^{page 209 note *} Memoires de l'Academie des Sciences, 1772.

^{page 225 note *} The small portion of sulphate of soda obtained with the sulphate of magnesia, in the second analysis, may have been formed by the action of the alcohol, which, though employed much less extensively than in the first, was still introduced to a certain extent. Or it might originate from other circumstances independent of this; for a similar result, I have been informed, sometimes occurs in the large way, sulphate of soda being procured in boiling down the bittern of sea water to obtain its sulphate of magnesia, or in purifying this sulphate. The circumstances on which this depends, it may be difficult to assign with perfect precision; but it probably arises from the relative quantities, of the different salts, and their tendency to crystallization, as influenced by the state of concentration, and the temperature. That both of these have a considerable effect on the combinations established in a compound saline solution, has been sufficiently shewn by the experiments of Berthollet and others. A striking proof of it was derived from the very salts which are the subject of the present observations, in a singular case of affinity, first observed by Scheele, and afterwads confirmed by Gren: that of muriate of magnesia and sulphate of soda, which decompose each other in a concentrated solution at a high temperaturte, producing muriate of soda and sulphate of magnesia; but, at temperatures below 32°, the reverse effect takes place, muriate of soda and sulphate of magnesia re-acting, and being converted into sulphate of soda and muriate of magnesia. This singular case is evidently owing to the relation of the solubility of these salts to temperature. Muriate of soda has its solubility little increased by heat, of course little diminished by cold; sulphate of soda is in this respect precisely the reverse; hence, at an elevated temperature, muriate of soda is the less soluble salt; and this determines its formation and separation from a compound solution, containing its elements; at a low temperature, again, sulphate of soda is the less soluble salt; and this equally determines its formation, of course occasions the reverse decompositions. Now, according to the proportion of saline ingredients, and according to the state of concentration, and the temperature favouring the tendency of certain salts to crystallization more than others; it is easy to conceive, that in a compound solution, different combinations may be established, as these circumstances vary, and thus products may be obtained, under certain conditions, which are not obtained under others. Although sulphate of magnesia, therefore, is usually obtained by evaporation from sea-water, sulphate of soda, at some stages of the operation or under peculiar circumstances, with regard either to relative quantity of the elements, or to temperature, may likewise be formed.

^{page 233 note *} There is sometimes obtained in the large way, from the products of the evaporation of sea-water, a triple salt, which has not been noticed by chemists, but which appears to be of definite composition, and is distinguished by peculiar properties,—a Sulphate of Magnesia and Soda. It is formed in purifying the sulphate of magnesia procured by the first evaporation from the bittern of sea-salt. In this process the sulphate, which is impure, both from the intermixture of muriate of soda and muriate of magnesia, and perhaps, also, of sulphate of soda, is dissolved in water, and by evaporation and cooling is obtained crystallised; a fresh quantity of it is added to the residual liquor, and by the necessary evaporation and cooling, a new crystallization is produced; this is repeated for a third or fourth time; and it is in these latter crystallizations that this triple salt is formed, frequently in considerable quantity, and usually at a high temperature, being precipitated even in the boiler. It crystallises in rhombs, at first irregular and semitransparent; but by solution in water, and a second crystallization, is obtained in more regular rhombs, truncated on the acute angles, on the obtuse angles and edges, and on the terminal edges, considering the rhomb as a four-sided prism, and transparent. The crystals are permanent in the air; they are soluble in little more than three times their weight of water, at the temperature of 60° they do not undergo the watery fusion from heat, but suffer decrepitation. In these properties, this salt differs entirely from sulphate of soda, or sulphate of magnesia.

To determine its composition, 20 grains reduced to powder were exposed to heat, raised gradually nearly to redness; they lost from the escape of water 5.6 grains. The residual powder was dissolved in water, and muriate of barytes was added as long as any precipitation was produced. The precipitate dried at a red heat, weighed 23.9 grains, equivalent to sulphuric acid 8.2 grains. To the clear liquor carbonate of ammonia was added, which did not impair the transparency; phosphoric acid was then dropped in, which produced a copious precipitation. The precipitate, calcined at a red-heat, weighed 5.3 grains, equivalent to 2.1 of magnesia, or 6.4 of sulphate of magnesia; the residual liquor being evaporated to dryness, the dry mass was submitted to heat, gradually raised, as long as any vapours exhaled; it afforded, by solution in water and evaporation, muriate of soda in cubes, which, after exposure to a red heat, weighed 6.4 grains, equivalent to 7.8 grains of sulphate of soda. 100 grains of the salt, therefore, afford of

It afforded also a slight trace of muriatic acid; its solution being in a very slight degree rendered turbid by nitrate of silver, probably owing to the intermixture of a little muriate of soda, as an extraneous ingredient. This accounts for the proportion of sulphuric acid, as inferred from the quantities of the bases, being a little larger than that directly obtained by the precipitation by muriate of barytes.

The difference of crystalline form, as well as other differences of properties in the salt from those, either of sulphate of soda or sulphate of magnesia, sufficiently prove that it is not merely an intermixture of the two, but that it is of definite composition. It deserves to be remarked, too, that it has not the same relation to water that either of these salts has, or any mean between them; the quantity of its water of crystallization being considerably less. Its taste is much less disagreeable than that of sulphate of soda or sulphate of magnesia; it might therefore probably be introduced with advantage as a purgative salt, especially as it could be procured at a low price; and from its composition, it would afford a very good substitute for the aperient mineral waters, which usually owe their activity to sulphate of soda and sulphate of magnesia.

^{page 236 note *} I shall have to state in a subsequent paper, the peculiar advantages attending this method of estimating the magnesia.

^{page 239 note *} In another analysis of sea-water, in which subcarbonate of ammonia was employed to precipitate the magnesia, a solution of it being added to the water concentrated by evaporation, the clear liquor, after the subsidence of the precipitate being evaporated to dryness, the saline matter being exposed to heat, to dissipate the muriate of ammonia; being re-dissolved in water, the subcarbonate of ammonia again added, and this repeated for a third, and even a fourth time, the results gave the following proportions of the elements,

Lime,

The principal difference here, is the proportion of magnesia being somewhat smaller, evidently owing to its precipitation by the carbonate of ammonia, even with the aid of the methods employed to promote it, being imperfect.

^{page 241 note *} Annales de Chimie, t. 20.

^{page 241 note †} Philosophical Transactions, 1810.

^{page 242 note *} Geological Transactions, vol. ii.