according to the temperature at which the crystals are formed. Thus common salt in crystallising from its solution in water at the ordinary or at a higher temperature does not contain water of crystallisation. But if its separation from the solution takes place at a low temperature, namely below -5°, then the crystals contain 38 parts of water in 100 parts. Crystals of the same substance which separate out at different temperatures may contain different amounts of water of crystallisation. This proves to us that a solid dissolved in water may form various compounds with it, differing in their properties and composition, and capable of appearing in a solid separate form like many ordinary definite compounds. This is indicated by the numerous properties and phenomena connected with solutions, and gives reason for thinking that there exist in solutions themselves such compounds of the substance dissolved, and the solvent or compounds similar to them, only in a liquid partly decomposed form. Even the colour of solutions may often confirm this opinion. Copper sulphate forms crystals having a blue colour and containing water of crystallisation. If the water of crystallisation be removed by heating the crystals to redness, a colourless anhydrous substance is obtained (a white powder). From this it may be seen that the blue colour belongs to the compound of the copper salt with water. Solutions of copper sulphate are all blue, and consequently they contain a compound similar to the compound formed by the salt with its water of crystallisation. Crystals of cobalt chloride when dissolved in an anhydrous liquid—like alcohol, for instance—give a blue solution, but when they are dissolved in water a red solution is obtained. Crystals from the aqueous solution, according to Professor Potilitzin, contain six times as much water (CoCl2,6H2O) for a given weight of the salt, as those violet crystals (CoCl2,H2O) which are formed by the evaporation of an alcoholic solution.
That solutions contain particular compounds with water is further shown by the phenomena of supersaturated solutions, of so-called cryohydrates, of solutions of certain acids having constant boiling points, and the properties of compounds containing water of crystallisation whose data it is indispensable to keep in view in the consideration of solutions.
Supersaturated solutions exhibit the following phenomena:—On the refrigeration of a saturated solution of certain salts, if the liquid be brought under certain conditions, the excess of the solid may sometimes