60 Rev. C. 8S. Lyman on the Auroral Bow of June, 1852. 
e time, as seen from East Windsor, the lower edge had an 
aitinide of 61° and the upper of 67° 
The plane (perpendicular the direction of the belt) in which 
these altitudes are taken, lies in a direction S. 8° or 9° W. of the 
true meridian, which angle is on the opposite side of the meridian 
from the maguetic variation. The bow, therefore, instead of 
crossing the magnetic meridian at right angles, made an angle 
With itabout 15° less than a right angle 
n like manner, Professor Twining, (Am. Jour., vol. xxxily pp 
223, 227,) found the meridian of the Auroral bows of May and 
August, 1836, to lie on the opposite side of the true meridian itr 
stead of on the same side with the magnetic variation 
East Windsor bears from Nev v Haven, according ‘to the best 
maps, about N. 23° 30’ E., certo forty English enrites: The dis 
tance reckoned in the direction of the plane in which the altitudes 
are estimated, is thirty-eight and a half miles. 
Before making further use of these data, it may be well to div 
rect attention more particularly to the apparent width of the belt. 
This, at New Haven, is clearly defined by Mr. Herrick to be not 
far from 2°. He could hardly be much mistaken in estimating 
one-fifth of the distance between Arcturus and Mirac. This dis- 
tance is 10° 17’. As seen from East Windsor, the width could 
not have been less than 6 or 7 degrees. On this point Mr. Wat- 
son remarks in a stibsequent letter. ‘For some minutes the belt 
was nearly stationary over the head of Serpens, and its width 
for some distance from the meridian very nearly uniform. I 
then, before any marked motion oceurred, noted its breadth 
pretty accurately by the stars mentioned, the five stars plainly 
discernible in the head of Serpens, the two westernmost in @ 
line with, and a little east of south of « Corone. The north 
ern star was then nearly upon, or a litle within, the northerm 
edge of the belt, and the greatest north and sonth distance of 
the stars was nine-tenths of the width of the belt. The widt 
was rather greater than less than that I have indicated. » The 
distance between @ and ¢ Serpentis, the northernmost and south — 
ernmost of the stars referred to, is 5° 42’... This would make 
the width of the belt not less than 69. Mr. Hyde’s observa 
tions confirm this estimate. Prof. Brocklesby’s idleness a siti 
ilar width as seen from Hartford. 
If now, with a base of 38} miles, and the altitudes are 
above for New Haven and East Windsor, (viz.: 72° and 7 
the lower and upper edges respectively, as seen at the ‘it 
place, and 61° and 67° for the same at the latter place,) we draw 
a diagram like the one on the opposite page, it is manifest that 
a section of the atrroral belt must lie within the elongated trape- 
zium A, B,C, D. And as each of the four sides of the trapeai- 
um is tangent to the belt, its position must be somewhat that of 
