32: 
DR. C. CHREE: ATMOSPHERIC ELECTRIC POTENTIAL RESULTS AT KEAY 
Connection of Atmospheric Potential with Meteorological Elements. 
§ 18 . There are so many possible ways in which a connection may subsist between 
electric potential and a meteorological element that it is difficult to know how to 
attack the problem. Suppose, for instance, that a particular June had a very high 
mean temperature, then there might be an effect either on the mean value of P, 
or on the amplitude of the daily changes, or on both. Possibly, however, there 
might be no appreciable effect on either quantity, and yet an appreciable difference 
might exist between the electric phenomena of the hotter and colder days of the 
average June. The former phenomenon would usually be interpreted to mean that 
no direct connection exists between potential gradient and temperature; the latter 
would naturally be interpreted in the opposite sense ; and yet neither interpretation 
would be justifiable without further investigation. A very warm June would differ 
from the average June in other respects besides temperature, and a true direct 
temperature effect might be practically annulled by some other effect. Again, during 
an ordinary June, a hot and a cold day usually present notable differences in other 
meteorological elements besides temperature. 
The first method I have applied is to divide the 7 Januarys, 7 Februarys, and so 
on, into two groups, one of three the other of four, according to the size of the mean 
potential gradient, and to find the difference between the meteorological elements for 
these two groups, taking first the selected days only, and then all the days of the 
month. The sub-division of each seven months was done from consideration of the 
mean potential only, the choice being decided by whether the biggest gap in the 
values of P occurred between the third and fourth, or between the fourth and fifth 
on the list, when the months were arranged in the order of the size of P. It is 
necessary to deal with the months of the year separately. The combination of days 
from all seasons of the year inevitably leads to error. The cold days, for instance, 
then come mainly from winter, and the hot days from summer, and the effects of the 
regular annual variation get mixed up with any true temperature effect. 
Table XII. gives the results of the comparison; s stands for the mean from the 
selected days only, a for that from all days of the month. The excess is shown for 
P as well as for the meteorological elements, as the size of the potential difference 
should be taken into account in considering the part that chance may play in the 
results. 
