About an attempt to climb to the top of Chimborazo

Abstract

Alexander von Humboldt wrote the final version of “Über einen Versuch den Gipfel des Chimborazo zu ersteigen” [About an Attempt to Climb to the Top of Chimborazo] almost 50 years after his return from the Americas. He published it in his Kleinere Schriften [Shorter Writings] in 1853, a scan of which is available from Google Books. As Humboldt himself mentions, he based a good part of this essay on his travel diary from 23 June 1802. Although he followed his diary fairly closely, Humboldt, as was his wont, updated the essay to include recent information. The earlier version from 1838 lacks the later, revised essay's copious endnotes.

Keywords:

• Alexander von Humboldt
• Chimborazo
• Quito, Ecuador
• volcanoes
• mountaineering
• geology

Notes

1. Humboldt described his ascent of Chimborazo in a lecture he delivered to the Association of Naturalists in Jena, Germany, on 26 September 1838. The earlier version, entitled “Über zwei Versuche den Chimborazo zu besteigen” [About Two Attempts to Ascend Chimborazo], was published in Heinrich Christian Schumacher's Jahrbuch für 1837 (Stuttgart: Cotta, 1837), 176–206. A French version followed in 1838: “Notice sur deux tentatives d'ascension du Chimborazo” in Annales de Chimie et de Physique 69 (1838): 401–34. The 1837 essay was first translated by Dr Martin Barry, a member of the Royal College of Surgeons of Edinburgh who had studied in Heidelberg and resided in Paris in 1827, in “Jameson's Journal” – the Edinburgh New Philosophical Journal, edited by Robert Jameson – and reprinted in four installments in the section “The Contemporary Traveller” in the Mirror of Literature, Amusement and Instruction 31 (1838): 92–8, 134–6, 163–5, 181–4. This London weekly was devoted to “select extracts from new and expensive works.” The French text of Humboldt's diary is available in Margot Faak's Reise auf dem Río Magdalena, durch die Anden und Mexico, part 1 (Berlin: Akademie Verlag, 1986), 215–25. For a Spanish translation, see Segundo E. Moreno Yánez, Diarios de viaje en la audiencia de Quito (Quito: Oxy, 2005), 189–204. [Trans.]

2. The bracketed page numbers refer to the pagination of Humboldt's German text in Kleinere Schriften (Vol. 1. Stuttgart: Cotta, 1853), 133–73; any words and phrases in other languages, notably Spanish, French, and Latin, have not been translated unless Humboldt himself also rendered them in German. Parenthetical page numbers at the beginning of each note are Humboldt's references to the pages of his essay. Occasional math errors have been tacitly corrected. One pied de Paris = 324.8406 millimeters. [Trans.]

3. [163] (P. 133) This is the place to bring together the numbers that, according to the current state of our hypsometric knowledge (Spring 1850), express the culmination points of the mountain chains on both continents. Since, in addition to fluctuating opinions, careless reductions of measurements have become the reason for such very different information in books and maps, I offer here the most important altitudes in British feet, toises, and meters.

The highest peaks of India were identified more than 70 to 80 years later than those of the American cordilleras. Not until the years 1819 to 1825 did people figure out, by way of combining famous works by British travelers ([Brian Houghton] Hodgson, Webb, Herbert, William Lloyd, the brothers Gerard [Alexander and James]), that in the part of the Himalaya chain that stretches from east to west, one had to recognize Dhawalagiri (white mountain) and Jawahir as the tallest peaks. To Dhawalagiri (lat. 28° 40', long. 80° 59' east of Paris) were attributed 26,345 pieds de Paris = 4391 toises = 8558 meters = 28,977 British feet; to Jawahir (lat. 30° 22', long. 77° 37') 24,160 pieds de Paris = 4027 toises = 7848 meters = 25,749 British feet. The measurement of Dhawalagiri, which the Tyrolean Jesuit [Joseph] Tieffenthaler had already marked on his map of the Himalayas in 1766 under the name Montes Albi quis Indis, nive obsiti, is less certain and erroneously explained in letters from Colebrooke to me (see Asie Centrale, vol. 1, pp. 281–90, and Kosmos, vol. 1, p. 41). Letters that I received from my friend Dr Joseph Hooker, the knowledgeable botanist of the last expedition to the South Pole, from Darjeeling in Sikkim-Himalaya (summer 1848), informed me that in the meridian of Sikkim between Dhawalagiri and [Gunung] Samalari, between Bhutan and Nepal, Colonel Baugh, director of the Trigonometric Survey of India, had measured a mountain, Kinchinjunga or Kintschin-Dschunga, with great precision, a mountain whose western snow peak was at 26,439 pieds de Paris = 4406 toises = 8588 meters = 28,278 British feet above sea [164] level. The eastern snow-capped peak is 25,356 pieds de Paris = 4226 toises = 8236 meters = 27,826 British feet high (compare Journal of the Asiatic Society of Bengal, Nov. 1848, vol. 17, part 2, p. 577). The notable colossus Kinchinjunga is shown on the title engraving of the magnificent work by Joseph Hooker, The Rhododendrons of Sikkim-Himalaya, 1849. This mountain is 379 toises higher than Jawahir, and it has been measured with such diligence that the seven results of Mr [Andrew] Waugh's trigonometric calculation from different points only fluctuated between 28,125.7 and 28,212.8. The base line that was measured in the plains was 36,685 British feet long. Later measuring of Dhawalagiri caused the Rhododendra monograph to declare Kinchinjunga as the higher of the two mountains. It appears, however, that there is such a small difference between these two and a third gigantic peak, Deodangha, that one is uncertain whether that difference may not be the result of an erroneous measurement. All three mountains are surely a little above 28,000 feet (above 26,272 pieds de Paris). “Mr Waugh concludes,” Dr Joseph Hooker wrote to me from Darjeeling on 26 April 1849, “that there can be but little difference between Dhawalagiri, Kinchinjinga [sic] and Deodangha, that no other peaks approach these.”

For 18 years, from 1830 to 1848, the following were regarded as the highest points of the cordilleras of the New Continent: Nevado de Sorata, the southern peak of this snow-capped mountain, a bit south of the village of Sorata (or Esquibel) in the eastern chain of Bolivia; and Nevado de Illimani, west of the mission of Yrupana, also in the eastern chain of Bolivia, the one that is most distant from the coast. In those days, the followings elevations were attributed to the two mountains: Sorata (south lat. 15° 51′ and longitude west of Paris 70° 54') 23,692 pieds de Paris = 3940 toises = 7696 meters = 25,250 British feet; Illimani (lat. 16° 39', longitude 70° 9') 22,519 pieds de Paris = 3753 toises = 7315 meters = 23,999 British feet. Pentland, who had long been the political representative of the British government in the free state of Bolivia, had made these hypsometric observations in 1827 and sent them to Mr Arago to publish them in the Annuaire du Bureau des Longitudes pour 1830 (p. 323). Since then, they have unfortunately been disseminated in all sorts of languages, in all writings on mountain elevations, and in many hypsometric mountain profiles. Since the appearance of the large, beautiful map of the basin of the Laguna de Titicaca, however, which Mr Pentland published in London in June 1848 (title: La Laguna de Titicaca and the Valleys [165] of Yucay, Collao and Desagüadero in Peru and Bolivia), we have learned that the above information about the heights of Sorata and Illimani are inflated by 3716 and 2675 pieds de Paris. The map grants Sorata 21,286 and Illimani 21,149 British feet, that is, only 19,972 and 19,843 pieds de Paris (3328 and 3307 toises), respectively. Mr Pentland gathered these results during a second stay in Bolivia, when he recalculated the trigonometric measurements from 1838. Therefore, for 18 years, from 1829 to 1848, it was erroneously claimed that Chimborazo, for which I had found 20,100 pieds de Paris = 3350 toises = 6530 meters = 21,422 British feet, was a full 3592 pieds de Paris (or 3827 British feet) lower than Sorata; even that the latter had only 2653 pieds de Paris (or 2828 British feet) less than Dhawalagiri, the highest peak of Himalayas. I myself have contributed much to the dissemination of these faulty opinions. We know now that Sorata is 126 feet lower than Chimborazo, and 6371 pieds de Paris (or 6791 British feet) lower than Dhawalagiri. Of the two trigonometric measurements of the Peak of Tenerife, which [Jean-Charles] Borda had made during the two expeditions of 1771 and 1776 (the first one with [Alexandre-Guy] Pingré, the second with Chastenet de Puységur), the first one was also in error by 1224 pieds de Paris. An angle had been accidentally recorded as 33' rather than 53'. After their research on the island in 1742, Borda and Pingré, surmising information from elevation angles at an assumed distance, gave the Peak of Tenerife a mere 1701 toises above sea level. The excellent trigonometric operation of 1776 yielded 1905 toises, while the barometric measurement that was recalculated according to [Pierre Simon, Marquis de] Laplace's formula, produced 1976 toises. Borda's earlier error was thus 1200 1/9 feet of the whole calculation, while, in the case of Sorata, the error was 3700 1/5 feet of the elevation only (compare my Voyage aux régions équinoxiales, vol. 1, pp. 277–83, where I first published fragments of an unpublished manuscript by Borda, which is preserved in the Dépôt de la Marine in Paris).

But even though Sorata and Illimani are lower than Chimborazo, it probably remains impossible to declare the latter the highest point on the entire New Continent. In August 1835, the officers of the Expedition and the Beagle, led by Captain [Robert] Fitzroy, measured Nevado de Aconcagua (lat. 32° 39') in the north-east of Valparaiso by using elevation angles, and they determined that it was between 23,000 and 23,400 British feet high. If one estimates this Nevado's elevation at 23,200 feet (or 21,767 pieds de Paris), [166] it would be 1667 pieds de Paris higher than Chimborazo (Narrative of the Voyages of the Adventure and Beagle, vol. 2, 1839; Proceedings of the Second Expedition, under the Command of Capt. Fitzroy, p. 481; Darwin, Journal of Researches, 1845, pp. 253 and 291). According to newer calculations of the same angles by Pentland, Aconcagua is supposed to have a height of 23,010 British feet = 22,434 pieds de Paris = 3739 toises (Mary Somerville, Physical Geography, 1849, vol. 2, p. 425). The mountain would thus be 2334 pieds de Paris higher than Chimborazo. Pentland's beautiful map shows four other peaks in that same western cordillera of Bolivia, east of Arica, between lat. 18° 7' and 18° 25', all of which exceed the height of Chimborazo. Those mountains are Sajama, Parinacota, Gualateiri, and Pomarape. The tallest of them (Sajama) supposedly has 20,971, the lowest (Pomarape) 20,360 pieds de Paris. Sajama would be 1463 pieds de Paris lower than Nevado de Aconcagua, but 871 feet higher than Chimborazo. I do not find it unimportant periodically to articulate in numbers what we know or believe about the form of the surface of our planet. Regrettably, I think, the highest points of the massive rises are isolated phenomena, even if they, like the fruitless climbs of high snow-capped mountains, fascinate people no end.

4. (P. 135) My trigonometric measurements of the elevation of Chimborazo above the level of the South Sea occurred in June of 1803 on the pumice-covered high plains of Tapia not far from the new city of Riobamba, between the church of La Merced and the monastery of St Augustine. The base line was 1702.40 = 874 toises long. The third segment of this line was measured three times. The distance from the endpoint A of the base line to the mountain's summit turned out to be 30,662.73; the horizontal distance, 30,437.40; in segments of the circle 16' 27.65". The elevation angle, freed from refraction and measured with the sextant on the artificial horizon, was A 6° 48' 58.20"; the resulting elevation of the summit of Chimborazo above the plain of the base line was 3639.35 = 1867.25 toises. According to my barometric readings, then, the high plateau of Tapia is 2891.2 = 1482.8 toises above sea level (Boussingault found it to be 11 toises less in a different season, when the heat decrease of the stacked air layers was different). Consequently, the entire elevation of Chimborazo is 6530.5 = 3350 toises, or 20,100 pieds de Paris.

According to Laplace's refraction formula from his [167] Mécanique céleste, Chimborazo would be at 3637.75 (with the effect of refraction) and 3645.32 (without refraction). In order for the result of the total elevation to shift by 21.4 meters, the error of the base line would have to be 10 meters. If the elevation angle were off by 10′′, that would impact the total height calculation by only 1.5 meters (for details on the individual parts of the entire calculation, compare Oltmanns in my Recueil d'observations astronomiques, introduction, vol. 1, pp. lxxii–lxxiv). My conclusions about the elevation of Chimborazo fall between the determinations of La Condamine and Don Jorge Juan; it approaches the latter's by 30 toises. If one considers the complications that conclusions about elevation confronted at a time when temperature corrections were applied to barometric measurements either not at all or with the completely wrong methods – and yet it was necessary to reduce geodetic operations from a height of 1350 to 1500 toises to sea level, as in the case of the measured base line between Caraburu and Oyambaro or Quito – one can explain the great differences in the results that derived from the same observations by both the French and the Spanish astronomer. Other combinations led to different hypsometric determinations. Bouguer and La Condamine give Chimborazo 3220 toises; Don Jorge Juan and his colleagues 3380 toises. The elevation of Quito, which, however, La Condamine and Bouguer had already underestimated by 32 toises – even Boussingault himself by 36 toises (216 feet) – does not directly affect these differences, because the elevation of the snow-capped mountains does not depend on it. Rather, it depends on the reduction to sea level of the measured base line, between Caraburu and Oyambaro in the plains of Yaruqui, by way of a series of triangles whose position points are mostly between 1800 and 2200 toises high. Don Jorge Juan himself delivers conspicuous proof of the unreliability of such complicated combinations, when, on the basis of various hypotheses, he reports the length of the base line of Caraburu at 1155, 1214, 1268, and 1283 toises (differences of 678 feet). Elevation angles of the summit of Chimborazo were taken from four surveying stations, nearest the mountain and yet only at 4° 19' 55" in Mulmul; but Mulmul itself could only be connected with the base line in Yaruqui (at a distance of 22 geographical miles) through triangles and the requisite series of signals. We have only a very imperfect notion of how the reduction to sea level, of this base line and of all the signals, occurs in order simultaneously to determine Chimborazo's absolute elevation. One only learns in general terms [168] that what was used for that reduction was the Cacumen lapideum of Pichincha and the two pyramids of Nevado de Iliniza, which are visible far toward the coastline; I reproduced the pyramids in Vues des Cordillères, Plate XXXV. The French academicians already assumed, however, that the Rucu-Pichincha was 2491 to 2432.69 toises (414 feet) too low. “Je ne pouvois partager, en août 1740, avec M. Bouguer,” writes La Condamine, “les fatigues d'une course pénible et laborieuse de près de deux mois, dans la Province d'Esmeraldas, pour déterminer, dans un lieu dont la hauteur au dessus de la mer fût connue, celle de quelques-unes de nos montagnes, afin de pouvoir réduire au niveau de la surface de la mer la valeur du degré que nous avions mesuré sur le haut de la Cordelière. L'observatoire de M. Bouguer (le point d'où il pouvoit voir Iliniça) étoit établi dans l'Isle de l'Inca sur la rivière d'Esmeraldas. – En mars 1741 j'étois occupé d'un travail peu agréable sur la hauteur absolue des montagnes. J'étois bien sûr que le travail de M. Bouguer a l'Isle de l'Inca et les angles observes a Papa-ourcou près du Cotopaxi comme au Quinche, où nous avions opéré ensemble, n'avoit pas besoin de vérification, et d'autant moins, que cent toises d'erreur sur la hauteur des montagnes n'auroient pas change de deux toises la longueur du degré. La multiplicité des éléments de cette supputation, et le long circuit qu'il falloit faire pour atteindre le but, ne me rebutèrent point: je fis le calcul tout au long; et après un travail opiniâtre je trouvai la distance de l'observatoire de l'Isle de l'Inca au sommet d'Iliniça, la hauteur de cette montagne et celle de Pitchincha les mêmes, à 2 o 3 toises (!) près, que M. Bouguer” (La Condamine, Journal du voyage à l'équateur, pp. 94 and 111). It is hinted, in the Mesure des trois premiers degrés du méridien dans l'hémisphère austral (p. 52), that the height of Inca Island, above the level of the South Sea or the mouth of the Río de las Esmeraldas, has been determined only by estimating the incline and the distance, and that La Condamine and Bouguer differ in the respective estimates by 12 toises height. How much easier is direct geodetic measuring, using one or two elevation angles from point locations of a well-measured base line, whether pointed toward the summit or in the direction of known deviation! The barometric formulas that we now have adjust the base line to sea level with great reliability, in order to translate relative into absolute elevation. Bouguer himself seemed to sense the unreliability of his complicated elevation [169] calculations; when he complains about the effect of refraction on the many depression angles, he adds that elevation could not have been calculated with the same precision as the distance of the signals (Figure de la terre, pp. 119–22 and 167). Even if the two-month work in the forest plains of the Isla del Inca did not yield particularly reliable hypsometric results, Bouguer still has the great satisfaction that, after Pascal Mariotte and Halley, he was the first to develop a true and convenient barometric formula, even if it was imperfect. Many years had to pass until the barometric coefficient for the temperature of mercury and the air also included geographic latitude and the decrease of gravity in a formula like Laplace and Ramond's!

The doubts that I have seeded here about Bouguer and La Condamine's measurements of Chimborazo's elevation are solely the result of the observation of the entirety of the process, not of an overly great confidence in my own outcome. For half a century, I have expressed, in the liveliest of terms, the wish that Chimborazo be remeasured geodetically by an experienced observer with precise instruments and by using a carefully determined base line. As Oltmanns has already noted, my measurements would assume that a difference of 100 toises in the final result was due to an error of 10' 54" in the angle between the endpoints of the base line and 21' 58" in the elevation angle. If the reason were refraction, then the difference would instead have had to have been increased to -1.39 of the arc between one station and the summit, rather than being -0.042. Will anyone ever carry a barometer up to this summit, as their bold spirit of adventure has inspired physicists to do during the past decades in the cases of the Finsteraarhorn, Jungfrau, and Schreckhorn?

5. (P. 136) The highest elevation for a phanerogamic plant which Colonel Hall found on the slope of Chimborazo was for Saxifraga Boussingaulti, at 2466 toises (14,796 feet); but this was at a time when the eternal snow line had been lower (see my Asie centrale, vol. 3, p. 262). We collected the following plants between 14,000 and 15,000 feet: cryptogams: Stereoczulon botryosum (very different from S. paschale); Lecidea atrovirens; Gyrophora rugosa; Bryum argenteuml; Polytrichum juniperinum; Grimmia longirostris at an elevation of 2380 toises; Jungermannia setacea (Hooker); and Gynostomum jucaleum; of phanerogams: Gentiana rupicola and G. cernua; Culcitium rufescens; C. nivale (substituting for the thick wooly [170] espeletia of the Páramos and Cordilleras of New Granada); Lysopomia reniformis; Ranunculus Gusmanni; three Calceolaria (C. saxatilis, C. rosmarinifolia, and C. hysopifolia); the crucifers Draba Bonplandiana, Eudema nubigena, and Arabis andicola, which are very rare in the tropics; in lower areas only between 10,000 and 11,000 feet: Arenaria serpens; Andromachia nubigena (a new species closely related to Senecio); Dumerilia paniculata. Among the many above-mentioned compositae (family of the Synantheria), one stands out on Chimborazo: the beautiful Bacharis gnidiifolia, one of 54 new types of Bacharis that we found and described. See Synopsis plantarum quas (quas in itinere ad plagam aequinoctialem orbis novi) collegerunt Al. de Humboldt et Am. Bonpland, by C.S. Kunth (in octavo), 1823, vol. 2, pp. 376–88, and our Nova genera et species plantarum (folio), vol. 4, pp. 48–68. Immediately after our return from Mexico, Sir William Hooker described some of the mosses, to which we paid particular attention, in Muscis exoticis. Contrary to long-standing opinion, there were among them many true European species, for example, Brynum argenteum, Sphagnum acutifolium, Polytrichum juniperinum, Trichestomum polyphyllum, Neckera crispa, Funaria hygrometrica, etc. I also believe that I must repeat here a fact that is important for the geography of plants and the dissemination of forms, namely, that in the colder regions of the tropics, Musci frondosi occurs by no means just as an alpine plant. We found moss beds in some very shaded places in the hot region, a few hundred feet above sea level, beds of such a fresh, abundant growth that it was just like in my Nordic fatherland. “Est enimen incredibilis numerus muscorum, lichenum et fungorum, non solum in cacumine Andium, aëre frigido circumfuso, sed etiam in calidis et opacissimis sylvis, ubi, sub luco viridente, plantae agamiae irriguam obtegunt terram. Exempla praebent regiones ferventissimae ad ripam fluminis Magdalenae, Hondam inter et Aegyptiaeam, sylvae Orinocenses propter Esmeraldam et Manldaracam, littora maris Antillarum prope ostia fluminis Sinu, ubi fere totum per annum aëris temperies inter 23° et 25° Reaum. consistit” (Humboldt, De distributione geografica plantarum, p. 29). I have depicted Chimborazo's vegetation and that of the nearby snow-capped mountains in a large image (Atlas géographique et physique de la relation historique, Plate IX) that encompasses the climates one stacked above the other, from sea level [171] to 15,000 feet elevation, and shows around 400 plants in their own characteristic regions (hypsometric positions).

6. (P. 137) My barometric measurement showed 2890, Boussingault's 2870 meters. My friend determined the mean temperature of the high plains of Tapia according to the heat of the earth at 16.4 °C.

7. (P. 139) I recall the Mexica (Aztec) legend that is linked with the flattened pyramid of Cholula (Cholollan) slightly to the west of La Puebla de los Angeles. In the important manuscript by the Dominican monk Pedro de los Ríos, who copied hieroglyphic paintings in New Spain in 1566, I discovered the following passage in the Vatican Library (I translate from the Spanish text): “Before the great flood (apachihuiliztli), …, the land of Anahuac was inhabited by giants (Tzocuillixeque). All those who did not perish were turned into fish, with the exception of seven, who took refuge in caves. When the waters subsided (in the fourth age of the world), one of these giants, Xelhua, known as the architect, went to Cholollan, where, in (monumental) memory of Tlaloc Mountain, which had served as a shelter for him and six of his brothers, he built an artificial hill in the shape of a pyramid. He had bricks made in the province of Tlalmanalco, at the foot of the Sierra de Cocotl, and, in order to transport them to Cholula, he placed a line of men (many miles long) who passed them along by hand. The gods were incensed by the sight of this structure, whose top was meant to reach the clouds; enraged by Xelhua's audacity, they cast fire upon the pyramid (the god-dwelling, teocalli). Many workers perished, the work was discontinued, and it was afterwards consecrated to the god of the air, Quetzalcoatl.” At the time of Cortés's expedition, the Cholulans preserved a stone that had fallen, shrouded in a globe of fire, from the heavens onto the top of the pyramid; this aerolite resembled a toad. (See my Vues des Cordillères [octavo edition], vol. 1, p. 114, Plate III, and Essai politique sur la Nouvelle-Espagne, vol. 2 [second ed. from 1827], p. 151; also Prescott, Conquest of Mexico, vol. 3, p. 380.)

8. (P. 146) The sand flea, called la chique by the French colonists of the Antilles, burrows into the human skin and causes infections, because the egg sack of the fertilized female swells up considerably. What is physiologically remarkable is that only newly arrived whites and blacks are afflicted by this insect, from which I had so often suffered; Indians (indigenous Americans) and almost all American-born Spanish Creoles are not.

9. (P. 150) Mechanik der menschlichen Gehwerkzeuge, [172] 1836, para. 64, pp. 147–60. More recent experiments by the brothers Weber in Berlin have reliably confirmed that the leg is held in the pan of the pelvis by the pressure of the atmospheric air.

10. (P. 156) My own observations, partly geodetic, partly barometric measurements (the former are starred), show that the height of the lower edge of the eternal snow line in the cordilleras of Quito (between 0° and 1 1/2° southern latitude) is at 2472 toises, or 4816 meters. This number is the arithmetic average of measurements carried out between February and June 1802, but which include small fluctuations caused, in such proximity to the equator, by the season itself.

On Antisana* .............. 2493 toises

" Cotopaxi*............... 2490  "

" Chimborazo*.......... 2471  "

" Huahua-Pichincha.. 2460  "

" Rucu-Pichincha...... 2455  "

" Corazón.................. 2458 " 

Boussingault found in 1831:

On Antisana.......... 4871 meters = 2499 toises

" Chimborazo..... 4868 meters =2455 "

" Cotopaxi.......... 4804 meters = 2464 "

The mean is 2453 toises (4720 meters); the difference from my own results is only 19 toises. The small oscillation of the lower edge of the snow line, and the few changes in the temperature of layered air masses in the tropics, makes the snow line on the slopes of the cordilleras, seen at such great altitude, appear in perfect horizontality – a remarkable view for the European traveler. In the Swiss Alps, a number of irregularities in the surface of the ground (clefts and slight unevenness in the valley) disturb this impression of horizontality. In the temperate zone, especially in very northern latitudes, the line appears not cleanly cut but as if it were broken up, marred by the phenomenon of glaciers, which depends on the temperature. Wherever, in the tropics, several snow-capped mountains (Nevados) can be seen together in groupings, the horizontality gives the uneducated peasants among the indigenous peoples a very good idea about the relative height of neighboring peaks. They recognize as highest those mountains whose eternally snow-covered masses rise the most above the lower edge of the snow line. Long before any measurements were made in the cordilleras, the natives (los Indios [173] del país) knew that Capac-Urcu and Chimborazo were the tallest mountains in the region. Temporary snowfall, which produces the same regularity for miles downward and the impression of a similar horizontality, also leads to a correct judgment of elevations where mountains are lower than the normal height of the perpetual snow line (14,830 feet).

According to my research, the first snow-capped mountains seen on the New Continent were the Sierra de Citarma (now called Sierra de Santa Marta) east of Cartagena de Indias, at 11° northern latitude. The expedition of Colmenares from 1510 first spread this news in Spain, along with the idea of “how close to the equator colossal mountains had to be in order still to show eternal snow.” One already recognized then the rising of the snow line from the pole to the tropics. Actual measurements of the height of the snow line were carried out by Bouguer and La Condamine between 1736 and 1742; that is, earlier than such similarly precise measurements were made in the Alps and the Pyrenees. Bouguer, who had incomplete but quite correct notions about the origins of the mountain cold and the effect of elevation on decreases in temperature (Figure de la terre, pp. xlvi–lii), set out to “déterminer la hauteur de la surface courbe qui passe par le bas de la neige sur toutes les montagnes du Globe.” He reports for the equator 2434 toises, for 28 1/4° at most 1950 toises, below 43° latitude (in France and Chile) 1500 to 1600 toises. These figures are less imprecise for the northern hemisphere than one should have thought. On the marble tablet that is preserved in the university building in Quito, and which I found there fully intact, one also reads: “Altitudo acutioris ac lapidei cacuminis nive plerumque operli 2432 hexap. Paris., ut et nivis infernae permanentis in montibus nivosis.” If, because of the error, one adds 32 toises to the elevation of the city of Quito, one gets 2462 toises and, through a great number of accidental combinations up to +9 toises (54 feet), the height that both Boussingault and I found (see my Asie centrale, vol. 3, pp. 251–6).

11. (P. 156) Arago in Annuaire du Bureau des Longitudes pour 1830, p. 331, and Asie centrale, vol. 3, pp. 273–81.

12. (P. 156) About the difference in the height of the snow line on the northern and southern slopes of the Himalayas, which I have proven since 1820, see Ansichten der Natur, 1840 ed., vol. 1, p. 126; Asie centrale, vol. 3, pp. 293–326; Joseph Hooker, On the Elevation of the Great Table Land of Thibet, 1850, p. 6 [174]; [R.A.] Strachan, “On the Snow-Line in the Himalaya,” Journal of the Asiatic Society of Bengal (April 1849, no. 29). More recent observations have irrefutably confirmed the difference between the Indian and the Tibetan slopes; but the extent of the difference in these latitudes from 30° to 31° in the temperate zone appears not to be the same for different seasons. It is difficult to distinguish the edges of sporadic snowfall from the eternal snow line; and such sporadic snowfall does not characteristically occur at the same time on the southern and the northern slopes. My earliest data were 12,180 pieds de Paris for the south and 15,600 for the north; the difference is 3420 feet; Hodgson and Joseph Hooker: 14,073 in the south, 18,764 in the north, a difference of 4691 feet; Strachey: 14,543 in the south, 17,358 in the north, a difference of 2815 feet. My own result lies between the latter two. In a letter that my friend Dr Joseph Hooker wrote to me, this time no longer from Darjeeling but from Tangu, one could read: south 14,073 feet, north 15,006 feet; difference 933 feet. The small elevation of the southern, Indian, slope which is mentioned here likely points either to the great effect of sporadic snowfall or to the local situation of the pass itself through which the journey led.

13. (P. 159) On the trigonometric measurement of Chimborazo, see above, pp. 166–9.

14. (P. 160) On Capac-Urcu and the legend of its collapse, see my Géographie des plantes, p. 119, and the essay in this volume [Kleinere Schriften] that follows “Boussingault's Besteigung des Chimborazo.”

15. (P. 160) The following analysis of summit rock from Chimborazo, which I had hacked off at an elevation of 2530 toises (15,180 feet), was sent to me by a superb geognost, Mr Hermann Abich, to whom we owe a thorough knowledge of the Caucasus:

Silica........................ 3136 grams =  65.09%

Clay......................... 0.770 grams =  15.98%

Iron oxide................. 0.278 grams =  5.77%

Limestone................ 0.126 grams =  2.61%

Talc.......................... 0.198 grams =  4.10%

Potassium................. 0.096 grams =  1.99%

Sodium..................... 0.215 grams = 4.46%

Volatile substances and chloride.............. 0.019 grams =  0.41%

     4.8381 grams =  100%