and in the southern as far as the mountains (altas savanas) of ocumare. after the remarkable interruption of the littoral cordillera in the province of barcelona, granitic rocks begin to appear in the island of marguerita and in the isthmus of araya, and continue, perhaps, towards the boca del drago; but on the east of the meridian of cape codera the northern chain only is granitic (of micaceous slate); the southern chain is entirely composed of secondary limestone and sandstone.

if, in the granitic series, where a very complex formation, we would distinguish mineralogically between the rocks of granite, gneiss, and mica-slate, it must be borne in mind that coarse-grained granite, not passing to gneiss, is very rare in this country. it belongs peculiarly to the mountains that bound the basin of the lake of valencia towards the north; for in the islands of that lake, in the mountains near the villa de cura, and in the whole northern chain, between the meridian of vittoria and cape codera, gneiss predominates, sometimes alternating with granite, or passing to mica-slate. mica-slate is the most frequent rock in the peninsula of araya and the group of macanao, which forms the western part of the island of marguerita. on the west of maniquarez the mica-slate of the peninsula of araya loses by degrees its semi-metallic lustre; it is charged with carbon, and becomes a clay-slate (thonschiefer) even an ampelite (alaunschiefer). beds of granular limestone are most common in the primitive northern chain; and it is somewhat remarkable that they are found in gneiss, and not in mica-slate.

we find at the back of this granitic, or rather mica-slate-gneiss soil of the southern chain, on the south of the villa de cura, a transition stratum, composed of greenstone, amphibolic serpentine, micaceous limestone, and green and carburetted slate. the most southern limit of this district is marked by volcanic rocks. between parapara, ortiz and the cerro de flores (latitude 9° 28′ to 9° 34′; longitude 70° 2′ to 70° 15′) phonolites and amygdaloids are found on the very border of the basin of the llanos, that vast inland sea which once filled the whole space between the cordilleras of venezuela and parime. according to the observations of major long and dr. james, trap-formations (bulleuses dolerites and amygdaloids with pyroxene) also border the plains or basin of the mississippi, towards the west, at the declivity of the rocky mountains. the ancient pyrogenic rocks which i found near parapara where they rise in mounds with rounded summits, are the more remarkable as no others have hitherto been discovered in the whole eastern part of south america. the close connection observed in the strata of parapara, between greenstone, amphibolic serpentine, and amygdaloids containing crystals of pyroxene; the form of the morros of san juan, which rise like cylinders above the table-land; the granular texture of their limestone, surrounded by trap rocks, are objects worthy the attention of the geologist who has studied in the southern tyrol the effects produced by the contact of poroxenic porphyries.*

[* leopold von buch. tableau geologique du tyrol page 17. m. boussingault states that these singular morros de san juan, which furnish a limestone with crystalline grains, and thermal springs, are hollow, and contain immense grottos filled with stalactites, which appear to have been anciently inhabited by the natives.]

the calcareous soil of the littoral cordillera prevails most on the east of cape unare, in the southern chain; it extends to the gulf of paria, opposite the island of trinidad, where we find gypsum of guire, containing sulphur. i have been informed that in the northern chain also, in the montana de paria, and near carupana, secondary calcareous formations are found, and that they only begin to show themselves on the east of the ridge of rock called the cerro de meapire, which joins the calcareous group of guacharo to the mica-slate group of the peninsula of araya; but i have not had an opportunity of ascertaining the accuracy of this information. the calcareous stratum of the southern chain is composed of two formations which appear to be very distinct the one from the other: namely limestone of cumanacoa and that of caripe. when i was on the spot the former appeared to me to have some analogy with zechstein, or alpine limestone; the latter with jura limestone; i even thought that the granular gypsum of guire might be that which belongs in europe to zechstein, or is placed between zechstein and variegated sandstone. strata of quartzose sandstone, alternating with slaty clay, cover the limestone of cumanacoa, cerro del imposible, turimiquiri, guarda de san agustin, and the jura limestone in the province of barcelona (aguas calientes). according to their position these sandstones may be considered as belonging to the formation of green sandstone, or sandstone with lignites below chalk. but if, as i thought i observed at cocollar, sandstone forms strata in the alpine limestone before it is superposed, it appears doubtful whether the sandstone of the imposible, and of aguas calientes, constitute one series. muriatiferous clay (with petroleum and lamellar gypsum) covers the western part of the peninsula of araya, opposite to the town of cumana, and in the centre of the island of marguerita. this clay appears to lie immediately over the mica-slate, and under the calcareous breccia of the tertiary strata. i cannot decide whether araya, which is rich in disseminated muriate of soda, belongs to the sandstone formation of the imposible, which from its position may be compared to variegated sandstone (red marl).

there is no doubt that fragments of tertiary strata surround the castle and town of cumana (castillo de san antonio) and they also appear at the south-western extremity of the peninsula of araya (cerro de la vela et del barigon); at the ridge of the cerro de meapire, near cariaco; at cabo blanco, on the west of la guayra, and on the shore of porto cabello; they are consequently found at the foot of the two slopes of the northern chain of the cordillera of venezuela. this tertiary stratum is composed of alternate beds of calcareous conglomerate, compact limestone, marl, and clay, containing selenite and lamellar gypsum. the whole system (of very recent beds) appears to me to constitute but one formation, which is found at the cerro de la popa, near carthagena, and in the islands of guadaloupe and martinico.

such is the geological distribution of strata in the mountainous part of venezuela, in the group of the parime and in the littoral cordillera. we have now to characterize the formations of the llanos (or of the basin of the lower orinoco and the apure); but it is not easy to determine the order of their superposition, because in this region ravines or beds of torrents and deep wells dug by the hands of man are entirely wanting. the formations of the llanos are, first, a sandstone or conglomerate, with rounded fragments of quartz, lydian stone, and kieselschiefer, united by a ferruginous clayey cement, extremely tenacious, olive-brown, sometimes of a vivid red; second, a compact limestone (between tisnao and calabozo) which, by its smooth fracture and lithographic aspect, approaches the jura limestone: third, alternate strata of marl and lamellar gypsum (mesa de san diego, ortiz, cachipo). these three formations appeared to me to succeed each other in the order i have just described, the sandstone inclining in a concave position, northward, on the transition-slates of malpasso, and southward, on the gneiss-granite of parime. as the gypsum often immediately covers the sandstone of calabozo, which appeared to me, on the spot, to be identical with our red sandstone, i am uncertain of the age of its formation. the secondary rocks of the llanos of cumana, barcelona and caracas occupy a space of more than 5000 square leagues. their continuity is the more remarkable, as they appear to have no existence, at least on the east of the meridian of porto cabello (70° 37′) in the whole basin of the amazon not covered by granitic sands. the causes which have favoured the accumulation of calcareous matter in the eastern region of the coast chain, in the llanos of venezuela (from 10 1/2 to 8° north), cannot have operated nearer the equator, in the group of the mountains of the parime and in the plains of the rio negro and the amazon (latitude 1 degree north to 1 degree south). the latter plains, however, furnish some ledges of fragmentary rocks on the south-west of san fernando de atabapo, as well as on the south-east, in the lower part of the rio negro and the rio branco. i saw in the plains of jaen de bracamoros a sandstone which alternates with ledges of sand and conglomerate nodules of porphyry and lydian stone. mm. spix and martius affirm that the banks of the rio negro on the south of the equator are composed of variegated sandstone; those of the rio branco, jupura and apoporis of quadersandstein; and those of the amazon, on several points, of ferruginous sandstone.* it remains to examine if (as i am inclined to suppose) the limestone and gypsum formations of the eastern part of the littoral cordillera of venezuela differ entirely from those of the llanos, and to what series belongs that rocky wall* named the galera, which bounds the steppes of calabozo towards the north? the basin of the steppes is itself the bottom of a sea destitute of islands; it is only on the south of the apure, between that river and the meta, near the western bank of the sierra, that a few hills appear, as monte parure, la galera de sinaruco and the cerritos de san vicente. with the exception of the fragments of tertiary strata above mentioned there is, from the equator to the parallel of 10° north (between the meridian of sierra nevada de merida and the coast of guiana), if not an absence, at least a scarcity of those petrifactions, which strikes an observer recently arrived from europe.

[* braunes eisenschussiges sandstein–conglomerat (iron-sand of the english geologists, between the jura limestone and green sandstone.) mm. spix and martius found on rocks of quadersandstein, between the apoporis and the japura, the same sculptures which we have pointed out from the essequibo to the plains of cassiquiare, and which seem to prove the migrations of a people more advanced in civilization than the indians who now inhabit those countries.]

[* is this wall a succession of rocks of dolomite or a dyke of quadersandstein, like the devil’s wall (teufelsmauer), at the foot of the hartz? calcareous shelves (coral banks), either ledges of sandstone (effects of the revulsion of the waves) or volcanic eruptions, are commonly found on the borders of great plains, that is, on the shores of ancient inland seas. the llanos of venezuela furnish examples of such eruptions near para(?) like harudje (mons ater, plin.) on the northern boundary of the african desert (the sahara). hills of sandstone rising like towers, walls and fortified castles and offering great analogy to quadersandstein, bound the american desert towards the west, on the south of arkansas.]

the maxima of the height of the different formations diminish regularly in the country we are describing with their relative ages. these maxima, for gneiss-granite (peak of duida in the group of parime, silla de caracas in the coast chain) are from 1300 to 1350 toises; for the limestone of cumanacoa (summit or cucurucho of turimiquiri), 1050 toises; for the limestone of caripe (mountains surrounding the table-land of the guarda de san augustin), 750 toises; for the sandstone alternating with the limestone of cumanacoa (cuchilla de guanaguana), 550 toises; for the tertiary strata (punta araya), 200 toises.

the tract of country of which i am here describing the geological constitution is distinguished by the astonishing regularity observed in the direction of the strata of which the rocks of different eras are composed. i have already often pointed the attention of my readers to a geognostic law, one of the few that can be verified by precise measurements. occupied since the year 1792 by the parallelism, or rather the loxodromism of the strata, examining the direction and inclination of the primitive and transition beds, from the coast of genoa across the chain of the bochetta, the plains of lombardy, the alps of saint gothard, the table-land of swabia, the mountains of bareuth, and the plains of northern germany, i was struck with the extreme frequency, if not the uniformity, of the horary directions 3 and 4 of the compass of freiberg (direction from south-west to north-east). this research, which i thought might lead to important discoveries relating to the structure of the globe, had then such attractions for me that it was one of the most powerful incentives of my voyage to the equator. my own observations, together with those of many able geologists, convince me that there exists in no hemisphere a general and absolute uniformity of direction; but that in regions of very considerable extent, sometimes over several thousand square leagues, we observe that the direction and (though more rarely) the inclination have been determined by a system of particular forces. we discover at great distances a parallelism (loxodromism) of the strata, a direction of which the type is manifest amidst partial perturbations and which often remains the same in primitive and transition strata. a fact which must have struck palasson and saussure is that in general the direction of the strata, even in those which are far distant from the principal ridges, is identical with the direction of mountain chains; that is to say, with their longitudinal axis.

venezuela is one of the countries in which the parallelism of the strata of gneiss-granite, mica-slate and clay-slate, is most strongly marked. the general direction of these strata is north 50° east, and the general inclination from 60 to 70° north-west. thus i observed them on a length of more than a hundred leagues, in the littoral chain of venezuela; in the stratified granite of las trincheras at porto cabello; in the gneiss of the islands of the lake of valencia, and in the vicinity of the villa de cura; in the transition-slate and greenstone on the north of parapara; in the road from la guayra to the town of caracas, and through all the sierra de avila in cape codera; and in the mica-slate and clay-slate of the peninsula of araya. the same direction from north-east to south-west, and this inclination to north-west, are also manifest, although less decidedly, in the limestones of cumanacoa at cuchivano and between guanaguana and caripe. the exceptions to this general law are extremely rare in the gneiss-granite of the littoral cordillera; it may even be affirmed that the inverse direction (from south-east to north-west) often bears with it the inclination towards south-west.

as that part of the group of the sierra parime over which i passed contains much more granite* than gneiss, and other rocks distinctly stratified, the direction of the layers could be observed in this group only on a small number of points; but i was often struck in this region with the continuity of the phenomenon of loxodromism. the amphibolic slates of angostura run north 45° east, like the gneiss of guapasoso which forms the bed of the atabapo, and like the mica-slate of the peninsula of araya, though there is a distance of 160 leagues between the limits of those rocks.

[* only the granite of the baragon is stratified, as well as crossed by veins of granite: the direction of the beds is north 20° west]

the direction of the strata, of which we have just noticed the wonderful uniformity, is not entirely parallel with the longitudinal axes of the two coast chains, and the chain of parime. the strata generally cut the former of those chains at an angle of 35°, and their inclination towards the north-west becomes one of the most powerful causes of the aridity which prevails on the southern declivity* of the mountains of the coast. may we conclude that the direction of the eastern cordillera of new grenada, which is nearly north 45° east from santa fe de bogota, to beyond the sierra nevada de merida, and of which the littoral chain is but a continuation, has had an influence on the direction (hor. 3 to 4) of the strata in venezuela? that region presents a very remarkable loxodromism with the strata of mica-slate, grauwacke, and the orthoceratite limestone of the alleghenies, and that vast extent of country (latitude 56 to 68°) lately visited by captain franklin. the direction north-east to south-west prevails in every part of north america, as in europe in the fitchtelgebirge of franconia, in taunus, westerwald, and eifel; in the ardennes, the vosges, in cotentin, in scotland and in the tarentaise at the south-west extremity of the alps. if the strata of rocks in venezuela do not exactly follow the direction of the nearest cordillera, that of the shore, the parallelism between the axis of one chain, and the strata of the formations that compose it, are manifest in the brazil group.*

[* this southern declivity is however less rapid than the northern.]

[* the strata of the primitive and intermediary rocks of brazil run very regularly, like the cordillera of villarica (serra do espinhaco) hor. 1.4 or hor. 2 of the compass of freiberg (north 28° east.)]

section 3.

nature of the rocks. relative age and superposition of the formations. primitive, transition, secondary, tertiary, and volcanic strata.

the preceding section has developed the geographical limits of the formations, the extent of the direction of the zones of gneiss-granite, mica-slate-gneiss, clay-slate, sandstone and intermediary limestone, which come successively to light. we will now indicate succinctly the nature and relative age of these formations. to avoid confounding facts with geologic opinions i shall describe these formations, without dividing them, according to the method generally followed, into five groups — primitive, transition, secondary, tertiary and volcanic rocks. i was fortunate enough to discover the types of each group in a region where, before i visited it, no rock had been named. the great inconvenience of the old classification is that of obliging the geologist to establish fixed demarcations, while he is in doubt, if not respecting the spot or the immediate superposition, at least respecting the number of the formations which are not developed. how can we in many circumstances determine the analogy existing between a limestone with but few petrifactions and an intermediary limestone and zechstein, or between a sandstone superposed on a primitive rock and a variegated sandstone and quadersandstein, or finally, between muriatiferous clay and the red marl of england, or the gem-salt of the tertiary strata of italy? when we reflect on the immense progress made within twenty-five years in the knowledge of the superposition of rocks, it will not appear surprising that my present opinion on the relative age of the formations of equinoctial america is not identically the same with what i advanced in 1800. to boast of a stability of opinion in geology is to boast of an extreme indolence of mind; it is to remain stationary amidst those who go forward. what we observe in any one part of the earth on the composition of rocks, their subordinate strata and the order of their position are facts immutably true, and independent of the progress of positive geology in other countries; while the systematic names applied to any particular formation of america are founded only on the supposed analogies between the formations of america and those of europe. now those names cannot remain the same if, after further examination, the objects of comparison have not retained the same place in the geologic series; if the most able geologists now take for transition-limestone and green sandstone, what they took formerly for zechstein and variegated sandstone. i believe the surest means by which geologic descriptions may be made to survive the change which the science undergoes in proportion to its progress, will be to substitute provisionally in the description of formations, for the systematic names of red sandstone, variegated sandstone, zechstein and jura limestone, names derived from american localities, as sandstone of the llanos, limestone of cumanacoa and caripe, and to separate the enumeration of facts relative to the superposition of soils, from the discussion on the analogy of those soils with those of the old world.*

[* positive geography being nothing but a question of the series or succession (either simple or periodical) of certain terms represented by the formations, it may be necessary, in order to understand the discussions contained in the third section of this memoir, to enumerate succinctly the table of formations considered in the most general point of view.

1. strata commonly called primitive; granite, gneiss and mica-slate (or gneiss oscillating between granite and mica-slate); very little primitive clay-slate; weisstein with serpentine; granite with disseminated amphibole; amphibolic slate; veins and small layers of greenstone.

2. transition strata, composed of fragmentary rocks (grauwacke), calcareous slate and greenstone, earliest remains of organized existence: bamboos, madrepores, producta, trilobites, orthoceratites, evamphalites. complex and parallel formations; (a) alternate beds of grey and stratified limestone, anthracitic mica-slate, anhydrous gypsum and grauwacke; (b) clay-slate, black limestone, grauwacke with greenstone, syenite, transition-granite and porphyries with a base of compact felspar; (c) euphotides, sometimes pure and covered with jasper, sometimes mixed with amphibole, hyperstein and grey limestone; (d) pyroxenic porphyries with amygdaloides and zirconian syenites.

3. secondary strata, presenting a much smaller number of monocotyledonous plants; (a) co-ordinate and almost contemporary formations with red sandstone (rothe todtes liegende), quartz-porphyry and fern-coal. these strata are less connected by alternation than by opposition. the porphyries issue (like the trachytes of the andes) in domes from the bosom of intermediary rocks. porphyritic breccias which envelope the quartzose porphyries. (b) zechstein or alpine limestone with marly, bituminous slate, fetid limestone and variegated gypsum (productus aculeatus). (c) variegated sandstone (bunter sandstein) with frequent beds of limestone; false oolites; the upper beds are of variegated marl, often muriatiferous (red marl, salzthon) with hydrated gypsum and fetid limestone. the gem-salt oscillates from zechstein to muschelkalk. (d) limestone of gottingen or muschelkalk alternating towards the top with white sandstone or brittle sandstein. (ammonitis nodosus, encrinites, mytilus socialis): clayey marl is found at the two extremities of muschelkalk. (e) white sandstone, brittle sandstein, alternating with lias, or limestone with graphites; a quantity of dicotyledonous mixed with monocotyledonous plants. (f) jura limestone of complex formation; a quantity of sandy intercalated marl. we most frequently observe, counting from below upwards; lias (marly limestone with gryphites), oolites, limestone with polypi, slaty limestone with fish, crustacea, and globules of oxide of iron (amonites planulatus, gryphaea arcuata). (g) secondary sandstone with lignites; iron sand; wealden clay; greensand or green sandstone; (h) chlorite; tufted and white chalk; (planerkalk, limestone of verona.)

4. tertiary strata, showing a much smaller number of dicotyledonous plants. (a) clay and tertiary sandstone with lignites; plastic clay; mollasse and nagelfluhe, sometimes alternating where chalk is wanting, with the last beds of jura limestone; amber. (b) limestone of paris or coarse limestone, limestone with circles, limestone of bolca, limestone of london, sandy limestone of bognor; lignites. (c) silicious limestone and gypsum with fossil bones alternating with marl. (d) sandstone of fontainebleau. (e) lacustrine soil with porous millstone grit. (e) alluvial deposits.)

1. co-ordinate formations of granite, gneiss and mica-slate.

there are countries (in france, the vicinity of lyons; in germany, freiberg, naundorf) where the formations of granite and gneiss are extremely distinct; there are others, on the contrary, where the geologic limits between those formations are slightly marked, and where granite, gneiss and mica-slate appear to alternate by layers or pass often from one to the other. these alternations and transitions appeared to me less common in the littoral cordillera of venezuela than in the sierra parime. we recognise successively, in the former of these two systems of mountains, above all in the chain nearest the coast, as predominating rocks from west to east, granite (longitude 70 to 71°), gneiss (longitude 68 1/2 to 70°), and mica-slate (longitude 65 3/4 to 66 1/2°); but considering altogether the geologic constitution of the coast and the sierra parime, we prefer to treat of granite, gneiss and mica-slate, if not as one formation, at least as three co-ordinate formations closely linked together. the primitive clay-slate (urthonschiefer) is subordinate to mica-slate, of which it is only a modification. it no more forms an independent stratum in the new continent, than in the pyrenees and the alps.

(a) granite which does not pass to gneiss is most common in the western part of the coast-chain between turmero, valencia and porto cabello, as well as in the circle of the sierra parime, near the encaramada, and at the peak of duida. at the rincon del diablo, between mariara and hacienda de cura, and at chuao, it is coarse-grained, and contains fine crystals of felspar, 1 1/2 inches long. it is divided in prisms by perpendicular vents, or stratified regularly like secondary limestone, at las trincheras, the strait of baraguan in the valley of the orinoco, and near guapasoso, on the banks of the atabapo. the stratified granite of las trincheras, giving birth to very hot springs (from 90.5° centigrade), appears from the inclination of its layers to be superposed on gneiss which is seen further southward in the islands of the lake of valencia; but conjectures of superposition founded only on the hypothesis of an indefinite prolongation of the strata are doubtful; and possibly the granite masses which form a small particular zone in the northern range of the littoral cordillera, between 70° 3′ and 70° 50′ longitude, were upheaved in piercing the gneiss. the latter rock is prevalent, both in descending from the rincon del diablo southward to the hot-springs of mariara, and towards the banks of the lake of valencia, and in advancing on the east towards the group of buenavista, the silla of caracas and cape codera. in the region of the littoral chain of venezuela, where granite seems to constitute an independent formation from 15 to 16 leagues in length, i saw no foreign or subordinate layers of gneiss, mica-slate or primitive limestone.*

[* primitive limestone, everywhere so common in mica-slate and gneiss, is found in the granite of the pyrenees, at port d’oo, and in the mountains of labourd.]

the sierra parime is one of the most extensive granitic strata existing on the globe;* but the granite, which is seen alike bare on the flanks of the mountains and in the plains by which they are joined, often passes into gneiss. granite is most commonly found in its granular composition and independent formation, near encaramada, at the strait of baraguan, and in the vicinity of the mission of the esmeralda. it often contains, like the granites of the rocky mountains (latitude 38 to 40°), the pyrenees and southern tyrol, amphibolic crystals,* disseminated in the mass, but without passing to syenite. those modifications are observed on the banks of the orinoco, the cassiquiare, the atabapo, and the tuamini. the blocks heaped together, which are found in europe on the ridge of granitic mountains (the riesengebirge in silesia, the ochsenkopf in franconia), are especially remarkable in the north-west part of the sierra parime, between caycara, the encaramada and uruana, in the cataracts of the maypures and at the mouth of the rio vichada. it is doubtful whether these masses, which are of cylindrical form, parallelopipedons rounded on the edge, or balls of 40 to 50 feet in diameter, are the effect of a slow decomposition, or of a violent and instantaneous upheaving. the granite of the south-eastern part of sierra parime sometimes passes to pegmatite,* composed of laminary felspar, enclosed in curved masses of crystalline quartz. i saw gneiss only in subordinate layers;* but, between javita, san carlos del rio negro, and the peak of duida, the granite is traversed by numerous veins of different ages, abounding with rock-crystal, black tourmalin and pyrites. it appears that these open veins become more common on the east of the peak of duida, in the sierra pacaraina, especially between xurumu and rupunuri (tributaries of the rio branco and the essequibo), where hortsmann discovered, instead of diamonds* and emeralds, a mine (four) of rock-crystal.

[* to prove the extent of the continuity of this granitic stratum, it will suffice to observe that m. leschenault de la tour collected in the bars of the river mana, in french guiana, the same gneiss-granites (with a little amphibole) which i observed three hundred leagues more to the west, near the confluence of the orinoco and the guaviare.]

[* i did not observe this mixture of amphibole in the granite of the littoral chain of venezuela except at the summit of the silla of caracas.]

[* schrift-granit. it is a simple modification of the composition and texture of granite, and not a subordinate layer. it must not be confounded with the real pegmatite, generally destitute of mica, or with the geographic stones (piedras mapajas) of the orinoco, which contain streaks of dark green mica irregularly disposed.]

[* the magnetic sands of the rivers that furrow the granitic chain of the encaramada seem to denote the proximity of amphibolic or chloritic slate (hornblende or chloritschiefer), either in layers in the granite, or superposed on that rock.]

[* these legends of diamonds are very ancient on the coast of paria. petrus martyr relates that, at the beginning of the sixteenth century, a spaniard named andres morales bought of a young indian of the coast of paria admantem mire pretiosum, duos infantis digiti articulos longum, magni autem pollicis articulum aequantem crassitudine, acutum utrobique et costis octo pulchre formatis constantem. [a diamond of marvellous value, as long as two joints of an infant’s finger, and as thick as one of the joints of its thumb, sharp on both sides, and of a beautiful octagonal shape.] this pretended adamas juvenis pariensis resisted the action of lime. petrus martyr distinguishes it from topaz by adding offenderunt et topazios in littore, [they pay no heed to topazes on the coast] that is of paria, saint marta and veragua. see oceanica dec. 3 lib. 4 page 53.]

(b) gneiss predominates along the littoral cordillera of venezuela, with the appearance of an independent formation, in the northern chain from cerro del chuao, and the meridian of choroni, as far as cape codera; and in the southern chain, from the meridian of guigne to the mouth of the rio tuy. cape codera, the great mass of the silla of galipano, and the land between guayra and caracas, the table-land of buenavista, the islands of the lake of valencia, the mountains between guigne, maria magdalena and the cerro do chacao are composed of gneiss;* yet amidst this soil of gneiss, inclosed mica-slate re-appears, often talcous in the valle de caurimare, and in the ancient provincia de los mariches; at cabo blanco, west of la guayra; near caracas and antimano, and above all, between the tableland of buenavista and the valleys of aragua, in the montana de las cocuyzas, and at hacienda del tuy. between the limits here assigned to gneiss, as a predominant rock (longitude 68 1/2 to 70 1/2°), gneiss passes sometimes to mica-slate, while the appearance of a transition to granite is only found on the summit of the silla of caracas.* it would require a more careful examination than i was able to devote to the subject, to ascertain whether the granite of the peak of st. gothard, and of the silla of caracas, really lies over mica-slate and gneiss, or if it has merely pierced those rocks, rising in the form of needles or domes. the gneiss of the littoral cordillera, in the province of caracas, contains almost exclusively garnets, rutile titanite and graphite, disseminated in the whole mass of the rock, shelves of granular limestone, and some metalliferous veins. i shall not decide whether the granitiferous serpentine of the table-land of buenavista is inclosed in gneiss, or whether, superposed upon that rock, it does not rather belong to a formation of weisstein (heptinite) similar to that of penig and mittweyde in saxony.

[* i have been assured that the islands orchila and los frailes are also composed of gneiss; curacao and bonaire are calcareous. is the island of oruba (in which nuggets of native gold of considerable size have been found) primitive?]

[* the silla is a mountain of gneiss like adams peak in the island of ceylon, and of nearly the same height.]

in that part of the sierra parime which m. bonpland and myself visited, gneiss forms a less marked zone, and oscillates more frequently towards granite than mica-slate. i found no garnets in the gneiss of parime. there is no doubt that the gneiss-granite of the orinoco is slightly auriferous on some points.

(c) mica-slate, with clay-slate (thonschiefer), forms a continuous stratum in the northern chain of the littoral cordillera, from the point of araya, beyond the meridian of cariaco, as well as in the island of marguerita. it contains, in the peninsula of araya, garnets disseminated in the mass, cyanite and, when it passes to clayey-slate, small layers of native alum. mica-slate constituting an independent formation must be distinguished from mica-slate subordinate to a stratum of gneiss, on the east of cape codera. the mica-slate subordinate to gneiss presents, in the valley of tuy, shelves of primitive limestone and small strata of graphic ampelite (zeicheschiefer); between cabo blanco and catia layers of chloritic, granitiferous slate, and slaty amphibole; and between caracas and antimano, the more remarkable phenomenon of veins of gneiss inclosing balls of granitiferous diorite (grunstein).

in the sierra parime, mica-slate predominates only in the most eastern part, where its lustre has led to strange errors.

the amphibolic slate of angostura, and masses of diorite in balls, with concentric layers, near muitaco, appear to be superposed, not on mica-slate, but immediately on gneiss-granite. i could not, however, distinctly ascertain whether a part of this pyritous diorite was not enclosed on the banks of the orinoco, as it is at the bottom of the sea near cabo blanco, and at the montana de avila, in the rock which it covers. very large veins, with an irregular direction, often assume the aspect of short layers; and the balls of diorite heaped together in hillocks may, like many cones of basalt, issue from the crevices.

mica-slate, chloritic slate and the rocks of slaty amphibole contain magnetic sand in the tropical regions of venezuela, as in the most northern regions of europe. the gannets are there almost equally disseminated in the gneiss (caracas), the mica-slate (peninsula of araya), the serpentine (buenavista), the chloritic slate (cabo blanco), and the diorite or greenstone (antimano). these garnets re-appear in the trachytic porphyries that crown the celebrated metalliferous mountain of potosi, and in the black and pyroxenic masses of the small volcano of yana–urca, at the back of chimborazo.

petroleum (and this phenomenon is well worthy of attention) issues from a soil of mica-slate in the gulf of cariaco. further east, on the banks of the arco, and near cariaco, it seems to gush from secondary limestone formations, but probably that happens only because those formations repose on mica-slate. the hot springs of venezuela have also their origin in, or rather below, the primitive rocks. they issue from granite (las trincheras), gneiss (mariara and onoto) and the calcareous and arenaceous rocks that cover the primitive rocks (morros de san juan, bergantin, cariaco). the earthquakes and subterraneous detonations of which the seat has been erroneously sought in the calcareous mountains of cumana have been felt with most violence in the granitic soils of caracas and the orinoco. igneous phenomena (if their existence be really well certified) are attributed by the people to the granitic peaks of duida and guaraco, and also to the calcareous mountain of cuchivano.

from these observations it results that gneiss-granite predominates in the immense group of the mountains of the parime, as mica-slate-gneiss prevails in the cordillera of the coast; that in the two systems the granitic soil, unmixed with gneiss and mica-slate, occupies but a very small extent of country; and that in the coast-chain the formations of clayey slate (thonschiefer), mica-slate, gneiss and granite succeed each other in such a manner on the same line from east to west (presenting a very uniform and regular inclination of their strata towards the north-west), that, according to the hypothesis of a subterraneous prolongation of the strata, the granite of las trincheras and the rincon del diablo may be superposed on the gneiss of the villa de cura, of buenavista and caracas; and the gneiss superposed in its turn on the mica-slate and clay-slate of maniquarez and chuparuparu in the peninsula of araya. this hypothesis of a prolongation of every rock, in some sort indefinite, founded on the angle of inclination presented by the strata appearing at the surface, is not admissible; and according to similar equally vague reasoning we should be forced to consider the primitive rocks of the alps of switzerland as superposed on the formation of the compact limestone of achsenberg, and that [transition, or identical with zechstein?] in turn, as being superposed on the molassus of the tertiary strata.

2. formation of the clay-slate (thonschiefer) of malpasso.

if, in the sketch of the formations of venezuela, i had followed the received division into primitive, intermediary, secondary and tertiary strata, i might be doubtful what place the last stratum of mica-slate in the peninsula of araya should occupy. this stratum, in the ravine (aroyo) of robalo, passes insensibly in a carburetted and shining slate, into a real ampelite. the direction and inclination of the stratum remain the same, and the thonschiefer, which takes the look of a transition-rock, is but a modification of the primitive mica-slate of maniquarez, containing garnets, cyanite, and rutile titanite. these insensible passages from primitive to transition strata by clay-slate, which becomes carburetted at the same time that it presents a concordant position with mica-slate and gneiss, have also been observed several times in europe by celebrated geologists. the existence of an independent formation of primitive slate (urthonschiefer) may even be doubted, that is, of a formation which is not joined below by strata containing some vestiges of monocotyledonous plants.

the small thonschiefer bed of malpasso (in the southern chain of the littoral cordillera) is separated from mica-slate-gneiss by a co-ordinate formation of serpentine and diorite. it is divided into two shelves, of which the upper presents green steatitous slate mixed with amphibole, and the lower, dark-blue slate, extremely fissile, and traversed by numerous veins of quartz. i could discover no fragmentary stratum (grauwacke) nor kieselschiefer nor chiastolite. the kieselschiefer belongs in those countries to a limestone formation. i have seen fine specimens of the chiastolite (macle) which the indians wore as amulets and which came from the sierra nevada de merida. this substance is probably found in transition-slate, for mm. rivero and boussingault observed rocks of clay-slate at the height of 2120 toises, in the paramo of mucuchies, on going from truxillo to merida.*

[* in galicia, in spain, i saw the thonschiefer containing chiastholite alternate with grauwacke; but the chiastolite unquestionably belongs also to rocks which all geologists have hitherto called primitive rocks, to mica-schists intercalated like layers in granite, and to an independent stratum of mica-slate.]

3. formation of serpentine and diorite (green-stone of juncalito.)

we have indicated above a layer of granitiferous serpentine inclosed in the gneiss of buenavista, or perhaps superposed on that rock; we here find a real stratum of serpentine alternating with diorite, and extending from the ravine of tucutunemo as far as juncalito. diorite forms the great mass of this stratum; it is of a dark green colour, granular, with small grains, and destitute of quartz; its mass is formed of small crystals of felspar intermixed with crystals of amphibole. this rock of diorite is covered at its surface, by the effect of decomposition, with a yellowish crust, like that of basalts and dolerites. serpentine, of a dull olive-green and smooth fracture, mixed with bluish steatite and amphibole, presents, like almost all the co-ordinate formations of diorite and serpentine (in silesia, at fichtelgebirge, in the valley of baigorry, in the pyrenees, in the island of cyprus and in the copper mountains of circumpolar america),* traces of copper. where the diorite, partly globular, approaches the green slate of malpasso, real beds of green slate are found inclosed in diorite. the fine saussurite which we saw in the upper orinoco in the hands of the indians, seems to indicate the existence of a soil of euphotide, superposed on gneiss-granite, or amphibolic slate, in the eastern part of the sierra parime.

[* franklin’s journey to the polar sea page 529.]

4. granular and micaceous limestone of the morros of san juan.

the morros of san juan rise like ruinous towers in a soil of diorite. they are formed of a cavernous greyish green limestone of crystalline texture, mixed with some spangles of mica, and are destitute of shells. we see in them masses of hardened clay, black, fissile, charged with iron, and covered with a crust, yellow from decomposition, like basalts and amphiboles. a compact limestone containing vestiges of shells adjoins this granular limestone of the morros of san juan which is hollow within. probably on a further examination of the extraordinary strata between villa de cura and ortiz, of which i had time only to collect some few specimens, many phenomena may be discovered analogous to those which leopold von buch has lately described in south tyrol. m. boussingault, in a memoir which he has recently addressed to me, calls the rock of the morros a problematic calcariferous gneiss. this expression seems to prove that the plates of mica take in some parts a uniform direction, as in the greenish dolomite of val toccia.

5. felspathic sandstone of the orinoco.

the gneiss-granite of the sierra parime is covered in some few places (between the encaramada and the strait of baraguan and in the island of guachaco) in its western part with an olive-brown sandstone, containing grains of quartz and fragments of felspar, joined by an extremely compact clayey cement. this cement, where it abounds, has a conchoidal fracture and passes to jasper. it is crossed by small veins of brown iron-ore, which separate into very thin plates or scales. the presence of felspar seems to indicate that this small formation of sandstone (the sole secondary formation hitherto known in the sierra parime) belongs to red sandstone or coal.* i hesitate to class it with the sandstone of the llanos, the relative antiquity of which appears to me to be less satisfactorily verified.

[* broken and intact crystals of feldspar are found in the todte liegende coal-sandstone of thuringia. i observed in mexico a very singular agglomerated felspar formation superposed upon (perhaps inclosed in) red sandstone, near guanaxuato.]

6. formation of the sandstone of the llanos of calabozo.

i arrange the various formations in the order which i fancied i could discern on the spot. the carburetted slate (thonschiefer) of the peninsula of araya connects the primitive rocks of gneiss-granite and mica-slate-gneiss with the transition strata (blue and green slate, diorite, serpentine mixed with amphibole and granular greenish-grey limestone) of malpasso, tucutunemo and san juan. on the south the sandstone of the llanos rests on this transition strata; it is destitute of shells and composed, like the savannahs of calabozo, of rounded fragments of quartz,* kieselschiefer and lydian stone, cemented by a ferruginous olive-brown clay. we there find fragments of wood, in great part monocotyledonous, and masses of brown iron-ore. some strata, as in the mesa de paja, present grains of very fine quartz; i saw no fragments of porphyry or limestone. those immense beds of sandstone that cover the llanos of the lower orinoco and the amazon well deserve the attention of travellers. in appearance they approximate to the pudding-stones of the molassus stratum, in which calcareous vestiges are also often wanting, as at schottwyl and diesbach in switzerland; but they appeared to me by their position to have more relation to red sandstone. nowhere can they be confounded with the grauwackes (fragmentary transition-rocks) which mm. boussingault and rivero found along the cordilleras of new grenada, bordering the steppes on the west. does the want of fragments of granite, gneiss and porphyry, and the frequency of petrified wood,* sometimes dicotyledonous, indicate that those sandstones belong to the more recent formations which fill the plains between the cordillera of the parime and the coast cordillera, as the molassus of switzerland fills the space between the jura and the alps? it is not easy, when several formations are not perfectly developed, to determine the age of arenaceous rocks. the most able geologists do not concur in opinion respecting the sandstone of the black forest and of the whole country south-west of the thuringer waldgebirge. m. boussingault, who passed through a part of the steppes of venezuela long after me, is of opinion that the sandstone of the llanos of san carlos, that of the valley of san antonio de cucuta and the table-lands of barquisimeto, tocuyo, merida and truxillo belong to a formation of old red sandstone or coal. there is in fact real coal near carache, south-west of the paramo de las rosas.

[* in germany sandstones which belong unquestionably to red sandstone contain also (near weiderstadt, in thuringia) nodules, and rounded fragments. i shall not cite the pudding-stone subordinate to the red sandstone of the pyrenees because the age of that sandstone destitute of coal may be disputed. layers of very large rounded nodules of quartz are inclosed in the coal sandstone of thuringia, and in upper silesia.]

[* the people of the country attribute those woods to the alcornoco, bowdichia virgilioides (see nova gen. et spec. plant. volume 3 page 377), and to the chaparro bovo, rhopala complicata. it is believed in venezuela as in egypt that petrified wood is formed in our times. i found this dicotyledonous petrified wood only at the surface of the soil and not inclosed in the sandstone of the llanos. m. caillaud made the same observation on going to the oasis of siwa. the trunks of trees, ninety feet long, inclosed in the red sandstone of kifhauser (in saxony), are, according to the recent researches of von buch, divided into joints, and are certainly monocotyledonous.]

before a part of the immense plains of america was geologically examined, it might have been supposed that their uniform and continued horizontality was caused by alluvial soils, or at least by arenaceous tertiary strata. the sands which in the baltic provinces and in all the north of germany, cover coarse limestone and chalk, seem to justify these systematic ideas, which have been extended to the sahara and the steppes of asia. but the observations which we have been able to collect sufficiently prove that both in the old and the new world, both plains, steppes, and deserts contain numerous formations of different eras, and that these formations often appear without being covered by alluvial deposits. jura limestone, gem-salt (plains of the meta and patagonia) and coal-sandstone are found in the llanos of south america; quadersandstein,* a saliferous soil, beds of coal,* and limestone with trilobites,* fill the vast plains of louisiana and canada. in examining the specimens collected by the indefatigable caillaud in the lybian desert and the oasis of siwa, we recognize sandstone similar to that of thebes; fragments of petrified dicotyledonous wood (from thirty to forty feet long), with rudiments of branches and medullary concentric layers, coming perhaps from tertiary sandstone with lignites;* chalk with spatangi and anachytes, jura limestone with nummulites partly agatized; another fine-grained limestone* employed in the construction of the temple of jupiter ammon (omm–beydah); and gem-salt with sulphur and bitumen. these examples sufficiently prove that the plains (llanos), steppes and deserts have not that uniform tertiary formation which has been too generally supposed. do the fine pieces of riband-jasper, or egyptian pebbles, which m. bonpland picked up in the savannahs of barcelona (near curataquiche), belong to the sandstone of the llanos of calabozo or to a stratum superposed on that sandstone? the former of these suppositions would approach, according to the analogy of the observations made by m. roziere in egypt, the sandstone of calabozo, or tertiary nagelfluhe.

[* the forms of these rocks in walls and pyramids, or divided in rhomboid blocks, seems no doubt to indicate quadersandstein; but the sandstone of the eastern declivity of the rocky mountains in which the learned traveller mr. james found salt-springs (licks), strata of gypsum and no coal, appear rather to belong to variegated sandstone (buntersandstein).]

[* this coal immediately covers, as in belgium, the grauwacke, or transition-sandstone.]

[* in the plains of the upper missouri the limestone is immediately covered by a secondary limestone with turritulites, believed to be jurassic, while a limestone with grypheae, rich in lead-ore and which i should have believed to be still more ancient than oolitic limestone, and analogous to lias, is described by mr. james as lying above the most recent formation of sandstone. has this superposition been well ascertained?]

[* formation of molassus.]

[* m. von buch very reasonably inquires whether this statuary limestone, which resembles parian marble, and limestone become granular by contact with the systematic granite of predazzo, is a modification of the limestone with nummulites, of siwa. the primitive rocks from which the fine-grained marble was believed to be extracted, if there be no deception in its granular appearance, are far distant from the oasis of siwa.]

7. formation of the compact limestone of cumanacoa.

a bluish-grey compact limestone, almost destitute of petrifactions, and frequently intersected by small veins of carburetted lime, forms mountains with very abrupt ridges. these layers have the same direction and the same inclination as the mica-slate of araya. where the flank of the limestone mountains of new andalusia is very steep we observe, as at achsenberg, near altdorf in switzerland, layers that are singularly arched or turned. the tints of the limestone of cumanacoa vary from darkish grey to bluish white and sometimes pass from compact to granular. it contains, as substances accidentally disseminated in the mass, brown iron-ore, spathic iron, even rock-crystal. as subordinate layers it contains (1) numerous strata of carburetted and slaty marl with pyrites; (2) quartzose sandstone, alternating with very thin strata of clayey slate; (3) gypsum with sulphur near guire in the golfo triste on the coast of paria. as i did not examine on the spot the position of this yellowish-white fine-grained gypsum i cannot determine with any certainty its relative age.

([footnote not indicated:] this sandstone contains springs. in general it only covers the limestone of cumanacoa, but it appeared to me to be sometimes enclosed.)

the only petrifactions of shells which i found in this limestone formation consist of a heap of turbinites and trochites, on the flank of turimiquiri, at more than 680 toises high, and an ammonite seven inches in diameter, in the montana de santa maria, north-north-west of caripe. i nowhere saw the limestone of cumanacoa (of which i treat specially in this article) resting on the sandstone of the llanos; if there be any such superposition it must be found on descending the table-land of cocollar towards the mesa de amana. on the southern coast of the gulf of cariaco the limestone formation probably covers, without the interposition of another rock, a mica-slate which passes to carburetted clay-slate. in the northern part of the gulf i distinctly saw this clayey formation at the depth of two or three fathoms in the sea. the submarine hot springs appeared to me to gush from mica-slate like the petroleum of maniquarez. if any doubts remain as to the rock on which the limestone of cumanacoa is immediately superposed, there is none respecting the rocks which cover it, such as (1) the tertiary limestone of cumana near punta delgada and at cerro de meapire; (2) the sandstone of quetepe and turimiquiri, which, forming layers also in the limestone of cumanacoa, belongs properly to the latter soil; the limestone of caripe which we have often identified in the course of this work with jura limestone, and of which we shall speak in the following article.

8. formation of the compact limestone of caripe.

descending the cuchillo de guanaguana towards the convent of caripe, we find another more recent formation, white, with a smooth or slightly conchoidal fracture, and divided in very thin layers, which succeeds to the bluish grey limestone formation of cumanacoa. i call this in the first instance the limestone formation of caripe, on account of the cavern of that name, inhabited by thousands of nocturnal birds. this limestone appeared to me identical (1) with the limestone of the morro de barcelona and the chimanas islands, which contains small layers of black kieselschiefer (slaty jasper) without veins of quartz, and breaking into fragments of parallelopiped form; (2) with the whitish grey limestone with smooth fracture of tisnao, which seems to cover the sandstone of the llanos. we find the formation of caripe in the island of cuba (between the havannah and batabano and between the port of trinidad and rio guaurabo), as well in the small cayman islands.

i have hitherto described the secondary limestone formations of the littoral chain without giving them the systematic names which may connect them with the formations of europe. during my stay in america i took the limestone of cumanacoa for zechstein or alpine limestone, and that of caripe for jura limestone. the carburetted and slightly bituminous marl of cumanacoa, analogous to the strata of bituminous slate, which are very numerous* in the alps of southern bavaria, appeared to me to characterize the former of these formations; while the dazzling whiteness of the cavernous stratum of caripe, and the form of those shelves of rocks rising in walls and cornices, forcibly reminded me of the jura limestone of streitberg in franconia, or of oitzow and krzessowic in upper silesia. there is in venezuela a suppression of the different strata which, in the old continent, separate zechstein from jura limestone. the sandstone of cocollar, which sometimes covers the limestone of cumanacoa, may be considered as variegated sandstone; but it is more probable that in alternating by layers with the limestone of cumanacoa, it is sometimes thrown to the upper limit of the formation to which it belongs. the zechstein of europe also contains a very quartzose sandstone. the two limestone strata of cumanacoa and caripe succeed immediately each other, like alpine and jura limestone, on the western declivity of the mexican table-land, between sopilote, mescala and tehuilotepec. these formations, perhaps, pass from one to the other, so that the latter may be only an upper shelf of zechstein. this immediate covering, this suppression of interposed soils, this simplicity of structure and absence of oolitic strata, have been equally observed in upper silesia and in the pyrenees. on the other hand the immediate superposition of the limestone of cumanacoa on mica-slate and transition clay-slate — the rarity of the petrifactions which have not yet been sufficiently examined — the strata of silex passing to lydian stone, may lead to the belief that the soils of cumanacoa and caripe are of much more ancient formation than the secondary rocks. we must not be surprised that the doubts which arise in the mind of the geologist when endeavouring to decide on the relative age of the limestone of the high mountains in the pyrenees, the apennines (south of the lake of perugia) and in the swiss alps, should extend to the limestone strata of the high mountains of new andalusia, and everywhere in america where the presence of red sandstone is not distinctly recognized.

[* i found them also in the peruvian andes near montau, at the height of 1600 toises.]

9. sandstone of the bergantin.

between nueva barcelona and the cerro del bergantin a quartzose sandstone covers the jura limestone of cumanacoa. is it an arenaceous rock analogous to green sandstone, or does it belong to the sandstone of cocollar? in the latter case its presence seems to prove still more clearly that the limestones of cumanacoa and caripe are only two parts of the same system, alternating with sandstone, sometimes quartzose, sometimes slaty.

10. gypsum of the llanos of venezuela.

deposits of lamellar gypsum, containing numerous strata of marl, are found in patches on the steppes of caracas and barcelona; for instance, in the table-land of san diego, between ortiz and the mesa de paja; and near the mission of cachipo. they appeared to me to cover the jura limestone of tisnao, which is analogous to that of caripe, where we find it mixed with masses of fibrous gypsum. i have not given the name formation either to the sandstone of the orinoco, of cocollar, of bergantin or to the gypsum of the llanos, because nothing as yet proves the independence of those arenaceous and gypsous soils. i think it will one day be ascertained that the gypsum of the llanos covers not only the jura limestone of the llanos, but that it is sometimes enclosed in it like the gypsum of the golfo triste on the east of the alpine limestone of cumanacoa. the great masses of sulphur found in the layers, almost entirely clayey, of the steppes (at guayuta, valley of san bonifacio, buen pastor, confluence of the rio pao with the orinoco) may possibly belong to the marl of the gypsum of ortiz. these clayey beds are more worthy of attention since the interesting observations of von buch and several other celebrated geologists respecting the cavernosity of gypsum, the irregularity of the inclination of its strata and its parallel position with the two declivities of the hartz and the upheaved chain of the alps; while the simultaneous presence of sulphur, oligist iron and the sulphurous acid vapours which precede the formation of sulphuric acid, seem to manifest the action of forces placed at a great depth in the interior of the globe.

11. formation of muriatiferous clay (with bitumen and lamellar gypsum) of the peninsula of araya.

this soil presents a striking analogy with salzthon or leberstein (muriatiferous clay) which i have found accompanying gem-salt in every zone. in the salt-pits of araya (haraia) it attracted the attention of peter martyr d’anghiera at the beginning of the sixteenth century. it probably facilitated the rupture of the earth and the formation of the gulf of cariaco. this clay is of a smoky colour, impregnated with petroleum, mingled with lamellar and lenticular gypsum and sometimes traversed by small veins of fibrous gypsum. it incloses angular and less friable masses of dark brown clay with a slaty and sometimes conchoidal fracture. muriate of soda is found in particles invisible to the naked eye. the relations of position or superposition between this soil and the tertiary rocks does not appear sufficiently clear to enable me to pronounce with certainty on this element, the most important of positive geology. the co-ordinate layers of gem-salt, muriatiferous clay and gypsum present the same difficulties in both hemispheres; these masses, the forms of which are very irregular, everywhere exhibit traces of great commotions. they are scarcely ever covered by independent formations; and after having been long believed, in europe, that gem-salt was exclusively peculiar to alpine and transition limestone, it is now still more generally admitted, either from reasoning founded on analogy or from suppositions on the prolongation of the strata, that the true location of gem-salt is found in variegated sandstone (buntersandstein). sometimes gem-salt appears to oscillate between variegated sandstone and muschelkalk.

i made two excursions on the peninsula of araya. in the first i was inclined to consider the muriatiferous clay as subordinate to the conglomerate (evidently of tertiary formation) of the barigon and of the mountain of the castle of cumana, because a little to the north of that castle i had found shelves of hardened clay containing lamellar gypsum inclosed in the tertiary strata. i believed that the muriatiferous clay might alternate with the calcareous conglomerate of barigon; and near the fishermen’s huts situated opposite macanao, conglomerate rocks appeared to me to pierce through the strata of clay. during a second excursion to maniquarez and the aluminiferous slates of chaparuparu, the connexion between tertiary strata and bituminous clay seemed to me somewhat problematical. i examined more particularly the penas negras near the cerro de la vela, east-south-east of the ruined castle of araya. the limestone of the penas is compact, bluish grey and almost destitute of petrifactions. it appeared to me to be much more ancient than the tertiary conglomerate of barigon, and i saw it covering, in concordant position, a slaty clay, somewhat analogous to muriatiferous clay. i was greatly interested in comparing this latter formation with the strata of carburetted marl contained in the alpine limestone of cumanacoa. according to the opinions now most generally received, the rock of the penas negras may be considered as representing muschelkalk (limestone of gottingen); and the saliferous and bituminous clay of araya, as representing variegated sandstone; but these problems can only be solved when the mines of those countries are worked. those geologists who are of opinion that the gem-salt of italy penetrates into a stratum above the jura limestone, and even the chalk, may be led to mistake the limestone of the penas negras for one of the strata of compact limestone without grains of quartz and petrifactions, which are frequently found amidst the tertiary conglomerate of barigon and of the castillo de cumana; the saliferous clay of araya would appear to them analogous to the plastic clay of paris,* or to the clayey shelves (dief et tourtia) of secondary sandstone with lignites, containing salt-springs, in belgium and westphalia. however difficult it may be to distinguish separately the strata of marl and clay belonging to variegated sandstone, muschelkalk, quadersandstein, jura limestone, secondary sandstone with lignites (green and iron sand) and the tertiary strata lying above chalk, i believe that the bitumen which everywhere accompanies gem-salt, and most frequently salt-springs, characterizes the muriatiferous clay of the peninsula of araya and the island of marguerita, as linked with formations lying below the tertiary strata. i do not say that they are anterior to that formation, for since the publication of m. von buch’s observations on the tyrol, we must no longer consider what is below, in space, as necessarily anterior, relatively to the epoch of its formation.

[* tertiary sandstone with lignites, or molassus of argovia.]

bitumen and petroleum still issue from the mica-slate; these substances are ejected whenever the soil is shaken by a subterranean force (between cumana, cariaco and the golfo triste). now, in the peninsula of araya, and in the island of marguerita, saliferous clay impregnated with bitumen is met with in connexion with this early formation, nearly as gem-salt appears in calabria in flakes, in basins inclosed in strata of granite and gneiss. do these circumstances serve to support that ingenious system, according to which all the co-ordinate formations of gypsum, sulphur, bitumen and gem-salt (constantly anhydrous) result from floods passing across the crevices which have traversed the oxidated crust of our planet, and penetrating to the seat of volcanic action. the enormous masses of muriate of soda recently thrown up by vesuvius,* the small veins of that salt which i have often seen traverse the most recently ejected lavas, and of which the origin (by sublimation) appears similar to that of oligist iron deposited in the same vents,* the layers of gem-salt and saliferous clay of the trachytic soil in the plains of peru and around the volcano of the andes of quito are well worthy the attention of geologists who would discuss the origin of formations. in the present sketch i confine myself to the mere enumeration of the phenomena of position, indicating, at the same time, some theoretic views, by which observers in more advantageous circumstances than i was myself may direct their researches.

[* the ejected masses in 1822 were so considerable that the inhabitants of some villages round vesuvius collected them for domestic purposes.]

[* gay–lussac on the action of volcanoes in the annales de chimie volume 22 page 418.]

12. agglomerate limestone of the barigon, of the castle of cumana, and of the vicinity of porto cabello.

this is a very complex formation, presenting that mixture and that periodical return of compact limestone, quartzose sandstone and conglomerates (limestone breccia) which in every zone peculiarly characterises the tertiary strata. it forms the mountain of the castle of san antonio near the town of cumana, the south-west extremity of the peninsula of araya, the cerro meapire, south of caraco and the vicinity of porto cabello. it contains (1) a compact limestone, generally of a whitish grey, or yellowish white (cerro del barigon), some very thin layers of which are entirely destitute of petrifactions, while others are filled with cardites, ostracites, pectens and vestiges of lithophyte polypi: (2) a breccia in which an innumerable number of pelagic shells are found mixed with grains of quartz agglutinated by a cement of carbonate of lime: (3) a calcareous sandstone with very fine rounded grains of quartz (punta arenas, west of the village of maniquarez) and containing masses of brown iron ore: (4) banks of marl and slaty clay, containing no spangles of mica, but enclosing selenite and lamellar gypsum. these banks of clay appeared to me constantly to form the lower strata. there also belongs to this tertiary stratum the limestone tufa (fresh-water formation) of the valleys of aragua near vittoria, and the fragmentary rock of cabo blanco, westward of the port of la guayra. i must not designate the latter by the name of nagelfluhe, because that term indicates rounded fragments, while the fragments of cabo blanco are generally angular, and composed of gneiss, hyaline quartz and chloritic slate, joined by a limestone cement. this cement contains magnetic sand,* madrepores, and vestiges of bivalve sea shells. the different fragments of tertiary strata which i found in the littoral cordillera of venezuela, on the two slopes of the northern chain, seem to be superposed near cumana (between bordones and punta delgada); in the cerro of meapire; on the [alpine] limestone of cumanacoa; between porto cabello and the rio guayguaza; as well as in the valleys of aragua; on granite; on the western declivity of the hill formed by cabo blanco, on gneiss; and in the peninsula of araya, on saliferous clay. but this is perhaps merely the effect of apposition.* if we would range the different members of the tertiary series according to the age of their formation we ought, i believe, to regard the breccia of cabo blanco with fragments of primitive rocks as the most ancient, and make it be succeeded by the arenaceous limestone of the castle of cumana, without horned silex, yet somewhat analogous to the coarse limestone of paris, and the fresh-water soil of victoria. the clayey gypsum, mixed with calcareous breccia with madrepores, cardites and oysters, which i found between carthagena and the cerro de la popa, and the equally recent limestones of guadalope and barbadoes (limestones filled with seashells resembling those now existing in the caribbean sea) prove that the latest deposited strata of the tertiary formation extend far towards the west and north.

[* this magnetic sand no doubt owes its origin to chloritous slate, which, in these latitudes, forms the bed of the sea.]

[* an-nicht auflagerung, according to the precise language of the geologists of my country.]

these recent formations, so rich in vestiges of organized bodies, furnish a vast field of observation to those who are familiar with the zoological character of rocks. to examine these vestiges in strata superposed as by steps, one above another, is to study the fauna of different ages and to compare them together. the geography of animals marks out limits in space, according to the diversity of climates, which determine the actual state of vegetation on our planet. the geology of organized bodies, on the contrary, is a fragment of the history of nature, taking the word history in its proper acceptation: it describes the inhabitants of the earth according to succession of time. we may study genera and species in museums, but the fauna of different ages, the predominance of certain shells, the numerical relations which characterize the animal kingdom and the vegetation of a place or of a period, should be studied in sight of those formations. it has long appeared to me that in the tropics as well as in the temperate zone the species of univalve shells are much more numerous than bivalves. from this superiority in number the organic fossil world furnishes, in every latitude, a further analogy with the intertropical shells that now live at the bottom of the ocean. in fact, m. defrance, in a work* full of new and ingenious ideas, not only recognizes this preponderance of the univalves in the number of the species, but also observes that out of 5500 fossil univalve, bivalve and multivalve shells, contained in his rich collections, there are 3066 univalve, 2108 bivalve, and 326 multivalve; the univalve fossils are therefore to the bivalve as three to two.

[* table of organized fossil bodies, 1824.]

13. formation of pyroxenic amygdaloid and phonolite, between ortiz and cerro de flores.

i place pyroxenic amygdaloid and phonolite (porphyrschiefer) at the end of the formations of venezuela, not as being the only rocks which i consider as pyrogenous, but as those of which the volcanic origin is probably posterior to the tertiary strata. this conclusion is not deduced from the observations i made at the southern declivity of the littoral cordillera, between the morros of san juan, parapara and the llanos of calabozo. in that region local circumstances would possibly lead us to regard the amygdaloids of ortiz as linked to a system of transition rocks (amphibolic serpentine, diorite, and carburetted slate of malpasso); but the eruption of the trachytes across rocks posterior to the chalk (in the euganean mountains and other parts of europe) joined to the phenomenon of total absence of fragments of pyroxenic porphyry, trachyte, basalt and phonolite (the fragments of these rocks appear only in tufas or conglomerates which belong essentially to basaltic formations or surround the most recent volcanoes. every volcanic formation is enveloped in breccia, which is the effect of the eruption itself.), in the conglomerates or fragmentary rocks anterior to the recent tertiary strata, renders it probable that the appearance of trap rocks at the surface of the earth is the effect of one of the last revolutions of our planet, even where the eruption has taken place by crevices (veins) which cross gneiss-granite, or the transition rocks not covered by secondary and tertiary formations.

the small volcanic stratum of ortiz (latitude 9° 28′ to 9° 36′) formed the ancient shore of the vast basin of the llanos of venezuela: it is composed on the points where i could examine it of only two kinds of rocks, namely, amygdaloid and phonolite. the greyish blue amygdaloid contains fendilated crystals of pyroxene and mesotype. it forms balls with concentric layers of which the flattened centre is nearly as hard as basalt. neither olivine nor amphibole can be distinguished. before it shows itself as a separate stratum, rising in small conic hills, the amygdaloid seems to alternate by layers with the diorite, which we have mentioned above as mixed with carburetted slate and amphibolic serpentine. these close relations of rocks so different in appearance and so likely to embarrass the observer give great interest to the vicinity of ortiz. if the masses of diorite and amygdaloid, which appear to us to be layers, are very large veins, they may be supposed to have been formed and upheaved simultaneously. we are now acquainted with two formations of amygdaloid; one, the most common, is subordinate to the basalt: the other, much more rare,* belongs to the pyroxenic porphyry.* the amygdaloid of ortiz approaches, by its oryctognostic characters, to the former of those formations, and we are almost surprised to find it joining, not basalt, but phonolite,* an eminently felspathic rock, in which we find some crystals of amphibole, but pyroxene very rarely, and never any olivine. the cerro de flores is a hill covered with tabulary blocks of greenish grey phonolite, enclosing long crystals (not fendillated) of vitreous felspar, altogether analogous to the phonolite of mittelgebirge. it is surrounded by pyroxenic amygdaloid; it would no doubt be seen below, issuing immediately from gneiss-granite, like the phonolite of biliner stein, in bohemia, which contains fragments of gneiss embedded in its mass.

[* we find examples of the latter in norway (vardekullen, near skeen), in the mountains of the thuringerwald; in south tyrol; at hefeld in the hartz, at bolanos in mexico etc.]

[* black porphyries of m. von buch.]

[* there are phonolites of basaltic strata (the most anciently known) and phonolites of trachytic strata (andes of mexico). the former are generally above the basalts; and the extraordinary development of felspar in that union, and the want of pyroxene, have always appeared to me very remarkable phenomena.]

does there exist in south america another group of rocks, which may be preferably designated by the name of volcanic rocks, and which are as distinct from the chain of the andes, and advance as far towards the east as the group that bounds the steppes of calabozo? of this i doubt, at least in that part of the continent situated north of the amazon. i have often directed attention to the absence of pyroxenic porphyry, trachyte, basalt and lavas (i range these formations according to their relative age) in the whole of america eastward of the cordilleras. the existence even of trachyte has not yet been verified in the sierra nevada de merida which links the andes and the littoral chain of venezuela. it would seem as if volcanic fire, after the formation of primitive rocks, could not pierce into eastern america. possibly the scarcity of argentiferous veins observed in those countries may be owing to the absence of more recent volcanic phenomena. m. eschwege saw at brazil some layers (veins?) of diorite, but neither trachyte, basalt, dolerite, nor amygdaloid; and he was therefore much surprised to see, in the vicinity of rio janeiro, an insulated mass of phonolite, exactly similar to that of bohemia, piercing through gneiss. i am inclined to believe that america, on the east of the andes, would have burning volcanoes if, near the shore of venezuela, guiana and brazil, the series of primitive rocks were broken by trachytes, for these, by their fendillation and open crevices, seem to establish that permanent communication between the surface of the soil and the interior of the globe, which is the indispensable condition of the existence of a volcano. if we direct our course from the coast of paria by the gneiss-granite of the silla of caracas, the red sandstone of barquisimeto and tocuyo, the slaty mountains of the sierra nevada de merida, and the eastern cordillera of cundinamarca to popayan and pasto, taking the direction of west-south-west, we find in the vicinity of those towns the first volcanic vents of the andes still burning, those which are the most northerly of all south america; and it may be remarked that those craters are found where the cordilleras begin to present trachytes, at a distance of eighteen or twenty-five leagues from the present coast of the pacific ocean.* permanent communications, or at least communications frequently renewed, between the atmosphere and the interior of the globe, have been preserved only along that immense crevice on which the cordilleras have been upheaved; but subterranean volcanic forces are not less active in eastern america, shaking the soil of the littoral cordillera of venezuela and of the parime group. in describing the phenomena which accompanied the great earthquake of caracas,* on the 26th march, 1812, i mentioned the detonations heard at different periods in the mountains (altogether granitic) of the orinoco. the elastic forces which agitate the ground, the still-burning volcanoes, the hot sulphurous springs, sometimes containing fluoric acid, the presence of asphaltum and naphtha in primitive strata, all point to the interior of our planet, the high temperature of which is perceived even in mines of little depth, and which, from the times of heraclitus of ephesus, and anaxagoras of clazomenae, to the plutonic theory of modern days, has been considered as the seat of all great disturbances of the globe.

[* i believe the first hypotheses respecting the relation between the burning of volcanoes and the proximity of the sea are contained in aetna dialogus, a very eloquent though little-known work by cardinal bembo.]

[* i stated in another place the influence of that great catastrophe on the counter-revolution which the royalist party succeeded in bringing about at that time in venezuela. it is impossible to conceive anything more curious than the negociation opened on the 5th of april, by the republican government, established at valencia in the valleys of aragua, with archbishop prat (don narciso coll y prat), to engage him to publish a pastoral letter calculated to tranquilize the people respecting the wrath of the deity. the archbishop was permitted to say that this wrath was merited on account of the disorder of morals; but he was enjoined to declare positively that politics and systematic opinions on the new social order had nothing in common with it. archbishop prat lost his liberty after this singular correspondence.]

the sketch i have just traced contains all the formations known in that part of europe which has served as the type of positive geology. it is the fruit of sixteen months’ labour, often interrupted by other occupations. formations of quartzose porphyry, pyroxenic porphyry and trachyte, of grauwacke, muschelkalk and quadersandstein, which are frequent towards the west, have not yet been seen in venezuela; but it may be also observed that in the system of secondary rocks of the old continent muschelkalk and quadersandstein are not always clearly developed, and are often, by the frequency of their marls, confounded with the lower layers of jura limestone. the muschelkalk is almost a lias with encrinites; and quadersandstein (for there are doubtless many above the lias or limestone with gryphites) seems to me to represent the arenaceous layers of the lower shelves of jura limestone.

i have thought it right to give at some length this geologic description of south america, not only on account of the novel interest which the study of the formations in the equinoctial regions is calculated to excite, but also on account of the honourable efforts which have recently been made in europe to verify and extend the working of the mines in the cordilleras of columbia, mexico, chile and buenos ayres. vast sums of money have been invested for the attainment of this useful end. in proportion as public confidence has enlarged and consolidated those enterprises, from which both continents may derive solid advantage, it becomes the duty of persons who have acquired a local knowledge of these countries to publish information calculated to create a just appreciation of the relative wealth and position of the mines in different parts of spanish america. the success of a company for the working of mines, and that of works undertaken by the order of free governments, is far from depending solely on the improvement of the machines employed for draining off the water, and extracting the mineral, on the regular and economical distribution of the subterraneous works, or the improvements in preparation, amalgamation, and melting: success depends also on a thorough knowledge of the different superposed strata. the practice of the science of mining is closely linked with the progress of geology; and it would be easy to prove that many millions of piastres have been rashly expended in south america from complete ignorance of the nature of the formations, and the position of the rocks, in directing the preliminary researches. at the present time it is not precious metals solely which should fix the attention of new mining companies; the multiplication of steam-engines renders it indispensable, wherever wood is not abundant or easy of transport, to seek at the same time to discover coal and lignites. in this point of view the precise knowledge of the red sandstone, coal-sandstone, quadersandstein and molassus (tertiary formation of lignites), often covered with basalt and dolerite, is of great practical importance. it is difficult for a european miner, recently arrived, to judge of a country presenting so novel an aspect, and when the same formations cover an immense extent. i hope that the present work, as well as my political essay on new spain, and my work on the position of rocks in the two hemispheres, will contribute to diminish those obstacles. they may be said to contain the earliest geologic information respecting places whose subterraneous wealth attracts the attention of commercial nations; and they will assist in the classification of the more precise notions which later researches may add to my labours.

the republic of colombia, in its present limits, furnishes a vast field for the enterprising spirit of the miner. gold, platinum, silver, mercury, copper, gem-salt, sulphur and alum may become objects of important workings. the production of gold alone amounted, before the outbreak of the political dissensions, on the average, to 4700 kilogrammes (20,500 marks of castile) per annum. this is nearly half the quantity furnished by all spanish america, a quantity which has an influence the more powerful on the variable proportions between the value of gold and silver, as the extraction of the former metal has diminished at brazil, for forty years past, with surprising rapidity. the quint (a tax which the government raises on gold-washings) which in the capitania of minas geraes was, in 1756, 1761 and 1767, from 118, 102 and 85 arobas of gold (of 14 3/5 kilogrammes), has fallen, during 1800, 1813 and 1818, to 30, 20 and 9 arobas; an arob of gold having, at rio janeiro, the value of 15,000 cruzados. according to these estimates the produce of gold in brazil, making deductions for fraudulent exportation, was, in the middle of the eighteenth century, the years of the greatest prosperity of the gold-washings, 6600 kilogrammes, and in our days, from 1817 to 1820, 600 kilogrammes less. in the province of san paulo the extraction of gold has entirely ceased; in the province of goyaz, it was 803 kilogrammes in 1793 and in 1819 scarcely 75. in the province of mato grosso it is almost nothing; and m. eschwege is of opinion that the whole produce of gold in brazil does not amount at present to more than 600,000 cruzados (scarcely 440 kilogrammes). i dwell on these particulars because, in confounding the different periods of the riches and poverty of the gold-washings of brazil, it is still affirmed in works treating of the commerce of the precious metals, that a quantity of gold equivalent to four millions of piastres (5800 kilogrammes of gold*) flows into europe annually from portuguese america. this produce is still estimated (mean year 1810 to 1821) at 1,736,000 piastres; while, according to official documents in my possession, the average of the quint of those ten years amounted only to 15 arrobas, or a product quint of 1095 kilogrammes, or 755,000 piastres. mr. john allen reminded the committee of the bullion report, in his critical notes on the table of m. brongniart, that the decrease of the produce of the gold-washings of brazil had been extremely rapid since 1794; and the notions given by m. auguste de saint hilaire indicate the same desertion of the gold-mines of brazil. those who were miners have become cultivators. the value of an arroba of gold is 15,000 brazilian cruzados (each cruzado being 50 sous). according to m. franzini the portuguese onca is equal to 0.028 of a kilogramme, and 8 oncas make 1 mark; 2 marks make 1 arratel, and 32 arratels 1 arroba.) if, in commercial value, gold in grains prevails, in the republic of columbia, over the value of other metals, the latter are not on that account less worthy to fix the attention of government and of individuals. the argentiferous mines of santa anna, manta, santo christo de las laxas, pamplona, sapo and la vega de sapia afford great hope. the facility of the communications between the coast of columbia and that of europe imparts the same interest to the copper-mines of venezuela and new grenada. metals are a merchandize purchased at the price of labour and an advance of capital; thus forming in the countries where they are produced a portion of commercial wealth; while their extraction gives an impetus to industry in the most barren and mountainous districts.

[* this error is twofold: it is probable that brazilian gold, paying the quint, has not, during the last forty years, risen to 5500 kilogrammes. i heretofore shared this error in common with writers on political economy, in admitting that the quint in 1810 was still (instead of 26 arrobas or 379 kilogrammes) 51,200 portuguese ounces, or 1433 kilogrammes; which supposed a product of 7165 kilogrammes. the very correct information afforded by two portuguese manuscripts on the gold-washings of minas geraes, minas novas and goyaz, in the bullion report for the house of commons, 1810, acc. page 29, goes as far only as 1794, when the quinto do ouro of brazil was 53 arrobas, which indicates a produce of more than 3900 kilogrammes paying the quint. in mr. tooke’s important work, on high and low prices part 2 page 2]