178 Scientific Proceedings, Royal Dublin Society. 
and on the outside— 
F2 Q? 
oer ag on K 
Hence Pat=pb', 
or the pressures vary inversely as the squares of the areas on 
which they act. 
From this it is easy to calculate the tension in the shell— 
a? 
T= PS . 
Now, if we call 6=b—a the thickness of the shell, we get— 
Qa, and consequently 
v3 
=a 
So that when charged to a given potential, the tension of the 
shell varies inversely as the square of its thickness. M. Duter 
observed (see Comptes rendu, Vol. 87, p.1036) that the enlargement 
varied inversely as the square of the thickness of the accumulator, 
and as the extension of the shell would be generally proportional 
to its tension, it is reasonable to conclude that the effect M. Duter 
observed is due to the cause I mention. 
Part of the effect in actual] accumulators may be due to defor- 
mations of the surface, which tend to make it more spherical. 
In the case, however, of an ellipsoidal surface, the tensions on the 
ends are so very much greater than elsewhere that it does not 
seem likely that it would tend to become more spherical. 
T 
Nortst.— M. D. J. Korteweg, in the Comptes rendu, No. 7, 17th 
February, 1879 (the same day as the above communication was read), 
gives the same explanation as I have given, and, besides, he mentions 
that he has made experiments showing that it is quantatively a sufli- 
cient one. ‘ 
