of Instability of Electrified Drops, ete. 81 
that for metal points with a radius greater than 0:025 cm. the 
voltages at which a discharge begins are proportional to the square 
roots of the radii of the points, which is the same relation as that 
obtained above for the beginning of instability with liquids, although 
for points smaller than the value given the voltages necessary for 
starting the current are larger than this relation would indicate. 
After a discharge has started from a surface, a large increase in 
the voltage may not bring on instability, because while a current 
_is flowing the electric intensity at the surface varies but little with 
the voltage used (Joc. cit., p. 85). 
Since the potential necessary for a discharge depends upon the 
pressure of the gas while the potential for the beginning of in- 
stability does not, it is possible to have instability begin first in 
a gas at atmospheric pressure and to have a discharge begin first 
in the same gas at reduced pressures. 
19. With water surfaces in air at atmospheric pressure, the 
potential at which a discharge starts is only a few per cent. higher 
than the potential at which instability sets in, and as the potential 
is gradually raised above this latter value the instability at the 
_ surface stops suddenly and a steady current flows from the surface. 
This indicates that either the surface tension is increased by the 
current or the electric intensity at the surface falls slightly as 
soon as the current starts. There is some evidence for both views. 
A sort of electrical discharge is present with water surfaces as 
soon as instability begins as is shown by a faint luminosity seen 
_ extending some distance from the point, but apparently this dis- 
charge is brought on by the changes of shape caused by the in- 
_ stability and may be confined altogether to the surfaces of the flying 
drops. 
The conical form of discharge surface, seen when carbonic acid 
is used with a water surface or when liquids of low surface tension 
are used, makes its appearance over a certain range of voltages in 
cases where the instability voltage is considerably lower than the 
discharge voltage. As the liquid is being pulled out rapidly from 
the tip end of the meniscus in the form of threads, the surface 
—— 
distribution over the rest of the meniscus becomes altered so that 
equilibrium necessitates the conical form of surface. This form of 
surface should be obtainable also with a water surface in air, if the 
discharge potential were raised sufficiently. The latter condition 
can be produced by increasing the pressure of the air, and on trial 
it was found that the conical form came on as predicted when the 
pressure of the air was increased but 10 cms. of mercury above 
that of the atmosphere. 
_ The presence of a little aleohol vapour in the discharge vessel 
also made the conical form of surface appear, by lowering the surface 
tension of the water. 
