In Table I are cited data on the vapor pressures of the common metallic chlorides. The temperatures given are in the region of magmatic temperatures, except where a high vapor pressure is developed at a much lower temperature.

Because of the volatility of chlorides, and because direct observation of volcanic sublimates shows their presence, it is deemed probable that chlorides are the most important form of combination in which metals are sublimed from the magma; fluorides to a somewhat less degree. The high vapor pressures of the chlorides of some of the most common metals— Zn, Cd, Pb, Ni, Sn, As, Sb, and Bi—should be noted. FeCl3 and AlCl3 are among the highest, and because of this and because of the mass action effect caused by their large concentration in the magma, they might be expected to be volatilized in the greatest quantities.

With Fe this evidently holds, but with Al the tendency of AlCl3 to react with H2O to form oxide exerts a restraining influence, which may prevent Al from being carried far, or may diminish to a great degree its escape from the magma, unless the ratio of HCl to water vapor is unusually high.

The very low volatility of CaCl2 is puzzling in view of considerable evidence that indicates that calcium salts have been volatilized and deposited as volcanic sublimates. This has considerable theoretical importance. Maier found no detectable pressure of CaCl2 at 1137.6°, within the 4 mm. limit of uncertainty of the method. On the other hand Fouque, employing a dynamic method, obtained results of different aspect. A mixture of CaSO4 and NaCl in a platinum boat within a platinum tube was heated to bright red in a current of water vapor, and apparently a considerable amount of lime in some form was volatilized from the boat and redeposited in the cooler part of the tube. The arrangement of the apparatus and the conditions of the experiment are described, and there is no obvious criticism to be made. Fouque regarded the reactions as involving a double decomposition of CaSO4 and NaCl, with volatilization of CaCl2.

If we accept the results for what they seem to mean, it is perhaps not astonishing that a substance of very low vapor pressure, such as CaCl2, should be carried along in appreciable quantities in a stream of water vapor. There is also a possibility that a double chloride of calcium and sodium may be formed, of higher volatility than CaCl2. Maier found reason to believe that NaCl and AgCl form such a compound, of vapor pressure greater than the sum of vapor pressures of the pure materials. Still another possibility is that though no calcium compound is itself effectively volatile, water vapor at a high temperature may exert a direct solvent action. Unless some such explanation is valid it is difficult to account for reported occurrences of calcium chloride and calcium sulphate as volcanic sublimates under conditions which do not appear explainable by the action of acid gases upon solid rock.

F. von Wolff gives Scacchi as authority for finding CaCl2 in large amounts, with chlorides of K, Na, and Mn, as sublimation products at the Vesuvius eruption of 1872. Koto, describing the great outbreak of Sakura-jima in 1914, speaks of the “universality and abundance of salmiac and gypsum, especially the first, among incrustations and sublimates in the juvenile lavas and the ventholes in Sakura-jima.” Palmieri, recording his observations of the Vesuvian eruption of 1872, speaks of the examination of a bomb, 4 to 5 meters in diameter, carried by the lava stream. Under the outer shell of the bomb, which still emitted smoke and HCl, he found micaceous iron and chloride of lime “almost pure.” At fumaroles on the brim of the crater he found sublimates of chloride of iron, together with chlorides of sodium, magnesium, and calcium. “This last chloride was frequent even among the sublimations of the fumaroles of the lavas. ” Lacroix, in similar situations to those described by Palmieri, that is, in fumaroles at the rim of the crater of Vesuvius, after the eruption of 1906, found the most abundant solid products to be “those which one finds in all the eruptions of Vesuvius; they consist of chlorides of iron, potassium, sodium, calcium, magnesium, etc.” The temperature was above 350° C. The chlorides were locally covered with realgar. Roth gives, as the most generally sublimed chlorides in volcanoes and solfataras, sodium chloride, the chlorides of iron, and ammonium chloride; in addition are chlorides of potassium, calcium, magnesium, lead, copper, cobalt, nickel, tin, and aluminum.

In the stratified deposits of lapilli and ash in Katmai Valley, derived from the 1912 eruption of Mount Katmai, the writer found in 1923 minute crystals of gypsum impregnating one of the layers. The amount of CaSO4, calculated from a determination of SO3 in the average pumiceous material of the layer, was 0.92 per cent. These loose strata are composed of the explosive ejecta from the volcano, which lay as they fell, and had not been acted upon subsequently by fumaroles. The site is many miles distant from the fumarolic areas of the Valley of Ten Thousand Smokes. It seems as if the gypsum represented a lime compound which had been evolved from the pumice during its flight from the crater. An alternative is that the concentration of acid vapors in the atmosphere around the flying pumice was so great that they acted upon it and extracted lime. In view of the high resistance of siliceous pumice to chemical action of this kind, this is difficult to believe, though the possibility must be recognized.

The volatility of the halide compounds of silicon was not determined in Maier’s work. Silicon forms with chlorine a number of compounds, easily decomposed by water. The most important is SiCl4, of which the boiling point is given as 59° C. at 760 mm. With fluorine only one compound, SiF4, is at all usual or well known, though apparently at least one other exists, SiF4 is a colorless gas, which is condensed to a liquid at —105.5° under 9 atmospheres pressure. Its critical temperature is — 1.5°. This compound is hydrolyzed fairly easily by water, but, as every analyst knows, SiO2 may be entirely volatilized by heating with HF in the presence of a moderate amount of water. Fluorine in a magma should therefore be effective in removing silica.

Greig, Merwin, and Shepherd found in numerous experiments that when the powder of any natural rock (even magnetite ore) has been heated in a platinum container within a silica tube to a temperature of 800° to 1200°, silica, in very appreciable quantity, has been caused to migrate, and has been redeposited on the platinum as a crust of cristobalite. This result is evidently due chiefly to the presence of a small residue of volatiles in the rock, for when the rock has been previously highly heated, the transfer of silica is imperceptible. If the original volatiles are expelled and a little water is introduced, there is again a considerable transfer and deposition of cristobalite on heating, as under the first conditions. Likewise, when powdered feldspar (Ab1An1) is placed in a platinum crucible and heated to 1175° with a little water in a silica tube, silica is transferred to the feldspar powder and forms a glassy crust, while in the absence of water the feldspar remains unsintered. These experiments seem to show that water vapor at high temperature is a solvent of silica.

At times fluorite appears in contact-metamorphic aureoles and similar situations in testimony of the presence of fluorine in emanations, but it need not always be deposited though fluorine be present; moreover, it is a mineral that is not insoluble in pure water and its solubility is increased by the presence of CO2, ammonium salts, or alkalis; and it is therefore likely to be removed by later solutions. Fluorine probably is considerably less abundant than chlorine in magmas, but its determination by analysis is difficult and uncertain, and frequently it has not been looked for. As a constituent of micas, humites, and several other minerals, it is widely distributed in small amount in igneous rocks and contact zones.

Chlorine forms very few combinations that are not highly soluble in aqueous solutions. One of the few minerals in which it becomes fixed is scapolite, and this is sometimes present in large quantity in contact rocks. Emmons and Calkins calculate from the amount of scapolite in the metamorphosed rocks adjacent to the Philipsburg batholith that a quantity of chlorine was given off, measurable in cubic miles at ordinary temperature and pressure.

Halide compounds are regarded in this paper as of chief importance in the chemistry of magmatic volatiles, because of their intrinsic volatility and because there is definite evidence of their participation in the formation of fumarolic deposits. A somewhat different process, consisting of the solution of mineral matter which is not itself volatile, in highly heated water vapor as in an aqueous liquid, has been frequently suggested. To what extent this may be invoked is uncertain, though there is evidence that it is a factor, and possibly an important one. Because of the small amount of evidence, however, it seems best not to emphasize the possibility. Even if water vapor possesses this solvent property the principles involved in volatilization and in reactions among volatiles are not greatly affected. Somewhat less importance is then to be attached to the volatility of halide compounds and somewhat greater importance to the solvent action of water vapor on mineral matter, but otherwise little modification is required.

There may be also, under magmatic conditions, other volatile combinations about which little or nothing is known. Shepherd found, in analytical work on gases evolved from rocks, that some silicon-bearing gas, perhaps a hydride of silicon, but of different properties from known hydride, was given off “whenever silica or a silicate is heated in the presence of hydrogen to about 1,100° or higher. It would seem to be of sufficient geological significance to warrant further study.”