The coincidences relied on in the foregoing observations depend, of course, upon (1) the accuracy of the observations themselves, and (2) the subsequent reduction of the lines for comparison. Assuming the correctness of the latter, what have we as to the former? Two of Professor Young’s positions of coronal lines, as stated, seem to have far too much of the ± element to make them sufficiently accurate. Pierce’s auroral observation does not state how the lines were positioned. As they _all_ end with a cypher, the suspicion naturally arises that the measurements did not extend beyond the first three places of the figures, and, if so, could not be used for accurate comparison. The auroral lines, too, are generally rather wide and nebulous, and not easy of comparison with sharper ones. CHAPTER IX. SUPPOSED CAUSES OF THE AURORA. [Sidenote: Supposed causes of the Aurora. Sulphurous vapours. Magnetic effluvia.] At first the Aurora was described to be sulphurous vapours issuing from the earth; and Musschenbroek pointed out that certain chemical mixtures sent forth a phosphorescent vapour, in some respects resembling the Aurora. Dr. Halley originally proposed a similar theory, but ultimately concluded that the Aurora might be occasioned by the circulation of the magnetic effluvia of the earth from one pole to another. [Sidenote: Zodiacal light.] M. de Mairan, in 1721, in a treatise, ascribed the Aurora to the impulse of the zodiacal light upon the atmosphere of the earth. [Sidenote: Luminous particles of our atmosphere.] Euler combated this theory, and ascribed the Aurora to the luminous particles of our atmosphere driven beyond its limits by the light of the sun, and sometimes ascending to the height of several thousand miles. [Sidenote: Electric fluid _in vacuo_ resembles Aurora.] Mr. Hawksbee very early showed that the electric fluid assumes, _in vacuo_ or in highly rarefied atmosphere, an appearance resembling the Aurora. Mr. Canton contrived an imitation of the Aurora by means of electricity transmitted through the Torricellian vacuum in a long glass tube, and showed that such a tube would continue to display strong flashes of light for 24 hours and longer without fresh excitation. [Sidenote: Experiment with electrical machine and exhausted receiver.] In the ‘Edinburgh Encyclopædia,’ date 1830, is mentioned an experiment in which an electrical machine and air-pump are so disposed that strong sparks pass from the machine to the receiver of the air-pump. [Sidenote: Dr. Franklin’s theory.] As the exhaustion proceeds the electricity forces itself through the receiver in a visible stream, at first of a deep purple colour; “but as the exhaustion advances it changes to blue, and at length to an intense white, _with which the whole receiver becomes completely filled_.” [It will be noticed that this experiment bears a close resemblance to Prof. Ångström’s exhausted flask referred to later in treating of the spectrum of the Aurora.]