492 
Proceedings of the Royal Society 
could detect no difference amounting to 1 per cent, which was about 
the uncertainty in my best experiments. 
The earliest measurements in absolute units, from which the 
dielectric strength of air could be deduced, were those of Sir William 
Thomson (1860), who measured the electromotive force between 
two parallel plates required to produce a spark varying in length 
from *0025 cm. up to T5 cm. De La Kue and Muller (1877), 
Macfarlane (1878), and T. B. Bailie (Comptes Rendus , Jan. 1882, 
pp. 38 and 130), have followed in his footsteps. The following table 
contains some selections from their results : — 
Spark Lengths 
in Centimetres. 
Potential Required to produce the spark in 
Electrostatic Units, 
C. G. S. 
Thomson. 
De La Rue 
and Muller. 
Macfarlane. 
Bailie. 
Water 
Electrometer. 
•01 
2-33 
... 
... 
3-16 
... 
•05 
7-28 
7-69 
7*48 
8-71 
... 
10 
13*30 
14-47 
11-56 
14-67 
... 
•30 
... 
34-44 
25-56 
35-35 
... 
•61 
... 
45-61 
64-81 
55-00 
It will be seen that the agreement between the different experi- 
menters is far from perfect. To those familiar with the manifold 
difficulties attending measurements of this kind, this is not surpris- 
ing. I may mention that the result in the last column was obtained 
last summer by Dr. Macfarlane and myself, by means of a new form 
of absolute electrometer which I was led to devise during a course 
of experiments of which we hope to give an account hereafter. 
The instrument consists simply of an insulated plane disc, suspended 
parallel to a surface of water connected with the earth. When the 
disc is raised to a high potential, the water rises up and forms a 
plateau, whose height is measured by observing with a micrometer 
eyepiece the displacement of the image of a fixed point seen by 
reflection at the surface of the water. From this we can calculate 
at once the difference of potential between the disc and the water in 
absolute units. The instrument is very direct and ready in its ap- 
