thing that of these aluminium sulphates the fully saturated one, Al_{2}(SO_{4})_{3}, is the most stable, for even on long boiling of its solution in water it suffers no change, but the more basic is the sulphate the less stable it becomes, and so the more easily it decomposes on heating or boiling its solution, giving a deposit or precipitate of a still more basic sulphate, or of hydrated alumina itself, Al_{2}(OH)_{6}, until we arrive at the salt Al_{2}(SO_{4})_{2}(OH)_{2}, which is quite unstable on boiling; Al_{2}(SO_{4})(OH)_{4} would be more unstable still. This behaviour may be easily shown experimentally. We will dissolve some "cake alum" or normal sulphate of alumina, Al_{2}(SO_{4})_{3}, in water, and boil some of the solution. No deposit or precipitate is produced; the salt is stable. To another portion of the solution we will add some caustic soda, NaOH, in order to rob the normal sulphate of alumina of some of its sulphuric acid. This makes the sulphate of alumina basic, and the more basic, the more caustic soda is added, the sodium (Na) of the caustic soda combining with the SO_{4} of the sulphate of alumina to form sulphate of soda (Na_{2}SO_{4}), whilst the hydroxyl (OH) of the caustic soda takes the position previously occupied by the SO_{4}. But this increase of basicity also means decrease of stability, for on boiling the solution, which now contains a basic sulphate of alumina, a precipitate is formed, a result which also follows if more caustic soda is added, production of still more basic salts or of hydrated alumina, Al_{2}(OH)_{6}, taking place in either case.