184 



SCIENCE 



[N. S. Vol. LIV. No. 1392 



During the last twenty years I have known 

 of bright auroral displays being observed in 

 Arizona on only a few occasions. One of 

 these was on June 15, 1915, another one on 

 March 22, 1920, and a third on the 14th of the 

 present May. The first and the third of these 

 were observed at the Lowell Observatory, but 

 not the second one as it came during unfavor- 

 able observing weather at Flagstaff. Of the 

 two I observed, that of May 14 was much the 

 more brilliant and wonderful (and this doubt- 

 less was also a finer display than that of March 

 22, 1920). It was recognized about 8:30 

 o'clock, and rapidly increased in brightness, 

 soon displaying streamers and bright and dark 

 cloud masses with the curtain or drapery fea- 

 tures. These continued bright for some time 

 and then in more or less subdued intensity dur- 

 ing about an hour when the northern sky be- 

 gan again to show the arch and drapery effects 

 rising and it was soon evident that another 

 outburst was developing. This one then also 

 progressed rapidly and at its height near eleven 

 o'clock it was even more remarkable than the 

 first. 



The rays (for convergence of these see ob- 

 servations of Dr. Russell and Mr. Lampland 

 given herewith) and the cloud forms were 

 present in all parts of the sky at eleven o'clock. 

 Some of these in the southern sky attracted 

 my attention particularly by undergoing strik- 

 ing fluctuations in brightness. These would 

 be bright for a few minutes, would then fade 

 nearly or quite to invisibility, brighten again 

 and fade, repeatedly, in the same position. 

 There were a few small ordinary vapor clouds 

 scattered low in the east and northeast and 

 one or two in the southwest, which as reference 

 objects brought out clearly the very different 

 and fleeting behavior of the auroral clouds, 

 both the luminous and the dark, which im- 

 pressed me as being the remnants of the dis- 

 membered arch and draperies that spread over 

 the whole sky. The spreading of the auroral 

 canopy southward over the sky was most strik- 

 ing as it swung forward through the east and 

 the west. The color displayed was most notice- 

 able in the northwest where a red tone usually 

 prevailed and was at times quite strong. A 



less intense red tint was sometimes very evi- 

 dent in the northeast. I did not see the blue 

 tones sometimes reported for northern auroral 

 displays. The most of the light was of the 

 intermediate colors, silver with more or less 

 greenish yellow, the silvery tone being notice- 

 able for the higher streamers and the green 

 and yellow tones increasing somewhat with the 

 zenith distance, and particularly was the 

 northern sky generally yellow with some green 

 and less often some red tones. 



My efforts were directed toward getting 

 spectrographic observations of the aurora. 

 Unfortunately no properly adjusted spectro- 

 graph was in readiness, but owing to the 

 length of the display two slit spectrographs 

 could be set up and three useful spectrograms 

 secured. Two of these were with a single 

 prism and a three and a quarter inch focus 

 camera, and the third with a very dense prism 

 and a 15 inch focus camera. The first of 

 the small scale spectra on an isochromatic 

 plate, is the stronger and shows about fifteen 

 lines and band heads between A 3800 and 

 the chief aurora line at A 5578.0. Those near 

 A 3914 and A 4276 were especially strong and 

 appeared to be the less refrangible edges of 

 flutings (band heads) — this is probably also 

 true of the radiation near A 4650. The other 

 small scale spectrum, on Ilford Panchromatic 

 plate, shows the stronger of these lines, in- 

 cluding the line A 5578 and a line in the red 

 whose wave-length according to my prelimi- 

 nary measures is near A 6320. 



The negative made with the higher disper- 

 sion instrument, on Ilford red plate, records 

 only faintly the chief aurora line, measures 

 of which give the wave-length as 5577.8. 

 This is in satisfactory agreement with the 

 value 5578.05, which I found some years ago 

 for this line from higher dispersion plates ex- 

 posed to the permanent auroral light of the 

 sky. This observation is of interest in that 

 it leaves no further doubt that the wave-length 

 is the same, within the errors of observation, 

 whether the auroral light is that permanently 

 scattered over the sky or is that of a violent 

 storm display. The red line A 6320 had been 

 previously observed visually but this is I be- 



