348 



late, and to have sufficient light to render the earlier of the faint 

 aurorse invisible; about the end of the first quarter, it does not set 

 till midnight, and thus shines throughout the period of the occurrence 

 of five-sixths of the aurorse ; afterwards it increases in brightness, 

 and the maximum effect in extinguishing faint aurorse is evidently 

 attained at opposition, when the moon begins to rise late enough to 

 allow the earlier aurora; to be visible ; towards the end of the third 

 quarter, when the moon does not rise till midnight, it is also evident 

 that the number of faint aurorse rendered invisible mast be very 

 small. From the beginning of the fourth quarter, therefore, till 

 conjunction, the numbers seen will obey nearly the true law of fre- 

 quency ; and as the visible maximum occurred before the end of the 

 third quarter, the true maximum must have occurred even nearer 

 to opposition. On the whole, it appears very certain, that the hy- 

 pothesis of an actual maximum of frequency at opposition, and mi- 

 nimum at conjunction, is satisfied by the previous numbers of aurorse, 

 seen under the conditions of the varying duration of moonlight for 

 the hours of maximum frequency. This hypothesis is in unison 

 with the law of magnetic distui'bance, which is a maximum at oppo- 

 sition, and a minimum at conjunction. 



Note on the Theory of the Aurora. 



Although temptations to frame hypotheses have been avoided 

 hitherto, I cannot refrain from repeating here the opinion, that the 

 phenomena of the aurora borealis are chiefly optical. 



After watching the various phases of the aurora for some years, 

 the hypothesis of self-luminous beams and arches appeared to me 

 unsatisfactory ; and the strongest argument in its favour, that ob- 

 tained from the computed height of the auroral arches, seemed of a 

 very doubtful character. I was quite prepared, therefore, to adopt 

 the idea, first I believe proposed by M. Morlet to the French Aca- 

 demy, in May 1847, that the auroral arch is an optical phenomenon 

 of position. M. Morlet has pointed out that the arch appears gene- 

 rally as a segment of a circle ; whereas, in these latitudes, it ought 

 invariably to appear as the segment of an ellipse, if the hypothesis 

 be true of a real luminous ring, with its centre on the continuation 

 of the magnetic pole. He has also, among many other very obvious 

 objections to that hypothesis, shewn that the summit of the arch is 

 generally in the magnetic meridian of the place, the plane of which 



