2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I ID 



METHODS 



The investigations, of which we shall describe on the following 

 pages only a few results, are characterized by a strict distinction and 

 completely separate statistical treatment of those different kinds of 

 solar-radiation eruptions which are known to influence the ionosphere. 

 All computations were made exclusively on the basis of daily values 

 of the solar, ionospheric, and meteorological elements. The method 

 used throughout the whole work was the so-called "superposed-epoch" 

 method. The application of this method takes place as follows : At 

 first a certain number of well-defined key dates are selected from 

 a series of observations, as long as possible, of that element (e.g., 

 geomagnetic activity), which is regarded hypothetically as controlling 

 another element. Then a mean value "n" is obtained by averaging 

 arithmetically all those values of the element, assumed as controlled 

 (e.g., sea-level air pressure), which belong to these key days. The 

 same process is repeated for several days which precede the key days 

 ("n— I," "n — 2," etc.) and for several days which follow the key 

 days ("n-{-i," "n-|-2," etc). In such a way there is obtained the 

 typical average behavior — free from the accidents of the individual 

 case — of the chosen meteorological element, before, during, and after 

 the time of the solar or ionospheric event which is assumed to control 

 this element. From this description it can be seen that the "super- 

 posed-epoch" method is a simplified correlation method which has 

 the advantage that it is applicable to the most widely varying forms 

 of correlations without modification. Random variations are elimi- 

 nated automatically by this method, if the number of key days is 

 sufficiently high. A criterion of reality consists in dividing the whole 

 statistics in parts of equal size, e.g., even years and odd years, 

 summer and winter, and comparing the resulting average curves of 

 these parts with each other. The selection of the key days is often 

 done by taking the five highest and the five lowest values of every 

 month. 



I. RELATIONSHIP BETWEEN SOLAR PARTICLE INVASIONS AND 

 SEA-LEVEL PRESSURE 



A. Origin of Particle Invasions and of Ionospheric Storms 



The particle invasions are generated by eruptive processes in the 

 so-called M-regions of the sun (J. Bartels (6)). They are not closely 

 connected with sunspots, photospheric faculae, bright chromospheric 

 flocculi or plages, or bright chromospheric eruptions or solar flares. 

 The earth is affected by these invasions only if the place of the M- 



