As the electric streams, which constitute these northern lights, seem to be repelled or radiated from an accumulation of that fluid in the north, and not attracted like the fireballs; this accounts for the diffusion of their light, as well as the silence of their passage; while their variety of colours, and the permanency of them, and even the breadth of them in different places, may depend on their setting on fire the mixture of inflammable and common air through which they pass; as seems to happen in the transit of the fire-balls. It was observed by Dr. Priestley that the electric shock taken through inflammable air was red, in common air it is blueish; to these circumstances perhaps some of the colours of the northern lights may bear analogy; though the density of the medium through which light is seen must principally vary its colour, as is well explained by Mr. Morgan. Phil. Trans. Vol. LXXV. Hence lightning is red when seen through a dark cloud, or near the horizon; because the more refrangible rays cannot permeate so dense a medium. But the shooting stars consist of white light, as they are generally seen on clear nights, and nearly vertical: in other situations their light is probably too faint to come to us. But as in some remarkable appearances of the northern lights, as in March, 1716, all the prismatic colours were seen quickly to succeed each other, these appear to have been owing to real combustion; as the density of the interposed medium could not be supposed to change so frequently; and therefore these colours must have been owing to different degrees of heat according to Mr. Morgan's theory of combustion. In Smith's Optics, p. 69. the prismatic colours, and optical deceptions of the northern lights are described by Mr. Cotes. The Torricellian vacuum, if perfectly free from air, is said by Mr. Morgan and others to be a perfect non-conductor. This circumstance therefore would preclude the electric streams from rising above the atmosphere. But as Mr. Morgan did not try to pass an electric shock through a vacuum, and as air, or something containing air, surrounding the transit of electricity may be necessary to the production of light, the conclusion may perhaps still be dubious. If however the streams of the northern lights were supposed to rise above our atmosphere, they would only be visible at each extremity of their course; where they emerge from, or are again immerged into the atmosphere; but not in their journey through the vacuum; for the absence of electric light in a vacuum is sufficiently proved by the common experiment of shaking a barometer in the dark; the electricity, produced by the friction of the mercury in the glass at its top, is luminous if the barometer has a little air in it; but there is no light if the vacuum be complete.