I exposed a small quantity of a pure vegetable fixed alkali to the air, in a broad and shallow vessel, for the space of two months; after which I found a number of solid crystals, which resembled a neutral salt so much as to retain their form pretty well in the air, and to produce a considerable degree of cold when dissolved in water. Their taste was much milder than that of ordinary salt of tartar; and yet they seemed to be composed only of the alkali, and of a larger quantity of air than is usually contained in that salt, and which had been attracted from the atmosphere: for they still joined very readily with any acid, but with a more violent effervescence than ordinary; and they could not be mixed with the smallest portion of vinegar, or of the sedative salt, without emitting a sensible quantity of air.

As it now appeared that several alkaline substances have an attraction for fixed air, I tried a few experiments to learn the relative strength of their several attractions.

Twenty four grains of magnesia in fine powder were mixed with five ounces of the caustic ley in a small vial, which was immediately corked and shaken frequently for four hours. The ley was then poured off, and the magnesia washed with repeated affusions of water, and dried. It had lost about the half of its weight, and when reduced to a fine powder was readily dissolved by acids with an effervescence which was hardly perceivable: the alkali had therefore extracted its air. I also threw some fresh magnesia into the ley which had been poured off, and thereby rendered it perfectly mild and similar to a solution of salt of tartar; so that it effervesced briskly with acids.

With an ounce of the mild spirit of salt ammoniac, I mixed a dram of magnesia in very fine powder which had been previously deprived of its air by fire; and observing that the magnesia had a tendency to concrete into a solid mass, I shook the vial very frequently. After some days the powder was increased to more than double its former bulk; and when the vial was opened, the alkaline spirit emitted a most intolerably pungent smell. It likewise floated upon water, but was not perfectly caustic; for it still yielded some air when mixed with acids, and also rendered lime-water turbid: neither of which would probably have happened if I had used a greater quantity of magnesia, or had allowed the mixture to remain a longer time in the vial. I now washed out the whole of the mixture into a bowl, and dryed the magnesia until it lost all smell of the alkali. It weighed a dram and fifty eight grains, effervesced violently with acids, and therefore contained a large quantity of air, which had been drawn from the alkali by a stronger attraction.

Having formerly shewn, that magnesia saturated with air separates an acid from a calcarious earth, which it is not able to do after being deprived of its air by fire; I now suspected that the air was the cause of this separation, because I found that it was joined to the calcarious earth at the same time that the acid was joined to the earth of magnesia; and imagined that a pure calcarious earth might possibly have a stronger attraction for acids than a earth of magnesia.

I therefore dissolved two drams of magnesia in the marine acid, and thus obtained a compound of an acid and of the pure earth of this substance; for the air which was at first attached to it, was expelled during the dissolution. I then added thirty grains of strong quick-lime in exceeding fine powder, shook the mixture well, and filtrated it. The powder remaining in the paper, after being well washed, was found to be a magnesia, which, as I expected, was destitute of air; for it was dissolved by the vitriolic acid without effervescence. And the filtrated liquor contained the lime united to the acid; for upon dropping spirit of vitriol into it, a white powder was immediately formed.

We must therefore acknowledge a stronger attraction between the calcarious earths and acids than between these and magnesia: but how does it then happen, that, if magnesia saturated with air be mixed with a compound of acid and calcarious earth, these two last, which attract one another the most strongly, do not remain united; but the acid is joined to the magnesia, and the calcarious earth to the air which it attracts much more weakly than it does the acid? Is it because the sum of the forces which tend to join the magnesia to the acid and the calcarious earth to the air, is greater than the sum of the forces which tend to join the calcarious earth to the acid, and the magnesia to the air: and because there is a repulsion between the acid and air, and between the two earths; or they are somehow kept asunder in such a manner as hinders any three of them from being united together?

The first part of this supposition is favoured by our experiments, which seem to shew a greater difference between the forces wherewith the calcarious earth and magnesia attract fixed air, than between those which dispose them to unite with the acid. The repulsions however hinted in the second are perhaps more doubtful, tho' they are suggested in many other instances of decomposition; but the bounds of my present purpose will not allow me to enter upon this subject, which is one of the most extensive in chemistry.

We meet also with a difficulty with respect to the volatile alkali similar to the above. Thus a calcarious earth that is pure or free of air has a much stronger attraction for acids than a pure volatile alkali, as is evident when we mix quick-lime with salt ammoniac; for the alkali is then immediately detached from the acid: and agreeably to this I found, upon trial, that a pure or caustic volatile alkali does not separate a calcarious earth from an acid. Yet, if we mix a mild volatile alkali, which is a compound of alkali and air, with a compound of acid and calcarious earth, these two last, which attract one another most strongly, do not remain united; but the acid is joined to the alkali and the earth to the air, as happens in the precipitation of a calcarious earth from an acid, by means of the common or mild volatile alkali.

I remember likewise a parallel instance with regard to quick-silver. This metal has an attraction for the vitriolic acid, and when joined to it appears under the form of turbith mineral: but this attraction is weaker than that of the fixed alkali for the same acid; for if we mix a dissolved salt of tartar with turbith mineral, the turbith is converted into a brown powder, and the alkali into vitriolated tartar; which change happens the sooner, if the pure or caustic alkali is used. Yet, if to a compound of quick-silver and the nitrous acid, we add a compound of the fixed alkali and the vitriolic acid, or a vitriolated tartar, and digest the mixture with a strong heat, the vitriolic acid does not remain with the alkali, but is joined to the quick-silver which it attracts more weakly, composing with it a turbith mineral; while the alkali is joined to the nitrous acid which it likeways attracts more weakly than it does the vitriolic, and is converted into salt-petre.