105 
AND ON THE ELECTRIC EIELD OF THUNDERSTORMS. 
discharge are reached. Zeleny, # who has made a very interesting series of 
investigations on the stability of electrified liquid surfaces, found that in air at 
atmospheric pressure the potential required to cause a discharge from the surface of 
a drop of water at the end of a capillary tube exceeds, though only by a few per 
cent., that required to produce instability and disruption of the drops. He points 
out that it would follow from his experiments that a discharge of minute electrified 
drops, constituting an upward shower, would take place from the edges of the wet 
leaves of a tree in a thunderstorm, before the electric force at the surface of the 
tree reached the sparking limits. 
It seems not unlikely that under certain circumstances a similar process may occur 
in a cloud, droplets suffering disruption where the field approaches the sparking limit. 
Consider a developing cumulus head in which a charge is accumulating, and suppose 
that the radial electric force near the edge of the cloud-head reaches the value 
required to cause disruption of the drops before it reaches the sparking limit. The 
induced charges on the two halves of the drop will then be separated and will tend 
to travel in opposite directions along a line of force. 
The magnitude of the induced charge on each half of a spherical drop of radius a 
in a field F is o-nO^VjAir = fF a 2 , and when F approaches the sparking value this 
will generally greatly exceed any resultant charge of the drop. The charge per cubic 
centimetre of water for each half of the drop = fE u a 2 /§7ra 3 = 9F„/87ra; if F„ = 100 
and a = 1 mm. the charge per cubic centimetre of water for each half of the drop is 
360 E.S.U. Thus if the original drop of 1 mm. in radius were divided into two 
oppositely charged halves, the force acting on each of the new drops would in a field 
of 100 E.S.U. amount to 36 times its weight. 
If the drop is drawn out by the action of the field into an ellipsoidal or cylindrical 
form before disruption, the induced charges will be considerably greater. Separation 
of the charges by division of the drop will thus give rise to oppositely charged 
portions each having a charge much greater than that of the original drop. The two 
portions will tend to travel in opposite directions along the line of force with velocities 
greatly exceeding the original radial velocity of the drop from which they were 
derived. 
The outward moving products of disruption of the drops in the head of a cumulo¬ 
nimbus cloud may possibly constitute “ false cirrus.” These particles are more likely 
to freeze than those constituting the original head of the cloud; (l) because the 
stretching of the water drop into a filament itself causes cooling ; ( 2 ) the conversion 
of a water filament into an ice crystal is not accompanied by a large increase of surface, 
and one of the main obstacles in the way of the freezing of small drops is thus 
removed; and (3) the charged particles are still further cooled through being driven 
by the action of the field into the colder and drier air outside the original cloud. 
Ice needles formed under the above conditions would not only be charged but also 
* Zeleny, ‘ Proc. Cambridge Phil. Soc.,’ vol. 18, p. 71, 1914. 
VOL. CCXXI.-A. q 
