AND ON THE ELECTRIC FIELD OF THUNDERSTORMS. 
75 
earthed supporting rod is small, and thus the potential at the level of the earth- 
connected sphere can be calculated from the charge upon it. The new method of 
mounting the sphere is shown in fig. 1. 
The sphere, 30 cm. in diameter, is supported on an 
earth-connected brass tube B, 2 cm. in diameter, from 
which it is insulated by sulphur-coated ebonite E ; 
insulators are indicated in the figure by the dotted 
areas. The tube is inserted within a wider one C which 
extends from the top to the bottom of the sphere and 
which is open below. The supporting tube B is rigidly 
fixed in a hole bored through the screw cap which 
closes the upper end of an iron pipe P, 5 cm. in 
external diameter and 427 cm. long, which can be 
turned about its lower end from the vertical to a nearly 
horizontal position as described in the former paper. 
The length of the brass tube from the top of the iron 
pipe to the bottom of the sphere is 38 cm. The Fi G I 
connection between the sphere and the electrometer is 
made by means of a tightly stretched wire W supported 
by quartz insulators. The wire is not attached directly 
to the sphere but to a brass disc D insulated from the 
supporting tube and fitting loosely within the wider tube 
C inside the sphere. The sphere is fixed to the disc by 
means of a screw which projects from its inner surface 
and can thus readily be removed to give access to the 
insulation. 
When the sphere is exposed by raising the iron pipe 
to its vertical position the height of its centre above 
the ground is 480 cm. 
The sphere when lowered is received in a metal-lined 
earthed box resting on the ground; a tightly fitting 
cover, also metal lined and earthed, protects the sphere 
from the atmospheric electrical field and from the 
weather. The charge on the earthed sphere in this position is taken as zero. 
The charge Q on an earthed sphere of radius R at a height h above level ground 
is assumed to be such that Q/R— Q/2h + Y = 0, where Y is the undisturbed air 
potential at the height h. The presence of the neighbouring hut exerts a disturbing 
influence which however is not large : the correction to be applied has been estimated 
by imagining the hut to be replaced by a conducting hemisphere large enough to 
enclose it. The vertical potential gradient over level ground being assumed uniform 
throughout a height exceeding that of the hemisphere, the lowering of potential at 
M 2 
