characterise the easterly deflections at one station are transferred to the westerly deflections at another, and vice versâ, as though they lay on opposite sides of the principal meridian of disturbance; also at some stations the principal hours of disturbance fall during the day, and at others during the night, producing what Humboldt calls the nocturnal episode in the diurnal variation of the needle; which nocturnal episode is not found where the principal hours of disturbance occur during the day; whereas on the old theory which referred the diurnal motion of the needle to the changes of temperature produced by the passage of the sun from east to west, and again from west to east of the meridian, there must of necessity be a double progression everywhere in the twenty-four hours, one maximum and one minimum falling during what may be termed the night hours. The elimination of the disturbances is thus shown to be absolutely necessary for restoring to the diurnal variation its normal character. In some localities, indeed, it is hard to recognise the well-known features of the diurnal curve when the disturbances are left in. A remarkable instance of this occurs at Point Barrow. One of the most striking features of the undisturbed diurnal variation in declination is the regularity with which, in the Northern hemisphere, the needle reaches its westerly extreme at about 1 P.M., and Point Barrow forms no exception to this rule when the disturbances are eliminated, whereas, when they are retained, the greatest westerly deflection is sometimes not attained till between 10 and 11 P.M.! It may be stated that no point on the earth's surface has yet been found where magnetic disturbance exhibits so much activity as at Point Barrow, and where there is so constant a display of its almost universal concomitant, the aurora borealis.* We here take occasion to

* When H.M.S. 'Plover' (Captain Maguire) was stationed at Point Barrow for nearly two years, between 1852-4, the aurora was observed six nights out of seven (in round numbers) during the months December, January, and February, for two successive years (1852-3).

The connexion between the aurora borealis and magnetic disturbance was, we believe, first observed by Hiorter at Upsal, 1741, though it had been suggested as probable by Halley as early as 1716,-another instance of the remarkable penetration of this great man. The effect of the aurora upon the magnetic needle has been examined at great length by Arago in his 'Meteorological Essays.' The principal results of his investigations may be summed up thus-(1.) The aurora affects the needle even in places where it is not visible. (2.) In general, the west declination increases before the appearance of the aurora, and sometimes even continues to increase after its appearance; then the

repeat what we have said of the extreme importance of the discovery which has shown the connexion between terrestrial magnetic disturbance and the physical aspect of the sun, as opening up a new field of research by suggesting the possi bility of relations hitherto unsuspected between our planet and the other bodies of the solar system; and we are persuaded that here if any where will be found the clue which will enable us to disentangle the various phenomena, and assign each perturbation to its proper source. Already the increased attention paid to solar physics has resulted in a more perfect knowledge of the period and laws of increase and decrease of the solar spots. The labours of Messrs. De la Rue, Balfour Stewart, and Loewy, together with those of Professor Wolf at Berne, have determined the mean length of the sunspot period as being rather more than eleven years (11.1 is Professor Wolf's determination). It also appears that this period is not equally divided between the times of increasing and decreasing activity of spots, the mean period from minimum to maximum being about 3 years, whilst that from the maximum to the next minimum is about 7, or more than double; the ratio of the two periods being 2-15. From this law, which seems to connect each period of sun-spot increase with the next period of decrease, combined with the fact that the shorter the period the greater the energy of spot-production, whilst a lengthening of the period is compensated for by diminished activity, seems, in the opinion of Mr. Balfour Stewart, to point to the conclusion that the energy of spotproduction, wherever it may be situated, is for every period constant. Of the ultimate source of this energy we are at present ignorant; but our knowledge has been advanced one step farther, to the point of establishing a connexion between the times of maxima and minima of spots and the configuration of the solar system. M. Fritz, who has investigated the comparative influences of the planets both singly and together, finds the greatest coincidence of maxima of spots with the

oscillations become very large, and the needle begins to return towards the east, and only stops when it has passed its normal position, which it does not usually regain for some hours. It appears from the observations of MM. Lottin and Bravais at Bossekop, in 1838-9, that the degree of disturbance of the needle varies with the intensity of the aurora. When this is faint and generally diffused, there is often no disturbance; but when the coruscations are very brilliant and coloured, the oscillations of the needle will sometimes range through several degrees. The needle will often predict the aurora by an abnormal motion westwards during the whole day.

time when Jupiter and Saturn are in quadrature, and the greatest coincidence of minima when they are in conjunction. There is also a minor coincidence of maxima when Jupiter and Venus are in quadrature. According to Mr. Balfour Stewart, who has very carefully examined the behaviour of sun-spots with regard to increase and decrease as they cross the sun's disk, it appears that with respect to the planets Venus and Mercury, the size of a spot attains (on the average) its maximum on the side which is turned away from these planets, and its minimum when it is in their neighbourhood. Mr. Stewart also states that he has found no indication of any such behaviour with respect to the planet Jupiter.

The late Professor Hornstein of Prague announced, not long since, the discovery that each of the three magnetic elements is subject to a periodical variation of about 26 days, which he regards as the effect of the sun's rotation; the true periodic time of which, as deduced from the above period, is 24.55 days, very nearly agreeing with the time of rotation of spots in the sun's equator as derived from astronomical observations.

It may well happen that more extended inquiry will induce a modification of some of the laws just enunciated, which, in the present state of our knowledge, it must be admitted are simply empirical. Nevertheless, that magneticians should have turned their energies in this direction, we cannot but regard as a hopeful augury for the future. But whilst the sun and planets are claiming so large a share of our attention, it would be hard if our nearest neighbour and satellite, the moon, were left out in the cold. That the moon produces sensible though small variations in the magnetic elements, was first announced to the Bohemian Society of Sciences by M. Kreil in 1841, but it is only of late years that lunar action has received the attention it merits, and even yet the inquiry cannot be said to be by any means exhausted. The most remarkable feature in the lunar diurnal variation is the double progression for each of the three elements in the twenty-four hours, the maxima and minima being separated by about equal intervals of six hours each; the actual turning hours being different at different stations, but never far distant from the hours of upper

* The Astronomer Royal, in a recent communication to the Royal Society, states that a careful examination of the observations of several years at Greenwich has not enabled him to detect any period of the kind indicated by Professor Hornstein, whose investigation is limited

to observations made in 1870.

and lower culmination. It appears, however, from the researches of Mr. Chambers at Bombay and Dr. Neumayer at Melbourne, that we get a very imperfect idea of the real magnetic action of the moon from the mean annual variation. Mr. Chambers thinks it necessary to distinguish between the cases where the sun and moon have the same or opposite declinations, again subdividing each of these into others according as the moon's declination is increasing, north or south, or decreasing; and in each of these sixteen cases he determines the variation for each of the four lunar quarters, making in all sixty-four varieties for consideration. Each of these presents distinct, and sometimes opposite features, which become obliterated in the mean annual variation, but in nearly every case, the leading feature of the lunar variation-the double progression is clearly exhibited.

Dr. Neumayer, though his grouping is less minute than that of Mr. Chambers, agrees with him as to the necessity of considering the moon's declination as well as that of the sun; as when they are both on the same side of the Equator the variation is more regular than when their declinations are of opposite kinds. In every case the maximum of easterly deflection occurs near the time of the lower transit, and the minimum about six hours afterwards, with secondary maxima and minima at about one hour and nineteen hours respectively. In the year 1863 Dr. Bache announced the existence of a semi-annual inequality in the lunar-diurnal variation of declination, as shown by the observations at Girard College, Philadelphia; the amplitudes of the deflections (both east and west) being less during the six months from October to March than from April to September, while the times of maxima and minima fall nearly an hour earlier during the former than during the latter period. A semi-annual inequality has also been detected by General Sabine in the lunar-diurnal variation both at Kew and Hobarton. The results at Kew are in accordance with those at Philadelphia as regards the amplitudes, but not as regards the turning hours, these falling rather earlier apparently in the April-September than in the OctoberMarch period. Whilst at Hobarton the results are opposed to those at Kew in both particulars. It is worthy of remark that no trace of the so-called decennial period, which affects every inequality depending on the sun, has yet been found in the lunar-diurnal variation. The shortness of the time, however, during which observations of the requisite accuracy have been carried on, forbids us to assume too confidently that no such period exists, and may hereafter be discovered.

It would, indeed, be strange if the careful examination, to which the magnetic elements are now subjected by so many able observers, should fail to detect, and that before long, any such inequality if it exist. With respect to the secular change the case is different. Of this it may be said, with as much truth as when Halley wrote two hundred years ago, that it is a secret reserved for the industry of future ages.' It is true that we know what Halley did not. In his time the direction of secular change (in this country at least) was westwards, and had been so from the time of the earliest recorded observation. There was nothing, therefore, as far as observation went, to preclude the supposition, that in process of time the needle would be directed to every point of the compass in turn. For us this question is set at rest. The westerly progression was arrested in this country in 1818, when the declination was 25° 30′ W. in London, since which time the secular change has been eastwards, the declination at Kew being now not quite 20° W. As, however, the last epoch of greatest easterly declination is unknown, we are still in ignorance as to the whole period of oscillation. Nor are we entitled to assume that the easterly retrogression will at all correspond to the previous progression towards the west. Otherwise, as we know that the declination was 0° in 1660, we might infer the epoch at which it will again vanish. Still less may we assume that the subsequent easterly motion will accord with the westerly, either in extent or duration. But though any such conjectures would be quite unwarranted in our present state of ignorance as to the sources and laws of the secular change, still the strictly progressive character of this change compels us to regard it as the expression of some determinate cause or causes. The question then arises, where are these to be found? Now, from whatever point on the earth's surface we contemplate the phenomena, we find ourselves in the presence of two distinct magnetic systems. This was first clearly recognised by Halley as a necessary consequence of even the scanty information then at his command, and the accumulated observations of two hundred years have corroborated in a very remarkable manner the conclusions at which he arrived,—that of these two systems, one was fixed, and the other in motion; and that the direction of the needle at any place resulted from the superposition of one of the systems on the other. It is well known that in order to get over the difficulty of a magnet having four poles,' as he expresses it, and to give a reasonable account' of the secular change, Halley imagined the earth