from the observations which were carried on for many years by Arago in the garden of the Paris Observatory, that very small differences of temperature were perceptible 30 feet below the surface. Bravais calculated one degree for about every 50 feet on the high northern latitude of Bossekop, in Finmark (69° 58′ N.L.). The difference between the highest and lowest annual temperature diminishes in proportion with the depth, and according to Fourrier this difference diminishes in a geometrical proportion as the depth increases in an arithmetical ratio.

The stratum of invariable temperature depends, in respect to its depth, conjointly upon the latitude of the place, the conductive power of the surrounding strata and the amount of difference of temperature between the hottest and the coldest seasons of the year. In the latitude of Paris (48° 50') the depth and temperature of the Caves de l'Observatoire (86 feet and 53°.30 F.) are usually regarded as affording the amount of depth and temperature of the invariable stratum. Since Cassini and Legentil in 1783 placed a very correct mercurial thermometer in these subterranean caves, which are portions of old stone quarries, the mercury in the tube has risen about 0°.4.39 Whether the cause of this rising is to be ascribed to an accidental alteration in the thermometrical scale which, however, was adjusted by Arago in 1817 with his usual care, or whether it indicates an actual increase of heat is still undecided. The mean temperature of the air at Paris is 510.478 F. Bravais is of opinion that the thermometer in the Caves de l'Observatoire stands below the limit of invariable temperature, although Cassini believes that he has found a difference of 25ths of a degree (Fahr.) between the winter and summer temperature, the higher tempe

in the Trans. of the Royal Soc. of Edinburgh, vol. xvi, 1849, pt. ii, p. 189.

39 All numbers referring to the temperature of the Caves de l'Observatoire have been taken from the work of Poisson, Théorie Mathématique de la Chaleur, pp. 415 and 462. The Annuaire Météorologique de la France, edited by Martins and Haeghens, 1849, p. 88, contains corrections by Gay-Lussac for Lavoisier's subterranean thermometer. The mean of three readings, from June till August, was 53°.95 F. for this thermometer, at a time when Gay-Lussac found the temperature to be 53°.32, which was therefore a difference of 0.63.

rature being found to prevail in the winter.40 If we now take the mean of many observations of the temperature of the soil between the parallels of Zurich (47° 22′) and Upsala (59° 51'), we obtain an increase of 1° F. for every 40 feet. Differences of latitude cannot produce a difference of more than 12 or 15 feet, which is not marked by any regular alteration from south to north, because the influence which the latitude undoubtedly exerts, is masked within these narrow limits by the influence of the conductive power of the soil, and by errors of observation.

As the terrestrial stratum in which we first cease to observe any alteration of temperature through the whole year lies, according to the theory of the distribution of heat, so much the nearer the surface, as the maxima and minima of the mean annual temperature approximate to one another, a consideration of this subject has led my friend Boussingault to the ingenious and convenient method of determining the mean temperature of a place within the tropical regions (especially between 10 degrees north and south of the equator) by observing a thermometer which has been baried 8 or 12 inches below the surface of the soil in some well protected spot. At different hours and different months of the year, as in the experiments of Captain Hall near the coast of the Choco in Tumaco, those at Salaza in Quito, and those of Boussingault in la Vega de Zupia, Marmato, and Anserma Nuevo in the Cauca valley, the temperature scarcely varied one-tenth of a degree; and almost within the same limits it was identical with the mean temperature of the air at those places in which it had been determined by horary observations. It was, moreover, very remarkable that this identity remained perfectly uniform, whether the thermometric soundings (of less than one foot in depth) were made on the torrid shores of Guayaquil and Payta, on the Pacific, or in an Indian village on the side of the volcano of Purace, which I found from my barometrical measurements to be situated at an elevation of 1356 toises, or 8671 feet above the sea. The mean temperatures differed by fully 25° F. at these different stations.41

40 Cassini, in the Mém. de l'Acad. des Sciences, 1786, p. 511.

41 Boussingault, sur la profondeur à laquelle on trouve dans la zone torride la couche de température invariable, in the Annales de Chimie et

I believe that special attention is due to two observations which I made on the mountains of Peru and Mexico, in mines which lie at a greater elevation than the summit of the Peak of Teneriffe, and are therefore the highest in which a thermometer has ever been placed. At a height of between 12,000 and 13,000 feet above the level of the sea I found the subterranean air 25° F. warmer than the external atmosphere. Thus, for instance, the little Peruvian town of Micuipampa lies, according to my astronomical and hypsometrical observations, in the latitude 6° 43′ S., and at an elevation of 1857 toises or 11,990 feet, at the base of Cerro de Gualgayoc, celebrated for the richness of its silver mines. The summit of this almost isolated fortress-like and picturesquely situated mountain rises 240 toises or 1504 feet higher than the streets of Micuipampa; the external air at a distance from the mouth of the pit of

de Physique, t. liii, 1833, pp. 225-247. Objections have been advanced by John Caldecott, the astronomer to the Rajah of Travancore, and by Captain Newbold, in India, against the method recommended in this memoir, although it has been employed in South America in many very accurate experiments. Caldecott found at Trevandrum (Edin. Transact. vol. xvi, part iii, pp. 379-393), that the temperature of the soil at a depth of three feet and more below the surface, (and therefore deeper than Boussingault's calculation,) was 85° and 86° F., while the mean temperature of the air was 80°.02. Newbold's experiments (Philos. Transact. for the year 1845, pt. i, p. 133), which were made at Bellary, lat. 15° 5', showed an increase of temperature of 4° F. between sunrise and 2 p.m. for one foot of depth, but at Cassargode, lat. 12° 29′, there was only an increase of 1°.30 F., under a cloudy sky. Is it quite certain that the thermometer in this case was sufficiently covered to protect it from the influence of the sun's rays? Compare also Forbes, Exper. on the Temp. of the Earth at different depths, in the Edin. Transart. vol. xvi, part ii, p. 189. Colonel A. Costa, the admirable historian of New Granada, has made a prolonged series of observations, which fully confirm Boussingault's statement, and which were completed, about a year ago, at Guadua, on the south-western side of the elevated plateau of Bogota, where the mean annual temperature is 43°.94 F. at the depth of one foot, and at a carefully protected spot. Boussingault thus refers to these experiments:-"The observations of Colonel A. Costa, whose extreme precision in everything which is connected with meteorology is well known to you, prove that when fully sheltered from all disturbing influences, the temperature within the tropics remains constant at a very small depth below the surface."

42 In reference to Gualgayoc (or Minas de Chota) and Micuipampa, see Humboldt, Recueil d'Observ. Astron. vol. i, p. 324.

43

the Mina del Purgatorio was 42°.26 F., but in the interior of the mine, which lies more than 2057 toises, or 13,154 feet above the sea, I saw that the thermometer everywhere indicated a temperature of 67°.64 F., there being thus a difference of 25°.38 F. The limestone rock was here perfectly dry, and very few men were working in the mine. In the Mina de Guadalupe, which lies at the same elevation, I found that the temperature of the internal air was 57°.9 F., showing therefore a difference of 15°.64 F. when compared with the external air. The water which flowed out from the very damp mine stood at 52°.34 F. The mean annual temperature of Micuipampa is probably not more than 45°.8 F. In Mexico, in the rich silver mines of Guanaxuato, I found in the Mina de Valenciana the external temperature in the neighbourhood of the Tiro Nuevo (which is 7590 feet above the sea) 70°.16 F., and the air in the deepest mines, for instance in the Planes de San Bernardo, 1630 feet below the opening of the shaft of Tiro Nuevo, fully 80°.6 F., which is about the mean temperature of the littoral region of the Gulf of Mexico. At a point 147 feet higher than the mouth of the Planes de San Bernardo, a spring of water issues from the transverse rock, in which the temperature is 84°.74 F. I determined the latitude of the mountain town of Guanaxuato to be 21° 0′ N., with a mean annual temperature varying between 60°.44 and 61°.26 F. The present is not a fitting place in which to advance conjectures, which it might be difficult to establish in relation to the causes of probably an entirely local rise of the subterranean temperature at mountain elevations, varying from 6000 to more than 12,000 feet.

A remarkable contrast is exhibited in the steppes of Northern Asia, by the conditions of the frozen soil, whose very existence was doubted, notwithstanding the early testimony of Gmelin and Pallas. It is only in recent times that correct views in relation to the distribution and thickness of the stratum of subterranean ice have been established by means of the admirable investigations of Erman, Baer, and Middendorff. In accordance with the descriptions given of Greenland by Cranz, of Spitzbergen by Martens and Phipps, 43 Essai Polit. sur le Roy. de la Nouv. Espagne (2ème ed., t. iii,

p. 201).

and of the coasts of the sea of Kara by Sujew, the whole of the most northern part of Siberia was described by too hasty a generalization as entirely devoid of vegetation, always frozen on the surface, and covered with perpetual snow, even in the plains. The extreme limit of vegetation in Northern Asia is not, as was long assumed, in the parallel of 67°, although sea -winds and the neighbourhood of the Bay of Obi make this estimate true for Obdorsk; for in the valley of the great River Lena, high trees grow as far north as the latitude of 71°. Even in the desolate islands of New Siberia, large herds of reindeer and countless lemmings find an adequate nourishment." Middendorff's two Siberian expeditions, which are distinguished by a spirit of keen observation, adventurous daring, and the greatest perseverance in a laborious undertaking, were extended from the year 1843 to 1846 as far north as the Taymir land in 75° 45′ lat., and south-east as far as the Upper Amoor and the Sea of Ochotsk. The former of these perilous undertakings led the learned investigator into a hitherto unvisited region, whose exploration was the more important in consequence of its being situated at equal distances from the eastern and western coasts of the old Continent. In addition to the distribution of organisms in high northern latitudes, as depending mainly upon climatic relations, it was directed by the St. Petersburgh Academy of Sciences that the accurate determination of the temperature of the ground and of the thickness of the subterranean frozen soil should be made the principal objects of the expedition. Observations were made in borings and mines at a depth of from 20 to 60 feet at more than twelve points (near Turuchansk, on the Jenisei, and on the Lena) at relative distances of from 1600 to 2000 geographical miles.

The most important seat of these geothermic observations was however Schergin's shaft at Jakutsk 62° 2′ N. lat.45 4 E. von Baer, in Middendorff's Reise in Sib., Bd. i, s. vii.

45 The merchant Fedor Schergin, cashier to the Russian-American Trading Company, began, in the year 1828, to dig a well in the courtyard of a house belonging to the company. As he had only found frozen earth and no water at the depth of 90 feet, which he reached in 1830, he determined to give up the attempt, until Admiral Wrangel, who passed through Jakutsk on his way to Sitcha, in Russian America, and who saw how interesting it would be, in a scientific point of view, to penetrate through this subterranean stratum of ice, induced Schergin