So long as men merely studied the appearances presented by the aurora, so long, in fact, as they merely regarded the phenomenon as a local display, they could form no adequate conception of its importance. The circumstance which first revealed something of the true character of the aurora was one which seemed to promise little. Arago was engaged in watching from day to day, and from year to year, the vibrations of the magnetic needle in the Paris Observatory. He traced the slow progress of the needle to its extreme westerly variation, and watched its course as it began to retrace its way towards the true north. He discovered the minute vibration which the needle makes each day across its mean position. He noticed that this vibration is variable in extent, and so he was led to watch it more closely. Thus he had occasion to observe more attentively than had yet been done the sudden irregularities which occasionally characterise the daily movements of the needle. All this seems to have nothing to do with the auroral streamers; but we now reach the important discovery which rewarded Arago’s patient watchfulness. In January 1819 he published a statement to the effect that the sudden changes of the magnetic needle are often associated with the occurrence of an aurora. I give the statement in his own words, as translated by General Sabine:—‘Auroras ought to be placed in the first rank among the causes which sometimes disturb the regular march of the diurnal changes of the magnetic needle. These do not, even in summer, exceed a quarter of a degree, but when an aurora appears, the magnetic needle is often seen to move in a few instants over several degrees.’ ‘During an aurora,‘ he adds, ‘one often sees in the northern region of the heavens luminous streamers of different colours shoot from all points of the horizon. The point in the sky to which these streamers converge is precisely the point to which a magnetised needle suspended by its centre of gravity directs itself.... It has, moreover, been shown that the concentric circular segments, almost similar in form to the rainbow, which are usually seen previous to the appearance of the luminous streamers, have their two extremities resting on two parts of the horizon which are equally distant from the direction towards which the needle turns; and the summit of each arc lies exactly in that direction. _From all this, it appears, incontestably, that there is an intimate connection between the causes of auroras and those of terrestrial magnetism._’