inferior limit of the perpetual snows.

It was near the Ice-Cavern (Cueva del Hielo), that, in the voyage of Laperouse, Messrs. Lamanon and Monges made their experiments on the temperature of boiling water. These naturalists found it 88.7 degrees, the barometer at nineteen inches one line. In the kingdom of New Grenada, at the chapel of Guadaloupe, near Santa-Fe de Bogota, I have seen water boil at 89.9 degrees, under a pressure of 19 inches 1.9 lines, At Tambores, in the province of Popayan, Senor Caldas found the heat of boiling water 89.5 degrees, the barometer being at 18 inches 11.6 lines. These results might lead us to suspect, that, in the experiment of M. Lamanon, the water had not reached the maximum of its temperature.

Day was beginning to dawn when we left the ice-cavern. We observed, during the twilight, a phenomenon which is not unusual on high mountains, but which the position of the volcano we were scaling rendered very striking. A layer of white and fleecy clouds concealed from us the sight of the ocean, and the lower region of the island. This layer did not appear above 800 toises high; the clouds were so uniformly spread, and kept so perfect a level, that they wore the appearance of a vast plain covered with snow. The colossal pyramid of the peak, the volcanic summits of Lancerota, of Forteventura, and the isle of Palma, were like rocks amidst this vast sea of vapours, and their black tints were in fine contrast with the whiteness of the clouds.

While we were climbing over the broken lavas of the Malpays, we perceived a very curious optical phenomenon, which lasted eight minutes. We thought we saw on the east side small rockets thrown into the air. Luminous points, about seven or eight degrees above the horizon, appeared first to move in a vertical direction; but their motion was gradually changed into a horizontal oscillation. Our fellow-travellers, our guides even, were astonished at this phenomenon, without our having made any remark on it to them. We thought, at first sight, that these luminous points, which floated in the air, indicated some new eruption of the great volcano of Lancerota; for we recollected that Bouguer and La Condamine, in scaling the volcano of Pichincha, were witnesses of the eruption of Cotopaxi. But the illusion soon ceased, and we found that the luminous points were the images of several stars magnified by the vapours. These images remained motionless at intervals, they then seemed to rise perpendicularly, descended sideways, and returned to the point whence they had departed. This motion lasted one or two seconds. Though we had no exact means of measuring the extent of the lateral shifting, we did not the less distinctly observe the path of the luminous point. It did not appear double from an effect of mirage, and left no trace of light behind. Bringing, with the telescope of a small sextant by Troughton, the stars into contact with the lofty summit of a mountain in Lancerota, I observed that the oscillation was constantly directed towards the same point, that is to say, towards that part of the horizon where the disk of the sun was to appear; and that, making allowance for the motion of the star in its declination, the image returned always to the same place. These appearances of lateral refraction ceased long before daylight rendered the stars quite invisible. I have faithfully related what we saw during the twilight, without undertaking to explain this extraordinary phenomenon, of which I published an account in Baron Zach's Astronomical Journal, twelve years ago. The motion of the vesicular vapours, caused by the rising of the sun; the mingling of several layers of air, the temperature and density of which were very different, no doubt contributed to produce an apparent movement of the stars in the horizontal direction. We see something similar in the strong undulations of the solar disk, when it cuts the horizon; but these undulations seldom exceed twenty seconds, while the lateral motion of the stars, observed at the peak, at more than 1800 toises, was easily distinguished by the naked eye, and seemed to exceed all that we have thought it possible to consider hitherto as the effect of the refraction of the light of the stars. On the top of the Andes, at Antisana, I observed the sun-rise, and passed the whole night at the height of 2100 toises, without noting any appearance resembling this phenomenon.

I was anxious to make an exact observation of the instant of sun-rising at an elevation so considerable as that we had reached on the peak of Teneriffe. No traveller, furnished with instruments, had as yet taken such an observation. I had a telescope and a chronometer, which I knew to be exceedingly correct. In the part where the sun was to appear the horizon was free from vapour. We perceived the upper limb at 4 hours 48 minutes 55 seconds apparent time, and what is very remarkable, the first luminous point of the disk appeared immediately in contact with the limit of the horizon, consequently we saw the true horizon; that is to say, a part of the sea farther distant than 43 leagues. It is proved by calculation that, under the same parallel in the plain, the rising would have begun at 5 hours 1 minute 50.4 seconds, or 11 minutes 51.3 seconds later than at the height of the peak. The difference observed was 12 minutes 55 seconds, which arose no doubt from the uncertainty of the refraction for a zenith distance, of which observations are wanting.

We were surprised at the extreme slowness with which the lower limb of the sun seemed to detach itself from the horizon. This limb was not visible till 4 hours 56 minutes 56 seconds. The disc of the sun, much flattened, was well defined; during the ascent there was neither double image nor lengthening of the lower limb. The duration of the sun's rising being triple that which we might have expected in this latitude, we must suppose that a fog-bank, very uniformly extended, concealed the true horizon, and followed the sun in its ascent. Notwithstanding the libration of the stars,* which we had observed towards the east, we could not attribute the slowness of the rising to an extraordinary refraction of the rays occasioned by the horizon of the sea; for it is precisely at the rising of the sun, as Le Gentil daily observed at Pondicherry, and as I have several times remarked at Cumana, that the horizon sinks, on account of the elevation of temperature in the stratum of the air which lies immediately over the surface of the ocean. (* A celebrated astronomer, Baron Zach, has compared this phenomenon of an apparent libration of the stars to that described in the Georgics (lib. 50 v. 365). But this passage relates only to the falling stars, which the ancients, (like the mariners of modern times) considered as a prognostic of wind.)

The road, which we were obliged to clear for ourselves across the Malpays, was extremely fatiguing. The ascent is steep, and the blocks of lava rolled from beneath our feet. I can compare this part of the road only to the Moraine of the Alps or that mass of pebbly stones which we find at the lower extremity of the glaciers. At the peak the lava, broken into sharp pieces, leaves hollows, in which we risked falling up to our waists. Unfortunately the listlessness of our guides contributed to increase the difficulty of this ascent. Unlike the guides of the valley of Chamouni, or the nimble-footed Guanches, who could, it is asserted, seize the rabbit or wild goat in its course, our Canarian guides were models of the phlegmatic. They had wished to persuade us on the preceding evening not to go beyond the station of the rocks. Every ten minutes they sat down to rest themselves, and when unobserved they threw away the specimens of obsidian and pumice-stone, which we had carefully collected. We discovered at length that none of them had ever visited the summit of the volcano.

After three hours' walking, we reached, at the extremity of the Malpays, a small plain, called La Rambleta, from the centre of which the Piton, or Sugar-loaf, takes its rise. On the side toward Orotava the mountain resembles those pyramids with steps that are seen at Fayoum and in Mexico; for the elevated plains of Retama and Rambleta form two tiers, the first of which is four times higher than the second. If we suppose the total height of the Peak to be 1904 toises, the Rambleta is 1820 toises above the level of the sea. Here are found those spiracles, which are called by the natives the Nostrils of the Peak (Narices del Pico). Watery and heated vapours issue at intervals from several crevices in the ground, and the thermometer rose to 43.2 degrees. M. Labillardiere had found the temperature of these vapours, eight years before us, 53.7 degrees; a difference which does not perhaps prove so much a diminution of activity in the volcano, as a local change in the heating of its internal surface. The vapours have no smell, and seem to be pure water. A short time before the great eruption of Mount Vesuvius, in 1805, M. Gay-Lussac and myself had observed that water, under the form of vapour, in the interior of the crater, did not redden paper which had been dipped in syrup of violets. I cannot, however, admit the bold hypothesis, according to which the Nostrils of the Peak are to be considered as the vents of an immense apparatus of distillation, the lower part of which is situated below the level of the sea. Since the time when volcanoes have been carefully studied, and the love of the marvellous has been less apparent in works on geology, well founded doubts have been raised respecting these direct and constant communications between the waters of the sea and the focus of the volcanic fire.* (* This question has been examined with much sagacity by M. Brieslak, in his "Introduzzione alla Geologia," tome 2 pages 302, 323, 347. Cotopaxi and Popocatepetl, which I saw ejecting smoke and ashes, in 1804, are farther from both the Pacific and the Gulf of the Antilles, than Grenoble is from the Mediterranean, and Orleans from the Atlantic. We must not consider the fact as merely accidental, that we have not yet discovered an active volcano more than 40 leagues distant from the ocean; but I consider the hypothesis, that the waters of the sea are absorbed, distilled, and decomposed by volcanoes, as very doubtful.) We may find a very simple explanation of a phenomenon, that has in it nothing very surprising. The peak is covered with snow during part of the year; we ourselves found it still so in the plain of Rambleta. Messrs. O'Donnel and Armstrong discovered in 1806 a very abundant spring in the Malpays, a hundred toises above the cavern of ice, which is perhaps fed partly by this snow. Everything consequently leads us to presume that the peak of Teneriffe, like the volcanoes of the Andes, and those of the island of Manilla, contains within itself great cavities, which are filled with atmospherical water, owing merely to filtration. The aqueous vapours exhaled by the Narices and crevices of the crater, are only those same waters heated by the interior surfaces down which they flow.

We had yet to scale the steepest part of the mountain, the Piton, which forms the summit. The slope of this small cone, covered with volcanic ashes, and fragments of pumice-stone, is so steep, that it would have been almost impossible to reach the top, had we not ascended by an old current of lava, the debris of which have resisted the ravages of time. These debris form a wall of scorious rock, which stretches into the midst of the loose ashes. We ascended the Piton by grasping these half-decomposed scoriae, which often broke in our hands. We employed nearly half an hour to scale a hill, the perpendicular height of which is scarcely ninety toises. Vesuvius, three times lower than the peak of Teneriffe, is terminated by a cone of ashes almost three times higher, but with a more accessible and easy slope. Of all the volcanoes which I have visited, that of Jorullo, in Mexico, is the only one that is more difficult to climb than the Peak, because the whole mountain is covered with loose ashes.