teristic spectra of lines or bands (these spectra being, however, different for portions lying at different depths), and that lower portion whose light is capable of giving a continuous spectrum.

Now, here we approach the great difficulty of interpreting the results of the spectroscopic analysis of the Sun. We have no means of learning whence that part of the light comes which gives the continuous spectrum. When we recognise certain dark lines, we know certainly that the corresponding element exists in the gaseous form at a lower temperature than the substance which gives the continuous spectrum. And so, also, we can interpret the appearance of bright lines. They show beyond question that the corresponding element exists in the gaseous form at a higher temperature than the substance which gives the continuous spectrum. But as regards that continuous spectrum itself we can form no such exact opinion. It must be remembered that a substance giving a continuous spectrum is not necessarily opaque to light from a substance at higher temperature also giving a continuous spectrum. It is capable of exercising a general absorption, but not necessarily (nor probably, under such conditions as exist in the Sun) of exercising an absorption at once general and complete. Hence we have no means of determining how great a depth of the solar substance is concerned in sending out the light which gives the continuous background of the spectrum. This light may come from the surface layers only-but it may be a shell whose thickness

forms no inconsiderable aliquot part of the Sun's diameter. And the reversal of the lines of certain elements, although it cannot take place at such excessive depths, may yet take place very far below the visible limits of the photosphere. For, as I have already shown, a depth of a few hundred miles would be wholly inappreciable in the most powerful telescope (spectroscopically armed), and so no peculiarities would be recognised as the result of processes taking place within such distances of the solar surface, however diligently the edge of the solar disc might be examined with the spectroscope. And, further, as respects the examination of the Sun's edge, on which so much stress has been laid, it is far from unlikely-if, indeed, it is not to be regarded as certain-that the visible edge of the solar disc lies considerably above the true limit of the photosphere. That light at the Sun's edge which seems to belong to the hemisphere of photospheric matter turned towards us comes probably from parts of the photosphere which lie really beyond the borders of that hemisphere, and are simply brought into view through the refractive power of the lower layers of the solar atmosphere. I am not here venturing a mere opinion or conjecture, though I profess no certainty of conviction in the matter. We have the evidence of

* This is in no way opposed to the evidence adduced by Sir John Herschel from the apparent uniformity of light derived from a glowing and transparent liquid not uniformly deep; for every part of such a liquid (at least in Sir John Herschel's experiment) is at appreciably the same temperature. If the lower layers could be heated to a higher temperature, the effect of depth would become apparent.

facts-observed by Carrington, Secchi, and othersshowing that the motions of the spots across the solar disc are really affected, as respects apparent rate, by the refractive power of the solar atmosphere. And it is impossible to doubt that if the apparent place of a spot can be affected in this way, then the solar regions beyond that hemisphere which is turned towards us at the moment, must be brought into view and form the real limits of the solar disc.*

I am sensible that I am not making definite statements as to the Sun's condition, but only stating difficulties. The difficulties are real, however, not imaginary, and ignoring them can serve no useful purpose.

When we turn to the details of the solar orb and its surroundings, we find the evidence slightly more definite; but still great difficulties surround us. In the course of the several chapters bearing on these matters nearly all the known facts which bear directly on the views we are to form respecting the Sun's physical constitution have been discussed as they have been described, so that but few words about the several solar features are here called for.

Taking the actual telescopic aspect of the solar spots

This really amounts to saying that the Earth, if viewed from the solar photosphere, would be visible above the solar horizon (as our Sun is after real sunset and before real sunrise) when the geometrical line to her passed below that horizon. We cannot doubt that the solar atmosphere would exert this refractive effect at least as powerfully as our own; and if the Earth could be seen in this way from parts of the Sun really turned away from her, then certainly those parts of the Sun must be visible from the Earth; for a visual ray passes along the same path from whichever of its extremities we suppose it to travel.

and their surroundings, a perplexing series of problems is suggested. These problems are indeed so perplexing as abundantly to justify the disagreement hitherto found among theorisers. Admitting the spots to be depressions, what is the real disposition of the matter which produces the appearances we call facula-penumbra-umbra-nucleus? The penumbra may belong to a lower layer; but the general arrangement according to which the willow-leaves on the penumbra point inwards towards the umbra seems to indicate a real connection between the penumbra and the faculous bordering. This arrangement is indeed sometimes so marked that one is led to imagine that the so-called willow-leaves are filamentous bodies which usually hang in a nearly vertical position and so appear nearly round, but when thrust aside during the formation of a spot hang nearly horizontally, the ends which had been lowest floating like streamers towards the region whence they had been removed. If we could but conceal from ourselves a large portion of the evidence we have (or else explain it away) this view might be insisted upon with pleasing confidence; but as a matter of fact it merely serves to indicate the impression produced by certain phenomena, and has at present no value whatever.*

A great difficulty lies in the fact that we have no

* Lest I should here be supposed to be too curtly criticising the views of others, let me hasten to say that the fancy thus summarily rejected is my own, and, so far as I know, as original as it is probably valueless. Yet I have not introduced it without a purpose. There is at least as much evidence in its favour as in favour of many theories which have been very confidently put forward.

clear evidence to show whether the Sun-spots are formed by forces acting from without or from within. Here I set on one side the theory that in a spot we see a region where a great heat has dissolved solid or liquid or cloudlike matter forming the photosphere, and that thus the intensely hot, but feebly radiating gaseous nucleus of the Sun (according to this theory) is disclosed. Kirchhoff has fairly disposed of this theory by showing that this intensely hot nucleus would be transparent to the light from the farther side of the Sun, and that therefore no spot could appear unless two openings on opposite sides of the Sun happened to be in the same visual direction.* refer now, not to this or similar theories, but to the definite problem, whether the seat of that action which leads to the formation of a spot lies below or above the level of the photosphere. The spectroscope shows that a spot is a region where certain gases exist at a lower temperature than in other parts of the Sun. But whether this low temperature results from the expansion of compressed gas erupted from the Sun, or from the fact that matter has reached the Sun from outer space, remains as yet altogether unknown.

* Fr. Secchi was the original propounder of this theory (not M. Faye, to whom it is usually ascribed). The theory really does account for many observed features of the solar spots, but it is none the less untenable. Fr. Secchi's answer to Kirchhoff's objection would seem to indicate that he has not recognised the exact force of that objection. He says it is not true that a gaseous nucleus would be perfectly transparent to rays from the further side of the Sun, for we see that our own atmosphere absorbs light as well as heat. Kirchhoff's argument is that the solar nucleus would be transparent on account of its existing at a higher temperature than the photosphere-according to the theory at 'east which he deals with.

I