Extremely Dilute Acid and Alkali Solutions. 3 
simple salt like potassium chloride is a straight line, going through 
the origin. 
The form of the curve corresponding to the stage in the 
addition of the acid when the ‘neutralisation’ of the impurity is 
incomplete, will depend on the nature of the impurity. If it were 
an alkali (e.g. NaOH or NH,OH) we should expect that the 
conductivity of the water would be reduced (the net result of the 
addition of the first traces of H,SO, being the substitution of 
sulphate ions for hydroxyl ions). This reduction would continue 
until the neutralisation were complete, and the simple conductivity 
curve would immediately afterwards become normal. If the 
impurity were a carbonate (or carbonic acid), an actual dimi- 
nution of the conductivity of the water would not be expected, 
‘since the carbonic acid molecule (H,CO,) is more readily ionised 
‘than the water one (H.OH), and the velocity of the carbonate ion 
ds less than that of the hydroxyl ion. What we should expect, 
however, is that the simple conductivity curve would be concave 
(upwards) to begin with (while the carbonate ions were being 
suppressed), and that it would gradually become a straight line. 
Exactly similar effects would result in the case of the con- 
ductivity of correspondingly dilute alkali solutions if the impurity 
im the water were a free acid, or the salt of a very weak base. 
If the impurity in the water consisted of the salt of a strong 
acid or base, we should not obtain this form of association, and 
‘the form of the conductivity curve would not be influenced by any 
‘such impurity. 
_ Whetham and Paine have already applied this hypothesis to 
results obtained by them, and found it to explain the facts. We 
shall apply it to the results obtained in the present paper. It 
seems to be quite sufficient for providing a quantitative ex plana- 
tion of the acid and alkali conductivity curves without supposing 
these substances to behave in a manner essentially different from 
that of other electrolytes. 
II. 
We thought it would be worth while to measure the con- 
‘ductivities of acid and alkali solutions more dilute than those 
used by Kohlrausch and Whetham,—not so much for the sake of 
‘obtaining absolute values (the simple method employed was too 
erude to be trustworthy to that extent) as for determining to what 
lower limit the simple conductivity curve for an acid or alkali 
solution remains a straight line, and for determining approximately 
what form it assumes beyond. Solutions were prepared by the 
‘additions of a stock solution from a burette, and their conduc- 
tivities measured by simply transferring them to a conductivity cell. 
For the purpose of menue the conductivities of the 
1—2 
i: 
1 
4 
