LONG-WAVE RADIO TELEPHONE TRANSMISSION 685 



of magnetic activity and solar limb-prominences than in the sunspot 

 and ultra-violet curves. The curve of reception at Washington, D. C, 

 of signals from several European telegraph stations on longer wave- 

 lengths (15 to 30 kc.) followed the Houlton and Cupar curves fairly 

 well except in 1931. 



The decrease in long-wave field strength apparently lagged a year or 

 two behind the decrease in sunspot-numbers. This may perhaps be 

 explained by the lower heliographic latitude of the sunspots near the 

 minimum of the solar cycle, offsetting the decrease in number of spots. 



Correlation by Months 



Monthly averages of the same phenomena as shown in Fig. 2 are 

 plotted in Fig. 3, omitting the curve of radio telegraph reception at 

 Washington. There seems to be little or no obvious correlation 

 between the field strength curves and the other curves except for the 

 one of magnetic activity. Sunspots in the central portion of the solar 

 disc are believed to have the major influence on terrestrial phe- 

 nomena,^' 5. 6, 7 g^j^fj therefore the sunspot-numbers employed in this 

 analysis are those for the central zone * of the sun. Solar limb- 

 prominences would hardly be expected to have considerable influence f 

 on the terrestrial phenomena, although some relation is suggested by 

 Fig. 2. Monthly averages of ultra-violet radiation also ofTer little 

 explanation of the variations of long-wave fields. 



Magnetic activity easily gives the best correlation with long-wave 

 fields on a month-to-month basis. A scatter diagram was constructed 

 in which the West-East and East-West monthly averages of daylight 

 long-wave radio transmission were correlated with terrestrial magnetic 

 activity for the years 1927-1932. For WNL the coefficient of correla- 

 tion 8 is 0.526 ± 0.059; for GBT and GBY, 0.747 ± 0.067, indicating 

 a high degree of correlation. Such correlation means either that 

 magnetic activity afTects radio transmission or that both are affected 

 similarly by a common cause of disturbance.^ 



The coefficient of correlation of magnetic activity with the monthly 

 averages of WNL and 2XS at New Southgate in 1923 and 1924, at 

 Chedzoy in 1925, and at Wroughton in 1926, 192/, and 1928 is 0.38 ± 

 0.10. 



* From 1917 to 1928, inclusive, the "central zone" of the sun was defined as that 

 part of the sun's surface included between two meridians situated 30° on either side 

 of the central meridian. (The central meridian is that meridian of the sun which 

 bisects the sun's disc.) Beginning in the year 1929 the central zone is defined as the 

 area on the sun's disc enclosed by a central circle having a diameter half that of the 

 disc. 



t Solar limb-prominences may have a more direct relation to sunspots without 

 having any significant relation to terrestrial phenomena. 



