known hj the Name cf the Aurora Borealis. 181 



ravs again reflected on the earth by the atm(5sphere at a certain 

 angle: therefore tlievvvill be seen from different places at nearly 

 the same angle, only affected by the angle which the reflected 

 rays make with the horizon in proceeding from the latter to- 

 wards the zenith; being with that exception similar to the rain- 

 bow, which, however the spectator may change his place, pre- 

 serves its relative situation to him. 



If 1 am at all right, I should suppose that an approximation to 

 the height of these phaenomena might be found by something 

 like the following method, which would apply more particularly 

 to the arch, if such a phaenomenon should again make its ap- 

 pearance, because its motion is slow and uniform, and seems only 

 to depend on the motion of the earth relative to the sun. Take 

 the angle of height and the bearing of the middle or highest 

 point of the arch, noting the time ; repeat this two or three times 

 at intervals till it disappear, and it would be ?o much the better 

 if these observations were made at two or three different parts 

 of the country as distant as possible from each other. It could 

 then be found, if a line passing through the centre of the sun 

 and the observer coincided with the different bearings at the 

 times they were taken. Find also the boundary of light and dark- 

 ness and the above-mentioned lines, or rather vertical planes, at 

 the times noted. A little on the light or north side of this 

 boundary must be the place from which the arch is reflected, 

 and which from the regularity of the figure seems to be the ocean 

 in a liquid state, or with a uniform surface of ice, or the latter 

 covered with snow. 



From the two times on which this phajnomenon was seen so 

 nearly coinciding, it is probable that it caimot i)e produced ex- 

 cept when the earth and sun are in the same relative position. 



The times alluded to, it will be recollected, were near the au- 

 tumnal equinox, from about 8 till 10 o'clock in the evening; and 

 it is scarcely to be supposed that the ocean could be frozen at 

 that time of the year in the direction of the sun between those 

 hours. But the surface of the water seems quite adequate to 

 reflect the ravs of the sun with all the splendour displayed by 

 that beautifui phaenomenon, when we consider the great obliquity 

 of the incident rays, and their reflection into the dark serene 

 atmosphere which on those evenings favoured their exhibition. 

 The al)()ve ol)servations being made, and making allowance for 

 refraction in both the incident and reflected rays, probably in- 

 creased by tlie earth's attraction on the latter, (which 1 am in- 

 clined to think is considerable, and principally causes the curve 

 observable in streamers,) it will be seen at once if the pheno- 

 mena are produced by the (;ause I have assigned. The reflected 

 rav being traced i^a above suggested, and being intersected at 



the 



