18g ANNUAL REPORT SMITHSONIAN INSTITUTION, 1937 



directly overhead. For the assumed conditions, the center would be 

 about 1,800 feet long. Some centers are obviously much greater in 

 extent. 



Our present knowledge of the shape and extent of lightning dis- 

 charge centers is too meager to establish a "usual" size for centers, if 

 such a size actually occurs. They must vaiy considerably from storm 

 to storm. It is important that their range be determined, for upon 

 their shape and size depends the nature of the electrical field near the 

 cloud. This electrical field in turn probably hiitiates the lightning 

 stroke. 



When the discharge center is of limited dimensions compared with 

 its height above ground, the electrical field will be much more intense 

 near the cloud than near the ground. In such a case, it would not be 

 expected that earthed objects would have much effect in initiating a 

 lightning stroke. If, on the other hand, the dimensions of the center 

 should be large compared with the height of the center above ground, 

 much more importance would have to be attached to the effect of 

 grounded objects in causing stroke?. This information is obviously of 

 considerable importance in lightning protection problems. 



The polarity of the discharge center near the front of a storm accord- 

 ing to Simpson's theory should be positive, while according to Elster 

 and Geitel's and Wilson's theories it should usually be negative. 

 According to Banerji it should consist of a negative front center 

 followed by a positive center. Although it has been shown that 

 95 percent of the strokes to transmission lines are negative, yet meas- 

 urements of electrical field changes resulting from lightning strokes 

 indicate a more nearly equal distribution of positive and negative 

 strokes. Indirect measurements by Norinder ^^ of the polarity and 

 magnitude of lightning currents, made by measuring the electromag- 

 netic fields caused by strokes, likewise show a considerably higher 

 percentage of positive strokes than is indicated by measurements of 

 currents in transmission towers. Further work seems necessary to 

 determine the relative frequency of occurrence and activity of positive 

 and negative discharge centers. 



The measurements made of lightning strokes to transmission lines, 

 using magnetic Imks as a means of measurement, have indicated in 

 many cases reversal of current. These results, together with those of 

 Norinder, coupled wnth a consideration of the probable location of 

 cloud charges, strongly suggest the possibility of reversal of polarity 

 between successive discharges constituting a multiple stroke. This 

 reversal of polarity may well result from various tappings of positive 

 and negative accumulations of charges within the cloud. 



" Norinder, H., Lightning currents and their variations, Journ. Franklin Inst., vol. 220, pp. 69-92, July 

 1935. 



