NO. 4 sun's variation and temperatures — ABBOT 3 



In this prevailing atmosphere of skepticism I sought to buttress the 

 conclusions by correlating other evidences of solar change than the 

 solar-constant measures v^ith weather changes. These discussions are 

 given in two later papers.- In these papers I showed that changes in 

 the areas of solar calcium flocculi observed at Ebro, Spain, and changes 

 in the critical frequency of the ionic layer, Fg, in the atmosphere, ob- 

 served at Huancayo, Peru, and Watheroo, Australia, had similar rela- 

 tions with terrestrial temperature changes to those connected with the 

 solar constant. In none of these studies preceding 1946 had I been 

 led to the discovery of a regular short period in solar variation, though 

 longer periods from 8 months up had been discovered.^ 



2. DISCOVERY OF THE 6.6456-DAY PERIOD 



1 was intrigued by noticing that the temperature effects seemed to 

 commence 2 or 3 days before the beginning of a change in the solar 

 constant of radiation. As stated above, I had traced the temperature 

 effects of a single solar change for as much as 17 days. It now oc- 

 curred to me to see whether any appreciable effects lingered still 

 longer. I therefore continued a tabulation like that shown in figure 2 

 of my paper "Correlations of Solar Variation with Washington 

 Weather" (fig. i of the present paper) for 20 days farther forward. 

 To my great surprise large features of weather change continued un- 

 diminished to the end. I then tabulated 20 days farther backward, 

 so that I now had a tabulation of 65 successive days. Throughout the 

 whole interval, large features of temperature change occurred with 

 substantially equal prominence. On plotting the results in a curve, 

 there appeared to be a regular period of about 6f days in the march 

 of temperature changes. 



In order to fix the exact length of the period, I then obtained and 

 tabulated departures from normal temperatures at Washington, for 

 the months of May and November, from the year 1910 to 1945. Sev- 

 eral periods were tried. It proved that with a period of 6.6456 days 

 there would be no progressive shift at all in the phases of the curves 

 from 1910 to 1945, either for Alay or November. As these 2 months 

 are a half year apart, no period much longer or much shorter could 

 suit both of them. Nevertheless individual years presented the fea- 

 tures sometimes i to 3 days early, sometimes i to 3 days late. The 

 four or five recurrences in a single month of any year seemed often 

 to show no departures from regularity of phase among themselves, 



- Smithsonian Misc. Coll., vol. 104, No. 5, 1944, and vol. 104, No. 13, 1945. 



2 Annals, Smithsonian Astrophysical Observatory, vol. 6, p. 181, 1942. 



