212 AURORA BOREALIS. 



saw seven; Jauuary 3 and 21, six arches; January 7, five arclies; September 30, 

 October 15, and December 25, four arches. The number of cases of three 

 arches seen simultaneously was considerable; and examples of two arches were 

 extremely frequent. 



Breadth of auroral arches. — This element is not the same when an arch 

 appears near the horizon as when it is seen near the zenith. At Bossekop, in 

 Scandinavia, for arches seen in the north at altitudes less than 60^, from a mean 

 of twenty observations the average breadth of the arches was 7". Near the 

 zenith and between the limits of 30^ zenith distance either north or south, 

 the mean of fifteen measurements gave a breadth of 25°. For arches seen south 

 of the zenith at altitudes less than 60**, the mean of nine observations gave a 

 breadth of 8^. 



The following example shows the change in the breadth of an arch during 

 its progress across the sky. December 16, the arch being in the south, at an 

 elevation of 52°, its breadth was 5"; a little later the arch passed the zenith 

 and had an altitude of 70^ towards the north; its breadth was then 40°. The 

 arch continued to sink, and when its altitude was 61°, its breadth was only 28° ; 

 ami at an altitude of 41° its breadth was about 6°. If the distance of an 

 arch from the earth remained constant during its movement of translation, and 

 the arch were of the form of a solid ring whose section was a circle, its breadth 

 when in the zenith should be double that at an elevation of 30°. But the ave- 

 rage of many measurements gave its breadth in the former case three or four 

 times as great as in the latter; showing that the greatest breadth of a section of 

 the ring is parallel to the earth. 



Anomalous forms of arches. — Sometimes an auroral arch consists of rays 

 an-anged in irregular and sinuous bands of various and variable curvature, pre- 

 senting the appearance of the undulations of a ribbon or flag waving in the 

 breeze. Sometimes the appearance is that of a brilliant curtain whose folds are 

 agitated by the wind. These folds sometimes become very numerous and 

 complex, and the arch assumes the form of a long sheet of rays returning into 

 itself', the folds enveloping each other and presenting an immense variety of the 

 most graceful curves. Sometimes these curves are continually changing, and 

 develop themselves like the folds of a serpent. 



It is evident, thercf ire, that a variety of disturbing causes may prevent an 

 auroral arch from taking up a position perpendicular to the magnetic meridian. 



Movements of auroral arches. — An auroral arch does not maintain, invariably, 

 a fixed position. It is frequently displaced, and is transported, parallel to itself, 

 from north to south, or from south to north. An arch which first appears near 

 the northei"n horizon sometimes rises gradually, attains the zenith, descends 

 towards the southern horizon, remains there for a time stationary, and then, 

 perhaps, retraces its course. The observations made by Lottin, at Bossekop, 

 presented 60 cases in which auroral arches moved from north to south, and 39 

 cases from south to north. There were 25 nights upon which only the first of 

 these movements was observed; 11 nights upon which only the opposite motion 

 was observed; and 17 nights upon which both movements were observed, suc- 

 cessively. Thus the motion from north to south appears to have been about 

 twice as common as that from south to north. In tlie United States, from a 

 considerable collection of observations it is inferred that the motion from north 

 to south is about ten times as frequent as the motion from south to north.* 



Sometimes there is a movement of the arch from west to east, or from east to 

 west. Sometimes while the height of the arch remains constant, the entire arch, 

 peems to turn around the vertical, either in the direction of the diurnal motion, 

 or in t'.ie opposite direction. Lottin and Bravais observed three cases in vv^hich 



* American Journal of Science, n. s., v. 34, pp. 41—15. 



