4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IIO 



Spot number is >40, and years with low sun activity, where the an- 

 nual mean of the relative sunspot numbers is ^40 ; the years with even 

 and the years with odd numbers ; and, finally, the different seasons. In 

 the pursuit of similar geophysical investigations, e.g., those performed 

 by J. M. Stagg (8), we divided the year in three parts, as follows: 

 winter (November, December, January, and February), spring + 

 autumn (March, April, September, and October), and summer (May, 

 June, July, and August). 



The sea-level pressure data used in that part of the investigation 

 were observed during an uninterrupted series of 11,688 single days. 

 The beginning on January i, 1906, and the end on December 31, 1937, 

 of that series were established by the fact that for the time before 

 January i, 1906, there were not available sufficiently reliable geo- 

 magnetic character numbers, and for the time after December 31, 

 1937, there were not yet published daily mean values of sea-level 

 pressure for the stations that we used. 



As a very detailed elaboration of the data from Potsdam (Ger- 

 many) and Stykkisholm (Iceland) had revealed that a conspicuous 

 relationship between ionospheric storms and sea-level pressure could 

 be demonstrated only for the years with low sun activity, when r 

 is ^40, and even then only during the winter months, the studies were 

 confined to the 16 years with low sun activity (1910-1914, 1920-1924, 

 and 1930-1935) and to the winter months November, December, Jan- 

 uary, and February. In addition to the above-mentioned, the data 

 of eight further stations were investigated. These places (De Bilt, 

 Karlsruhe, Vienna, Breslau, Konigsberg, Warsaw, Lemberg, and 

 Kiev), as well as Potsdam, are all situated within the zone 45° to 55° 

 N. latitude and 5° to 30° E. longitude. Computed by means of the 

 superposed-epoch method, figure i shows for these stations the aver- 

 age behavior of sea-level pressure as related to all those days (320) 

 when the ionosphere was particularly disturbed, and to all those days 

 (320) when the ionosphere was particularly undisturbed. Besides 

 this, the curves demonstrate the average behavior of sea-level pressure 

 on those 3 days which precede the key days, and on those 11 days 

 which follow the key days. It can be seen from figure i that at all 

 stations the sea-level pressure is lower than normal after the days 

 when the ionosphere is particularly disturbed, with a minimum value 

 3 days after, and higher than normal after the days when the iono- 

 sphere is particularly undisturbed, with maximum value 3 to 4 days 

 after. The different behavior of sea-level pressure after ionospheric 

 storms and after ionospheric calms is especially clearly demonstrated 

 by means of difference curves, "disturbed minus undisturbed," which 



