1906-7.] The Electric Conductivity of Sea- Water. 237 
Temperature Corrections . — Several of the waters were examined at 
different temperatures in the above way, and it was found that all of them 
behaved similarly. For a mean temperature difference of 1° C. it was 
found that the mean difference in the minimum deviation was 30" of arc. 
The individual measurements agreed well among themselves, and the 
temperature correction is practically identical with that obtained by Mr 
Manley, whose value was 31" for 1°C. This correction was accordingly 
applied. It would have taken time to work out the refractive indices of 
the samples, and the writer agrees with Mr Manley * that, for the end in 
view, the minimum deviation gives exactly the same information as the 
refractive index does. The angles of minimum deviation are, therefore, 
compared in place of the refractive indices. 
Errors of Experiment . — The error of observation on a single measure- 
ment might be as much as 30" ; but since at least two independent measure- 
ments were made with each sample of water, the probable error is not more 
than 10" to 15", and the agreement of the values obtained from different 
measurements of the same water showed that 10" may be considered a 
reasonable approximation to the probable error. 
The following table gives the values for the various constants and the 
chlorine in grams per litre. The latter, together with the relative density 
A, at 18° (by the pyknometer), were determined by Mr Manley for his own 
work and kindly supplied by him. 
TABLE I. 
Serial 
Number. 
Chlorine, 
Grams 
per Litre. 
Relative 
Density 
A 18- 
Specific Con- 
ductivity l Ip 
at 25° in terms 
of the Legal 
Ohm. 
Angle of 
Minimum Devia- 
tion at 24° by 
a Prism having 
^1 = 60° 6' 10". 
1 
20*651 
1*02764 
0-054858 
86925" 
2 
*651 
74 
4787 
913 
3 
*756 
83 
5018 
931 
4 
•686 
85 
5333 
925 
5 
•756 
83 
5137 
931 
6 
•756 
91 
5137 
934 
7 
•756 
84 
5137 
926 
8 
•615 
70 
4922 
919 
9 
•615 
74 
4964 
922 
10 
•721 
81 
4922 
919 
11 
•721 
87 
5069 
932 
12 
•721 
86 
5175 
940 
13 
21*215 
1-02849 
6117 
967 
* Proc. Roy. Soc. Edin ., 1900. 
