148 Atmospheric Origin of the Aurora, 6fc. 



As they are all preceded by a depression of atmospheric tem- 

 perature below the mean, and by an augmentation of pressure 

 greater than that which precedes the fall of snow or rain at times 

 when none of these meteors have recently appeared, there is ad- 

 ditional evidence of the similarity of their origin. 



Proposition 7th. Theoretical inference in relation to their alti- 

 tude. 



We may infer from the last two propositions that a magnetic 

 cirrous cloud is higher than a halo, but lower than a lunar column, 

 and the latter lower than the aurora borealis. 



Proposition 8th. Pi^actical infet^ence vnth regard to the prog- 

 nostication of storms. 



The foregoing propositions which relate to pressure and tempe- 

 rature may suggest a rule for predicting storms much earlier than 

 by other methods ; inasmuch as these changes, and especially 

 that of the barometer, take place even more generally than those 

 opposite changes which often occur within the twenty four hours 

 immediately preceding the storm, and which have been observed 

 by others, and generally regarded as among the surest indications.' 



The above propositions are deduced from tables here omitted, 

 and are founded upon the observation of forty auroras, twenty 

 two auroral clouds, seen in the day time, seventeen halos, and 

 four luminous columns. The propositions in relation to the last 

 and more rare phenomenon, the author considered as entitled to 

 less confidence on account of the small number observed. Yet 

 the optical theory which he gave of it in which he attributes it 

 to a mixture of horizontal, specularly-reflecting, crystalline plates, 

 with masses which are more amorphous and which produce a 

 reflection virtually radiant, he considers as complete and satisfac- 

 tory, and corroborated by his observations on the crystals which 

 subsequently descended. The author has observed the aurora in 

 connexion with the above and other meteorological phenomena 

 of the same, the preceding and the succeeding days, and endeav- 

 ored to trace their respective and relative changes, and as far as 

 the subject admitted, by the statistical and numerical method. 

 This is a fertile field, and comparatively unoccupied. 



In the 3d, or class of inferences, he has endeavored to show a 

 connexion between the aurora boi'ealis and the crystallization of 

 snow. 



