of the popular interest which auroras have aroused is due to the varied and often changing forms that they assume. Their protean shapes have caused people of earlier times to see in them bloody flames, armies in the clash of battle, or furiously riding Valkyries. They may appear as faint lights without defined shape; if more distinct, they are seen to have the form of rays converging upon some point in the sky, or of more or less clearly defined arcs, or of hands which fold over on themselves like a curtain, and are called draped auroras. M. Angot describes many variations of these typical shapes, and presents plates on which some of the most interesting are depicted. The vertical rays of which most auroras seem to be made up move both horizontally and vertically, and as they are usually in constant motion an aurora may readily change from one typical form to another. While the light from most auroras is white, the rays are frequently tinged with yellow, and are sometimes red at the lower extremity and green at the upper. The nature of the auroral light is not established, although the spectroscope and polariscope indicate that it is emitted by luminous gases. It seems probable that a slight rustling or crackling sound accompanies auroral displays, but our author finds no credible evidence of any odor. While many auroras of small extent appear only as local phenomena in high latitudes, others are visible to within twenty degrees of the equator. It seems to be usual for an aurora australis to occur simultaneously with an aurora borealis, notably on February 4, 1872, when the globe, with the exception of an equatorial zone of about forty degrees, was enveloped in polar lights. The periodicity of auroras has been studied with the result of establishing a diurnal and an annual period, and a period of a little more than eleven years. Less exactly determined are the periods of about twenty-eight days, and of about fifty-five years and a half. Those of twenty-eight days and eleven years seem to connect the auroras with sun spots. The relations of the aurora with meteorological phenomena and with terrestrial magnetism have also been investigated. The data obtained from researches on the foregoing questions have given rise to many theories as to the cause of auroras. Our author states several hypotheses that have been made clearly untenable by recent advances of knowledge. One of these is the idea of Mairan that auroras occur when the earth passes through the cloud of matter that produces the zodiacal light, some of this matter falling into our atmosphere and becoming ignited. The reflection of sunlight from particles of ice in the atmosphere is another cause suggested, and still another regards the light as a sort of fluorescence. Our author treats with more respect, although positively rejecting it, the theory first definitely stated by Dalton, that the light is given off from silent electric discharges between the upper and lower strata of the atmosphere, these discharges being conducted through ferruginous dust falling upon the earth from space. He gives also several electric theories, among which he regards that of Edlund as the most satisfactory. Edlund starts from the phenomena of unipolar induction—the production of currents in a metallic sheath surrounding a magnet when the sheath is rapidly revolved. The general phenomena of terrestrial magnetism justify regarding the earth as such a sheath. Electricity, according to this theory, is constantly being carried by molecules of air from the equator to the poles, where it accumulates and from time to time returns to the ground by slow discharges which produce auroras. M.
