82 _ Mr Zeleny, On the Conditions 
20. The facts noted in §§ 5 and 7 that with water in carbonic | 
acid and with alcohol in air or carbonic acid, the oscillations of the 
meniscus began at a potential in each of the cases which was 
independent of the pressure of the gas over a considerable range, 
are explained by the discharge potential being much higher at 
atmospheric pressure than the instability potential, so that even 
at much lower pressures the discharge voltage was not reduced 
enough to be below the potential for instability. 
The successive individual discharges, mentioned in § 2, which — 
are observed at increasing potentials when a discharge is first 
started from an old water surface, are doubtless caused by the © 
surface having a low surface tension owing to contamination. As _ 
pieces of the liquid are pulled away, the contamination gradually — 
disappears and a higher and higher voltage is necessary to produce — 
instability. The falling of the meniscus after each of these dis- 
charges shows that an increase in the surface tension has taken 
place. 
At a later stage where the surface undergoes continuous 
oscillations at a given voltage, a different cause doubtless enters, 
the surface becoming stable intermittently owing to electric 
shielding by the material pulled out of its surface, i 
Sporadic discharges like those considered above are often 
noticed with metal points at voltages far below the final discharge 
potential, and it is possible that something of the same nature ~ 
takes place there as in these cases, a part of the adsorbed layer 
of gas and water being pulled off from the surface at each dis- 
charge. 
21. The expression given in § 15 for the potential at which 
instability begins applies to cases where the medium between the 
point and plane is a gas, and thus having a dielectric constant of 
approximately unity. If the experiment were done in a non- 
conducting fluid not miscible with the liquid of the drop, the 
expression would have to include &, the dielectric constant of 
the new medium, and would be writtenkV?=CaT. Evidently, the 
dielectric constant could be determined in this way by measuring — 
the surface tension between the two liquids and the voltage at 
which instability starts. 
22. It follows from the results recorded in this paper that | 
before a potential is reached which would result in lightning ~ 
striking a tree, for example, the leaves, if wet, will discharge water 
from their edges in the form of minute electrified drops which 
will form an upward shower tending to neutralize the electric 
field. But since the discharge potential is so near to the in- 
stability potential for water, it is not likely that this phenomenon 
plays any important part in the atmospheric electricity during 
storms. 
