chlorine. On the addition of a reagent, usually silver nitrate, it is found that the solution gives a reaction characteristic for chlorine, inasmuch as a curdy precipitate of silver chloride is formed. It can easily be proved that this reaction will not detect every chlorine atom, but only those existing as ions, for a number of substances containing chlorine, such as potassium chlorate, monochloracetic acid, and other organic and inorganic compounds in which the presence of chlorine can otherwise be proved, do not give this characteristic reaction. These substances do not dissociate to give chlorine ions, \^^len a substance, such as mercuric chloride, gives only a small amount of chlorine ion, then on addition of silver nitrate this is removed from the solution as silver chloride. By the removal of the chlorine ion the dissociation equilibrium is disturbed, and in order to re-establish it more chlorine ion must be formed at the expense of undissociated mercuric chloride molecules. Consequently, one part of the chlorine after another is precipitated, until the whole of it is completely removed from the solution. If, however, the number of chlorine ions at the beginning is extremely small, it may happen that on the addition of silver nitrate the solubility product of silver chloride is not reached, and con- sequently there will be no precipitation of this substance. The reagent silver nitrate can nevertheless be used to detect the presence of a certain extremely small quantity of chlorine ion in solution. This and other chemical means are often very delicate, and are therefore of great use in determining the occurrence of ions in solutions in which the electric con- ductivity gives no certain result on account of the presence of other ions.
The same sort of behaviour is exhibited by most of the chemical reagents generally used in oixlinary "wet" analysis.
An example, to which I {12) called attention in 1884, is the behaviour of the ferrocyanides, which, although they contain iron, do not give the reactions characteristic of iron, or rather of the iron ions. Cases like this were pre^dously
�� �
