320 DR. C. CHREE: ATMOSPHERIC ELECTRIC POTENTIAL RESULTS AT IvEW 
the largest over the smallest hourly value in individual days, the noil-cyclic effect not 
being eliminated. The fifth column shows the ratio borne by the mean range thus 
found for the month or season to the corresponding mean value of P. 
The third and sixth columns similarly relate to the non-cyclic effects treated 
numerically. 
The unit in the first three columns is, as in Tables I. to VI., 1 volt per metre. 
The last column gives the ratios borne by the ranges in the second column to the 
corresponding amplitudes of the regular diurnal inequality as given in Table III. 
The monthly means in the first three columns of Table X. were really calculated to 
one significant figure beyond that recorded, and these more exact results were 
employed in deducing the seasonal means and the ratios. 
Several interesting and novel results appear. 
The mean of the largest and smallest hourly values exceeds in every case the true 
mean from the 24 hours; but the excess is not large, and, though slightly less in 
winter than in the other seasons, varies comparatively little throughout the year. 
The mean of the individual daily ranges is, on the average, no less than 2% times 
the amplitude of the regular diurnal inequality. 
The ratio borne by the irregular to the regular range is distinctly greater in the 
winter and equinoctial than in the summer months, showing a fairly well marked 
maximum in October and November. 
Taking the year as a whole, the mean of the individual daily ranges exceeds the 
mean value of P by about 10 per cent.—a somewhat remarkable fact; the ratio borne 
by the former quantity to the latter falls slightly below unity in three of the winter 
months, but seems nearly uniform from March to September. 
The (numerical) non-cyclic effect resembles the mean of the individual daily ranges 
in being largest, absolutely considered, in winter, while showing the largest ratio to 
the mean value of P in the equinoctial season. The large size of the average non- 
cyclic effect, when taken independent of sign, is an important element in determining 
the significance to be attached to the results obtained in Table II. where it is treated 
algebraically. 
Special Meteorological Features of the Selected Days. 
§ 17. In confining ourselves to 10 selected days a month we have followed a 
precedent set by the annual selection of 5 magnetically “ quiet ” days a*month by the 
Astronomer Iioyal. In the case of the magnetically quiet days, so far as is known at 
present, there is no systematic departure from average conditions in anything but the 
phenomena of Terrestrial Magnetism. In the present case, however, it is otherwise. 
We have eliminated all the rainy days, which constitute a considerable proportion of 
the days in most winter months, and not infrequently even in summer months. The 
meteorological conditions on the average selected day thus represent a climate which 
at some seasons of some years is far from being representative of Kew. It seemed 
