BRIGHT CLOUDS ON A DARK NIGHT SKY. 



35 



3. Of Aurora Spectroscopically observed in Edinburgh at the Astronomers' 



House in 1S82 and 1883. 



The concluding remark in Mr Webster's letter, that the date of his stated Aurora, April 8, 1882, 

 was followed by a considerable depression of the atmosphere, is perfectly true ; for the record of 

 Barometer (uncorrected), wind, rainfall, and clouds, kept at the Royal Observatory, Edinburgh, in 

 connection with Time signalling, runs thus from April 5 to April 15, 1882 : — 



Day. 



Barometer. 



Wind. 



Miles per 



hour. 



Wind. 

 Direction. 



Rainfall, 

 Depth of. 



Clouds. 

 to 10. 



April. 



Inches. 







Inch. 





5 



29-99 



3 



N.E. 



•000 



10 



6 



30-10 



2 



N.E. 



•013 



6 



7 



3023 



4 



N.E. 



■000 







8 



30-22 



2 



N.E. 



•000 







9 



30-02 



2 



N.E. 



•000 



3 



10 



29-77 



3 



N.E. 



•000 







11 



29-58 



2 



N.N.E. 



•000 



10 



12 



29-45 



3 



S.S.E. 



•000 



9 



13 



29-06 



8 



E. 



•035 



10 



14 



28-83 



10. 



N.E. 



•424 



10 



15 



2929 



1 



N. 



•230 



10 



That circumstance, however, does not prove that the luminosity Mr Webster observed in the sky 

 at Gordon Castle was a real Aurora. It certainly could not have been one of the grander Aurora?, or 

 it would have been noted at some of the 23 other stations ; would have imprinted itself indelibly on 

 the observer's memory, and would have marked itself on the continuous photographic curves of the 

 magnetic needles at Greenwich. 



Whether it was, however, a faint Aurora, or perhaps some other luminous manifestation, I do not wish 

 to suggest a word either for or against ; but if the observer could have said that he spectroscoped it 

 with a pocket spectroscope, and saw the Auroral Citron line, — I should have accepted the testimony 

 immediately. For I myself have never yet spectroscoped any decided Aurora, without its showing 

 that line ; and have never seen that line in any other light, though I have, on one occasion, elicited 

 the line out of a full-Moon, mid-night sky, when at the time there was no appearance of Aurora or 

 any of its usual accompaniments, — but where, an hour before, there had been some unmistakable 

 needle-shaped jets and darts of its light on the Northern horizon. 



The opportunities I have had for good numerical observations of the Auroral spectroscopic line, 

 during the last 1 5 months have not been many ; but may form a useful addendum to the present 

 paper, — as follows : — 



O October 22, 1882, at ll h p.m. ; the moon ten days old, and the sky bright with moonlight; 

 Aurora seen as a faint but regularly-shaped arc on the N.W. horizon ; spectrum place of its Citron line 

 = 45 511 of Wave Number per Brit. Inch. 



h October 28, 1882, at midnight; past full-Moon; sky clear and frosty. No Aurora visible to 

 the naked eye at the time, though there had been earlier in the evening. The place of Aurora line, 

 doubtfully observed, came out 45 350 (?) W.N. Place. 



$ November 14, 1882. Aurora lasted through much of the night, but chiefly behind clouds. 



? November 17, 1882. A grand Aurora; but barely seen here by reason of clouds, smoke, and 

 direction. 



h November 25, 1882, ll h 30 m p.m.; Moon just past the full. No Aurora to naked eye, 

 but its line was clearly seen in the spectroscope, followed by a broad band of continuous spectrum of 

 the Moonlight. Place of the line observed at 45 593; while the continuous band of Moonlight 

 began at 46 610, culminated in intensity near 48 200, and ended gradually near 52 610 W.N. Place. 



([ March 26, 1883, at 10 h p.m.; a rather bright Auroral arc northwards, its central portion 

 rising 10° or 15° high, and having dark shade underneath. Its Citron line's place measured 45 558 

 W.N. Place. 



$ March 27, 1883, at 10 h p.m.; a long low quiescent, or blandly-shining faint arch of Aurora 

 on the northern horizon. Its Citron line's spectrum place measured 45 542 W.N. Place. 



VOL. XXXII. PART I. F 



