204 



NA TURE 



[August i i, 1923 



except that it is of shorter period and seems to vary 

 from cycle to cycle, Dr. Chree's correlation coefficient 

 for the Kew series 191 1 to 1921 is increased from 

 0'55 to 077, and for the less regular series 1898 

 to 1909 from 0-24 to 0'62. 



Table i.— Relation between Annual Potential Gradient of 

 Atmospheric Electricity and Annual Sun-Spot Activity. 



It will be instructive to show (Table 2) for a favour- 

 able case, Ebro, the application of formula (i). We 

 have for this station, if AP<,= (P„-P„) represents 

 the observed quantity and AP^ similarly the 

 computed quantity, expressed in percentages of 

 P„= H2-6 volts per metre : 



P,- P„=o-3i3(S- 39-6) -273(T- 1916-5), 



or AP<,=o-3i3AS -273AT. . . . (2) 



Table 2. — Comparison of Observed and Computed Values of 

 Potential Gradient at Ebro Observatory, 1911-1921. 



Let us take, for example, the change in the observed 

 values of the potential gradient P„ between 191 7 

 (year of maximum sun-spot activity) and 1921, as 

 shown in Table 2. The observed change (decrease) 

 at Ebro amounted to 39 per cent. ; the computed 

 decrease, according to formula (2), is 36 per cent. ; 

 25 per cent, being contributed by the sun-spot term 

 and II per cent, by the cycle or AT-term (see 

 columns 9, 10, 11, and 12). That this remarkable 

 decrease at Ebro between 191 7 and 1921 was a 

 world-wide phenomenon, besides being corroborated 

 by Eskdalemuir and Kew, is further shown by the 

 fact that the results of the Carnegie potential -gradient 

 observations on the oceans, all instrumental and 

 reduction constants having been most carefully con- 

 trolled throughout the various cruises of the Carnegie, 

 gave a mean value for 1921-5 -which was about 30 

 per cent, less than the corresponding value for 1917-5. 



There are no such correspondingly large changes 

 in terrestrial magnetism during a sun-spot cycle 

 as have just been shown to occur in atmospheric 

 electricity. According to my 1918 investigations ' 

 an increase of 100 in the sun-spot number would 

 correspond to a decrease in the intensity of magnetisa- 

 tion of the earth of about o-i per cent., whereas 

 for an increase of 100 in the sun-spot number the 

 normal potential gradient of atmospheric electricity 

 was increased about 30 per cent. 



The investigations thus far would indicate that 



• Terr. Mag. and Aim. Elect., vol. 23 (1918), p. 63. 



the electric conductivity X docs not vary with > 



spot activity in the same marked degree as ■ 



the potential gradient P; hence, as the curn-iri 



density of the vertical current is i =XP, we may 



say, with high degree of certainty, that the vertical 



conduction current increa-* s 



with increased sun-sp>ot a< \ - 



ivity at the rate of al)out 



3 per cent, per lo sun-sjxjt 



numbers. The bearing of 



this most interesting fact 



upon theories of the origin 



and maintenance of the 



earth's negative electric 



charge needs no elaboration 



here. 



But of ecjual importance 

 with the sun-spot effect 

 to theories is the cycle effect, which indicates that 

 the earth's negative charge, even if all periodic 

 variations and sun-spot effects are eliminated, is 

 not maintained constant but may progressively, from 

 year to year, show in one cycle a steady diminution, 

 and in another a steady increase. During the 

 present cycle, beginning with the year of minimum 

 spottedness, 1913, the total decrease may be such 

 that the potential gradient at Ebro in the approach- 

 ing year of minimum, 1923 

 or 1924, may be about 30 

 per cent, less than in the 

 minimum year of 191 3, when 

 the observed value was no 

 volts per metre. But, as al- 

 ready intimated, the cycle 

 effect may not always be a 

 diminishing one. I hope to 

 study the variations in t 

 more exhaustively by utilis- 

 ing all past data obtained 

 with the requisite care. 



Lack of space will not 

 permit describing here in 

 detail the various examina- 

 tions already made con- 

 cerning the effect of sun-spot activity on the periodic 

 variations of the potential gradient of atmospheric 

 electricity. Dr. Chree in his paper cites the results 

 drawn by me from a Fourier analysis of the diurnal 

 variation of P at Ebro for the period 1910 to 1920, 

 and he finds a correlation coefficient of 0-72 ; if 

 account be taken of the cycle effect, which is also 

 evident here, as the diurnal variation is a function 

 of the absolute value of P, then the correlation 

 coefficient is 0-96. The analysis has been extended 

 so as to include the data for 1921, which have become 

 available since the 1922 paper. 



In my 192 1 investigations, which Dr. Chree 

 apparently overlooked, I investigated the relation- 

 ship between the range of the diurnal variation of 

 the potential gradient at Ebro and sun-spot activity, 

 and found that the sun-spot variation in the diurnal 

 range between minimum and maximum wcis about 

 25 per cent., and that it increased with increased 

 sun-spot activity.* With the aid of a similar formula 

 to (i), and taking the Ebro series 1911-1921, the 

 value of s turns out to be -rO-31 for the mean of 

 the year ; i.e. an increase in the sun-spot number 

 of 100 between minimum and maximum, which 

 was about the case for the present cycle, was accom- 

 panied by an increase of 31 per cent, in the diurnal 

 range at Ebro. A similar result is found for the 

 Kremsmiinster series, 1903-1910. The average corre- 



* Terr. Mag. and Attn. Elec, vol. 26 (1921), p. 68, conclusion b, and 

 Fig. VII., fifth curve. 



NO. 2806, VOL. 112] 



