during auroras. Andrée, in 1882–'83, at Spitzbergen, with a Mascart electrometer, found that before an aurora appeared the positive electric potential of the air diminished rapidly, and even became negative, as usually happens when it rains; but as soon as the aurora appears the potential takes as before a high positive value. But it must be remembered that the electric potential of the atmosphere, especially at a short distance from the earth, varies constantly in an abrupt and irregular manner; so that it would be imprudent to assume any coincidence, even during an aurora.
Auroras seem to depend for their form and position in space on the distribution of magnetism on the surface of the globe. The appearances of auroras coincide in the majority of cases with magnetic disturbances on the earth's surface. The auroral arc generally has its summit near the magnetic meridian, where the compass needle points. The force which governs the auroras appears to be the same which the magnetic needle obeys. In general, the point where the rays converge in auroral crowns, the center of the crown, is near the magnetic zenith. The variation, in fact, in such instances is but one degree, which may be allowed as the amount of probable error in observations.
The conclusion is, therefore, that the earth's magnetic forces certainly play the most important part in the auroral display. The arcs or "bands" arc nearly perpendicular to the magnetic meridian, and the direction of the rays parallel to the magnetic needle. The deviations from these rules are due to atmospheric conditions of temperature and humidity. The electric discharges which constitute the aurora borealis encounter all sorts of conditions of the atmospheric strata which are unequally conductive; hence lack of symmetry in the auroral displays. Meteorological conditions very likely influence not the production of the aurora, but its form and position. There is an intimate relation, as we have already pointed out, between the weather and these magnetic disturbances.
The magnetic needle sometimes begins to be agitated an hour before the appearance of an aurora. The magnetic disturbances last a long while, often for twenty-four hours. Motionless arcs and faint auroras do not affect the magnetic needle sensibly, while, during an active aurora, which is apparently nearer, the needle is greatly agitated, especially when the great red and green rays flash suddenly like lightning.
Arago has shown that if the aurora seems to be absent during magnetic disturbances, it is often because it is too distant, or below our horizon, and visible only in more northern latitudes. In the arctic and antarctic regions there is small deviation of the magnetic needle, but this is considered due to the great height of these auroras,
