water into its two elements. At a later stage, the gases varied incessantly in composition, and the combustible elements at last entirely disappeared. The emitted gases consisted chiefly of free hydrogen, marsh-gas, hydrochloric, sulphuric, and carbonic acids, sulphuretted hydrogen, and nitrogen. Water played a prominent part in the different phases of the eruption. In the state of vapor it was present in all the emissions of volatile matters whatever was their temperature, and it might be regarded as the immediate cause of the explosions. No dry smoke-vent was observed. Coming in its liquid state from hot springs, the temperature and flow of the water varied with the state of the sea.
Chloride of iron, expelled as a vapor, was found associated with hydrochloric acid; but the hydrochlorate of ammonia, commonly abundant in eruptions, was almost wanting. This fact lends support to the opinion that the ammoniacal product of volcanoes is of organic origin, and is brought by the atmosphere over vents charged with hydrochloric acid. The distance and the small extent of cultivated lands explain the rarity of ammonia in the air at Santorin.
Over the central eruptive mouths, the points of the volcano where the incandescence was most lively, the spectroscope showed the presence of volatilized salts of soda and potash. After the eruptive manifestations had ceased, the salts which were deposited around the orifices of the vents were collected and analyzed, and appeared to consist largely of the chlorides of sodium and potassium, the sulphates of soda and potash, and the carbonates of soda and magnesia. Sea-water after evaporation leaves a residuum of analogous composition. These results are interesting in their bearing on the theory that the water of the sea is the immediate ordinary agent of the eruptions. All of the elements usually occurring at volcanic vents were found to be present at the points which were the seat of a volatilization of alkaline salts, lending support to the opinion which had been previously drawn from studies made at Etna, that the vents at their highest temperature present at once all the chemical elements of the volatilized bodies in the volcanoes, and that the cooler vents are gradually impoverished as their temperature becomes insufficient for the reduction of the eruptive materials to vapors.
Some of the volatilized substances in the volcanic conduits are susceptible of reacting on each other and producing fixed compounds. Thus are engendered the hydrated oxides of iron, specular iron-ore, free sulphuric acid, alum, and sulphate of lime, which are met around the vents.
Certain crystallized silicates, however, generally originate under different conditions. Although they are formed of elements which we are in the habit of considering fixed, they are found in the volcanic vents, on the surface of the rocks, under such conditions that they could have been produced only through volatilization. Such silicates
