on the nights of those days, not only in these latitudes but at Rome; in the West Indies; on the tropics within 18° of the equator (where they hardly ever appear); nay, what is still more striking, in South America and in Australia,-where, at Melbourne, on the night of September 2, the greatest aurora ever seen there made its appearance. These auroras were accompanied with unusually great electro-magnetic disturbances in every part of the world. In many places the telegraphic wires struck work. They had too many private messages of their own to convey. At Washington and Philadelphia, in America, the telegraphic signal-men received severe electric shocks. At a station in Norway, the telegraphic apparatus was set fire to; and at Boston, in North America, a flame of fire followed the pen of Bain's electric telegraph (which writes down the message upon chemicallyprepared paper).'

It is demonstrated then that some association exists between the disturbance of the solar photosphere and the phenomena of terrestrial magnetism. What the nature of the association may be is not so clearly apparent. We have seen that the solar spot period has been supposed to be associated by Wolf with the motions of the planets, and we shall presently see that the phenomena of spots, -as their change of form and of size, the regions in which they appear, and so on,—have been

transcription). Although several years had elapsed since I read those statements, I do not find that my account required correction in a single essential particular.

conceived to depend on the planetary motions. If these views be correct, we should have to account for laws of association between the planetary motions, terrestrial magnetism, and solar spots. But which of these three orders of phenomena should be regarded as the cause of the others; or whether the association be of the nature of cause and effect, and not rather due to some as yet unknown common cause of the two latter classes of phenomena; or whether, finally, the planets' motions, without being the direct cause of the other phenomena, yet indirectly brings them about, remains to be determined.*

Let us next turn to Mr. Carrington's researches, intimately associated with Schwabe's, and among the most important of all the contributions which have

I would point out, however, that the consideration of the association here discussed led General Sabine, Professor Challis, and Dr. Stewart to inquire whether the coloured prominences of the Sun may not be due to solar auroras. Since we now know that the prominences are not of this nature, may we not transfer the suggestion to the solar corona and the zodiacal light? May not these be the true solar auroras? The latter, at any rate, gives the same spectrum as the aurora, and the former gives a spectrum closely analogous. If, as General Sabine suggests, an auroral outburst in the Sun may perhaps be responded to simultaneously by the different planets, so that the whole solar system would seem to thrill almost like a living being under the magnetic excitement,' it seems at least probable that the solar auroras would extend to distances enormously exceeding those assigned them when the prominences were taken to be such auroras. And as we have abundant evidence that terrestrial auroras occur where the atmosphere has an inconceivable rarity, it would seem possible that under the enormous action of the Sun, even the quasi-vacuum of the interplanetary spaces might be traversed by electrical discharges. The meteoric systems occupying more or less densely the whole of these spaces, would seem to afford the requisite channel for this electric action.

been made in recent times to our acquaintance with solar physics.

The object to which Mr. Carrington specially devoted his energies was the endeavour to detect regularity in the distribution of the spots, the determination of the true period of rotation of the body of the Sun, and the detection of systematic movements or currents of the surface, if such existed in a definable manner.' In these researches he perfectly succeeded.

It would be desirable if space permitted to describe Carrington's method of observation, and the exact and systematic processes by which he deduced his results. I must be content, however, so far as these points are concerned, to refer the curious reader to Carrington's voluminous and masterly treatise, in which also will be found an interesting account of the results obtained by former observers. Here I have room only for the conclusions to which he was led by his series of observations, which commenced in the year 1853 and ended in the year 1861.

He discovered, in the first place, that the discrepancies between the values formerly deduced for the Sun's rotation, arise from real differences in the velocities with which the spots move in different solar latitudes. Near the equator a spot moves at a rate indicating a more rapid rotation than in higher latitudes. Further, even among spots in the same latitude, proper motions may be recognised. These latter motions are to be regarded, however, as abnormal, and simply rendering unreliable such observations as are made on

P

but a few spots. The peculiarities affecting the motions of spots in different latitudes have been reduced by Mr. Carrington into a formula.* The following table gives the observed rates of rotation for different latitudes (the formula being based on these values):—

* This formula is as follows:-Let & be the angle through which a part of the Sun in latitude A rotates in one day. Then

It is remarkable that in all southern latitudes the observed daily mean rotation is less than in the corresponding northern latitudes. It is doubtful whether we have in this relation any indication of the true cause of the observed variations in the rate of rotation, or merely a peculiarity which would have disappeared in a longer series of observations. In favour of the former view, we have the consideration that the determination for each southern as well as for each northern latitude was independently effected, so that the coincidence of the results indicates the existence of some real cause. If Sir John Herschel is right in considering that the more rapid rotation near the solar equator implies the action of external matter in maintaining the rotation of the photosphere, it may be suggested that the northern surface of the Sun being directed somewhat more fully towards that region whither the Sun's

I shall have occasion to dwell further on upon the significance of the varying rotation-period deduced for different solar latitudes. This result of Mr. Carrington's labours cannot but be regarded as one of the most important contributions recently made to our knowledge of solar physics.

Spörer has re-examined the whole subject, taking into account later observations, and in particular those which have been made by Fr. Secchi. The following table includes the general elements of the solar rotation as deduced by Carrington and Spörer, and reduced by Secchi to the year 1869:

(No weight can be attached to the last two decimal figures in Spörer's value of the diurnal rotation and the rotation period.) In fig. 46 the varying presentation of the Sun towards the Earth on account of the inclination of his equator to the Earth's orbit is exhibited as

proper motion is carrying him (see the concluding chapter of this work, and the illustrative cuts), would probably be more exposed to the influence of this external action—'the frictional impulse of circulating planetary matter in process of subsidence into, and absorption by, the central body'- much as our northern hemisphere is saluted with a larger number of meteoric missiles from June to December, when the northern hemisphere is in advance, than from December to June, when this hemisphere is towards the more sheltered side of the Earth.