570 
CHEMISTRY: E. W. WASHBURN 
or less satisfactorily the observed relation between A and C over at 
least the lower portions of the concentration range. The lowest con- 
centration to which any observer has hitherto found it practicable to 
push accurate conductivity measurements is 0.0001 normal. (Meas- 
urements by KohlrauschO . This limitation of the concentration range 
open to measurement is due to the magnitude of the 'water correction' 
in the very dilute solutions, and to the uncertainty involved in applying 
it. Conductivity water which is approximately stable in contact with 
the atmosphere has a specific conductance of the order of magnitude 
of 1.10-^ reciprocal ohms, due to the impurities (chiefly CO2) which it 
dissolves from the air. The 'water correction' which must therefore 
be applied to the measured value of the conductivity of a dilute salt 
solution may amount at 0.0001 normal to as much as 10%, at 0.00001 
normal to as much as 100%, and at 0.000001 normal to as much as 
1000% of the conductivity of the salt itself. And, aside from the 
magnitude of the correction the proper method of applying it is ren- 
'dered uncertain owing to our lack of knowledge of the exact nature 
and amounts of the impurities present in the conductivity water, as 
well as to lack of sufficient constancy in the value of the conductivity 
of the water during the time which is required to prepare and measure 
the conductivity of the salt solution. 
The Ao values in use at present have, consequently, been obtained by 
extrapolation from conductivity data which in the most favorable 
cases do not extend below 0.0001 normal and which in most cases stop 
at 0.001 normal. The values of Ao selected by different observers 
using the same conductivity data are, therefore, almost as numerous 
as the different functions which have been proposed for representing 
the data over the range open to measurement, ranging, for example, 
in the case of KCl at 18° from Ao = 128.3 (Kraus and Bray^), and 
Ao = 128.5 (Wegscheider^) to Ao = 130.0 (Kohlrausch; Noyes and 
Falk^). Now unless the Ao value for a salt can be determined with a 
certainty of at least 0.01 to 0.02% we cannot hope to obtain any infor- 
mation concerning questions which involve a knowledge of the concen- 
tration of the im-ionized molecules of strong electrolytes in dilute 
solutions, such questions, for example, as the behavior of strong elec- 
trolytes with respect to the requirements of the Mass-Action Law. 
About eight years ago the writer became convinced that the only 
possible way of securing an answer to these much discussed questions 
was to obtain accurate conductivity data in the range of concentra- 
tions below 0.0001 normal, and that, in spite of the great difficulties 
which such an investigation presented, the questions at stake were of 
