134 



COSMOS. 



suits, deduced from 6000 observations, made from the middle 

 of May, 1806, to the end of June, 1807, the oscillation gene- 

 rally fluctuated only from one-half of a graduated interval to 

 the other half, amounting therefore only to 1' 12"; in indivi- 

 dual cases, and often when the weather was very stormy and 

 much rain was falling, the needle appeared to be either per- 

 fectly stationary, or to vary only 0.2 or 0.3 of a graduated 

 interval, that is to say, about 24" or 28". But on the occur- 

 rence of a magnetic storm, whose final and strongest mani- 

 festation is the Aurora borealis, the oscillations were either in 

 some cases only 14' and in others 38' in the arc, each one 

 being completed in from 1^- to 3 seconds of time. Fre- 

 quently, on account of the magnitude and inequality of the 

 oscillations, which far exceeded the scale parts of the tablet 

 in the direction of one or both of its sides, it was not pos- 

 sible to make any observation. 69 This, for instance, was the 



fi9 The month of September, 1806, was singularly rich in great mag- 

 netic disturbances. By way of illustration, I will give the following 

 extracts from my journal : 



Sept. 1806, from 4h. 36m. A.M. till 5h. 43m. A.M. 



If- 



f- 



4h. 40m. 7h. 2m. 



3h. 33m. 6h. 27m. 



3h. 4m. 6h. 2m. 



2h. 22m. 4h. 30m. 



2h. 12m. 4h. 3m. 



Ih. 55m. 5h. 27m. 



Oh. 3m. Ih. 22m. 



The disturbance last referred to was very small, and was succeeded by 

 the greatest quiet, which continued throughout the whole night, aud 

 until the following noon. 



f Sept. 1806, from lOh. 20m. P.M. till llh. 32m. P.M. 



This was a small disturbance, which was succeeded by great calm 

 until 5h. 6m. A.M. 3 T Sct.S 1806, about 2h. 46m. A.M. a great but 

 short magnetic storm, followed by perfect calm. Another equally great 

 magnetic disturbance about 4h. 30m. A.M. 



The great storm of ff September had been preceded by a still 

 greater disturbance from 7h. 8m. till 9h. llm. P.M. In the following 

 winter months there was only a very small number of storms, and these 

 could not be compared with the disturbances during the autumnal 

 equinox. I apply the term great storm to a condition in which the 

 needle makes oscillations of from 20 to 38 minutes, or passes beyond all 

 the scale parts of the segment, or when it is impossible to make any 

 observation. In small storms, the needle makes irregular oscillations of 

 from 5 to 8 minutes. 



