78 Mr Zeleny, On the Conditions 
14. Lastly, considering V and the volume of the ellipsoid, 
constant, instability sets in when 
i 
Weeacah ee (8), 
(1 + é) log | SS 2? 
and this expression, at the limit of a mr reduces to 
Vis L2traT oo ec i ve eee (9). 
It is seen that under the conditions here taken, when instability’ 
begins, the pressure inside the drop is greater than that outside! 
the drop, whereas under the conditions assumed in § 13 the opposite! 
is the case. 
In the experiments with hemispherical drops of water at the} 
ends of tubes it was observed that when instability commenced) 
there was an excess of pressure on the inside of the drops and this} 
was approximately equal to one-tenth of the whole pressure caused) 
by surface tension alone. With a low meniscus however, the! 
pressure inside the drops, under the same conditions, was found) 
to be less than that outside. | 
It may be said that the values of V* given by equations (6) and| 
(8) differ very little from those given by equations (7) and (9) respec- | 
tively, for values of e as high as 0:2 or more. + 
15. The preceding results, and some further considerations, | 
lead to the general expression (applicable to drops on the end of, 
a tube) | 
as giving the potential at which instability commences, C being a 
constant depending upon the shape of the drop and the manner in) 
which it undergoes axial changes. We may test the expression| 
experimentally by finding how the potential necessary for insta-| 
bility is dependent upon the size of the tube and upon the suri 
tension of the liquid used. 
Experimental Results. 
16. As the first example we will take some results obtained} 
for water and alcohol. The potentials at which these surfaces’ 
ceased to be unstable were found to be 4050 volts and 2350 volts. 
respectively. In agreement with our formula, the squares of 
these numbers are seen to be proportional to the surface tensions 
of the two liquids (taking 72 for the value of the surface tension 
of water and 24 for that of the alcohol used, ‘ites was nearly 
absolute). i 
As a further test of the surface tension factor in formula 10, | 
