AND ON THE ELECTRIC FIELD OF THUNDERSTORMS. 
85 
gradient being the greater in fig. 15, the two being approximately equal in fig. 16, and 
the second being the greater in fig. 17 ; in the last case the initial change is negative, 
in the two others positive. The duration of the double discharge is about one-fifth 
of a second in fig. 15, two-fifths of a second in fig. 16, and two-fifths of a second in 
fig. 17. 
Double discharges consisting of two successive sudden changes of potential 
gradient of the same sign are also not uncommon. A striking example is that of 
the last discharge shown in fig. 3, where a sudden positive change of potential 
of 840 volts per metre is followed 2'4 seconds later by a second change of the same 
sign amounting to 870 volts per metre. The discharge at 14li. 12m. 50s. (in the same 
fig. 3) was also really a double one of this type, the interval between the two 
components of magnitudes, 220 and 130 volts per metre, being about one-fifth of a 
second. 
What have been called above double discharges, it should be noted, are not 
necessarily discharges along the same track or even from the same thunder-cloud ; it 
may often be observed that lightning flashes from two different centres occur almost 
simultaneously. 
In the last three figures of Plate 5 are reproduced enlargements of records of 
multiple discharges, i.e., of records showing a rapid succession of changes of potential 
gradient of opposite sign. These were all obtained during the same thunderstorm, 
that of the afternoon of June 16, 1917. The first shows sudden changes of potential 
gradient of —9600, +4350 and —1500 volts per metre, the intervals between the 
reversals being about one-third of a second. The second shows sudden changes of 
potential gradiant amounting to —7100, +1700, —1700, +300, —1900, +700, —600, 
+ 1000, the total time occupied by the eight reversals being 2‘1 seconds. In the third 
the changes of potential gradient are —1600, +900, —1600, +1200, +700, —700 volts 
per metre, the total duration being 1'9 seconds. 
III. On the Prevailing Sign of the Sudden Changes Produced in the Potential 
Gradient by Lightning Flashes. 
The sudden changes produced in the potential gradient at the place of observation 
by the passage of lightning discharges have been more often positive than negative, 
i.e., the greater number have consisted in a sudden increase of a previously existing 
positive potential gradient or a diminution or reversal of a previously existing 
negative gradient; in other words they might be explained as being due to the 
discharge of a negatively charged cloud. Discharges producing such a positive 
change of potential gradient are called in what follows positive discharges. 
The number of positive discharges recorded in 1917 was 432, of negative discharges 
279. If the observations of 1914 and 1915 are included, the numbers are 528 and 
336, the ratio being 1’56. Of the ten days of thunder on which records were 
obtained in 1917, there were nine on which more positive than negative discharges 
