156 Atmospheric Origin of the Aurora^ &pc. 



dian, yet the author has not confounded them, but has shown 

 that the former differ from the latter in the absence of the dip. 



But the analogy is not restricted to position. It was soon de- 

 tected in the concomitant phenomena. I have shown, by tabular 

 views, that the thermometer usually begins to fall, and the baro- 

 meter to rise, several days before each, and rain or snow to de- 

 scend within one, two, or three days after them. In the cases 

 subsequently presented, in which the number of hours between 

 the aurora and storm has been carefully noted, I have usually 

 found that the time has been about thirty six hours, and that 

 there! is a curious exception in the case of two auroras on two or 

 three consecutive nights, in which case, the rain or snow is less 

 likely to descend, or is deferred till nearly the usual time after 

 the last. The same is true of the polarized clouds, and of halos ; 

 in both of which, vapor, which had unquestionably been precipi- 

 tated, is redissolved, or otherwise disposed of, during the time 

 and under the influence of the circumstances preparatory to or at- 

 tendant on the second exhibition. 



This interference of one aurora with the results of its prede- 

 cessor, opens a curious field of investigation, discloses a new anal- 

 ogy between this and meteors of a confessedly aqueous origin, 

 and refers to a general law the observed exceptions to the descent 

 of precipitated vapor which so generally takes place after an au- 

 rora. In almost every instance in which this has been deferred, 

 there have been traces of auroral action on the succeeding night, 

 though sometimes masked by the moon. The following rule has 

 had few exceptions, viz. If the evening of the day after an au- 

 rora is totally clear, no storm follows on the second day ; and 

 conversely, if no storm is to follow, this evening is totally or 

 nearly clear. This general clearness is itself one of the usual 

 attendants of auroral action ; and I have for many years observed, 

 that the morning following an aurora is, in this respect, remarka- 

 ble, as compared with other mornings. In this fact, and in the 

 unusual clearness of the night of the meteor' — with the exception 

 of some peculiar, transient clouds — we have proof of the influ- 

 ence of an aurora, or the circumstances Which precede and attend 

 it, in effecting the resolution or disappearance of visible vapor or 

 precipitations. • 



This enables us to explain or generalize the fact of the non- 

 appearance of the storm, of which the first of two consecutive 



