ON THE PERIODICITY OF THE AURORA BOREALIS. 347 



century and in the earlier part of the present century (allowance being made for 

 the increased number of observers and the multiplicity of records), -without a similar 

 cause existing, another explanation seems to be required for this periodicity. 



Some connection between the aurora and the indications of the barometer was 

 hinted at many years since by Madison.* Recently, the discussion of long series 

 of barometrical observations, in a particular way, has revealed a period -of fifty-six or 

 sixty years.t But this periodicity will not help us to explain the secular periods of 

 the aurora. Of little avail, also, for this purpose, have proved those secular changes 

 in temperature which the delicate scrutiny of multiplied observations manifests: for 

 example, the oscillation between warm winters and cold winters in a period of about 

 forty-one years, the cold reaching its height in 1830, and again (by prediction) in 

 1871.J 



* Philosophical Transactions of Philadelphia, II. 142. Trans. Irish Academy, II. 71, 72. 

 f Vierteljahrsschrift der Naturforsch. Gesellsch. in Zurich, VIII. 106. 

 X 'tVochenschrift fur Astronomie, Meteorologie, etc., IV. 162. 



