LONG-WAVE RADIO TELEPHONE TRANSMISSION 687 



are closely in phase. While the two curves of average magnetic 

 characterization (" C") are considered not so reliable * for this analysis 

 as the Ui curves, the same equinoctial maxima appear in all magnetic 

 curves. However, seasonal averages of solar limb-prominences, 

 sunspot-numbers, and ultra-violet radiation apparently have no sig- 

 nificance in relation to radio field strength. Bartels ^" found that there 

 is no real seasonal cycle of sunspot-numbers. Cupar and Houlton 

 observations were combined f for the years 1927 to 1933 to make the 

 dashed curve. 



West-East average transmission for eleven years is shown by the 

 solid curve at the bottom of Fig. 4. The February dip was recorded at 

 Cupar, Scotland, in 1931 only but was previously observed at the 

 receiving locations in England (also during four of the seven years at 

 Houlton). The logio Ui curve for 1872-1930 was derived from data 

 published by J. Bartels ^^ in a comprehensive analysis of magnetic data, 

 which he critically examined for the reality of seasonal variability. 

 The other logio Ui magnetic activity curves t cover the first ten of the 

 eleven years for which the radio data are available in one direction, 

 and the first six of the seven years for which radio data are available in 

 both directions, East-West and West-East. 



Correlation with Magnetic Storms 



The average daylight radio field strength of long-wave signals does 

 not reach a maximum until five to eight days § after the peak of a long 

 magnetic storm, nor does the average night-time field reach a minimum 

 until about the same time. This is shown (Fig. 5) by a study of 25 

 storms of four or five days duration in the four years from 1930 to 1933, 

 inclusive. The incidence of succeeding storms did not permit the 

 study to include a longer time after the storms had subsided, so that 

 less than one complete cycle of field variation is all that can be shown 

 with any confidence. Data for WNL, GET, and GEY were combined 

 to obtain the trend curves for the radio field strength. A study of 16 

 periods (in the years 1927-1933) of one and two months' duration, 

 including one to three significant magnetic storms, showed that the 

 difficulty involved in extending this kind of study to individual storms 

 lies primarily in the lack of highly accurate radio data. 



* See Appendix I. 



t See Appendix II. 



i These curves were drawn before the data for magnetic activity («i) for 1933 were 

 available. The 1933 data became available shortly before the paper was released for 

 publication but do not materially change the overall picture. 



§ C. N. Anderson '■'■ showed a similar delay in the maximum increase of daylight 

 field following severe magnetic storms in 1927 and 1928. I. J. Wymore '^ found a 

 delay of about two days for signals of somewhat longer wave-length. 



