a small portion of which is shown in the figure) is horizontal, so that the vertical lines are the dark lines of the spectrum. The horizontal lines indicate the regions of the spectrum corresponding to those parts of a spot where a general absorption takes place. It will be seen that where this general absorption is sufficient only to produce a degradation of brilliancy, all the lines in this part of the spectrum are visible. The

Illustrating the changes in certain lines in the spectra of Sun-spots.

F line (belonging to hydrogen) is, however, peculiarly affected across the whole region of the spot. At the upper and lower extremity we see it of its normal width, while over all the remaining breadth of the spectrum, except two small portions, it is much broader and has shaded edges. In one place it is bent; along another part of its length a narrow line of light is seen to be almost centrally placed upon it; and lastly in two places it appears bright and irregularly shaped.

Now, the facts here noticed (first observed I believe by Mr. Lockyer, and confirmed by Dr. Huggins, Fr. Secchi, and Professor Young) are thus (probably) to be interpreted. Beginning with the top of the line F in fig. 57, we have first the normal black line showing that in the part of the Sun included within the uppermost part of the slit, the hydrogen is in its ordinary condition as respects temperature. It is less heated than the matter whence the main portion of the solar light is radiated, and so absorbs that portion of the light which it has itself the power of radiating. Next we come to a bright hydrogen line of the normal width on a shaded background. The corresponding part of the Sun (that is the part next below the former within the slit) is, then, either somewhat reduced in temperature, or else partially covered by generally absorbing matter, over which there is a layer of hydrogen at the normal pressure, but more heated than the radiating region below. Hence in this place the hydrogen radiates more light than it absorbs, and the line F is rendered relatively bright. Next is a region where the hydrogen line is still bright but very much wider. Over the corresponding portion of the Sun, therefore, the hydrogen not only exists at a higher temperature but at a greater pressure. Then we come to the widened dark line, indicating that over the corresponding portion of the Sun there is hydrogen at a relatively low temperature and at an abnormally high pressure. The bend towards the red end of the spectrum indicates that the corresponding portion of the

hydrogen envelope is moving from the eye,* or, in other words, that there is in this part of the Sun a downrush of hydrogen. Where we see a relatively bright line superposed on the relatively dark one, we learn that above the compressed hydrogen at a relatively low temperature there is a layer (or tongue, or prominence) of more heated hydrogen. While, lastly, where we see the pointed dark line close to the bottom of the figure, we learn that above hydrogen as heated as the general radiating substance of the Sun, there is the usual layer of hydrogen at lower temperature, very shallow where the line is pointed, but deepening within a short distance to its normal condition.

Along a narrow strip, then, crossing the width of a solar spot, all these varieties of condition are thus recognisable. Nor is hydrogen the only element whose lines exhibit such peculiarities. The lines of sodium, magnesium, barium, and other elements, have been observed to exhibit similar indications of violent action, rapid motion, and remarkable changes of pressure. But perhaps the most striking of all the phenomena revealed by the spectroscope is the occurrence, in the spectra, of large spots, of lines and bands corresponding to those due to the presence of aqueous vapour in our own atmosphere. Fr. Secchi not only testifies to this, but

The vertical dotted lines 1, 2, 3, on either side of F indicate how far the line F should be shifted to indicate a velocity of 8, 16, and 24 geographical miles respectively from or towards the eye. The decimal figures between the vertical lines numbered 1, 1, indicate the length of the light-waves (in parts of a millimetre), corresponding to the part of the spectrum where the line is.

he describes experiments by which he convinced himself that these lines really belong to the spots, and not to our own atmosphere. He found that these 'water-lines' were not visible when the instrument was directed in clear weather to unspotted parts of the solar disc; but that as the instrument was shifted, the approach of a spot was clearly indicated by the appearance and gradually intensifying of these lines. When the sky was covered with thin clouds he saw the same lines towards whatever part of the Sun the instrument was directed; but they always appeared strongest in the spectrum of a spot.*

The second point which I wish to notice, in conclusion, is the evidence of the polariscope respecting the general condition of the solar photosphere. I do not feel justified in giving space to an account of the principles on which polariscopic analysis depends, because, to say the truth, the polariscope has thrown but little light on the subject of solar physics. I therefore merely state that light under certain conditions of emission, reflection, and refraction, acquires a peculiar property called polarisation, by which its capability for subsequent reflection or refraction is materially modified. We have here to deal with emission; and the special law which concerns us is this, that light emitted from an incandescent solid or liquid at a

* The reader is referred for fuller details than there is here space for to Dr. Schellen's work Die Spectralanalyse, already referred to. The English Edition lately published, under the able supervision of Dr. Huggins, is specially worthy of very careful study in all matters relating to the spectral analysis of the Sun.

very oblique angle is partially polarised in such sort that when incident on a plane at right angles with the angle of emanation, the polarised portion does not, like the rest, undergo reflection. Now, the light from near the edge of the Sun's disc shows no signs of having this particular quality. Hence, Arago and others have concluded that the solar photosphere cannot be formed of incandescent solid or liquid substance, but must necessarily be gaseous. We shall have occasion further on to consider the bearing of this evidence on the views we are to form respecting the Sun's physical constitution. It is necessary to note, however, that Sir John Herschel has called in question Arago's conclusion; and, without asserting that the solar photosphere must. necessarily be either solid or liquid, he has shown that the evidence of the polariscope is more than questionable, since the Sun can by no means be regarded as a smooth, uniform globe. Its surface is, in all probability, so rough and uneven that the light received from near the edge may come for the most part from surfaces nearly at right angles to the visual line.*

Here I conclude my survey of the solar surface. I have presented but such portions of the vast mass of material really available as seemed most instructive

The case may be compared with that of the Moon. If the Moon were a smooth, uniform globe, she ought, when full, to seem much darker near the edge than near the centre of her disc. That she does not is due to the inequalities of her surface; and Dr. Zöllner has been able to show from the observed luminosity of the Moon at different times that the probable average inclination of the lunar mountains is about 56 degrees.

R