Statistics on the occurrence of long continuing current as a functionof the number of strokes per flash, the stroke order, and the duration of the interstroke interval preceding the continuing current have been derived from electric field and other records of 346 flashes from five separate studies in Florida and two separate studies in New Mexico. Electric field records of 76 flashes were analyzed for the duration of the leaders preceding strokes initiating long continuing current and for the initial electric field peak of both strokes initiating long continuing current and strokes preceding those initiating long continuing current. Only 6 percent of single-stroke flashes contained long continuing current, while 49 percent of multiple-stroke flashes did; 1.4 percent of first strokes (in single and multiple-stroke flashes) were followed by long continuing current while this percentage varied between 3 and 15 percent for subsequent strokes of order 2 through 26.

The percentage of five-stroke flashes terminated by long continuing current is about three times lower than the percentage of two-, three-, four-, six-, seven-, or eight-stroke flashes. Further, the total number of five-stroke flashes is about two to three times lower than the number of two-, three-, four-, and six-stroke flashes. The continuing current occurrence statistics for Florida and for New Mexico were, in general, very similar. Leaders that were followed by long continuing current and those that were not showed no appreciable difference in duration. The geometric mean of the initial electric field peak for strokes preceding strokes initiating long continuing current was 1.5 times larger than for regular strokes; and a similar tendency was found for the charge lowered by strokes. The data presented are interpreted physically to indicate (1) that the lightning channel to ground is cumulatively conditioned by sequential strokes, (2) that the positive charge introduced into the upper portions of the channel to ground by large return strokes can influence some salient flash characteristics, and (3) that flashes draw their charge from potentially interconnected regions each of which can support, on average, 4 strokes. The probability of finding a long continuing current after a relatively short interstroke interval, in the range from 8 to 32 ms, was about 0.35, about 3 times higher than for all intervals.

The probability of finding a stroke initiating a long continuing current following a stroke with a relatively large initial electric field peak, in the range from 8 to 15 V/m (normalized to 100 km), was about 0.32, four times larger than for all strokes. When a stroke with a relatively large initial electric field peak of 5 V/m or higher was followed by a relatively short interstroke interval of 8 to 32 ms, the next stroke intiated a long continuing current in more than half the cases. The statistics on long continuing current occurrence as a function of return stroke initial electric field peak and interstroke interval duration may be applicable for detecting the occurrence of long continuing current using presently-operating long-range lightning location systems.