Now, the solar corona has been seen, during total eclipses of the sun, to extend to a distance at least equal to the sun’s diameter from the eclipsed orb. So that, assuming the corona to be a solar atmosphere, it would have a depth of about eight hundred and fifty thousand miles, and being also drawn toward the sun by his enormous attractive energy (exceeding more than twenty-seven times that of the earth), it could not fail to exert a pressure on his surface exceeding many thousand-fold that of our air upon the earth. In fact, such an atmosphere, let its outermost layers be as rare as we can conceive, would yet have its lower layers absolutely liquefied, if not solidified, by the enormous pressure to which they would be subjected. We can not, then, believe this corona to be a solar atmosphere. [Illustration: Fig. 33.—Tychonic System] Yet it is quite impossible to dissociate the corona, either wholly or in part, from the sun. I am aware that physicists of eminence have attempted to do this, and not only so, but to make of the zodiacal light a terrestrial phenomenon. But they have overlooked considerations which oppose themselves irresistibly to such a conclusion. In the first place, the mere fact that, during a total eclipse, the moon looks black, in the very heart of the corona, affords, when properly understood, the most conclusive evidence that the light of the corona comes from behind the moon. If the glare of our atmosphere could by any possibility account for the corona (which is not the case), then that glare should appear over the moon’s disk also. That this is so is proved by the fact that, when the glare really does cover the moon, as while the sun is but slightly eclipsed, the moon is not projected as a black disk on the background of the _sky_, though, where her outline crosses the sun, it appears black, by contrast with the intensity of his light.[25] The point seems, however, too obvious to need discussion. And, secondly, as Mr. Baxendell has pointed out, during totality the part of the earth’s atmosphere between the eye and the corona is not illuminated by the sun. Over a wide space all round the sun we are looking through an atmosphere which is completely dark. In fact, if the earth’s atmosphere alone were in question, we ought to see a dark or negative corona around the sun, the illuminated atmosphere only beginning to be faintly visible at a considerable angular distance from the sun. This argument, rightly understood, is altogether decisive of the question.[26] But the spectroscope has given certain very perplexing evidence respecting the light of the corona, and it remains that we should endeavor to see how that evidence bears on the interesting problem which the corona presents to our consideration.