319 
1912-13.] Atmospheric Electric Potential Results. 
level, and instead of discussing the variation of the potential at the point 
of observation, it is usual to discuss the variation of the ground value of 
the atmospheric potential gradient. 
With a view to finding the factor by which the potential on the tank 
must be multiplied in order to get the potential at one metre from the 
ground, observations were frequently taken on the grass in front of the 
laboratory. In making these observations of the absolute magnitude of the 
potential gradient, a Wilson * Universal electrometer was used. This 
instrument measures the charge induced on a plate, and from this the 
potential gradient in volts per metre can he calculated. The observations 
were usually made at intervals of a minute for about an hour. These were 
averaged, and the mean compared with the mean derived from the electro- 
graph record, and thus the factor was obtained which reduced the potential 
readings at the water-dropper to the corresponding ground value of 
potential gradient. A flame collector and a Kelvin multicellular electro- 
static voltmeter were also used. At first the flame collector was a spirit 
lamp on a glass stand, while later a Lutz -f flame collector with an ebonite 
stand has been used and has the advantage of being serviceable even in a 
fairly high wind. 
A slightly different method has also been employed to obtain the factor. 
Simultaneous observations of the potential on the tank and at a known 
height above the ground have been made in order to see how closely the one 
curve follows the other. Two observers took simultaneous readings of an 
Exner electroscope which was connected to the Lutz flame collector placed 
out in the open and of the Kelvin voltmeter connected to the electrograph 
tank. In figs. 1a and 1b the values of the readings of the flame collector 
electroscope and of the voltmeter are plotted against the corresponding 
times ; the first pair of curves was obtained on a day when there was 
comparatively little wind, although there was considerable variation in the 
value of the potential during the time of observation, while the second pair 
was obtained on a day on which there was a strong wind and rapid changes 
in the value of the potential. A comparison of the curves shows that the 
variation of potential on the tank follows very closely that given by the 
flame collector : it will be noticed that the tank shows a tendency to lag 
behind the flame collector. This lag is chiefly due to the difference in 
electrical capacity of the two systems and the relative efficiency of a flame 
and a water-dropper as collectors. The closeness of the agreement between 
the curves shows that the use of a factor to deduce from the value of the 
* Proc. Gamb. Phil. Soc., xm. iv. p. 184, 1905 ; also xiii. vi. p. 363, 1906. 
t Munch. Ber., xxxvi. p. 507, 1906. 
