stormy cloud was observed about 11 P.M. in the N.N.W., and lightning was distinctly visible in the dark cumulous mass. Around this mass extended glories of a phosphorescent whiteness, which melted away into the darkness of the starry sky. Round the cloud was a single and uninterrupted corona, and outside this, two fainter coronæ broken by rifts which corresponded with each other. After the cloud had sunk below the horizon the glories were still visible. The light could not have been due to the moon or any foreign cause. The rays showed great mobility, and a sort of vibration intermediate between that of the aurora and the brush discharge' of the electric machine." He goes on to say that— that certain kinds of cirro-stratus are also regarded by | vations on the night of the 6th September 1865-“ A meteorologists as sure indications of rain and wind. Dr Richardson frequently observed the lower surface of nebulous masses illuminated by polar lights,--a fact illustrative of the comparatively low situation of these aurora. Biot, also, in the island of Unst, observed many auroræ that could not be higher than the region of clouds. Sir John Franklin in like manner observed low aurora. "The important fact," says he, "of the existence of the aurora at a less elevation than that of dense clouds was evinced on two or three occasions this night (13th February 1821, at Fort Enterprise), and particularly at 11 hours 50 min., when a brilliant mass of light, variegated with the prismatic colours, passed between a uniform steady dense cloud and the earth, and in its progress completely concealed that portion of the cloud which the stream of light covered, until the coruscation had passed over it, when the cloud appeared as before." Captain Parry, as stated in his third voyage, observed auroræ near to the earth's surface. It is said that while Lieutenants Scherer and Ross and Captain Parry were admiring the extreme beauty of a polar light, they all simultaneously uttered an exclamation of surprise at seeing a bright ray of the aurora shoot suddenly downward from the general mass of light, and between them and the land, which was only 3000 yards distant. The ray or beam of the polar light thus passed within a distance of 3000 yards, or less than 2 miles, of them. Further, Mr Farquharson observed in Aberdeenshire an aurora borealis not more than 4000 feet above the level of the sea. Fitzroy believed that aurora in northern latitudes indicates and accompanies stormy weather at a distance, and that straining and cracking of the ice may cause the hissing and whizzing sounds. M. Silbermann (Comptes Rendus, lxviii. p. 1051) notes facts which strongly confirm the connection of aurora with some form of cirrus cloud He says (of the aurora of 15th April 1869),-" At 11 hours 16 min. the phenomenon disappeared in a singular fashion. It appeared as if the columns of the aurora were still visible, but the stars were hidden, and it soon became obvious that fan-like cirrus clouds, with their point of divergence in the north, had taken the place of the aurora. Between 1 and 2 in the morning these clouds had passed the zenith, and let fall a very fine rain. On stretching out the back of the hind one felt a pricking of cold, and now and then there were minute scintillations in the nearest strata of air, like a hail of tiny crystals of ice, which afterwards turned to a rain of larger and larger drops. At 4 o'clock in the morning the cirrus of the false aurora was still visible, but deformed towards the top, and presenting a flaky aspect. One interesting point is, that the cirrus never appeared to replace the aurora either from the right or the left, but to substitute itself for it, like the slow changes of a dioramic view." "I had previously observed a fall of small ice crystals on the 30th April 1865. At 6 P.M. Paris seemed enveloped in a cirrus of vertical fibres, recalling those of amianthus, and more or less wavy. It was a rain of little sparkling prisms. At the same time I heard a rustling or crepitation, and on extending my hand I felt a pricking sensation of cold, and distinguished the crystals which fell and melted immediately." In a later memoir (Ibid., p. 1120) he remarks that many storm-clouds throw out tufts of cirri from their tops, which extend over a great portion of the sky, and resolve themselves into a very fine and cold drizzle, which frequently degenerates into a warmer and more abundant rain. Usually the fibres are more or less sinuous, but in much rarer cases they become perfectly rectilinear, and surround the cloud like a glory, and occasionally shine with a sort of phosphorescence. As an illustration he quotes his obser "Luminous clouds have been frequently observed. There are many examples in Gilbert's Annals, and we may recall also the observations of Becaria, Deluc, the Abbé Rozier, Nicholson, and Colla. Mists also are occasionally luminous, as, for instance, that observed by Dr Verdeil at Lausanne in 1753, and by Dr Robinson in Ireland." A still more curious fact is mentioned by Sabine, who, during his magnetic survey, anchored some days at Loch Scavaig in Skye. This loch is surrounded by high and bare mountains, one of which was nearly always enveloped in a cloud, resulting from the vapours which almost constant west winds brought from the Atlantic. This cloud at nights was permanently self-luminous, and Sabine frequently saw rays similar to those of the aurora. He entirely repudiates the idea that the rays could be due to aurore beyond the mountain, and is sure that these phenomena, whatever their nature, were produced in the cloud itself. Silbermann asserts that aurora are preceded by the same general phenomena as thunderstorms, and concludes that everything had happened as if the aurora of 1859 and 1869 had been storm-clouds, which, instead of bursting in thunder, had been drawn into the upper parts of the atmosphere, and their vapour being crystallised in tiny prisms by the intense cold, the electricity had become luminous in flowing over these icy particles. This view is very strongly supported by the observation of Professor Piazzi Smyth that the monthly frequency of aurora varies inversely with that of thunderstorms. The following are his numbers of relative frequency, the means of all observations of the Scottish Meteorological Society prior to 1871 : 24.0 20.1 Mean of whole year... It must, however, be remembered that the observed frequency of aurora is much affected in Scotland by the continuous twilight during the summer months. If there be this connection between thunder-clouds and auroræ, it is not improbable that the "dark segment" is sometimes a real cloud or mist, situated at a height where the density of the air is too great for luminous discharge; and in several cases Silbermann has seen auroral rays rise from small clouds, which gradually melted entirely away, or left a small non-luminous nucleus when their electricity was discharged. If, as would certainly be the case in a mist, any portion risa ht. of the auroral light is reflected, whether it be its own or derived from some other body, it should be polarised; but so far polariscope observations are deficient, and give no certain information. It is difficult to separate the proper polarisation of the aurora from the mere atmospheric polarisation of the sky. Mr Ranyard, who appears to have used a double-imaged prism and Savart during the great aurora of Feb. 4, 1872, and also to have made some observations on that of Nov. 11, 1871, did not detect polarisation. On the other hand, Prof. Stephen Alexander, in his report on his expedition to Labrador (App. 21, U. S. Coast Survey Rep., 1860), found strong polarisation with a Savart, and, singularly enough, thought it strongest in the dark parts of the aurora. The observations were made in lat. about 60°, in the beginning of July, and near midnight, but he does not state whether there was twilight or any trace of air polarisation at the time, nor does he give the plane of polarisation. With regard to the height of auroræ, Sir W. R. Grove (Nature, vol. iii. p. 28) states that he saw an aurora some years ago at Chester in which the rays came between him and the houses; and Mr Ladd observed a similar case in which the lighthouse at Margate was visible through a ray. The evidence, however, appears strong that aurora is usually at a very great height. Dalton calculated the height of an auroral arch, which was seen as far north as Edinburgh, and as far south as Doncaster, and at most intermediate places, from its apparent aititude, as measured by its position in relation to the stars as seen from Kendal and Warrington, 83 miles apart. The resulting height was about 100 miles, and the position slightly south of Kendal. An observation at Jedburgh confirmed this, but some taken at Edinburgh placed it above Carlisle at a height of 150 miles. Dalton, however, considered the former reckoning the more trustworthy. Backhouse has made many calculations, and considers that the average height of aurora ranges from 50 to 100 miles, and numerous other observers have calculated similar heights. All these observations, however, are liable to the objection, that different observers may really have seen different arches, of which, as has been remarked, there are often several concentric ones. It is not likely that this was really the case in most instances, but it has, no doubt, sometimes occurred, and may account for the heights of 500 to 1000 miles calculated by early observers. This difficulty is met by a method proposed by Frof. H. A. Newton (Sill. Jour. of Sc., 2d ser. vol. xxxix. p. 286) for calculating the height by one observation of altitude and amplitude of an arch. It seems almost certain that the auroral arches are arcs of circles, of which the centre is the magnetic axis of the earth; or, at least, that they are nearly parallel to the earth's surface, and probably also to the narrow belt or ring surrounding the magnetic and astronomical poles, and passing through Faroe, the North Cape, and the north of Nova Zembla, which Loomis and Fritz have found to be the region of most frequent aurora. This being assumed, Prof. Newton finds that, d being the distance from the observer to the centre of curvature of the nearest part of this belt (which for England is situated about 75° N. lat., 50° W. long.), h the apparent altitude of the arch, 2a its amplitude on the horizon, x its height, R the earth's radius, and c the distance of the observer from the ends of the arch, calculated by other methods. It cannot well be objected that such altitudes are beyond the limits of our atmosphere, since Prof. A. S. Herschel (Nature, vol. iv. 504) gives the height of twenty meteors varying from 40 to 118 miles, with an average of about 70 miles, and it is almost certain that these bodies are rendered incandescent by atmospheric friction. Assuming 0° C. as the temperature at the earth's surface, and the absolute zero, - 273° C., as a minimum for the auroral region, the pressure would be about 0.2 millimetre (0.0078 inch) at a height of 100 kilometres (62 miles) above the earth's surface. This result, of course, assumes a good deal; but if correct, it implies a vacuum attainable with difficulty even with the Sprengel pump. The pressure may, however, be much greater in the path of the auroral beams, since, as Prof. A. S. Herschel suggests, electrical repulsion may carry air or other matter up to a great height. A similar effect is observed in the so-called vacuum tubes, in which the pressure becomes much greater in the narrow central part, while the discharge is passing. It is found that the apparent altitude of the auroral corona is always a little less than that indicated by the dipping needle, owing to the curvature of the lines of magnetic force, or, in other words, because its altitude corresponds with the inclination of the parallel of latitude over which it is actually situated; and Galle has suggested (Pogg. Ann., cxlvi. 133), that from this divergence the height may be calculated, and, indeed, gives a series of heights so determined, which do not differ materially from Prof. Newton's. It is, however, doubtful if the position of these coronæ, and consequently the value of the small angle (not more than 4° or 5°), admit of sufficiently accurate determination for such a use. observa Early observers, and especially Mr Canton, conjectured Spectrothat the aurora was an electric discharge in the rarefied scopic upper atmosphere, and the resemblance between it and the tions. phenomena exhibited by discharges in an air-pump vacuum confirmed the idea. Recent spectroscopic observations have thrown some little doubt on this conclusion, or at least have shown that there is still a mystery left unexplained. When the light of any glowing gas is analysed by the prism, it is found to consist of a series of coloured lines and bands, of which the number and position is dependent on the nature of the gas, and which is called its spectrum. The light of the aurora gives a spectrum usually consisting of a single line in the greenish yellow, which does not coincide with a principal line of any known substance, -a spectrum totally different from those of the gases of the atmosphere. Besides this line there is occasionally visible a sharp line in the red, and several fainter and more refrangible bands. The following table includes most of the principal determinations of the auroral lines, which have hitherto been published: (1), 5571 Vogel +2 ±0.92 5570 (2), (3). He gives the heights of twenty-eight aurora calculated by this method, ranging from 33 to 281 miles, with a mean of 130 miles. The method, of course, rests on the assumption that auroral arches are arcs of circles, but it is decidedly confirmatory both of this assumption and of the heights 5545 5569 5570 { Winlock Oettingen Struve N. German Polar Peirce Respighi Expedition 5573 5579 C. Piazzi Smyth 5600 Ellery Vogel remarks that the line at 5569, which is often the only one visible, as well as the faint band at 4667, become noticeably fainter when the red line is visible, while under the same circumstances that near 5189, as well as the red line, is very brilliant. This fact, which has also been noted by other observers, makes it almost certain that the auroral spectrum is not a simple one, but is derived either from two or more sources, or from the same source under very varying conditions. Angström says (Nature, x. 211) "It may be assumed that the spectrum of the aurora is composed of two different spectra, which, even although appearing sometimes simultaneously, have in all probability different origins. The one spectrum consists of the homogeneous yellow light which is so characteristic of the aurora, and which is found even in its weakest manifestation. The other spectrum consists of extremely feeble bands of light, which only in the stronger aurora attain such intensity as enables one to fix their position even approximately. As to the yellow line in the aurora, or the one-coloured spectrum, we are as little able now as when it was first observed to point out a corresponding line in any known spectrum. True, Fiazzi Smyth (Comptes Rendus, lxxiv. 597) has asserted that it corresponds to one of the bands in the spectrum of hydrocarbons; but a more exact observation shows that the line falls into a group of shaded bands, which belong to the spectrum, but almost midway between the second and third. Herr Vogel has observed that this line corresponds to a band in the spectrum of rarefied air (Pogg. Ann., cxlvi. 582). This is quite true, but in Angström's opinion is founded on a pure misconception. The spectrum of rarefied air has in the yellow-green part seven bands of nearly equal strength, and that the auroral line corresponds with the margin of one of these bands, which is not even the strongest, cannot be anything else than merely accidental." Angström's own view is that this line is due to fluorescence or phosphorescence, and he remarks that "since fluorescence is produced by the ultra-violet rays, an electric discharge may easily be imagined, which though in itself of feeble light, may be rich in ultra-violet rays, and therefore in a condition to cause a sufficiently strong fluorescence. It is also known that oxygen is phosphorescent, as also several of its compounds." We are, however, just as ignorant of any body which would give such a light by phosphorescence or fluorescence as by ignition, and it seems more probable that the light may be due to chemical action. It is assumed by Angström that water vapour is necessarily absent in the higher atmosphere on account of the cold, but when we remember that its molecular weight is lighter than that of oxygen in the proportion of 9 to 16, it is not unlikely that it may attain great elevations under the very low tensions that prevail at such heights, and it is possible also that both it and other bodies may, by electric repulsion in the auroral beams, be carried up much above the level which they would attain by gravity. If, then, electric discharges take place between the small sensible particles of water or ice in the form of mist or cirrus, as Silbermann has shown to be likely, surface decomposition would ensue, and it is highly probable that the nascent gases would combine with emission of light. It has been almost proved in the case of hydrogen phosphide that the very characteristic spectrum produced by its combustion is due neither to the elements nor to the products of combustion, but to some peculiar action at the instant of combination, and it is quite possible that, under such circumstances as above described, water might also give an entirely fresh spectrum. It is, perhaps, proper to mention that H. R. Procter found an apparent coincidence by often repeated direct comparison with a band frequently seen both in air and oxygen tubes, which he eventually succeeded in tracing with tolerable certainty to some form of hydrocarbon. The comparison spectroscopes were only of low dispersion, but on more accurate measurement of the carbon band it was found that, though more refrangible than the first band of citron acetylene (candle-flame), it was still less so than careful measurement assigns to the aurora. In addition, the band was shaded towards the violet, which is not the case with that of the aurora, though with feeble light it seemed like a line. If, leaving the citron line, we pass on to the feeble spectrum towards the violet, we shall obtain more hopeful coincidences. Angström thinks that three of the bands correspond with the three brightest bands of the violet aurora of the negative pole in rarefied air, and has tried to reproduce the conditions of the aurora on a small scale. He says "Into a flask, the bottom of which is covered with a layer of phosphoric anhydride, the platinum wires are introduced, and the air is pumped out to a tension of only a few millimetres. If the inductive current of a Ruhmkorff coil be then sent through the flask, the whole flask will be filled, as it were, with the violet light, which otherwise proceeds only from the negative pole, and from violet bands. If this spectrum be compared with that of the aurora, both electrodes a spectrum is obtained consisting chiefly of shaded Angström thinks the agreement between the former and some of the best established bands of the latter is satisfactory. netic by Angström and others. Finally, should the aurora be observed | oscillations when violent are called magnetic storms, and as it appears at a less height in the atmosphere, then are recognised it must be noted that auroral display never takes place both the hydrogen lines and also the strongest of the bands of the dark-banded air-spectrum. There are found also again nearly all except during such disturbances, although a large proporthe lines and light-bands of the weak aurora spectrum whose posi- tion of the most remarkable magnetic storms are unaccomtion has with any certainty been observed." panied by visible aurora. With regard to the red line, which is sometimes perfectly sharp and well defined, and occasionally, though very rarely, even as bright as the citron line, scarcely even a plausible theory has been hazarded. That it is not the C line of hydrogen is certain, as they have been directly compared, and are widely separated; and none of the air lines near its position are at all comparable to it in brightness. Vogel thinks it may "correspond with the first system of lines in the spectrum of nitrogen (6620 to 6213), and that probably only the bright part of this group of lines is visible on account of the extreme faintness of the aurora." This, however, cannot be the case, since the present writer has seen it both bright and sharp. Vogel points out that the line near 5189 closely corresponds to an oxygen line of that wave-length which is bright and constant under very different conditions of pressure and temperature. He states that the faint line near 5390 corresponds in like manner to a nitrogen line. He points out that, though the correspondences with the iron lines are very striking, but little weight can be laid on the fact, since many of the brightest lines of the iron spectrum do not appear. The following table gives the principal iron lines (Thalén) and the auroral ones; and it will be seen that the former are so abundant that coincidences could scarcely fail: Iron. Brightness. Aurora. | Iron. Brightness. Aurora. Iron. Brightness. Aurora. 6490 6 5546 10 5167 8 5161 6399 10 5429 10 5139 8 4383 10 4325 10 Angström asserted some years since that he had detected the principal line of the aurora in the spectrum of the zodiacal light, but he appears to have been misled by a faint aurora, for more recent observers, and notably Prof. C. Piazzi Smyth, Mr Backhouse, and A W. Wright (Sill. Jour. of Sc., viii. 39), have found that the spectrum of the zodiacal light is continuous and quite analogous to that of twilight or faint starshine, and polariscope observations prove that it is mostly reflected. The very faint line positioned by Alvan Clark at 5320 has been said by Winlock to coincide with the principal coronal line 5322. The position of the auroral line is uncertain; and even if it were accurate, a single doubtful coincidence with a faint line is not the least proof of identity. We have already remarked the manifest relation between tions. the forms and position of aurora and the earth's lines of magnetic force, and in addition to this have noted the disturbance of the magnetic needle during auroral displays. It is not, however, at such times only that the magnetic elements are subject to variation; the total force, declination, and inclination, all are constantly varying both regularly with the hours of the day and the seasons of the year, and irregularly at uncertain times. The irregular relation (at Fort Enterprise, 64° 30' N., 113° 10' W.), says of the magnetic needle,-"The motion communicated to it was neither sudden nor vibratory. Sometimes it was simultaneous with the formation of arches, prolongation of beams, or certain other changes of form or action of the aurora. But generally the effect of these phenomena upon the needle was not visible immediately, but in about half an hour or an hour the needle had attained its maximum of deviation. From this its return to its former position. was very gradual, seldom regaining it before the following morning, and frequently not until the afternoon, unless it was expedited by another arch of the aurora operating in a direction different from the former one." Franklin, who was one of the first observers of this "The arches of the aurora," he adds, "most commonly traverse the sky nearly at right angles to the magnetic meridian, but deviations from this direction, as has already been stated, were not rare; and I am inclined to consider that these different positions of the aurora have considerable influence on the direction of the needle. When an arch was nearly at right angles to the magnetic meridian, the motion of the needle was towards the west. This westward or about 59° to the west of the magnetic north, that is, when the motion was still greater when one extremity of the arch bore 301°, extremity of the arch approached from the west towards the magnetic north. A westerly motion also took place when the extremity of an arch was in the true north, or about 36° to the west of the magnetic north, but not in so great a degree as when its bearing was about 301. A contrary effect was produced when the same end of an arch originated to the southward of the magnetic west, viz., when it bore from 245° to 234°, and of course when its opposite extremity approached nearer to the magnetic north. In these cases the motion of the needle was towards the east. In one case only a complete arch was formed in the magnetic meridian, in another the beam shot up from the magnetic north to the zenith; and in both these cases the needle moved towards the west. "The needle was most disturbed on February 13th, P.M., at a time when the aurora was most distinctly seen passing between a stratum of clouds and the earth, or at least illuminating the face of the clouds opposed to the observer. This and several other appearances induced me to infer that the distance of the aurora from the earth varied on different nights, and produced a proportionate effect on the needle. When the light shone through a dense hazy atmosphere, when there was a halo round the moon, or when a small snow was falling, the disturbance was generally considerable; and on certain hazy, cloudy nights the needle frequently deviated in a considerable degree, although the aurora was not visible at the time. Our observations do not enable us to decide whether this ought to be attributed to an aurora concealed by a cloud or haze, or entirely to the state of the atmosphere. Similar deviations have been observed in the day-time, both in a clear and cloudy state of the sky, but more frequently in the latter case. An aurora sometimes approached the zenith without producing any change in the position of the needle, as was more generally the case; whilst at other times did not come near the zenith. The aurora was frequently seen a considerable alteration took place although the beams or arches without producing any perceptible effect on the needle. At such times its appearance was that of an arch, or an horizontal stream of dense yellowish light, with little or no internal motion. The disturbance in the needle was not always proportionate to the agitation of the aurora, but it was always greater when the quick motion and vivid light were observed to take place in a hazy atmosphere. In a few instances the motion of the needle was observed to commence at the instant a beam darted upwards from the horizon; and its former position was more quickly or slowly regained according to circumstances. If an arch was formed immediately afterwards, having its extremities placed on opposite sides of the magnetic north and south to the former one, the return of the needle was more speedy, and it generally went beyond the point from whence it first started." Speaking of the aurora of May 13, 1869, M. Lamont of Munich says (Comptes Rendus, lxviii. 1201) "1. During 40 years I have only seen seven or eight aurore at Munich, and this small number is insufficient for a study of the characters of the phenomenon. "2. Aurora, whether visible at Munich or not, are always accompanied by magnetic perturbations. "3. In the perturbations of declination which I have observed | magnet, the direction of its rotation being changed by for 28 years, I have been unable to recognise any general law. 4. The perturbations of horizontal intensity commence in general by an increase of that force, and finish always by a diminution, which lasts for two or three days. 66 5. In all perturbations there is a constant relation between changes of inclination and the simultaneous changes of horizontal intensity, such that an augmentation of intensity of corre sponds to a diminution of inclination of 8° 28 (for Munich). "6. In telegraphic wires we cannot observe the existence of a constant terrestrial current, since the conductivity of the soil is infinitely greater than that of the telegraphic wire, and it is only sudden changes that manifest themselves. In consequence, during magnetic perturbations in the galvanometer of a telegraphic wire, we only see irregular deflections to right or left, succeeding each other at intervals of a few minutes. "In 1850 and 1851 we made electrical observations from hour to hour, from 7 A. M. to 6 P.M., without being able to see any connection between the atmospheric electricity and the magnetic perturbations. Later I abandoned these observations, because the indications of the electrometers depended too much on local and accidental circumstances." It should be noted here that the horizontal component of magnetic force varies with the inclination as well as with the intensity of the total force, and the ratio noted above is almost exactly that which would be produced by a change in the inclination alone; and it would appear as if the actual horizontal force, independent of the inclination, was subject to comparatively little variation. This is not improbable, since variations in the horizontal force could correspond only to electro-magnetic easterly or westerly currents, while changes in declination, inclination, and vertical force might correspond to currents from the magnetic north and south, which there is reason to believe are most frequent in auroral displays. To give some idea of the extent of magnetic perturbations, we may mention that during the aurora of 13th May 1869, the declination at Greenwich varied 1° 25', while the vertical force experienced four successive maxima, and the greatest oscillation amounted to 0.04 of its total mean value. The horizontal force at the same time only varied 0.014 of its mean value. During the aurora of the 15th April of the same year the declination at Stonyhurst varied 2° 23′ 14′′ in nine minutes. The electric currents produced at such times in telegraph wires, though transient, are often very powerful. Loomis (Sill. Jour., vol. xxxii.) mentions cases where wires had been ignited, brilliant flashes produced, and combustible materials kindled by their discharge. It often happens that the ordinary signals are completely interrupted during their continuance. Electrical In addition to the resemblance between the auroral character phenomena and those of electric discharges in rarefied of aurora. gases which we have already mentioned, we have seen that auroral displays are accompanied by marked disturbances both in the direction and force of terrestrial magnetism. This fact is in itself almost proof of their electrical character, and, taken in conjunction with the strong "earth-currents" which are at such times produced in lines of telegraph, and with the manifest polarisation of the arches and rays with regard to the magnetic meridian, may be considered as conclusive that the aurora is some sort of electric discharge. There are still some points with regard to the origin of this electricity which are unexplained, and it is uncertain whether the magnetic disturbance causes the electrical phenomena, or vice versa. It has been shown by Prof. Plücker that when an electric discharge takes place through rarefied gas in the field of a magnet, it is concentrated in the magnetic curves, which are the only paths in which it can move without being disturbed by the magnet. This is well shown in De la Rive's well-known experiment, in which an electro-magnet is enclosed in an electric egg. As soon as the magnet is set in action, the discharge, which had before filled the egg, is concentrated into a defined band of light, which rotates steadily round the reversal either of the current or of the polarity of the magnet. If we suppose that the aurora is an electric discharge passing from one magnetic pole to the other, and following the terrestrial magnetic curves, we shall find that the theory agrees with observed facts even in its lesser details. In these latitudes the magnetic curves are sensibly straight and parallel, and are incline 1 S. E. at an angle of about 70° from the perpendicular, and, by the well-known laws of perspective, will appear to converge towards this point, as, in fact, the auroral streamers do. The streamers should move from east to west, or from west to east, according as the discharge is from north to south, or vice versa, and, in fact, they are in constant motion. Professor Loomis (Sill. Jour. of Sc., xxxiv. 45) gives a catalogue of forty-six cases of such movement, of which thirtyone were from E. to W. and only fifteen in the opposite direction; and as part of these apparent motions are due to a real motion from N. to S., he concludes that the actual motion of the streamers is from about N.N.E. to S.S.W. This would make the north pole the negative electrode, which is most likely usually the case. Prof. Loomis has shown that during auroral displays electrical currents traverse the earth's surface in the same general direction, though subject to great variation in intensity and even to reversal. Waves of magnetic disturbance are also propagated in the same direction (ibid., xxxii. 318). With regard to the arches it is evident that they are generally circles concentric to the magnetic pole, and it is very probable that they are analogous to the striæ often. seen in discharges in rarefied gases. Gassiot, quoted by B. V. Marsh (Sill. Jour., xxxi. 316, and Roy. Soc. Proc., vol. x. Nos. 38 and 39), describes an experiment with his great Grove's battery of 400 cells, in which the exhausted receiver was placed between the poles of the large electromagnet of the Royal Institution:-"On now exciting the magnet with a battery of ten cells, effulgent strata were drawn out from the positive pole, and passed along the under or upper surface of the receiver according to the direction of the current. On making the circuit of the magnet and breaking it immediately, the luminous strata rushed from the positive, and then retreated, cloud following cloud with a deliberate motion, and appearing as if swallowed up by the positive electrode." This, as Mr Marsh remarks, bears a very considerable resemblance to the conduct of the auroral arches, which almost invariably drift slowly southward; and we cannot do better than sum up his theory in his own words:-"The foregoing considerations seem to render it probable that the aurora is essentially an electric discharge between the magnetic poles of the earth leaving the immediate vicinity of the north magnetic pole in the form of clouds of electrified matter, which float southward through the atmosphere at a height of 40 miles or more from the earth, sometimes to a distance of more than 30° from the pole; that whilst they are thus moving forward, with a comparatively slow and steady motion, or sometimes even remaining almost stationary for a long time, bright streams of electricity are from time to time suddenly shot out from them in a nearly vertical direction, that is to say, in the magnetic curves corresponding to the points from which they originate; that these curves, ascending to a great height beyond the atmosphere, then bending more and more southward and downward until they finally reach corresponding points in the southern magnetic hemisphere, are the pathways by which the electric currents pass to their destination; and that for several hundred miles from the earth these curves are thus traced through space and illuminated with bright electric light;' and further, that the magnetism of the earth also causes these luminous currents and the electrified matter # |