AUROEA BOREALIS. 211 



observed at New Haven, Connecticut, March 27, 1781, touched the eastern 

 horizon at E. 2° S. by compass, and it touched the western horizon, at W. 20"' 

 N. by compass, indicating a deviation of 11"^ to the east of the magnetic me- 

 ridian. Observers in most parts of Europe, and also in the United iStates, bave 

 generally described auroral arches as perpendicular to the magnetic meridian. 

 It is desirable that an extended series of observations should be made, to deter- 

 mine the azimuth of each extremity of these arches. It is not improbable that 

 such observations would show not merely that auroral arches are occasionally 

 not perpendicular to the magnetic meridian, but that, for many localities, there 

 is a small constant deviation from this position. 



For 711 of auroral arches. — Auroral ajches are not arcs of great circles; that 

 is, they do not cut the horizon at points 180° from each other. This is shown 

 conclusively by the observations made by Lottin and Bravais in latitude 70°. 

 The following table shows the average result of 145 observations, arranged in 

 seven groups, according to the height of the arcs.* 



] st group height = 20°. 2 amplitude = 1 37°. 7 16 observations. 



2d group height 41". 4 amplitude 154°.4 20 observations. 



3d group height 7U°.l aiuplitudo ]63°.9 ]7 observations. 



4th group height 90° auiplitude 17r.°.4 41 obsei-vations. 



5th group height 108°..') amplitude 185°. 27 observations. 



6th group height 136°. 8 amplitude 184°.2 21 observations. 



7th group height 157°. 3 amplitude 190°.0 3 observations. 



The altitudes are supposed to be measured from the north hoi'izon, and the 

 amplitude is the angular distance between the two extremities of the arc, meas- 

 ured on the north side. 



If we divide the observations into three groups, we have — 



Northern arcs height = 44°.2 amplitude = ]52°.4 53 obseirations. 



Zenith arcs height 90°. amplitude 175°. 4 41 observations. 



Southern arcs height 123°.0 amphtude . 184°. 9 51 observutious. 



Carefid measurements made at five or six points of some of the most remarkable 

 arcs showed that, except near the horizon, they may be regarded as portions of 

 small circles parallel to the earth's surface. Near the horizon there is some- 

 times a sensible deviation i'rom this circular form, and the appearance is some- 

 times that of a portion of an ellipse, the extremities of the arch being bent 

 inward, as shown in 

 the annexed > figure. 

 Such appearances were 

 frequently noticed by 

 Lottin and Bravais in 

 Scandinavia. Hans- 

 teen says that, at Christiania, latitude 59° 54', he has twice seen an ellipse 

 almost entire.t 



Observations made at New Haven accord remarkably with the preceding. 

 Of 27 .'luroial arches whose average height was 8.^° the average amplitude 

 was 80'^. A single arch, whose altitude was 66^, had an amplitude of 165°. | 



During the splendid aurora of September 2, 1859, near Cape Horn, in lati- 

 tude 57° south, there was noticed a bright yellowish light foiming an ellipse 

 who.se diameters were as two to one, the centre of the ellijjse being elevated 

 about 15° above the south horizon § 



Number of axiroral arches. — On the 2d of January, 1839, Lottin and Bravais 

 saw nine different arches at the same time ; November 2 and January 24, they 



* Voyages en SCandinavie, pp. 466-478; 

 tMemoires de I'Acadeiuie de Belgique, t. 20, p. 119. 

 \ American .Journal of Science, n. s., v. 39, p. 289. 

 \ Ibid., V. 30, p. 89. 



