300 DR. C. CHREE: ATMOSPHERIC ELECTRIC POTENTIAL RESULTS AT KEW 
much shorter than under the old arrangement, thus reducing the risk of defective 
insulation. Again, the down tube from the tank to the horizontal discharge tube is 
about 5|- feet long, instead of a few inches as formerly, so that the pressure under 
which the jet issues is practically constant. 
Lastly, the jet is emitted parallel to the wall of the building, instead of perpendicular 
to it as before. This tends to increased uniformity in the conditions, as the potential 
varies rapidly m the direction perpendicular to the building, but little, if at all, 
in a horizontal direction taken parallel to it. The discharge tube, which projects 
from a window in the west wall, is situated about 10 feet to the north of the old tube, 
but nearly at the same level above the ground. 
A flower border used to run the whole length of the west wall, containing shrubs 
3 or 4 feet high. The northern half of this border was cleared of shrubs and laid 
down in grass, which has been kept short the whole year round. 1 hese changes were 
made with the object of securing that no appreciable fictitious element should be 
introduced into the annual change of potential gradient. 
A definite test of the insulation of the apparatus was introduced, and is applied 
each morning. It is not always possible to improve defective insulation immediately 
when discovered, but at the worst the test shows when a day s record is to be 
regarded with suspicion. 
The surface of the ground outside the building is not absolutely level, so the height 
of the spot where the jet breaks into drops can hardly be given with mathematical 
precision. It is very approximately 11 feet (3‘35 metres). The perpendicular distance 
from the general outline of the west wall is 50^- inches (127’3 cent inis.). 
§ 2. If these distances, the environment, and the insulation of the apparatus could 
be kept absolutely uniform, the potential recorded on the electrograms would only 
require multiplication by an absolute constant to become converted into potential 
gradient per metre of height in the open. This ideal is not, however, practically 
attainable, and experience showed the desirability of obtaining a multiplier by 
reference to absolute eye observations made regularly in the open. 
The eye observations are taken with a Kelvin portable electrometer, supported on 
a small level platform having three vertically projecting points, which prevent the 
electrometer from slipping. The platform is carried on a brass rod, fitting inside a 
hollow brass tube sunk in the ground. The rod can be raised in the tube and clamped 
at any convenient height. The original idea was to observe at two heights, a given 
distance apart. Having regard, however, to time and other considerations, it has 
been judged best not to vary the level, and the electrometer fuse has been used in a 
horizontal position at a fixed height above the ground. Thus the potential observed 
has referred to a fixed level, approximately L465 metres above the general level of 
the ground in the immediate vicinity. The ground is practically flat for a considerable 
space round the stand. The adjacent grass is kept short, and there is no tree, shrub, 
or building within what seems a reasonably satisfactory distance, 
