raised by the latent heat thus evolved, till, at last, the lower part of the cavity is filled with boiling water and the upper with steam under high pressure. The expansive force of the steam becomes, at length, so great, that the water is forced up the fissure or pipe E B, and runs over the rim of the basin. When the pressure is thus diminished, the steam in the upper part of the cavity A expands, until all the water D is driven into the pipe: and when this happens, the steam, being the lighter of the two fluids, rushes up Mamusers as No. 42.

Supposed reservoir and pipe of a Geyser in Iceland.*

through the water with great velocity. If the pipe be choked up artificially, even for a few minutes, a great increase of heat must take place, for it is prevented * From Sir George Mackenzie's Iceland.

from escaping in a latent form in steam; so that the water is made to boil more violently, and this brings on an eruption.

If we suppose that large subterranean cavities exist at the depth of some miles below the surface of the earth, in which melted lava accumulates, and that water penetrates into these, the steam thus generated may press upon lava and force it up the duct of a volcano, in the same manner as a column of water is driven up the pipe of a Geyser.

Agency of water in volcanos.—No theory seems at first more improbable, than that which represents water as affording an inexhaustible supply of fuel to the volcanic fires; yet, if subterraneous heat be derived from chemical action, as before hinted, and if electric currents in the crust of the earth may exert a slow decomposing power, the hypothesis is far from visionary.

It is a fact that must never be overlooked, when we are speculating on the probable causes of volcanos, that, while a great number are entirely submarine, the remainder are for the most part in islands or maritime tracts. There are a few exceptions, but some of these, as Dr. Daubeny observes, are near inland salt lakes, as in Central Tartary; while others form part of a train of volcanos the extremities of which are near

the sea. Thus Jorullo, in Mexico, though itself not less than forty leagues from the nearest ocean, appears to be connected with the volcano of Tuxtla on the one hand, and that of Colima on the other; the first bordering on the Atlantic, the latter on the Pacific Ocean. This communication is rendered more probable by the parallelism that exists between these and several intermediate volcanic hills.*

Sir H. Davy supposes that, when the sea is distant, as in the case of some of the South American volcanos, they may still be supplied with water from subterranean lakes; since, according to Humboldt, large quantities of fish are often thrown out during eruptions.*

It has been already stated, that the gases exhaled from volcanos, together with steam, are such as would result from the decomposition of salt water, and the fumes which escape from the Vesuvian lava have been observed to deposit common salt. † M. Boussingault, who has lately examined the elastic fluids evolved from the volcanos of equatorial America, informs us, that the same are given out by all the different vents, namely, aqueous vapour, in very large quantity, carbonic acid gas, sulphurous acid gas, and sometimes fumes of sulphur. The same naturalist found by analysis ‡, that all the thermal waters of the Cordilleras were charged with sulphuretted hydrogen gas.

M. Gay Lussac, while he avows his opinion that the decomposition of water contributes largely to volcanic action, calls attention, nevertheless, to the fact, that hydrogen has not been detected in a separate form among the gaseous products of volcanos; nor can it, he says, be present, for, in that case, it would be inflamed in the air by the red-hot stones thrown out during an eruption. Dr. Davy, also, in his account of Graham island, says, "I watched when the lightning was most vivid, and the eruption of the greatest degree of violence, to see if there was any inflammation oc* Phil. Trans., 1828. p. 250.

Davy, Phil. Trans., 1828. p. 244.

Ann. de Chim. et de Phys., tom. lii. p. 181.

casioned by this natural electric spark-any indication of the presence of inflammable gas; but in vain.”*

May not the hydrogen, Gay Lussac inquires, be combined with chlorine, and produce muriatic acid; for this gas has been observed to be evolved from Vesuvius-and the chlorine may have been derived from sea salt; which was, in fact, extracted by simple washing from the Vesuvian lava of 1822, in the proportion of nine per cent.+ But it was answered, that Sir H. Davy's experiments had shown, that hydrogen is not combustible when mixed with muriatic acid gas; so that if the latter was really evolved in large quantities, the hydrogen might be present without inflammation.

M. Gay Lussac, in the memoir just alluded to, expresses doubt as to the presence of sulphurous acid; but the abundant disengagement of this gas during eruptions is now ascertained: and thus all difficulty in regard to the absence of hydrogen in an inflammable state is removed. For, as Dr. Daubeny supposes, the hydrogen of decomposed water may unite with sulphur to form sulphuretted hydrogen gas, and this gas will then be mingled with the sulphurous acid, as it rises to the crater. It is shown, by experiment, that these gases mutually decompose each other when mixed, where steam is present: part of the hydrogen of the one immediately uniting with the oxygen of the other, to form water, while the excess of sulphurous acid alone escapes into the atmosphere. Sulphur is at the same time precipitated: but this would immediately be volatilized by the volcanic heat.

* Phil. Trans., 1832. p. 240.

Ann. de Chim. et de Phys., tom. xxii.

Quart. Journ. of Science, 1823. p. 132. note by editor.

It may, however, be asked, whether the flame of hydrogen would be visible during an eruption; as that gas, when inflamed in a pure state, burns with a very faint blue flame, which even in the night could hardly be perceptible by the side of red-hot and incandescent cinders. Its immediate conversion into water when inflamed in the atmosphere, might perhaps account for its not having been detected in a separate form.

When treating of springs and overflowing wells, I have stated that porous rocks are percolated by fresh water to great depths, and that sea-water probably penetrates in the same manner through the rocks which form the bed of the ocean. But, besides this universal circulation in regions not far from the surface, it must be supposed that, wherever earthquakes prevail, much larger bodies of water will be forced by the pressure of the ocean into fissures at greater depths, or swallowed up in chasms; in the same manner as, on the land, towns, houses, cattle, and trees are sometimes engulphed. It will be remembered, that these chasms often close again after houses have fallen into them; and, for the same reason, when water has penetrated to a mass of melted lava, the steam into which it is converted may often rush out at a different aperture from that by which the water entered. The frequent explosions caused by the generation of steam in the neighbourhood of the sea or of deep lakes, may shatter the solid crust of the earth, and allow the free escape of gases and lava, which, but for this cause, might never have reached the surface, and might only have given rise to earthquakes.

Importance of attending to the unseen volcanic phenomena.—In concluding these remarks on the causes of volcanos and earthquakes, I may observe, that spe