Now, though the appearance of an aurora is due to some special terrestrial action (however excited), yet the material substances between which the discharges take place must be assumed to be at all times present in the upper regions of air. In all probability, they are the particles of those meteors which the earth is continually encountering. And since we know that meteor-systems must be aggregated in far greater numbers near the sun than near the earth, we may regard the coronal light as due to electrical discharges excited by the sun’s action, and taking place between the members of such systems. Besides this light, however, there must necessarily be a large proportion of light reflected from these meteoric bodies. In this way the peculiar character of the coronal spectrum may be readily accounted for. We know, from the auroral spectrum, that the principal bright lines due to the electrical discharges would be precisely where we see bright lines in the coronal spectrum. But, besides these, there would be fainter bright lines corresponding to the various elements which exist in the meteoric masses. These elements, we know, are the same as those in the substance of the sun. Thus the bright lines would correspond in position with the dark lines of the solar spectrum. Hence, as light reflected by the meteors would give the ordinary solar spectrum, there would result from the combination a continuous spectrum, on which the bright lines first mentioned would be seen, as during the American eclipse. What the polariscope has told us respecting the corona is in accordance with this view. In the same way the quality of the Zodiacal Light admits of being perfectly accounted for, without resorting to the hypothesis that this phenomenon is a terrestrial one. The explanation thus put forward has at least the advantage of being founded on well-established relations. We know that the auroral light is associated with the earth’s magnetism, and that meteoric bodies are continually falling upon the earth’s atmosphere. We know, also, that the sun exerts magnetic influences a thousand-fold more intense than those of the earth, and that in his neighborhood there must be many million times more meteoric systems. But we have other and independent reasons, which must not be overlooked, for considering the corona to be of some such nature as I have suggested. Leverrier has shown that there probably exists in the neighborhood of the sun a family of bodies whose united mass suffices appreciably to affect the motions of the planet Mercury. It would not be safe to neglect considerations thus vouched for. Mr. Baxendell also has shown that certain periodic variations in the earth’s magnetism point to the existence of such a family of bodies; and he has been able to assign to them a position according well with that determined by Leverrier.