The position of the dark auroral segment varies with the place occupied by the observer, and its centre always corresponds with the magnetic meridian. In our Eastern States the auroral segment appears a little to the west of the north point; but as the observer proceeds westward it gradually approaches this point, and is due north when seen from the vicinity of Lake Winnipeg. At Point Barrow, in the extreme north-west of the United States, the aurora is observed in the east. In Melville Islands, Parry saw it in the south; while in Greenland it is directly in the west. It is stated that auroras are more numerous about the equinoxes than they are at any other seasons; and also, when the earth is in perigee, than when it is in apogee. An examination which I have made of a catalogue by Professor Loomis, comprising 4,137 auroras observed in the temperate zone of our hemisphere from 1776 to 1873, sustains this statement. During this period, one hundred more auroras were recorded during each of the months comprising the equinoxes, than during any other months of the year; while eighty more auroras were observed when the earth was in perigee, than when it was in apogee. But to establish the truth of this assertion on a solid basis, more observations in both hemispheres will be required. The aurora is not simply a terrestrial phenomenon, but is associated in some mysterious way with the conditions of the Sun's surface. It is a well-known fact that terrestrial magnetism is influenced directly by the Sun, which creates the diurnal oscillations of the magnetic needle. Between sunrise and two o'clock, the north pole of the needle moves towards the west in our northern hemisphere, and in the afternoon and evening it moves the other way. These daily oscillations of the needle are not uniform in extent; they have a period of regular increase and decrease. At a given place the daily oscillations of the magnetic needle increase and decrease with regularity during a period which is equal to 10⅓ years. As this period closely coincides with the Sun-spot period, the connection between the variation of the needle and these solar disturbances has been recognized.