470 REPORT—1904. 
these observers describe the results of their investigations with regard to the effect 
of the soil upon the loss of electrification in a neighbouring charged conductor. 
They had previously been led to the conclusion that the conductivity of the air 
is due to the presence of free ions, and the object of their research was to discover 
the source from which free ions arose. That they might proceed from the surface 
of the ground under suitable conditions there was no reason to doubt. 
In the case of certain calcareous soils, the conductivity of the adjoining air 
was very remarkable, and, in general, air which had been for a time in immediate 
contact with soil acquired increased conductivity. 
Partly for the purpose of repeating these experiments, which had attracted 
attention abroad, and in part for the purpose of discovering the nature of the rela- 
tion between the conductivity of the air and the principal atmospheric phenomena, 
a series of daily observations was made on the Kast Coast of Aberdeenshire from 
January to May of the present year. A statement of some part of the observa- 
tions recorded, together with a description of the instruments employed, has 
already been published (‘ Elec. Review,’ March and April 1904). 
It was, however, at the conclusion of an extended series of experiments that it 
became possible to ascertain whether, and to what extent, correlations existed. It - 
is believed that the present series is one of the first that has been made in this 
country, while it is probably the largest yet attempted anywhere. 
The number of observations is sufficiently great to ensure a tolerably reliable 
result from the application of the methods of reckoning correlation. More espe- 
cially is this so, because each experiment was checked by two observers, and 
repeated if any doubt arose. 
During the first five months of the year great variety of weather conditions 
was experienced. 
The daily experiment consisted in aspirating a measured volume of air over 
the leaves of a specially constructed electroscope which had been charged to a 
certain potential. Any leakage on the part of the insulators was first measured 
and subsequently allowed for. 
An illustrated description of the apparatus originally employed, and after- 
wards somewhat improved, will be found in the ‘ Elec. Review,’ March 1904. 
In the improved form of electroscope the leaves are unequal in size, so that the 
heavier one hardly moves when a small charge is imparted. Both leaves are 
suspended by the finest and best silk fibres over a diaphragm, which concentrates 
the current of aspirated air upon them. , 
Now it was evident as the experiments proceeded that some relation existed 
between the height of the barometer and the conductivity of the atmosphere, but 
yet on several occasions the electroscope showed insignificant loss of charge, not 
only when the pressure was high but also when it was low. Further, in such 
cases the barometer was, as arule, steady for a time. To this there were some 
exceptions, On the contrary, a greater fall of potential had from time to time 
been observed when the barometer was falling than when it was rising, and this 
occurred both with high and with low pressures. 
Hence it seemed almost likely that the conductivity of the air was dependent 
on the variation of barometric pressure. A calculation was accordingly made 
(following generally Professor Karl Pearson’s long method), and it resulted in 
the discovery of this fact, that the conductivity of the air is increased by a fall of 
pressure, and is correlated with daily variations of pressure to the extent of 37:4 
per cent. 
In the subjoined table the fall of the leaves of the electroscope (which were 
charged on the average to a divergence of about 1 cm.) is stated on the horizontal 
columns in 100ths of a mm. per tenth of cubic foot of air aspirated. 
The rise or fall of the barometer during the day preceding an observation is 
stated in twentieths of an inch on the vertical columns. 
The mean fall of the leaves for 114 observations was 11:26 hundredths of 
amm., and the §.D. for the series 11:73. 
The mean variation of pressure for the period was almost exactly zero, and the 
S.D. was 4°68, 
