The law of the decrease of heat with the increase of elevation at different latitudes is one of the most important subjects involved in the study of meteorological processes, of the geography of plants, of the theory of terrestrial refraction, and of the various hypotheses that relate to the determination of the height of the atmosphere. In the many mountain journeys which I have undertaken, both within and without the tropics, the investigation of this law has always formed a special object of my researches.*
[footnote] *Humboldt, 'Recueil d'Observations Astronomiques', t. i., p. 126-140; 'Relation Historique', t. i., p. 119, 141, 227; Biot, in 'Connaissance des Temps pour l'an' 1841, p. 90-109.
Since we have acquired a more accurate knowledge of the true relations of the distribution of heat on the surface of the earth, that is to say, of the inflections of isothermal and isotheral lines, and their unequal distance apart in the different eastern and western systems of temperature in Asia, Central Europe, and North America, we can no longer ask the general question, what fraction of the mean annual or summer temperature corresponds to the difference of one degree of geographical latitude, taken in the same meridian? In each system of 'isothermal' lines of equal curvature there reigns a p 328 close and necessary connection between three elements, namely, the decrease of heat in a vertical direction from below upward, the difference of temperature for every one degree of geographical latitude, and the uniformity in the mean temperature of a mountain station, and the latitude of a point situated at the level of the sea.
In the system of Eastern America, the mean annual temperature from the coast of Labrador to Boston changes 1.6ºdegrees for every degree of latitude; from Boston to Charleston about 1.7 degrees; from Charleston to the tropic of Cancer, in Cuba, the variation is less rapid, being only 1.2 degrees. In the tropics this diminution is so much greater, that from the Havana to Cumana the variation is less than 0.4 degrees for every degree of latitude.
The case is quite different in the isothermal system of Central Europe. Between the parallels of 38 degrees and 71 degrees I found that the decrease of temperature was very regularly 0.9degrees for every degree of latitude. But as, on the other hand, in Central Europe the decrease of heat is 1.8 degrees for about every 534 feet of vertical elevation, it follows that a difference of elevation of about 267 feet corresponds to the difference of one degree of latitude. The same mean annual temperature as that occurring at the Convent of St. Bernard, at an elevation of 8173 feet, in lat. 45 degrees 50' should therefore be met with at the level of the sea in lat. 75 degrees 50'.
In that part of the Cordilleras which falls within the tropics, the observations I made at various heights, at an elevation of upward of 19,000 feet, gave a decrease of 1 degree for every 341 feet; and my friend Boussingault found, thirty years afterward, as a mean result, 319 feet. By a comparison of places in the Cordilleras, lying at an equal elevation above the level of the sea, either on the declivities of the mountains or even on extensive elevated plateaux, I observed that in the latter there was an increase in the annual temperature varying from 2.7 degrees to 4.1 degrees. This difference would be still greater if it were not for the cooling effect of nocturnal radiation. As the different climates are arranged in successive strata, the one above the other, from the cacao woods of the valleys to the region of perpetual snow, and as the temperature in the tropics varies but little throughout the year, we may form to ourselves a tolerably correct representation of the climatic relations to which the inhabitants of the large cities in the Andes are subjected, by comparing these climates with the temperatures of particular months in the plains of France and Italy. While p 329 the heat which prevails daily on the woody shores of the Orinoco exceeds by 7.2 degrees that of the month of August at Palermo, we find, on ascending the chain of the Andes, at Popayan, at an elevation of 3826 feet, the temperature of the three summer months of Marseilles; at Quito, at an elevation of 9541 feet, that of the close of May at Paris; and on the Paramos, at a height of 11,510 feet, where only stunted Alpine shrubs grow, though flowers still bloom in abundance, that of the beginning of April at Paris. The intelligent observer, Peter Martyr de Aughiera, one of the friends of Christopher Columbus, seems to have been the first who recognized (in the expedition undertaken by Rodrigo Enrique Colmenares, in October, 1510) that the limit of perpetual snow continues to ascend as we approach the equator. We read, in the fine work 'De Rebus Oceanicis',* "the River Gaira comes from a mountain in the Sierra Nevada de Santa Maria, which, according to the testimony of the companions of Colmenares, is higher than any other mountain hitherto discovered.
[footnote] *Anglerius, 'De Rebus Oceanicis', Dec. xi., lib. ii., p. 140 (ed. Col., 1574). In the Sierra de Santa Marta, the highest point of which appears to exceed 19,000 feet (see my 'Relat. Hist.', t. ii., p. 214), there is a peak that is still called Pico de Gaira.
It must undoubtedly be so if 'it retain snow perpetually' in a zone which is not more than 10 degrees from the equinoctial line." The lower limit of perpetual snow, in a given latitude, is the lowest line at which snow continues during summer, or, in other words, it is the maximum of height to which the snow-line recedes in the course of the year. But this elevation must be distinguished from three other phenomena, namely, the annual fluctuation of the snow-line, the occurrence of sporadic falls of snow, and the existence of glaciers, which appear to be peculiar to the temperate and cold zones. This last phenomenon, since Saussure's immortal work on the Alps, has received much light, in recent times, from the labors of Venetz, Charpentier, and the intrepid and persevering observer Agassiz.
We know only the 'lower', and not the 'upper' limit of perpetual snow; for the mountains of the earth do not attain to those ethereal regions of the rarefied and dry strata of air, in which we may suppose, with Bouguer, that the vesicles of aqueous vapor are converted into crystals of ice, and thus rendered perceptible to our organs of sight. The lower limit of snow is not, however, a mere function of geographical latitude or of mean annual temperature; nor is it at the equator, or p 330 even, in the region of the tropics, that this limit attains its greatest elevation above the level of the sea. The phenomenon of which we are treating is extremely complicated, depending on the general relations of temperature and humidity, and on the form of the mountains. On submitting these relations to the test of special analysis, as we may be permitted to do from the number of determinations that have recently been made,* we shall find that the controlling causes are the differences in the temperature of different seasons of the year; the direction of the prevailing winds and their relations to this land and sea; the degree of dryness or humitidy in the upper strata of the air; the absolute thickness of the accumulated masses of fallen snow; the relation of the s-line to the total height of the mountain; the relative position of the latter in the chain to which it belongs, and the steepness of its declivity; the vicinity of either summits likewise perpetually covered with show; the expansion, position, and elevation of the plains from which the snow mountain rises as an isolated peak or as a portion of a chain; whether this plain be part of the sea-coast, or of the interior of a continent; whether it be covered with wood or waving grass; and whether, finally, it consist of a dry and rocky soil, or of a wet and marshy bottom.
[footnote] *See my table of the height of the line of perpetual snow, in both hemispheres, from 71 degrees 15' north lat. to 53 degrees 54' south lat., in my 'Asie Centrale', t. iii., p. 360.