298 



BELL SYSTEM TECHNICAL JOURNAL 



for the reason that more of the remote than of the near strokes of low in- 

 tensity escape observation. To find the actual incidence, a curve of the 

 apparent incidence versus the radius of the observation area may be extra- 

 polated to an area of zero radius. On account of the comparatively few 

 observations for small radii, however, such extrapolation is rather inaccu- 



75 

 70 

 65 

 60 

 55 



UJ 

 Q. 



■s. 



< 45 

 O 



40 



z 



(- 

 ? 35 



3 30 



UJ 25 



20 



15 



1.5 2 3 4 5 6 7 8 9 10 



NUMBER OF CURRENTS EXCEEDING ORDINATE 



Fig. 12 — Comparison between observed and theoretically expected number of cur- 

 rents exceeding given crest values during 108 thunderstorm days in a 21.5-mile section. 



rate. Improved accuracy is ob tamed by using theoretical expectancy 

 curves in the extrapolation as indicated by the curves in the figure. 

 These curves are derived on the assumption that the proportion of currents 

 exceeding a given crest value / is given by P (7) = e~ ^ where ^ is a constant 

 — a relation in substantial agreement with observations' — and that the 

 energy in the electromagnetic wave from the stroke current, as well as 

 that in the sound wave due to the thunder, is proportional to P/r^, where r 



