hydroborofluoric acid (HBF_{4}) according to the equation 4BF_{3}H_{4}O_{2} = 3HBF_{4} + BH_{3}O_{3} + 5H_{2}O.[16] This hydroborofluoric acid has its corresponding salts--for instance, KBF_{4}. On evaporating the aqueous solution this free acid decomposes, with the evolution of hydrofluoric acid, and a stable system is again obtained: 2HBF_{4} + 5H_{2}O = B_{2}F_{6}H_{10}O_{5} + 2HF. The resultant solution (containing 2BF_{3},5H_{2}O, sp. gr. 1·58), which is identical with that formed by the evaporation of a solution of boric acid with hydrofluoric acid, again only contains a compound of boron fluoride with water. Probably there are various other possible and more or less stable states of equilibrium and definite compounds of boron fluoride, hydrofluoric acid, and water. [13] Boron fluoride is frequently evolved on heating certain compounds occurring in nature containing both boron and fluorine. If calcium fluoride is heated with boric anhydride, calcium borate and boron fluoride are formed, and the latter, as a gas, is volatilised: 2B_{2}O_{3} + 3CaF_{2} = 2BF_{3} + Ca_{3}B_{2}O_{6}. The calcium borate, however, retains a certain amount of calcium fluoride. [14] In order to avoid the formation of silicon fluoride the decomposition should not be carried on in glass vessels, which contain silica, but in lead or platinum vessels. Boron fluoride by itself does not corrode glass, but the hydrofluoric acid liberated in the reaction may bring a part of the silica into reaction. Boron fluoride should be collected over mercury, as water acts on it, as we shall see afterwards. [14 bis] It appears to me that from this point of view it is possible to understand the apparently contradictory results of different investigators, especially those of Gay-Lussac (and Thénard), Davy, Berzelius, and Bazaroff. In the form in which the reaction of BF_{3} on water is given here, it is evident that the act of solution in water is accompanied by complex but direct chemical transformations, and I think that this example should prove the justness of those observations upon the nature of solutions which are given in Chapter I. [15] They are called fluoborates. They may be prepared directly from fluorides and borates. Such compounds of halogens with oxygen salts are known in nature (for instance, apatite and boracite), and may be artificially prepared. The composition of the fluoborates--for example, K_{4}BF_{3}O_{2}--may be expressed as that of a double salt, BO(OK),3KF. If an excess of water decomposes them (Bazaroff), this does not prove that they do not exist as such, for many double salts are decomposed by water.