Now we cannot tell what proportion holds in the case of the sun's disc, because we do not know how much light has been absorbed where we look at the middle of the disc. All we know is that whatever remains after such absorption is about twice as much as we receive from near the edge of the disc. It is easily seen that this knowledge is insufficient for our requirements. But there can be no question whatever that the total absorption near the edge exceeds many times that near the middle of the disc; and on very reasonable assumptions as to this excess, it may readily be shown that the absorbing atmosphere cannot exceed some five or six hundred miles in depth. Probably it is even shallower.

Now, there is a circumstance which perfectly accounts for the non-recognition by spectroscopists of an atmosphere relatively so shallow as this. Let it be remembered, in passing, that the average height of the sierra may be set at about five thousand miles; so that the atmosphere we are dealing with would be at the outside but one-fifth as high as that fine rim of red light with saw-like edge which astronomers detected around the eclipsed sun in the total eclipses of 1842, 1851, and 1860. Still it might be thought that patience only would be needed to detect the signs of such an atmosphere, shallow though it be. But there is a peculiarity of telescopic observation which renders the recognition of such an atmosphere, if of less than a certain depth, not difficult merely, but impossible. It may be well to exhibit the nature of the peculiarity at length, because it is of considerable interest to all who possess or use telescopes. I take an illustrative case, which seems, at first, to have little connection with my subject.

Every reader of this work has heard of the double stars, and I dare say most of those who read this particular article have seen many of these beautiful objects. It is known that some double stars are much closer than others, and we commonly hear it mentioned as a proof of the excellence of a telescope that it will divide such and such a double star. But it might seem that if a telescope of a certain size were constructed with extreme care, it should be capable of dividing any double star; because we might use an eye-piece of any magnifying power we pleased, and so, as it were, force apart the two star-images formed by the object-glass.

Instead of this being the case, however, there is a limit for every object-glass, beyond which no separation is possible; for this reason, simply, that the star-images formed by the object-glass are not points of light, as they would be if they correctly represented the stars of which they are the optical images. The larger the object-glass (assumed to be perfect in construction) the smaller is the star-image;1 but it has always a, definite size, and if this size is such that the two images of the stars forming a pair actually touch or overlap, we cannot separate them by using highly-magnifying eyepieces.

Now what is true of a star is true of every point of any object we examine with a telescope. The image of the point is always a circle of light, which, though minute, has yet appreciable dimensions. The image of the object is made up of all these circles, which necessarily overlap. Nor let the reader suppose that on this account telescopic observation is untrustworthy. Precisely the same peculiarity affects ordinary vision. There is no such thing as a perfect optical image of an object; though neither eyesight nor telescopic vision need be regarded as deceptive on this account. Our power of seeing minute details is limited by this peculiarity, but we are not actually deceived. If microscopic writing be shown us, for instance, we may find ourselves, after poring over it for some time, unable to make out its meaning, the letters seeming all blended together; but we know what our failure really means, and do not fall into the mistake of concluding that there are no details because the actual details are inscrutable.

1 A curious illustration of this is given by the fact that a certain astronomer of old, having reduced the aperture of his telescope to a mere pin-hole, announced that he was thus enabled to measure the real globes of the stars, for, instead of seeing the stars through his telescope as minute points of light, he now saw them with discs like the planets. He thought he was improving the defining qualities of his telescope, instead of altogether destroying them.

Let us apply this consideration to the sun, and more particularly to the appearance presented by the edge of the sun's disc. The image of every point of this edge is a small circle; the combination of all these small circles must produce a ring of light all round the true outline of the disc. If the sun's atmosphere did not reach beyond this ring, then no contrivance whatever could render the atmosphere discernible, let the telescope be ever so perfect and the observer ever so clear-sighted or skilful. Now, the actual extension of this ring will be greater or less according as the object-glass of the telescope is less or greater. It may readily be shown that neither Mr. Lockyer's telescope nor Fr. Secchi's could possibly show any signs of a solar atmosphere under two hundred miles in depth, while in all probability an atmosphere four or five times as deep would escape their scrutiny.

Are we then to remain altogether in ignorance of such an atmosphere, supposing that it actually exists, and that the dark lines in the solar spectrum are due to its absorptive power ? Is there no way of obviating the difficulty which has just been dealt with ?

So far as the method of observing the sun when uneclipsed is concerned, the answer to these questions must be negative; or, rather, it must be answered that our only hope of meeting the difficulty consists in increasing the size of the telescopes with which the sun is spectroscopically studied. And inasmuch as Dr. Huggins is preparing to apply the powers of a much larger telescope than either Mr. Lockyer's or Fr. Secchi's, we may possibly still hope to hear that the relatively shallow atmosphere can be studied when the sun is not eclipsed. For we may now speak of the existence of this atmosphere as a demonstrated fact. The difficulty which seemed to present insuperable obstacles to the observers who study the uneclipsed sun, has been overcome by the ingenuity of one of the most skilful of those very observers-Professor Young, of America-when studying the solar eclipse of last December.