The lower limit of the aurora cannot be said to have yet been accurately determined i but it is probable that this must in some cases be placed within a mile or two of the earth's surface, if, indeed, it does not sometimes ex tend almost to the ground itself. Trustworthy observers have in some cases seen tht beams between their own positions and a nearby mountain, or below the clouds, while careful measurements of the heights of low lying auroras have been secured. It is also very fre quently observed in countries where auroras are numerous that the phenomenon is accom parried by a crackling sound, somewhat resem bling the rustling of silk or straw, or the electrical sputtenngs accompanying wireless messages. Such sounds have been described in the Arctic regions as regularly increasing and decreasing in intensity as the brightness of the aurora waxed and waned. In such cases the aurora must have been very near the ground; it is not to be expected that any sound can be detected in lower latitudes, where the aurora occurs at a far greater altitude.
Cause of the Science has not been able to determine as yet with certainty and precision the cause of this very common phenomenon. When the luminous effect pro duced by the passage of electricity through the highly rarefied air of a vacuum tube was first observed, its resemblance to the aurora led to the view that the latter was produced by electric currents in the upper regions of the atmosphere. Although it is not impossible that such currents may be associated with the aurora, they do not adequately explain its light, and are apparently inadequate to explain its rays. Yet there is no doubt that the aurora is associated in some way with the magnetism of the earth. The coinci dence of the rays of the aurora with the direc tion of the magnetic needle is one proof of this relation; another proof is found in the rela tion of the aurora to magnetic storms. The general rule is, that a very brilliant aurora is associated with such a storm, disturbing the magnetic needle not only at the point where the aurora is visible, but perhaps over the whole earth. Earth currents so strong that from time to time a line of telegraph may be run by them are also occasional accompaniments of a bril liant aurora, hut although the intimate connec tion of the two phenomena is so well estab lished, the exact relationship is yet to be worked out. In recent years Arrhenius has propounded
a theory, based on the very probable fact that the sun emits a flood of corpuscles of a nature similar to that of the mysterious rays known as X-rays, cathode rays, electrons, etc., the investi gation of which has occupied so large a place in recent physics. His theory is, that these cor puscles on approaching the earth are acted on in the direction of the lines of its magnetic force round which they describe helices. While nothing as yet has been found to disprove this theory, it is one which still needs much prov ing. What can be said with reasonable prob ability is that the aurora is caused by irregular emanations of corpuscles from the sun, which are stopped in the upper regions of our atmos phere.
Periodicity of the Records of auroras extending back two centuries or more show that they have been much more numerous at some periods than at others. Sometimes it has been supposed that they have been more numerous at intervals of 33 or some fixed number of years, but this has not yet been proved, nor has any law been determined by which we can definitely say at what times they appear in, the greatest number. But several periods gre observed which show that the ap pearance of the aurora is in some way con nected with the sun. The first instance of this is the fact that they are more numerous when the spots of the sun are more numerous. As there is an 11-year period in the spots on the sun, so there is a corresponding period in the aurora. There is also a semi-annual period in the frequency of the aurora, the greatest number being observed in March and Septem ber and the smallest in June and December. Arrhenius has connected this with the fact that in March and September the earth is over the region of the sun's surface in which spots are more numerous. Statistics also show that there are more auroras in the northern hemisphere when the moon is south of the equator than when she is north. This, how ever, is not proved to be a general law. There is also an observed period of 25.93 days in the frequency of the aurora. This is so near the time of the sun's rotation that it may be con nected with the latter. Consult Angot, 'The Aurora' (International Scientific Series).