as stated above, act in opposition to the electro-affinity) are probably very different, so that no simple relation between the degrees of ionisation of these three compounds can appear.

The application of this theory to the formation of complex ions gives more positive results. Complex ions may be regarded as having been formed by the union of an ion with a neutral molecule; e.g. CdI," is a compound of 2I' with CdI. If the neutral part is a salt, as in the case quoted, the complex ion will the more readily split up, the greater the electro-affinities of its atoms. So that, on this theory, a complex ion will be the more stable the smaller the electroaffinities of the atoms forming the neutral part. We find that the salts of the alkali metals (which have large electro-affinities) practically never appear as neutral parts of complex ions. The salts of the alkaline earths (which have electro-affinities smaller than those of the alkali metals) form the neutral parts of many complex ions which are stable in concentrated solution, but which dissociate in dilute solution, e.g. MgCl, the anion of carnallite. Corresponding to the still smaller electro-affinity of cadmium, we have the more stable complex ions in solutions of cadmium salts. And among the salts of the trivalent metals (with still smaller electro-affinities) are AlF, which is the neutral part of the cryolite anion (AlF)"; Fe(CN)2, the neutral part of the ferrocyanide anion (Fe(CN))"". The stability of this ion is largely due to the low electro-affinity of the (CN)' ion, for complex ions in which the neutral part is a compound of iron or aluminium, with an atom of large electro-affinity, are not so stable, eg. the complex anions present in solutions of the alums.

Since, on the electrolysis of a complex salt, the complex ions practically never appear at the electrodes, but generally the simpler ions, it follows that the simple ions give up their electricity more easily than complex ions, i.e. that the electroaffinity of a simple ion is less than that of a complex ion of which it is a part. Since, therefore, the electro-affinity of a simple ion becomes greater on taking up a neutral molecule, we should naturally expect that ions with small electro-affinities will combine with neutral molecules more readily than those

with large electro-affinities. This is borne out by the facts. E.g. a solution of potassium chloride dissolves very little silver chloride; a solution of potassium bromide dissolves more silver bromide; a solution of potassium iodide dissolves a considerable amount of silver iodide, and potassium cyanide dissolves an equivalent amount of silver cyanide, forming indeed a complex salt. The series, chlorine, bromine, iodine, cyanogen, are in descending order of their electro-affinities. The distinction between complex and double salts is thus due to the different electro-affinities of the atoms.

This slight sketch of the theory is enough to show that useful results are obtained by assigning to each atom a specific electro-affinity, or tendency to ionise. By this theory, or a modification of it, we may hope to build up a comprehensive theory of ionisation.

§ 5. PSEUDO ACIDS AND BASES.

A few years ago Hantzsch1 called attention to the fact that many neutral or slightly acid compounds form salts which behave as salts of strong acids, in that they are hydrolysed only to a very small extent. Such compounds are called pseudo acids, and it has been shown that in these cases salt formation is accompanied by a change in chemical constitution. For example, phenylnitromethane, CH,CH,.NO2, a neutral compound, gives a neutral strongly ionised sodium salt CH.CH N-ONa.

Another class of compounds is known, which form salts with acids, with intra-molecular change. These are called pseudo

bases. Thus, pararosaniline, (NH,C,H1)2COH gives

CH_NH,

a hydrochloride (parafuchsine), (NH2C6H4)2C = CH ̧=NH2C1, which corresponds to (NH,C,H1)2C = C ̧H ̧NH2OH, and not to pararosaniline itself.

1 Ber. d. deutsch. Chem. Gesell., 32. 576 (1899).

Among pseudo acids, two classes may be distinguished; (a) the pseudo acids proper, which give no ions; and (b) compounds giving the same anions as their salts, although these anions do not correspond in constitution with the non-ionised molecules. An example of the second class is nitroform, which in non-ionising solvents has the formula CH(NO2)3, but which in aqueous solution gives ions. ((NO2)2C=N—O)' and H, and which gives salts of type Ӧ

(NO2)2C=N-ONa. But since in aqueous solution pseudo

acids of the class (a) must to some, even if very small, extent, give anions corresponding to their salts for otherwise the formation of the salts in alkaline aqueous solution would be difficult to explain the distinction between (a) and (b) becomes one of degree and not of kind.

In a few cases it has been found possible to isolate the hydrogen compound corresponding to the salt, and these are found to be strongly ionised. The following numbers are for brom-phenylnitromethane at o

(v

=

Pseudo acid, BrCH.CH.NO2.

In saturated solution

2700 litres) the electric conductivity of the water was raised to only a small extent.

The effect of a slight change in constitution is seen to be very large. In all cases, the true acid is obtained from the salt by decomposition with a stronger acid. Even if the true acid cannot be isolated, its presence in solution can be proved, unless the velocity with which it goes over into the pseudo acid is very large. For in a dilute solution of equivalent quantities of a salt of a pseudo acid, and of hydrochloric acid, both are completely ionised, and the conductivity of the solution is equal to that of equivalent quantities of sodium

chloride, and the true acid corresponding to the salt, at the particular dilution. But the true acid will change over to the non-conducting pseudo acid with a velocity depending on its constitution. So that after a time, the conductivity of the solution will be that of the sodium chloride present. And at any time the excess of the conductivity of the solution over the conductivity of the sodium chloride present is a measure of the amount of true acid present.

The following numbers show the variation of conductivity with time, of a mixture of equal volumes of normal hydrochloric acid and normal sodium phenylnitromethane (0°)

32

The difference in the value for sodium chloride and the final value for the mixture of solutions is due to the presence of a small amount of nitrous acid caused by the partial decomposition of the compound. As an example of a pseudo base we may take methylphenylacridonium hydrate,1 which, according to its method of formation, should have the formula C&H

CH

OH

The con

8

salts, with acids, correspond to the first formula. ductivity at o° of a mixture of equal quantities of normal solutions of the hydrochloride of the base and sodium hydrate are given by the numbers—

1 Hantzsch and Kalb, Ber. d deutsch. Chem. Gesell., 32. 3109 (1899).

The large value of the initial conductivity shows that the base present in the solution (corresponding to the first of the above formula) is very strongly ionised.

These examples suffice to show the nature of these phenomena. For further work on this important class of compounds, the papers of Hantzsch and his pupils must be consulted.

§ 6. AMPHOTERIC ELECTROLYTES.

There are many compounds which act both as acids and bases, i e. which give both hydrogen and hydroxyl ions. Such are known as amphoteric electrolytes. For example, aluminium hydroxide dissolves in a solution of sodium hydrate to form a sodium salt Al(ONa). Since in this reaction, hydroxyl ions disappear from the solution, the aluminium hydroxide must furnish hydrogen ions. Similarly aluminium hydroxide dissolves in acids to give salts, and must therefore be capable of giving hydroxyl ions. This compound may therefore ionise according to either of the equations1

Similar phenomena have been observed with the oxides of gallium, zinc, lead, and tin.

The amido acids have been more fully investigated than other amphoteric electrolytes. The mode of ionisation of these compounds, however, differs from that of the compounds mentioned above. For the hydroxyl and hydrogen ions come from different groups in the molecule of an amido acid, and 1 These equations are illustrative only. See Hantzsch, Zeitschr. f. Anorg. Chemie, 30. 289 (1902).

2 Winkelblech, Zeitschr. phys. Chem., 36. 550 (1901).