AURORA BOREALIS. 221 



beams or columns, eveiywlicre nearly parallel to the direction of a magnetic 

 needle when freely suspended ; that is, in the United States, these beame were 

 nearly vertical, their upper extremities being inclined southward at angles vary- 

 ing from 15" to 30°, These beams were, therefore, about 500 miles in length; 

 and their diameters varied from five to ten and twenty miles, and perhaps, some- 

 times, they were still greater. 



The aurora of August 28, 1859, formed a belt of light of nearly equal extent, 

 ■ and, it pervaded the entire interval between the elevations of 46 and 534 miles 

 above the earth's surface. 



The height of a large number of auroras has been computed by similar 

 methods, and the average result for the ixpper limit deduced from 31 examples 

 is about 450 miles* 



Professor Potter, of London University, from a com{iariaon of a very large 

 number of uncommonly good observations, has determined the height of the 

 auroral arches of September 17 and October 12, 1833. His results were for the 

 mean heights of the upper edge of the arches from eight comparisons, 72 miles ; 

 mean height of the under edge of the arches from two comparisons, 63 miles.t 

 Dr. Dalton determined the height of the auroral arch of March 29, 1826, to bo 

 100 miles or upwards | 



From these and a multitude of similar results it is concluded that the aurora 

 seldom if ever appears at an elevation above the earth's surface less than about 

 45 miles, and that it extends upwards sometimes to an elevation of at least 500 

 miles. 



It is believed that these conclusions correspond substantially with the views 

 of those whose opinions on this subject are entitled to the greatest weight; 

 nevertheless, there are some, who contend that the aurora is sometimes seen at 

 elevations ,pf less than one mile above the earth's surface. In the Philosophical 

 Transactions of London for 1839, pp. 277-280, Professor James Farquarson, of 

 Scotland, has given an account 'of observations upon an auroral arch made at 

 two stations distant but a little more than a mile apart, from which it was con- 

 cluded that the height of the lower edge of the arch was only 2,481 feet, or less 

 than half a mile. It is difEcult to decide wherein consisted the fallacy of this 

 determination. One observation was made at 7'' and another at 7^' 5'". The 

 observations at the two stations indicated an apparent parallax of 7°; the arch 

 was about 12° broad; clouds were visible during bothrobservations ; and "at 7^ 

 10™ the whole sky had become too much obscured to admit of longer continued 

 cotcmporaneous observations." It is possible that this cloud hid a portion of the 

 arch from one of the stations while it was visible at the other, so that the two 

 observers were not viewing the same object ; or the object of observation might 

 have be«n simply a thin cloud illumined by auroral light. I believe Mr. Far- 

 quarson's conclusions to be erroneous, because they differ so widely from those 

 of more practiced observers under at least equally favorable circumstances. 



Simdar observations for the determination of the parallax of the aurora 

 were made by the French observers in Scandinavia in 1839. On the 9th of 

 January M. Bravais left Bossekop and went to Jupvig, distant less than ten 

 miles, for the purpose of observing the height of auroral arches, while M. Lottin 

 remained at Bossekop to make similar and simultaneous observations. The 

 altitudes of the arches were measured with a theodolite, and the times were noted 

 by a chronometer. On comparing the observations it was discovered that the 

 aurora always presented nearly the same appearance at the two stations.§ Care- 



M 



" Auuals of Philosophy, December, 1814, p. 431. 



t Transactions of Cambridge Philosophical Society, v. 8, pp. 322-325. 



X Philosophical Transactions Royal Society 1828, p. :<;98. 



$ Voyages en Scandinavia, p. 537 . 



