306 Proceedings of Royal Society of Edinburgh. [sess. 
In absence of oxygen, aqueous solutions of hydrogen iodide are 
not decomposed by sunlight (15). 
Potassium iodide, when dry, is not affected by light. Moist 
potassium iodide exposed to air and light is oxidised (16) : 
4KI + 2H 2 0 + 0 2 = 4KOH + 2I 2 . 
Here the solution of potassium hydrate is a better electrolytic con- 
ductor than the corresponding solution of potassium iodide. 
Concentrated solutions of nitric acid are decomposed when 
exposed to the action of light, according to the equation : 
2HN0 3 = H 2 0 + N 2 0 4 + ° • 
The acid acquires a yellow colour, owing to the dissolved 
nitrogen peroxide. 
At first sight, this might seem an exception to the rule, as it 
involves the decomposition of an electrolyte without the formation 
of any better conductor, but it must be remembered that the elec- 
trolytic conductivity of nitric acid increases on dilution with water 
up to a certain maximum ; on further dilution, the conductivity 
rapidly falls away. According to the determinations of Kohl- 
rausch and Grotian (17), and E. Bouti (18), the maximum conduc- 
tivity of nitric acid is reached when the solution has been diluted 
until it contains about 32 per cent, of nitric acid. 
Now, nitric acid becomes less subject to the action of light as it 
is diluted, and if this action be related to increase of conductivity, 
we should expect to find that the action of light would cease to be 
readily observable at a dilution somewhat less than but not far from 
that corresponding to maximum conductivity. The most definite 
statement on this point which I have hitherto found is contained 
in Wurtz’s Didionnaire de Chemie ; it is there stated that the 
action of light on the acid is the more rapid the greater the con- 
centration, and that an acid having a specific gravity of P30 ceases 
to be coloured by light. Nitric acid of specific gravity P30 con- 
tains about 47 per cent. HN0 3 ; so that in this case also the rule 
holds, as will be readily seen by reference to the curve of conduc- 
tivity of nitric acid. 
Numerous other more or less complex cases might be quoted. 
