composing the arch, to revolve round the magnetic pole of the earth, giving them the motion from east to west, or from west to east, which the components of the arch are observed to have.”
The principal difficulties and deficiencies of this hypothesis, which was first suggested by De la Rive, are that it makes no attempt to account for the origin of such an electrical discharge, and that it is difficult to understand how an electric current can traverse vast spaces of the almost perfect vacuum which must exist at the distance from the earth (many hundreds of miles) which is attained by the magnetic curves, since, in the best vacuums of our Sprengel pumps, discharge will not take place even across the interval of a few centimetres. It is not, however, certain that stellar space is an insulator, and it is possible, moreover, that the auroral currents do not follow the magnetic curves through their whole course, since electric discharge is always in the path of least resistance, and this is modified not only by the magnetic forces, but by atmospheric density, and it is possible that on attaining a certain height the current may proceed horizontally on a stratum of least resistance. It need create no surprise that the discharge is generally invisible in the intermediate zone of low latitudes, since this is well accounted for not only by the large surface over which it is spread at great heights, but because this part of its course is at right angles to the line of sight, while in higher latitudes we look at the streamers almost “end-on,” and thus have before our eyes a very great depth of luminous gases. It is common enough, too, in discharges in rarefied gases to see the two poles surrounded by luminous auræ, while the intermediate space is almost or quite dark, or consists of luminous disks or striae separated by dark spaces. It seems probable that this “glow” discharge in rarefied gases is really a sort of electrical convection, which is propagated comparatively slowly, and from particle to particle; and that the striæ are surfaces at which the difference of potential of the moving molecules is so great as to cause discharge between them, while in the intermediate dark spaces the electric force is carried mechanically and silently by the particles moving in regular currents under the repulsive and attractive forces of electrification. On this hypothesis the auroral discharge becomes comprehensible, since we have only to suppose that the electricity is carried mechanically, as it were, through the vacuous spaces, which, if they contain no matter to conduct electricity, can contain none to impede the motion of the molecules. It is, moreover, by no means certain that the bright rays indicate actual currents. They may simply consist of matter rendered luminous in the arches, and projected by magnetic or electrical repulsion in the curves of magnetic force, since Varley (Roy. Soc. Proc., xix. 236) shows that when a glow discharge in a vacuum tube is brought within the field of a powerful magnet, the magnetic curves are illuminated beyond the electrodes between which the discharge is taking place as well as within the path of the current; and also that this illumination is caused by moving particles of matter, since it deflected a balanced plate of talc on which it was caused to impinge. It has also been shown that in electrical discharges in air at ordinary pressures, while the spark itself was unaffected by the magnet, it was surrounded by a luminous cloud or aura, which was drawn into the magnetic curves, and which might also be separated from the spark by blowing upon it. It is evident, therefore, that any mechanical force may separate the luminous particles from the electric discharge which produces them.
Geogra-
phical dis-
tribution.With regard to the geographical distribution of aurora, Prof. Loomis (Sill. Jour., xxxi.) has laid down a series of zones of equal auroral frequency, and in Petermann’s Mittheilungen for October 1874, Prof. Fritz has given a chart embodying the results of his extensive researches on the same subject. He finds, like Prof. Loomis, that the frequency of auroral display does not continue to increase to the pole, but reaches a maximum in a zone which, for the northern hemisphere, passes through the Faroe Islands, reaches its most southern point, about 57°, nearly south of Greenland, passes over Nain on the Labrador coast, then tends northwards, across Hudson’s Bay (60° N. lat.), and through great Bear Lake, and leaves the American continent slightly south of Point Barrow. It then skirts the northern coast of Asia, reaching its most northerly point, about 76° N., near Cape Taimyr, passing through the north of Nova Zembla, and skirting the N.W. coast of Norway. Not only are auroral displays less frequent in Iceland and Greenland than further south, but it is found that while south of this zone auroræ appear usually to the north of the observer, north of it they are generally to the south, and within it, north or south indifferently. South of this lie other zones approximately parallel to it, and of constantly diminishing frequency. That in which the average yearly number of auroræ is 100 passes through the Drontheim, the Orkneys, and the Hebrides, and reaches the American coast just north of Newfoundland. South of this the frequency diminishes rather rapidly. At Edinburgh the annual average is 30, at York 10, in Normandy 5; while at Gibraltar the average is about 1 in ten years.
These curves, which Prof. Fritz calls isochasmen, are nearly normal to the magnetic meridians, and bear a close relation to the curves of equal magnetic inclination, especially with those laid down by Hansteen in 1730, while they noticeably diverge in some places from those of Sabine of 1840. They also approximate to the isobaric curves of Schouw, and Prof. Fritz remarks that the curves of greater frequency tend towards the region of lowest atmospheric pressure. It is not unlikely that there may be such a connection, since Prof. Airy has showed a relation between barometric and magnetic disturbances.
It will be noticed that, eastward from England, the isochasmic curves tend rapidly northward, Archangel being only on the same auroral parallel as Newcastle. Prof. Fritz points out that they bear some relation to the limit of perpetual ice, tending most southward where, as in North America, the ice limit comes furthest south. He also endeavours to establish some connection between the periods of maximum of auroræ and those of the formation of ice, and considers ice as one of the most important local causes which influence their distribution. He quotes a curious fact mentioned by several Arctic voyagers, that aurora was most frequently seen when open water was in sight, and usually rather in the direction of the water than of the magnetic north. In this connection it may be well to remind our readers that the water of the Arctic regions is always warmer than the ice fields, and must cause upward currents of damp air. For the southern hemisphere there are not yet sufficient observations to make any determination of geographical distribution.
With regard to distribution in time Loomis and Fritz Distribu-
tion in
timeand Wolf have shown that there are periodical maxima about every ten or eleven years, and that these maxima coincide both with those of sun spots, and of magnetic disturbance. The following are Fritz and Wolf’s dates of maxima;—
| Sun Spots | Auroræ |
| 1706 | 1707 |
| 1718 | 1721 |
| 1728 | 1728 |
| 1739 | 1738 |
| 1750 | 1749 |
| 1761 | 1760 |
| 1770 | 1769 |
| 1779 | 1779 |
| 1788 | 1788 |
| 1804 | 1804 |
| 1817 | 1816 |
| 1830 | 1830 |
| 1837 | 1839 |
| 1848 | 1848 |
| 1860 | 1860 |
| 1871 | 1872 |
