The aurora borealis, or northern dawn, is very ingeniously accounted for by Dr. Franklin on principles of electricity. He premises the following electric phenomena: 1. that all new fallen snow has much positive electricity standing on its surface. 2. That about twelve degrees of latitude round the poles are covered with a crust of eternal ice, which is impervious to the electric fluid. 3. That the dense part of the atmosphere rises but a few miles high; and that in the rarer parts of it the electric fluid will pass to almost any distance. Hence he supposes there must be a great accumulation of positive electric matter on the fresh fallen snow in the polar regions; which, not being able to pass through the crust of ice into the earth, must rise into the rare air of the upper parts of our atmosphere, which will the least resist its passage; and passing towards the equator descend again into the denser atmosphere, and thence into the earth in silent streams. And that many of the appearances attending these lights are optical deceptions, owing to the situation of the eye that beholds them; which makes all ascending parallel lines appear to converge to a point. The idea, above explained in note on l. 123, of the existence of a sphere of inflammable gas over the aerial atmosphere would much favour this theory of Dr. Franklin; because in that case the dense aerial atmosphere would rise a much less height in the polar regions, diminishing almost to nothing at the pole itself; and thus give an easier passage to the ascent of the electric fluid. And from the great difference in the specific gravity of the two airs, and the velocity of the earth's rotation, there must be a place between the poles and the equator, where the superior atmosphere of inflammable gas would terminate; which would account for these streams of the aurora borealis not appearing near the equator; add to this that it is probable the electric fluid may be heavier than the magnetic one; and will thence by the rotation of the earth's surface ascend over the magnetic one by its centrifugal force; and may thus be induced to rise through the thin stratum of aerial atmosphere over the poles. See note on Canto II. l. 193. I shall have occasion again to mention this great accumulation of inflammable air over the poles; and to conjecture that these northern lights may be produced by the union of inflammable with common air, without the assistance of the electric spark to throw them into combustion.