on a luminous Arch . 103 
The direction of the arch here defcribed in that part near 
the zenith is W. 1 8° S. ; and if a line is drawn through Cam- 
bridge in this diredion, Kimbolton is 12,8 geographical miles 
north of it; and therefore, as the arch appeared 12 0 more 
fouth at Kimbolton than at Cambridge, the height of the arch 
above the furface of the earth muft be 6i| geographical or 71 
flatute miles. If we fuppofe that the middle of the arch really 
pafied through £ Aurigae, the height comes out 52 ftatute 
miles. On the whole, 1 fhould think, the height could hardly 
be lefs than 52 miles, and is not likely to have much exceeded 
7 I# 
The common aurora borealis has been fuppofed, with great 
reafon, to confift of parallel ftreams of light (hooting upwards, 
which, by the laws of perfpedive, appear to converge towards 
a point; and when any of thefe ftreams are over our heads, 
they appear adually to come to a point, and form a corona. 
Hence, from analogy, it feems not unlikely, that thefe lumi- 
nous arches may confift of parallel ftreams of light, difpofed 
fo as to form a long thin band, pretty broad in its upright di- 
redion, and ftretchedout horizontally to a great length one way, 
but thin in the oppofite diredion. If this is the cafe, they will 
appear narrow and well defined to an obferver placed in the 
plane of the band ; but to one placed at a little diftance from 
it, they will appear broader, fainter, and lefs well defined ; 
and when the obferver is removed to a great diftance from the 
plane, they will vanifh, or appear only as an obicure ill-defined 
light in the iky. 
There are tw r o circumftances which rather confirm this con* 
jedure : firft, that though we have an account of another 
arch befides this * having been feen at great diftances in the 
* That of Feb. 15, 1750. Phil. Tranf. XLVI. p. 47 ^» and 647. 
diredion 
