the Cape of Good Hope, by direction of the Admiralty, and placed under the command of Lieut. Moore, R.N., who had been one of the officers of the Terror' in the Antarctic expedition, and therefore well versed in conducting operations of the kind. The Pagoda' started from the Cape in January 1845, and returned in June, having accomplished its mission, and thus completed the survey of the Southern hemisphere. The observations from the four colonial observatories, as well as from the Antarctic Expedition, were regularly forwarded to this country, where, under the able supervision and untiring industry of General Sabine, they soon bore fruit in copious additions to our knowledge of this branch of physical study.

One of the first results of his labours was to clear away a fallacy respecting the diurnal variation in declination, which was entertained by nearly every magnetician at that time. It had long been known that in the middle Northern latitudes the needle is to the east of its mean position in the forenoon, and to the west of it in the afternoon; and that the opposite is the case in the middle Southern latitudes. Hence, it was argued, there must be some line on the earth's surface where the causes, whatever they are, which produce these opposite motions in the two hemispheres neutralise each other, and where, therefore, there is no diurnal variation. The next question was, Where is this line? Is it, as Arago asks, the geographical equator, or the line of no dip; or, as was imagined by some, the line of least intensity? The determination of this question had its influence in the choice of site for two at least of the colonial observatories-St. Helena and the Cape-as well as for that at Singapore, which was founded by the East India Company. St. Helena is near the point of least intensity on its meridian, whilst it is at some distance from the Equator and the line of no dip. The Cape, though rather farther from the line of least intensity, was chosen for the same reason; whilst Singapore was selected as being near both the Equator and the line of no dip. The first five years of observation enabled General Sabine to announce to the world the unexpected fact that this so-called line of no horary variation, about whose existence all had been agreed, was not only not found, but never would be found, for the simple reason that it did not exist. In a paper read before the Royal Society in 1847 he shows, from the observations at St. Helena, that at that station the motion of the needle accords with that observed in the Northern hemisphere during the period from April to September, whilst from October to March it exhibits the features of the Southern hemisphere. The change of direction

takes place soon after the sun crosses the Equator, the motion during what may be called the equinoctial months-i.e. March, April, September, and October-partaking more or less, from one day to another, of the character of both seasons; but at all other times the contrariety is decided. Subsequent investigations showed that a semi-annual variation similar to the above exists at all stations where observations have been made; i.e. that in the Northern hemisphere, from April to September the easterly motion of the forenoon, and the westerly motion of the afternoon, are increased; whilst the contrary takes place from October to March; and that in the Southern hemisphere the westerly motion of the forenoon and the easterly motion of the afternoon are decreased from April to September, and the contrary from October to March; showing this remarkable fact, that whilst the mean diurnal variation changes its direction in passing from one hemisphere to the other, yet the direction of the semi-annual variation remains the same; this direction depending on the position of the sun with respect to the Equator, and not to the zenith of the place of observation, since it changes sign soon after the Equinoxes. If it be asked why no reversal takes place in our latitudes, the answer is easy. We may regard the diurnal variation as the resultant of two variations, one of which we may call the mean solar-diurnal variation, and the other the semi-annual inequality; and the actual diurnal variation as it presents itself to our observation will result from the superposition of these two. Now, in the middle Northern latitudes the mean solar-diurnal variation ranges from 9' to 10', whilst that of the semi-annual inequality is only from 3' to 4'. The former will, therefore, be the dominant variation, the effect of the latter being merely to alter the amount of variation, increasing it during one half of the year, and decreasing it during the other half. As we approach the tropics the range of the mean solar-diurnal variation diminishes, whilst the semi-annual inequality remains constant in direction and amount. We shall, therefore, have stations where, as at St. Helena, the latter constitutes nearly the whole of the diurnal variation, and where, consequently, we meet with the semi-annual reversal which has been proved to exist at that station. A semi-annual variation has been found to exist likewise in the values of the dip and the intensity, depending on the relative positions of the earth and sun, and having its periods of maxima and minima nearly coincident with the solstices; and it appears that in both hemispheres the needle is most nearly vertical and the magnetic intensity is the greatest at the same time, i.e.

when the earth is nearest the sun and moves with the greatest velocity in its orbit. We need hardly remark that, on the theory which refers these changes to temperature, they ought to occur at opposite periods of the year in the two hemispheres, whereas the reverse is in fact the case.

These discoveries, besides having given us a clearer insight into the true nature of the diurnal variation, are of deep interest theoretically, from the indications which they contain of the cosmical features discernible in some at least of the phenomena of terrestrial magnetism, and hitherto quite unsuspected. These features were brought still more prominently into notice by the remarkable discovery, due to General Sabine, of the periodical character (when considered in their mean effects) of those irregular perturbations of the needle which had hitherto baffled all attempts to reduce them to law. It will be remembered that one of the principal objects sought by the establishment of the British colonial observatories was the study of these perturbations on the system commenced by the Göttingen Association. In furtherance of this object the observations taken at the different observatories were regularly forwarded to Woolwich for examination, arranged in monthly tables. Even a slight examination of these tables showed that at any particular station the disturbances did not occur arbitrarily, or with equal frequency at all hours; also that the hours principally affected were not the same at different stations; and moreover that whatever law they might follow, it was certainly not that of the regular diurnal variation. It therefore became necessary, in the first place, to devise some means of separating the disturbed observations from the rest. This, of course, could never be done completely; but by fixing a limit (constant for the same station), beyond which all observations should be considered as disturbed, a sufficient number could be set apart for an examination of the laws of disturbance; whilst it might be expected that from the remaining portion of the observations, thus freed from the effects of the larger disturbances, a more correct knowledge of the regular variations might be obtained than would have been the case had this elimination not been effected. The result of this experiment was first brought before the Royal Society, in January 1851, in a paper by General Sabine on the Pe'riodical laws discoverable in the mean effects of the larger magnetic disturbances.' The stations selected for comparison were Toronto and Hobarton, as being both extra-tropical, in opposite hemispheres, and having nearly the same North and South latitudes respectively. The disturbances discussed were

6

those of the declination; and it appeared that when these were divided in easterly and westerly-i.e. those increasing the easterly and westerly deflections respectively, and these again arranged according to the months of the year, or again according to the hours of the day-they showed at both stations periodical changes depending on seasons and hours, and therefore evidently pointing to the sun as their primary source. The years included in this first analysis were the three 1843-5. On extending the investigation to the next three years, another and still more noteworthy feature presented itself. It appeared that not only was there a variation in the disturbances from month to month and hour to hour, in any particular year, but that the aggregate amount of disturbance varied from one year to another. From 1843 to 1849 there had been a progressive annual increase of disturbance, and that to an extent which could not be supposed accidental; the amount in 1848-9 being more than double that in 1843-4. Concurrently with this increase of disturbance there had been a similar though smaller increase in the range of the diurnal oscillation of each of the three elements. But for this, the cosmical features of the phenomenon might have remained undiscovered for some time longer, as the period granted to the colonial observatories had expired and the hourly observations broken off. Fortunately there were records of the diurnal range in the declination for 1841-2. This range proved to be greater than that for 1843-4, which might, therefore, be fairly assumed to be the epoch of minimum range, and therefore, probably, of minimum disturbance; and it also appeared that the range for 1850-1 was less than that for 1848-9, thus giving that epoch the character of a maximum. This naturally raised the question whether there might not be a period of disturbance of which the interval from minimum to maximum was five years. Any doubt that might have been felt as to the answer to be given to this question was set at rest by the publication, about the same time, of Schwabe's table of the variations of the solar spots from 1826 to 1850, showing that during that interval these spots had shown a period of between ten and eleven years, having maxima in 1828, 1837, and 1848, and minima in 1833 and 1843, of which 1843 and 1848 were known to be periods of minimum and maximum magnetic disturbance respectively; and as the connexion between the disturbances and the sun was evident by their conformity to solar hours, there could no longer be any hesitation in assigning to them a period coincident with that of the solar spots. This discovery was first communicated to the Royal Society in March 1852;

and subsequent observation from that time to the present has shown that the coincidence then pointed out was not accidental, but that there is such a mutual interdependence between these two classes of apparently dissimilar phenomena as to compel us to regard them as the results of some common cause.

We look upon this discovery of the periodical character of the disturbances and their cosmical origin as, without exception, the greatest advance ever made towards a true understanding of the magnetic phenomena. Whether we regard it from a practical or a theoretical point of view, it is impossible to overestimate its value. Practically it has produced a complete revolution in the principles of the methods employed in dealing with magnetic observations. It was at one time supposed that by simply multiplying the number of observations, the irregular perturbations might be expected to neutralise one another to a certain extent, and that thus their effects would partially disappear from the result. This view of the matter, which would have been justifiable had the disturbances occurred arbitrarily, is quite inadmissible now that we know their periodical character. It is now certain that in order to arrive at a true knowledge of the regular variations, the disturbances must be eliminated, at least as far as practicable; and the method now universally adopted is that devised by General Sabine for the treatment of the colonial observations. The disturbances of each element have to be taken. separately, and again subdivided into six distinct categoriesthose which increase the easterly and those which increase the westerly deflection, those which increase and those which decrease the dip, and those which increase and those which decrease the intensity. Each of these classes is found to have distinct, and apparently independent laws, requiring separate study and analysis. Each element has its proper hours of principal disturbance, well defined at each particular station, but varying, apparently without limit, in passing from one station to another. To confine our attention to the most observed of the three elements, the declination; at some stations there is a predominance of easterly disturbance, at others of westerly. Moreover, when we come to form the curves of easterly and westerly disturbance, we find two forms, and two only, prevailing without exception at all places where the analysis has been made, one characterising the easterly and the other the westerly disturbances; indicating seemingly the existence either of different (and probably only two dif ferent) sources of disturbance or different forms of activity emanating from the same source, whilst the features which