186 



SCIENCE 



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



and never long at rest, and when momentarily 

 a favorable formation developed the intensity 

 fell ofF too rapidly to give suitable photo- 

 graphs for accurate determination of the con- 

 vergent. At times the convergent could be 

 readily determined when streamers from many 

 different directions were nearly or actually 

 concurrent and the position of this point 

 could be located with reference to comparison 

 stars. My own observations were as follows: 



These observations, and also Dr. Russell's, 

 show that generally the point of convergence 

 was near the magnetic meridian but between 

 5° and 6° south of the "magnetic zenith" 

 (coordinates of the magnetic zenith for Flag- 

 staff are: azimuth S. 15° W.; altitude 62°). 

 If the auroral streamers follow the lines of 

 force of the earth's magnetic field the higher 

 parts of the streamers might be expected to 

 show a deflection in the direction indicated 

 but no calculations have been made to see if 

 the magnitude of the apparent displacement 

 might be of the order observed. What the ef- 

 fect of parallax may be is also a point not to 

 be overlooked. There can be little doubt but 

 that the point of convergence was subject to 

 greater actual variation in azimuth than alti- 

 tude. In the magnetic disturbances that ac- 

 company auroras it is possible that changes in 

 the earth's magnetic field might be perceptible 

 in the course of the auroral streamers. 



At 10 :44 a series of parallel streamers, com- 

 ing into view almost directly overhead and ex- 

 tending east and west, were seen drifting very 

 rapidly towards the north, with undulating and 

 flickering motions flowing lengthwise through 

 them. They were visible for only a moment. 



the detail dissolving a short distance north of 

 the zenith, and one had the impression that the 

 phenomena taking place were comparatively 

 near. 



Streamers of both narrow and broad (at 

 times somewhat diffuse) types were present. 

 As a rule a greenish tint was most prominent 

 in both the streamers and the extensive lumi- 

 nous areas but a pinkish or ruddy color was 

 also much in evidence. At various times 

 pinkish or pale red streamers were seen, gen- 

 erally in the northeast or northwest. A broad 

 pinkish streamer in the northwest appeared 

 to be rather quiescent and remained visible 

 much longer than any of the other streamers 

 during the display. It was 3° or more in 

 width, extending perpendicularly upward 

 from the horizon about thirty degrees. About 

 12 :35 a superb display of both pinkish and 

 green streamers was visible for a few minutes 

 in the northeast, extending up about 45° from 

 the horizon, considerably inclined southward. 

 Throughout the display streamers of greatly 

 varying intensity playing upward from the 

 northern horizon were visible but these were 

 not as conspicuous as might have been ex- 

 pected in view of the brightness and activity 

 of the auroral light in other parts of the sky. 

 Occasionally these streamers were subject to 

 marked flickering, and some movement — a 

 slow lateral drift. Now and then dark lanes 

 occurred between the streamers, and in one 

 instance a very conspicuous dark rift had a 

 leisurely motion eastward. 



From the time the auroral light was first 

 made out in the waning twilight it was strong, 

 with greatly increased intensity during the 

 intervals of the outbursts mentioned, until 

 after 1 o'clock when it rapidly subsided into 

 a feeble glow along the northern horizon. It 

 was reported by someone stationed at a sheep 

 camp north of the San Francisco Peaks that 

 another outburst of streamers developed later 

 in the night. 



The auroral light must have been of great 

 intensity as the display was a magnificent 

 spectacle even when dimmed with the light 

 of the moon near the first quarter. 



On the day following the display two spots 



