APPENDICES.

APPENDIX A.
REFERENCES TO SOME WORKS AND ESSAYS ON THE AURORA.

(Most of these are cited in the ‘Edinburgh Encyclopædia’ and the ‘Encyclopædia Britannica.’)

Musschenbroek, Instit. Phys. c. 41.

‘Trai. Phys. et Hist. de l’Aurore Boréale,’ par M. de Mairan. Paris, 1754.

Beccaria, ‘Dell’Elettricismo Artif. e Nat.’ p. 221.

Smith’s ‘Optics,’ p. 69.

D’Alembert’s ‘Opuscules Mathématiques,’ vol. vi. p. 334.

‘Philosophical Transactions’ as under:—

Vol.	Pages
1716	406
1717	584, 586
1719	1099, 1101, 1104, 1107
1720	21
1721	180, 186
1723	300
1724	175
1726	128, 132, 150
1727	245, 301
1728	453
1729	137
1730	279
1731	53-55
1734	243, 291
1736	241
1740	368
1741	744, 839, 840, 843
1750	319, 345, 346, 499
1751	39, 126
1752	169
1753	85
1762	474, 479
1764	326, 332
1767	108
1769	86, 307
1770	532
1774	128
1781	228
1790	32, 47, 101

‘Miscell. Berolinens.’ 1710, vol. i. p. 131.

‘Comment. Petrop.’ tom. i. p. 351, tom. iv. p. 121.

‘Acta Petrop.’ 1780, vol. iv. p. 1.

‘Mem. Acad. Paris,’ 1747, pp. 363, 423; 1731; 1751.

‘Mem. Acad. Berl.’ 1710, vol. i. p. 131; 1747, p. 117.

Schwed. ‘Abhandlungen,’ 1752, p. 169; 1753, p. 85; 1764, pp. 200, 251.

Bergman, ‘Opusc.’ vol. v. p. 272.

‘Americ. Trans.’ vol. i. p. 404.

‘Mém. de Mathémat. et Phys.’ tom. viii. p. 180.

Rozier, vol. xiii. p. 409; vol. xv. p. 128; vol. xxxiii. p. 153.

Franklin’s Works, vol. ii.

Weidler, ‘De Aurora Boreale.’ 4to.

Nocetus, ‘De Iride et Aurora Boreale, cum Notis Boscovisch.’ Rome, 1747.

Chiminello, ‘Mem. Soc. Ital.’ vol. vii. p. 153.

Gilbert’s ‘Journal,’ vol. xv. p. 206; and (particularly) Dr. T. Young’s ‘Nat. Phil.’ vol. i. pp. 687, 716, and vol. ii. p. 488.

Wiedeburg, ‘Ueber die Nordlichter.’ Jena, 1771.

Hüpsch, ‘Untersuchung des Nordlichts.’ Cologne, 1778.

Van Swinden, ‘Recueil de Mémoires.’ Hague, 1784.

Wilke, ‘Von den neuesten Erklärungen des Nordlichts,’ Schwed. Mus. Wismar, 1783.

Dalton’s ‘Meteor. Observ.’ 1793, pp. 54, 153.

Loomis, ‘Sill. Journal,’ 2nd series, xxxii. p. 324; xxxiv. p. 34. The same, 3rd series, v. p. 245; B. V. Marsh, 3rd series, xxxi. p. 311.

Oettingen and Vogel, Pogg. Ann. cxlvi. pp. 284, 569.

Galle and Sirks, ibid. cxlvi. p. 133; cxlix. p. 112.

Silbermann, ‘Comptes Rendus,’ lxviii. pp. 1049, 1120, 1140, 1164.

Prof. Fritz, “Geog. Distrib.,” Petermann’s Mitth., Oct. 1874.

Zehfuss, ‘Physikalische Theorie.’ Adelman, Frankfort.

‘Nature,’ iii. pp. 6, 7, 28, 104, 126, 346, 348, 510; iv. pp. 209, 213, 345, 497, 505; x. 211 (Ångström).

‘Edinburgh Astronomical Observations,’ vol. xiv. 1870-1877.

‘English Mechanic,’ No. 461 (January 23, 1874), pp. 445-447; and No. 462, pp. 475, 476.

APPENDIX B.
EXTRACTS FROM THE MANUAL AND INSTRUCTIONS FOR THE (ENGLISH) ARCTIC EXPEDITION, 1875.

Note on Auroral Observations. By Prof. Stokes, Sec. R.S.

The frequency of the Aurora in Arctic regions affords peculiar facilities for the study of the general features of the phenomenon, as in case the observer thinks he has perceived any law he will probably soon, and repeatedly, have opportunities of confronting it with observation. The following points are worthy of attention:—

Streamers.—It is well known that, at least as a rule, the streamers are parallel to the dipping-needle, as is inferred from the observation that they form arcs of great circles passing through the magnetic zenith. It has been stated, however, that they have sometimes been seen curved. Should any thing of this kind be noticed, the observer ought to note the circumstances most carefully. He should notice particularly whether it is one and the same streamer that is curved, or whether the curvature is apparent only, and arises from the circumstance that a number of short, straight streamers start from bases so arranged that the luminosity as a whole presents the form of a curved band.

Have the streamers any lateral motion? and if so, is it from right to left or left to right, or sometimes one and sometimes the other, according to the quarter of the heavens in which the streamer is seen, or other circumstances? Again, if there be lateral motion, is it that the individual streamers move sideways, or that fresh streamers arise to one side of the former, or partly the one and partly the other? Do streamers, or does some portion of a system of streamers, appear to have any uniform relation to clouds, as if they sprang from them? Can stars be seen immediately under the base of streamers? Do streamers appear to have any definite relation to mountains? Are they ever seen between the observer and a mountain, so as to appear to be projected on it? This or any other indication of a low origin ought to be most carefully described.

When streamers form a corona, the character of it should be described.

Auroral Arches.—Are arches always perpendicular to the magnetic meridian? If incomplete, do they grow laterally? and if so, in what manner, and towards which side? Do they always move from north (magnetic) to south? and if so, is it by a southerly motion of the individual streamers, or by new streamers springing up to the south of the old ones? What (by estimation, or by reference to known stars) may be the breadth of the arch in different positions in its progress? Do arches appear to be nothing but congeries of streamers, or to have an independent existence? What relations, if any, have they to clouds? and if related, to what kind of clouds are they related?

Pulsations.—Do pulsations travel in any invariable direction? What time do they take to get from one part of the heavens to another? Are they running sheets of continuous light, or fixed patches which become luminous, or more luminous, in rapid succession? and if patches, do these appear to be foreshortened streamers? Are the same patches luminous in successive pulsations?

Sounds (?).—As some have suspected the Aurora to be accompanied by sound, the observer’s attention should be directed to this question when an Aurora is seen during a calm. If sound be suspected, the observer should endeavour, by changing his position, brushing off spicules of ice from the neighbourhood of the ears, his whiskers, &c., to ascertain whether it can be referred to the action of such wind as there is on some part of his dress or person. If it should clearly appear that it is not referable to the wind, then the circumstance of its occurrence, its character, its relation (if any) to bursts of light, should be most carefully noted.

These questions are prepared merely to lead the observer to direct his attention to various features of the phenomenon. Answers are not demanded, except in such cases as definite answers can be given; and the observer should keep his attention alive to observe and regard any other features which may appear to be of interest. It is desirable that drawings should be made of remarkable displays.

Observations with Sir William Thomson’s electrometer would be very interesting in connexion with the Aurora, especially a comparison of the readings before, during, and after a passage of the Aurora across the zenith.

Spectroscopic Observations. By Prof. G. G. Stokes, Sec. R.S.

Spectrum of the Aurora.

The spectrum of the Aurora contains a well-known conspicuous bright line in the yellowish green, which has been accurately observed. There are also other bright lines of greater refrangibility, the determination of the positions of which is more difficult on account of their faintness, and there are also one or more lines in the red, in red auroras.

Advantage should be taken of an unusually bright display to determine the positions of the fainter lines. That of the brightest lines, though well known, should be measured at the same time to control the observations. The character of the lines (i. e. whether they are strictly lines, showing images of the apparent breadth of the slit, or narrow bands, sharply defined or shaded-off) should also be stated.

Sometimes a faint gleam of light is seen at night in the sky, the origin of which (supposed from the presence of clouds) is doubtful. A spectroscope of the roughest description may in such cases be usefully employed to determine whether the light is auroral or not, as in the former case the auroral origin is detected by the chief bright line. The observer may thus be led to be on the look-out for a display which otherwise might have been missed.

It has been said, however, that the auroral light does not in all cases exhibit bright lines, but sometimes, at least in the eastern and western arch of the Aurora, shows a continuous spectrum. This statement should be confronted with observation, special care being taken that the auroral light be not confounded with light which, though seen in the same direction, is of a different origin, such, for example, as light from a bank of haze illuminated by the moon.

Sir Edward Sabine once observed an auroral arch to one side (say north) of the ship, which was in darkness. Presently the arch could no longer be seen, but there was a general diffuse light, so that a man at the mast-head could be seen. Later still, the ship was again in darkness, and an auroral arch was seen to the south.

Should any thing of the kind be observed, the whole of the circumstances ought to be carefully noted, and the spectroscope applied to the diffuse light.

Polarization of Light. By W. Spottiswoode, M.A., LL.D., Treas. R.S.

It has been suggested that the Aurora, inasmuch as it presents a structural character, may afford traces of polarization. Having reference to the fact that the striæ of the electric discharge in vacuum-tubes present no such feature, the probability of the suggestion may be doubted. But it will still be worth while to put the question to an experimental test.

If traces of polarization be detected, it must not at once be concluded that the light of the Aurora is polarized; for the Aurora may be seen on the background of a sky illuminated by the moon, or by the sun, if not too far below the horizon, and the light from either of these sources is, in general, more or less polarized; therefore, if the light of the Aurora is suspected to be polarized, the polariscope should be directed to an adjacent portion of clear sky, free from Aurora, but illuminated by the moon or sun as nearly as possible similar, and similarly situated to the former portion; and the observer must then judge whether the polarization first observed be merely due to the illumination of the sky.

The presence of polarization is to be determined:—

(1) With a Nicol’s prism, by observing the light through it by turning the prism round on its axis, and by examining whether the light appears brightest in some positions and least bright in others. If such be the case, the positions will be found to be at right angles to one another. The direction of “the plane of polarization” will be determined by that of the Nicol at either of these critical positions. The plane of polarization of the light transmitted by a Nicol, is parallel to the longer diagonal of the face; and, accordingly, the plane of polarization, or partial polarization, of the observed light is parallel to the longer diameter of the Nicol when the transmitted light is at its greatest intensity, or to the shorter when it is at its least.

(2) The observation with a double-image prism is similar to that with a Nicol. This instrument, as its name implies, gives the images which would be seen through the Nicol in two rectangular positions, both at once, so that they can be directly compared; and when in observing polarized light the instrument is turned so that one image is at a maximum, the other is simultaneously at a minimum. Both these methods of observation, (1) and (2), are especially suitable for faint light; because in such a case the eye is better able to appreciate differences of intensity than differences of colour.

(3) The observation with a biquartz differs from (1) only by holding a biquartz (a right-handed and a left-handed quartz cemented side by side) at a convenient distance beyond the Nicol, and by observing whether colour is or is not produced. If the Nicol be so turned that the two parts of the biquartz give the same colour (choose the neutral tint, teint de passage, rather than the yellow), we can detect a change in the position of the plane of polarization by a change in colour, one half verging towards red, the other towards blue. This observation is obviously applicable to a change in the plane, either at different parts of the phenomenon at the same time, or at the same parts at different times.

(4) We may use a Savart’s polariscope, which shows a series of coloured bands in the field of view. For two positions at right angles to one another corresponding to the two critical positions of a Nicol, these bands are most strongly developed; for two positions midway between the former the bands vanish. In the instruments here furnished, the plane of polarization of the observed light will be parallel to the bands when the central one is light, perpendicular to them when the central band is dark.

Instructions in the use of the Spectroscopes supplied to the Arctic Expedition. By J. Norman Lockyer, F.R.S.

Spectroscopic Work.

Scales prepared on Mr. Capron’s plan, together with forms for recording positions, also accompany the instrument.

A. In using these, carefully insert the principal solar lines in their places on the forms, as taken from a fine slit, and keep copies of this scale for use. If the slit opens only on one side, note on scale in which direction the lines widen out, whether towards red or violet. Also fill up some of these forms with gas and other spectra, as taken at leisure with the same instrument and scale.

When observing, close the slit (after first wide opening it) as much as light will permit, and then with pen or pencil record the lines as seen upon the micrometer-scale on the corresponding part of the form, and note at once relative intensities with Greek letters, α, β, &c. (or numbers).

Reduce at leisure line-places on scale to wave-lengths, and note as to each line the probable limits of instrumental error. B.

In case the auroral spectrum is so faint that the needle-point or micrometer-scale is invisible, half of the field of view may be covered with tinfoil, with a perfectly straight smooth edge running along the diameter of the field, in perfect focus, and parallel to the lines of the spectra. The reading-screw being set to 10, the bending-screw should then be adjusted so that the green line of the Aurora is just eclipsed behind the blackened edge of the tinfoil. A similar eclipse of other lines will give their positions.

In this instrument the reference-prism is brought into action by turning the slipping piece to which is fixed the two terminals. Care should be taken that the prism itself is adjusted before commencing observations, as it may be shaken out of position on the voyage. The tubes provided for the reference-spectra may be either fastened to the terminals or arranged in some other manner. The air-spectrum may also be used as a reference-spectrum. To get this, two wires should be screwed into the insulators, their ends being at such a distance apart and in such a position that the spectrum is well seen.

General Observations regarding the Spectrum of the Aurora [17].

C. Note appearance, colour, &c. of arc, streamers, corona, and patches of light.

Get compass positions of principal features, and note any change of magnetic intensity. If corona forms, take its position and apparent height.

Look out for phosphorescence of Aurora and adjacent clouds. Listen for reported sounds. Note any peculiarity of cloud scenery, prior to or pending the Aurora.

Sketch principal features of the display, and indicate on this sketch the parts spectroscopically examined.

Examine line in red specially in reference to its assumed connexion with telluric lines (little a group), and note as to its brightening in sympathy with any of the other lines.

Examine line in yellow-green (Ångström’s) as to brightness, width, and sharpness (or nebulosity) at the edges. Notice as to a peculiar flickering in this line sometimes seen; note also whether this line is brighter (or the reverse) with a fall of temperature. Note ozone papers at the time of Aurora.

Note whether the Auroræ can by their spectra be classed into distinct types or forms, and examine for different spectra as under:—

α. The auroral glow, pure and simple.

β. The white arc.

γ. The streamers and corona.

δ. Any phosphorescent or other patches of light, or light cloud in or near the Auroræ. D.

The information collected together in the ‘Manual’ should be carefully consulted, and the line of observations suggested by Ångström’s later work followed out. To do this, not only record the positions of any features you may observe in the spectrum, but endeavour to determine, if any, and if so which, of the features vary together. Compare, for instance, the two spectra of nitrogen in the Geissler tube supplied, by observing first the narrow and then the wider parts of the tube. It will be seen that the difference in colour and spectrum results simply from an addition to the spectrum in the shape of a series of channelled spaces in the more refrangible end in the case of the spectrum of the narrow portion.

Try to determine whether the difference between red and green Auroras may arise from such a cause as this, and which class has the simpler spectrum.

See whether indications of great auroral activity are associated with the widening or increased brilliancy of any of the auroral lines.

Remember that if auroral displays are due to gaseous particles thrown into vibration of electric disturbance, increased electric tension may either (1) dissociate those particles and thus give rise to a new spectrum, the one previously observed becoming dimmer; or (2) throw the particles into more intense vibration without dissociation, and thus give rise to new lines, those previously observed becoming brighter.

Careful records of auroral phenomena from both ships may enable the height of some, observed from both, to be determined. It will be very important that those the heights of which are determined by such means should be carefully observed by the spectroscope, in order to observe whether certain characteristics of the spectrum can be associated with the height of the Aurora.

APPENDIX C.
EXTRACTS FROM PARLIAMENTARY BLUE BOOK, CONTAINING THE “RESULTS DERIVED FROM THE ARCTIC EXPEDITION 1875-76.” (Eyre and Spottiswoode, 1878.)

Auroras observed 1875-1876, at Floebery Beach and Discovery Bay.

By Lieutenant A. C. Parr, R.N.

Though the auroral glow was often present, and served in some degree to lighten the darkness of the sky during the long winter, when the moon was absent, the actual appearances of the Aurora itself were few, and the nimbus worthy of any particular remark extremely small. Those which were stationary assumed the form of low arches, with streamers flashing up to them from the horizon, and usually to the eastward. But the more common form was for an arch to appear low down in some part of the sky where the glow was brightest; at first it was very faint and narrow, but as it rose gradually in the heavens it would increase both in size and intensity, till on arriving near the zenith, with its ends extending nearly to the horizon, it would be about the breadth of three or four rainbows, and its colour that of white fleecy clouds lit up by the rays of the full moon. On reaching this point, however, its course was nearly run; for after appearing to remain stationary, as little white gaps would suddenly rend the arch asunder, the portions thus detached seemed to roll together and concentrate all their brightness in the smaller space, and then gradually fade away and become extinct. Sometimes a very pale green would show itself in the more luminous patches, and once or twice there was a slight suspicion of red; but never was the whole sky illuminated by streams running in all directions, and forming coronæ, while these colours varied every moment.

When instead of the arch rising up from the horizon a streamer appeared, its origin was in the north. From the northern horizon it would stretch out towards the zenith, passing nearly overhead, and reaching to within a few degrees of the land to the south. In appearance they would be the same as the arches, but sometimes a second would grow out of the first, and on one occasion three were visible at the same time. They had lateral motion either from east to west, or west to east, but there was no flashing to brighten them, and they gradually faded away.

The time at which Auroras usually occurred was between 9 P.M. and midnight, the last display being on February 19th, commencing at 11 P.M. It was a beautifully clear night, without mist or haze of any description, and small stars visible close down to the horizon. At the above-named hour two arches made their appearance, and remained stationary; the lower one was the brighter, being of a pale green colour, its centre bearing E.S.E. (true), and having an altitude of about 5°, with a breadth of about twice that of a rainbow. The second arch was concentric with the first, and about 7° above it, but rather broader and fainter. These arches maintained their altitude, the upper one at about the same intensity, but that of the lower one varied considerably. It would gradually lighten up, then send flashes to the upper one, then break up and fade away; before, however, it had quite disappeared, flashes would come up to it from the horizon which seemed to endue it with new life, for the arch would be reformed, brighten up, and the same performance would be again repeated. This occurred three or four times in the course of three quarters of an hour; but the flashes from the horizon never extended beyond the lower arch, and those from the lower never went beyond the upper. During this display the citron-line was obtained very clearly with the spectroscope, but no other lines were visible.

On six or seven occasions Auroras were visible at the same time on board both the ‘Alert’ and ‘Discovery;’ but the absence of characteristic features makes it impossible to determine whether they were the same display, or merely two distinct ones which happened to occur at the same time. But as by far the larger number of those recorded in the one ship were not visible at the other, it was certainly only under exceptional conditions that they could be simultaneously observed at both stations, if, indeed, they ever were. Auroras seemed to appear indifferently both when there was wind and when it was calm, with either a high or low barometer, and seemed quite unconnected with the temperature, although on an occasion the thermometer was observed to fall 3° during the display, and to rise 2° almost immediately afterwards. But it was never seen illuminating the edges of clouds, as we saw it on the passage home, nor playing about the outline of the land, and never was there the slightest suspicion of sound being produced by it.

The opportunities for observing the spectrum of the Aurora in this position have been most unsatisfactory, as the displays were small in number and deficient in brilliancy.

The form they generally assumed was to rise like an arch from a portion of the horizon where there was a luminous glow, at first very faint, but gradually increasing in brilliancy till near the zenith, where it would remain stationary for a short time and then break up and disappear. Sometimes they would rise up as streamers, but only occasionally was more than one visible at a time, and they lasted for such a short time, that even if they had been bright it would have been very difficult to make satisfactory observations.

Very few showed any signs of colour, and those only the slightest tinge. Nearly all that were observed gave the citron-line with the small pocket spectroscope with more or less distinctness, though no signs of any other lines were ever seen; but on only two occasions was it bright enough to get the line with Nury’s spectroscope, and then only for such a short time that a satisfactory measure could not be obtained.

Then follows a descriptive list of the Auroræ seen, from which I have selected three of the finest, viz. January 2nd, February 14th, and February 19th, 1876.

January 2nd, 1876. Lieut. Parr. Floeberg Beach.—9 P.M. Streams of Aurora. Stars shining brightly.

Register. Discovery Bay.—9 P.M. Observed an Aurora like a pale band of light in the form of an arch whose centre was on the true meridian and 15° from the zenith. It shortly afterwards broke up into feathered edges, their direction being a little to the eastward of the zenith. The arch grew fainter, and shifted to the eastward of the meridian four points; the left extremity of the arch faded away, and the right assumed the shape of the folds of a curtain doubled over. The weather was clear and calm. The display lasted upwards of 30 minutes.

A spectroscope, one of Browning’s 8-in. direct-vision, was directed towards the Aurora, but the light was not sufficient to give any spectrum.

The temperature was -39°. Barometer 29·56 inches. No wind. Clouds stratus 2. Eight meteors were observed during the time the Aurora was visible.

February 14th. Register. Discovery Bay.—At 2 A.M. a faint Aurora passing across the heavens from S.E. to S.W. was observed, like an arch of a pale colour. It lasted only a short time, and was very indistinct. Temperature -47°. Barometer 30·44 inches. No wind or clouds.

Lieut. Aldrich. Floeberg Beach.—2 A.M. A faint Aurora towards the S.W. Weather calm. Cumulus-stratus clouds 3. Temperature -46°. 8 P.M. Faint flashes of Aurora in the E. and S.W.

Lieut. Aldrich and Lieut. Parr. Floeberg Beach.—11.50 P.M. A moderately bright arch of Aurora extended from due N. to about S.S.W., where it terminated close down to the horizon in a crook turned to the eastward. In a few moments a streamer flashed from the end of the crook parallel to the first and right across the heavens, its edges being quite sharp and parallel to each other. A third streamer shot up a minute afterwards, but did not extend more than 80° upwards. The streamers were visible for a very short time, the first remaining longest. The second-named arch gradually faded away till within a few degrees of the S.S.W. horizon, and (still being a continuation of the crook) bent round to the eastward, and towards the horizon, going on to what was left of the stump of the third arc. A lateral motion to the eastward now began, the whole body gradually turning round until it disappeared about due south. Stars were visible through it at its brightest, but not very distinctly. This is the most intense and variegated Aurora we have experienced, but scarcely any colours were to be seen. Temperature -51°. Barometer 30·43 inches, stationary. Calm weather. Clouds cumulus 1. Preceded and followed by calm weather.

Meteorological Register. Discovery Bay.—9.15 P.M. An Aurora was observed to the southward, spreading out like a fan in separate ways. It was faint. A few cirro-stratus clouds were visible, apparently between the observer and the Aurora. It lasted about 40 minutes, and then gradually faded away. Temperature -47°. Barometer 30·51 inches, stationary. No wind. Clouds cirro-stratus 4.

February 19th. Meteorological Report. Discovery Bay.—9.45 P.M. An Aurora like a fluted arch, with rays flashing towards the Pole, was observed spanning the hills from the south to the east. The direction of the lines of light from all parts of the arch was towards the zenith. Above the arch a pale band of colour appeared, like a secondary arch above the other. It appeared very much as if it was caused by the reflected light of the Aurora. The Aurora was bright for a few seconds, and then gradually died away. It lasted altogether about 30 minutes. The centre of the arch bore S.E., having an altitude of about 30°. The secondary arch was about 15° above the former. Both arches were of a pale light colour, the upper one very faint. Temperature -34°. Barometer 29·87 inches, rising rapidly. Weather calm. Misty. No clouds.

Lieut. Parr. Floeberg Beach.—An Aurora appeared shortly after 11 P.M., consisting of bright arch, whose centre bore about E.S.E., and had an altitude of about 5°, with a second broader and fainter arch about 7° above the first. These arches maintained their altitudes, the upper one at about the same intensity, but that of the lower one varied considerably. It would gradually brighten up, then send streamers up to the second, then break up into light patches, and gradually fade away. This happened three or four times during the 40 minutes that the display lasted. At times streamers would come up from the horizon to the lower arch, for it was a splendidly clear night, and seemed to brighten it up, but none of them extended beyond it. Neither did the streamers from the lower arch extend beyond the upper one. It was slightly green in colour when brightest, and the citron-line was well defined, but no others were visible. Temperature -46°. Barometer 29·95 inches, steady. Weather calm. Cumulus clouds 4. Misty.

Table of Dates when Auroras were observed by the Arctic Expedition, 1875-76.

Date.			H.M.S. ‘Alert,’ Floeberg Beach.	H.M.S. ‘Discovery,’ Discovery Bay.
1875,	October	25	11.45 P.M. Faint.	Cloudy.
”	”	26	10 P.M. Very faint.	10 P.M.
”	”	30	Sky obscured. Faint.	Ditto.
”	November	1	Ditto. Faint, but well marked.	Ditto.
”	”	2	9 to 10 P.M. Arches and streamers.	A few clouds.
”	”	21	Ditto. Bright streamer.	9 to 10 P.M. and 10 to 11 P.M.
”	”	22	2 P.M. and 8 P.M. Slight, red.	Clear sky.
”	”	25	9.30 A.M. Character not recorded.	Ditto.
”	”	26	10 A.M. Stream of light.	} A few clouds. 10 P.M.
”	”	26	Cloudy to 10 P.M., bright afterwards.	} A few clouds. 10 P.M.
”	”	27	Midnight. Slight.	11.40 P.M.
”	”	28	1 A.M. Bright streak.	Clear sky.
”	”	29	Cloudy, brighter at 11 A.M. Faint glow.	9.30 A.M.
”	”	30	A few clouds. Very faint.	4.30 A.M.
”	”	30	5 P.M., 8 P.M., and 10 P.M. Flashes.	5 P.M.
”	December	2	Evening. Streamers.	Clear sky.
”	”	3	1 A.M. Flashes.	Ditto.
”	”	3	Bright sky. Faint Aurora.	2.30 P.M.
”	”	16	10 P.M. Slight; showed citron-line.	11 P.M.
”	”	19	3 P.M. to 5 P.M., faint; and 9 to 10 P.M., moderately bright arc.	Very clear sky.
”	”	22	10 P.M. Slight.	Ditto.
”	”	23	6 P.M. Ditto.	Ditto.
”	”	24	Misty, a few stars visible. Arch.	9 A.M.
”	”	26	Very bright sky. Faint.	6 P.M.
”	”	29	Ditto. Very faint.	6.15 P.M.
”	”	31	4 P.M. Same.	Sky obscured.
1876,	January	1	5 P.M. and 11 P.M. Slight.	A few clouds.
”	”	2	9 P.M. Described and figured.	9 P.M.
”	”	17	Very bright sky. Very faint streamers.	9.25 A.M.
”	”	18	9.45 P.M. and 10.5 P.M. Character not recorded.	10.15 P.M.
”	”	19	Very bright sky. Faint.	9.45 P.M.
”	”	20	2 A.M. Slight.	2.30 A.M.
”	”	23	7.55 A.M. and 2 P.M. Slight.	8.45 P.M.
”	”	24	Bright sky. Slight flash.	2 A.M.
”	”	24	5 P.M. and 11.15 P.M. Faint Aurora.	Very clear sky.
”	”	27	2 A.M. to 3.45 A.M. Faint.	1 A.M. to 4 A.M.
”	”	27	Very bright sky. Faint double arch.	8.30 P.M.
”	”	28	6 P.M. and 7 to 9 P.M. Faint flashes.	7.20 P.M.
”	”	30	8 P.M. Streak.	7.50 to 9 P.M.
”	”	31	8.30 A.M. and 7.30 P.M. Very faint.	8.25 A.M., 5.30 P.M.
”	February	3	10 P.M. Slight flash.	Very clear sky.
”	”	11	Sky obscured. Very faint.	11 P.M.
”	”	13	11 P.M. Flashes.	Clear sky.
”	”	14	2 A.M., 9.15 to 10 P.M. Described and figured.	2 A.M. and 11.50 P.M.
”	”	19	9.45 P.M.	11 P.M.
”	”	20	2 A.M. Very faint.	2.30 A.M.
”	”	22	2 A.M. Character not recorded.	Very clear sky.
”	”	24	Bright sky. Very faint.	Midnight.
”	”	26	10 P.M. and 11 P.M. Faint flashes.	Sky obscured.

I have added to the above Table the character of the Aurora in each instance as taken from the fuller descriptions given.—J. R. C.

Auroras and Magnetic Disturbances.

The appearances of Auroras and the synchronous movements of the declinometer-magnet were subjects of special observation during the stay of the ‘Alert’ and ‘Discovery’ at their winter-quarters. The Table on page 187 gives the dates and hours when Auroras were visible. On all occasions they were observed to be faint, with none of those brilliant manifestations which are described by our own officers as seen at Point Barrow, and by the Austro-Hungarian Expedition at Franz-Josef Land, where the magnetical instruments were so sensibly disturbed.

These phenomena were not observed either in the ‘Alert’ or the ‘Discovery,’ especially no connexion between magnetical disturbances and the appearances of Auroras could be traced.

This is quite in accordance with the remarks of previous observers within the region comprehended between the meridians of 60° and 90° west, and north of the parallel of 73° north. For example:—

In the Phil. Trans. 1826, Part IV. p. 76, Capt. Parry and Lieut. Foster remark, in the discussion of their magnetical observations at Port Bowen:—“As far, however, as our own observations extended, we have reason to believe that on no occasion were the needles in the slightest degree affected by Aurora, meteors, or any other perceptible atmospheric phenomenon.”

Again, in the Smithsonian Contributions, vol. x., 1858, Mr. A. Schott, in his discussion of Dr. Kane’s observations at Van Rensselaer Harbour, in 1854, remarks—“In conformity with the supposed periodicity of this phenomenon as recognized by Professor Olmstead, no brilliant and complete Auroras have been seen; with an exception of very few, they may all be placed in his fourth class, to which the most simple forms of appearances have been referred.” The following statement is given in the same page as a footnote:—“The processes have no apparent connexion with the magnetic dip, and in no case did the needle of our unifilar indicate disturbance.”

The following description of the Aurora observed on 21st November, 1875, is given by Commander Markham and Lieut. Giffard, in their abstract of observations at Floeberg Beach:—

“Between 10 and 11 P.M. bright broad streamers of the Aurora appeared 10° or 15° above the north horizon, stretching through the zenith, and terminating in an irregular curve about 25° above the south horizon, bearing S.S.W. During the Aurora’s greatest brilliancy the magnet was observed during five minutes to be undisturbed.”

[Note.—I applied for a loan of the lithographic stones to enable me to give copies of the three diagrams of Auroræ referred to in the Arctic “Results;” but the Lords Commissioners of H.M. Treasury refused this, except on the terms of my paying one third of the original cost of production of such diagrams. I did not think it worth while to accept these conditions. Only one of the drawings has any special interest; and this is a “curtain” Aurora, similar to that figured on Plate II. of this work.—J. R. C.]

APPENDIX D.
THE AURORA AND OZONE.

Aurora and ozone. Dr. Allnatt’s notes and conclusions deduced therefrom.

While Part I. was in the press, Dr. Allnatt, formerly of Frant, and for many years the well-known meteorological contributor to ‘The Times’ newspaper, kindly placed at my disposal his large series of notes. Upon an examination of these we came to the following conclusions:—

1. That Auroral periods are also periods of comparative abundance of ozone.

2. That instances are by no means wanting in which an abnormal development of ozone appears to be coincident with the manifestation of an Aurora.

Year 1870 remarkable for sun-spots, auroræ, and ozone.

In reference to the first point, it is found, as the result of an examination of Dr. Allnatt’s notes, that particular years and months are notable at once for Auroræ and for ozone in abundance. 1870 was one of these years, and was specially recorded by Dr. Allnatt, in his ‘Summary for the Year,’ as remarkable for sun-spots, Auroræ, and ozone.

Particulars of some of the monthly records.

The month of February in that year was marked by intense cold and brilliant Auroræ. Atmospheric electricity was feeble, but ozone was, throughout the month, well developed; and there was no tangible period of antozone.

In the month of April of the same year, eight days consecutively (19th to 26th) were marked for ozone 10, the maximum of Dr. Allnatt’s scale.

In May of the same year there were magnificent Auroræ, and atmospheric electricity was intense. Ozone was scanty; but this was accounted for by the wind being generally E.N.E., ozone being mostly developed with a W. or S.W. wind, and a moist state of the atmosphere.

In August 1870 the unusually large number of 22 days were recorded for a maximum of ozone.

September 1870 was hardly less remarkable, with 19 days of maximum. It was recorded that there were splendid Auroræ during this month, and the solar spots were very large.

October 1870 had 20 days of maximum ozone, and November had several fine Auroræ and maxima of ozone noted. In fact, nearly every month in that year was referred to by Dr. Allnatt for displays of Aurora (of both Arctic and Antarctic forms) and for a development of ozone very considerably above the average.

Year 1871.

The year 1871 had more or less of the same character. In the month of October of that year, fine Auroræ were prevalent, and ozone was registered as at its maximum during 22 days.

There seems reason to conclude that if a systematic comparison of annual or other periods of Aurora and ozone development were made, it would result in disclosing a connexion (probably an intimate one) between the two phenomena.

Instances showing a connexion between a specific Aurora and an ozone maximum.

With reference to the second point, the following (among other) instances may be quoted, for the purpose of showing a connexion between a specific Aurora and an ozone maximum.

The Aurora of 24th September, 1870, was splendid and universal, being seen in Europe, Asia, Africa, America, and Australia. Ozone reached, on the morning of the 24th, 8 of the scale (the scale running from 1 to 10), and, on the morning of the 25th, 10, the maximum.

In October 1870 there were grand displays on the 14th, 20th, 22nd, 24th, and 25th, and ozone was correspondingly abundant, as is seen by the following Table:—

Date.			Aurora.	Ozone.
1870,	October	14th.	Aurora.	8	The display of the 24th was accompanied by the formation of a corona, and that of the 25th was splendidly seen in Edinburgh.
”	”	20th.	Aurora.	10
”	”	21st.	None seen.	5
”	”	22nd.	Aurora.	10
”	”	23rd.	None seen.	8
”	”	24th.	Aurora.	10
”	”	25th.	Aurora.	8

The foregoing figures somewhat point to the conclusion that ozone quantity rises and falls coincidently with the Aurora displays.

The following seems, however, a case still more strongly in point.

Date.			Wind.	Aurora.	Ozone.
1871,	January	25th.	E.S.E.	None seen.	0
”	”	26th.	N.N.W.	None seen.	2
”	”	27th.	E.S.E.	Aurora at night in N. and S. horizons.	10
”	”	28th.	E.	None seen.	8
”	”	29th.	S.E.	None seen.	2

It is curious, in examining the above Table, to note how the ozone rose, notwithstanding an east wind, from 0 on the 25th, and 2 on the 26th, to 10 on the 27th, when the Aurora appeared, and 8 on the 28th, when it might have lingered; and how it again descended to 2 on the 29th.

The case of the Aurora of 6th of October, 1869, when a broad belt of Aurora was in the north, is also an illustrative one, as will be seen by the following data:—

Date.			Wind.	Ozone.	Aurora.
1869,	October	5th.	S.S.W.	1	—
”	”	6th.	S.S.E.	5	Aurora.
”	”	7th.	S.S.W.	10	—
”	”	8th.	S.	10	—
”	”	9th.	S.E.	5	—

The Aurora of the night of the 6th was here represented by the ozone-paper of the morning of the 7th with a maximum of 10, which lasted till the 8th.

[It should be borne in mind, in examining these Tables, that the Aurora is of the night of the given date, while the ozone-papers are taken and recorded in the morning of the date quoted.]

Other instances.

We will now take instances where the ozone has not reached its maximum; but even in these cases a certain amount of rise and fall of the ozone development towards and from the Aurora is traceable.

Date.			Wind.	Ozone.	Aurora.
1871,	April	8th	S.S.E.	5	Aurora on 9th, but wind E. and unfavourable to ozone.
”	”	9th	S.S.E.	8
”	”	10th	S.E.	5
1871,	November	9th	N.	5	Aurora on all three nights.
”	”	10th	N.W.	8
”	”	11th	N.	5
1872,	February	3rd	S.W.	4	Aurora on night of the 4th represented by ozone-paper of morning of the 5th.
”	”	4th	S.S.W.	5
”	”	5th	S.W.	8
”	”	6th	S.W.	5

Other cases are, we are bound to say, found, in which ozone was either not remarkable for quantity, or positively fell during the Aurora, as, for instance, this:—

Date.			Wind.	Ozone.	Aurora.
1874,	March	16th	W.N.W.	6	Aurora on the 18th represented by test-paper of the 19th with only two degrees of discoloration.
”	”	17th	S.W.	6
”	”	18th	W.	5
”	”	19th	S.S.W.	2

It is, however, possible that such instances may be accounted for, either by some reaction on the test-papers after they have been coloured, or by some accidental antagonistic circumstance affecting the tests. The following is a case well illustrating this:—

Date.			Wind.	Ozone.	Aurora.
1874,	January	31st	N.N.W.	6	There was an Aurora on the night of the 2nd represented by the ozone-paper (4 only) on the morning of the 3rd.
”	February	1st	N.W.	8
”	”	2nd	N.W.	2
”	”	3rd	N.N.W.	4
”	”	4th	E.N.E.	8

This instance would seem strongly opposed to the theory of a connexion between Aurora and ozone but for the fact that on the 2nd, when the Aurora was seen at night, and on other days in the same month, Dr. Allnatt has recorded a strong wave of antozone to have swept over the whole of England, and blanched the ozone-papers, however deep their coloration might have previously been. Indeed, it is easy to understand that some antozonic influence may, at times, disturb the evidence of the test-papers, even in so elevated and apparently pure an atmosphere as that of Frant.

It may not be considered that the foregoing instances are enough to establish a case of ozone=Aurora; but there seems, at least, sufficient to base a requisition for further inquiry upon.

It would, too, be interesting to investigate whether Auroræ and ozone development are respectively localized. Mr. Ingall’s fine Aurora, seen at Champion Hill, S.E., July 18th, 1874 (antè, pp. 22 and 23), was not observed at Frant, and the ozonoscopes there were described as blanched by antozone.

APPENDIX E.
INQUIRIES INTO THE SPECTRUM OF THE AURORA.

By H. C. Vogel.[18]

The frequent appearance of the Aurora in the past winter, as well as this spring, has given me opportunity to institute exact inquiries into the spectrum of the Aurora. It is known that the nature of Auroræ is as yet but little explored. It has been considered necessary to abandon the former view—that they are discharges of the electricity collected at the poles—because it has been hitherto found impossible to bring the chief lines of the Aurora-spectrum into coincidence with the spectra of the atmospheric gases. Theoretical considerations, based on the great alterations to which the spectrum of the same gas is subject under varying conditions of temperature and density, have very recently led Zöllner to the opinion that probably the spectrum of the Aurora does not coincide with any known spectrum of the atmospheric gases, only because it is a spectrum of another form of our atmosphere hitherto incapable of artificial demonstration [19].

The following article will show how far I have succeeded, in conjunction with Dr. Lohse, in supporting this view by exact observations of the Aurora-spectrum itself, as well as by comparison with the spectra of the gases constituting the air.

The star-spectrum apparatus belonging to the 11-inch equatorial of the Bothkamp Observatory was used for these observations. It consists of a set of prisms à vision directe, five prisms with refracting angle 90°, slit, collimator, and observing telescope. The lowest eyepiece (magnifying four times) of this telescope was employed. The telescope is capable of being moved in such a way, by the aid of a micrometer-screw, that different portions of the spectrum can be brought into the centre of the field of vision. As fractions of the rotation of this screw are marked, the distances of the spectral lines can be readily found.

Repeated measurements of 100 lines of the solar spectrum have enabled me, upon the basis of Ångström’s Atlas (‘Spectre normal de Soleil’), to express the indications of the screw directly in wave-lengths.

In place of the cross wires originally introduced into the focus of the observing telescope, I have inserted a tiny polished steel cone, the very fine point of which reaches to the centre of the field of vision. The axis of this cone stands perpendicular to the length of the spectrum, therefore parallel with the spectral lines, and the setting of the point of the cone on the latter is accomplished with great sharpness. If the spectrum is very faint, or consists only of bright lines, the cone is lighted by a small lamp. For this purpose, opposite to the point of the cone, there is an opening in the telescope, through which, regulated by a blind, light can be thrown on the point. As the latter is polished, a fine line of light thus appears, which extends to the centre of the field of vision, and the brilliancy of which can be altered by withdrawing the lamp to a greater distance or lowering the blind, so that even the faintest lines of a spectrum can be brought with facility and certainty into coincidence with this line of light.

The head of the micrometer-screw is divided into 100 parts, and each part, in the neighbourhood of the Fraunhofer line F, answers to about ·00016 wave-length. The probable error of position on one of the well-marked lines in the sun’s spectrum amounts to about 0·008 of a turn of the screw with the lowest eyepiece of the telescope. I have subjected the screw itself to a thorough examination with reference to such range, as well as to periodical inequalities in the single worms of the screw, but could discover no error exceeding 0·01 of a turn of the screw. I have to mention, further, that after each observation in the position in which the instrument was used, readings followed on the sodium-lines, or on some of the hydrogen-lines, in order to eliminate errors which might arise in the unavoidable disturbance of any particular part of the spectral apparatus.

1. Observations of the Aurora.

1870, Oct. 25th.—A very bright Aurora. In the brightest parts, besides a very bright line between D and E, several other fainter lines were to be discerned, situated further towards the blue end of the spectrum. They appeared on a dimly-lighted ground, and stretched out over the Fraunhofer lines E and b to about midway between b and F. Towards the red end the spectrum was terminated by the bright line first mentioned. No measurements could be taken, as the apparatus had not yet undergone the above-mentioned alterations, and even the brightest line of the spectrum did not diffuse sufficient light to be able to perceive the fine cross wires. The red rays of the Aurora were not examined.

1871, Feb. 11th.—Towards ten o’clock appeared in the north-west a very bright light-bow of greenish colour as the edge of a dark segment. Even with a very narrow slit, the line between D and E could be well recognized and measured. The average of six readings gave 7·11 turns, equal to 5572 wave-length. In a small spectroscope of low dispersion which is arranged on Browning’s plan, a few more lines placed further towards the blue could be recognized (as in October). Towards the red end of the spectrum no lines were observable. The greatest development of the Aurora was about midnight. Magnificent rays rose to about 60° elevation; they had the same greenish colouring as the bow of light, and the appearance of the spectrum also was exactly the same. I again obtained two sets of measurements: the average of six readings in the first set gave 7·10 turns, 5572 wave-length; in another part of the heavens at the same time 7·10 was the result of four readings.

On Feb. 12, towards eight o’clock, the intensity of the Aurora was already great enough to allow measurements of the brightest line. The average of six readings gave 7·09 turns, or 5576 wave-length. Dr. Lohse took observations later, with the same apparatus, and found from six readings 7·12 turns, or 5569 wave-length.

Yet the appearance of the spectrum in the spectroscope of low dispersion was essentially distinct from that of February 11th. The green continuous spectrum was present; it extended from the bright Aurora-line to the lines b of the solar spectrum, and was traversed by some bright lines. Between band b and F, was another line standing alone, out beyond F, in the blue part of the spectrum, a clear bright stripe; and just before G a very faint broad band of light was perceived.

Amongst the rays which, later on, shot upwards, and were coloured red at their ends, another very intense red line appeared in the spectrum between C and D, yet placed nearer to C [20].

April 9th.—An exceedingly brilliant Aurora, of which the greater development took place in the early morning hours. Magnificent red sheaths rose up to the zenith. The spectrum was like that observed on February 12th, only much more intense, so that the lines could be seen and measured with the larger spectral apparatus. In the brightest part of the Aurora was the dark segment; the spectrum consisted of five lines in the green, and a somewhat indistinct broad line or band in the blue.

The red rays, on the other hand, allowed us to recognize seven lines, whilst the bright line again appeared in the red part of the spectrum. I could not again perceive the faint stripe observed on February 12th, in the vicinity of line G. The mean measurements of four readings on an average, for each line, gave:—

Turns of screw.	Probable errors.	Wave-length.	Probable errors.	Remarks.
4·62	·0037	6297	·00014	Very bright stripe.	On a faintly lighted ground.
7·12	9	5569	2	Brightest line of the spectrum; becomes noticeably fainter at appearance of the red line.
7·92	—	5390	—	Extremely faint line; unreliable observation.
8·71	21	5233	4	Moderately bright.
8·95	49	5189	9	This line is very bright when the red line appears at the same time, otherwise equal in brilliancy with the preceding one.
10·06	20	5004	3	Very bright line.
12·33	—	4694	—	Broad band of light, somewhat less brilliant in the middle.
12·59	22	4663	—	Broad band of light, somewhat less brilliant in the middle.
12·88	—	4629	3	Very faint in those parts of the Aurora in which the red line appears.

April 14th.—Faint Aurora; only the bright line in the green could be recognized in its spectrum. The mean of two readings gave 7·12 turns, or 5569 wave-length.

I append a table of the wave-lengths of the brightest line, as exactly measured on four evenings:—

1871,	February	11	5573
	”	12	5573
	April	9	5569
	”	14	5569

Therefore the average result (if only half-weight is allowed to the last observation, because it only depends upon two readings) gives for the wave-length of the brightest line 5571·3, with a probable error of ·000·92. According to Ångström [21], the wave-length of this line is 5567; according to Winlock [22], on the other hand, 5570.

2. On the Spectra of some Gases in Geissler’s Tubes, as well as on the Spectrum of the Atmospheric Air.

Numerous experiments have been made in order to find out some admitted connexion between the spectrum of the Aurora and the spectra of the principal gases composing the atmosphere. I limit myself to noticing some of the often-repeated observations in Plücker’s tubes, which contained oxygen, hydrogen, and nitrogen, as well as the observations of the spectrum of the air under different conditions. The experiments were made with a small inductive apparatus, in which the length of the spark between platinum points in ordinary air was 15 millims. at the most. As Zöllner (in the pamphlet mentioned) comes to the conclusion, that if the development of the light in the Aurora, according to the analogy of gases brought to glow in rarefied spaces, is of an electric nature, it must belong to very low temperature—in order to bring the gases enclosed in the tubes to glow at the lowest possible temperature, I have always employed such weak currents that the gas was only just steadily alight.

The following observations have been repeated often and at various times. The figures are averages of the indications of the micrometer-screw, so that the uncertainty of the figures will, in the rarest cases, amount to no more than 0·015 turn of the screw, and must be reckoned somewhat more highly only in the case of completely faint misty lines. The spectrum apparatus was that described above, and the slit was nearly the same in every experiment, and so narrow that the sodium-lines could be seen separated. The measurements, for the most part, extend only to the Fraunhofer line G, as I feared lest, through further turning the telescope by means of the micrometer-screw, too great a pressure might be exercised on the worms of the latter.

I. Oxygen.

a. In the narrow part of the Plücker tube.

Screw.	Wave-length.	Remarks.
3·97	6562	Moderately bright.
5·04	6146	Very bright.
6·98	5603	Very bright, misty towards the violet.
8·19	5332	Faint.
8·95	5189	Moderately bright.
10·97	4870	””
11·02	4863	Faint.
11·26	4829	Bright; misty towards the red end of the spectrum.
13·30	4583	Very faint.
14·05	4506	Moderately bright.
15·55	4372	””

b. In the wide part of the Plücker tube.

Screw.	Wave-length.	Remarks.
6·98	5603	Very faint.
8·95	5189	Very bright.
11·26	4829	Moderately bright.

The lines near 3·97 and 11·02 belong to hydrogen. Probably traces of aqueous vapour were present in the tube, which were decomposed by the galvanic current. These two lines are not to be found in a lower temperature in the broad part of the tube. It is striking that the red nitrogen-line near 5·04 is also missing there. In the narrow part of the tube the lines stand out in the green on a very dimly-lighted ground, whilst in the wider part they appear on a perfectly dark ground.

II. Hydrogen.

a. In the narrow part of the tube.

Screw.	Wave-length.	Remarks.
3·98	6558	Very bright.
6·16	5813	Moderately bright, on both sides very faint lines.
7·01	5596	Moderately bright.	}On a dimly lighted ground, which becomes fainter towards the violet.
7·18	5555	Moderately bright.
7·77	5422	Faint.
8·95	5189	Moderately bright.	}On a faint steadily bright ground.
10·03	5008	Faint.
10·55	4929	Moderately bright.
11·04	4861	Very bright.	}From 11·5 to 12·9 a bright ground, which towards the violet becomes very bright.
12·86	4632	Moderately bright.
13·32	4581	Very faint.	} On a dull ground.
14·05	4506	”
15·90	4342	Very bright.

b. In the broad part of the tube.

Screw.	Wave-length.	Remarks.
5·30	6063	Faint.
7·00	5598	Bright.
8·96	5187	Very bright.
11·28	4828	”
14·04	4507	Moderately bright.

The lines appeared on a perfectly dark ground.

The tube shows in the narrow part the hydrogen-spectrum of the first order; the lines in the green do not coincide with the lines of the nitrogen, though some lines belonging to nitrogen are found. Here, too, most probably small particles of aqueous vapour have been enclosed in the tube and are decomposed. Very striking is the spectrum in the broad part of the tube; nothing is to be seen of the bright shining lines Hα 3·98, Hβ 11·04, Hγ 15·90; on the other hand, four very bright lines and one quite faint one are in the red end of the spectrum, which appear, in opposition to the spectrum of the narrow part, not on a partially lighted, but on an entirely dark ground. The appearance is very striking if we bring the tube in front of the slit; and so, by degrees, at first the light in the narrow part, then the light at the connecting-point of the narrow and wide parts, and, finally, the light in the latter fall upon the slit. At the connecting-point of the wide ends of the tube the three well-known hydrogen-lines decrease in intensity, the continuous ground of some parts of the spectrum disappears, and a new line appears near 11·28, which has about the same brilliancy as Hβ.

A comparison with the spectrum of oxygen shows the bright lines which are in the spectrum in the wide end of the tube as belonging to that element. The heat evolved by the current appears insufficient to bring the hydrogen to glow, whilst by it the oxygen, which is of a more rarefied character, becomes incandescent. An alteration of the direction of the current has no influence on the appearance.

III. Nitrogen.

a. In the narrow part of the tube.

Screw.	Wave-length.	Remarks.
3·84	6620	Several faint, broad, close lines, increasing in brilliancy as they approach the violet end.
4·85	6213
5·30	6063	Broad bright lines, so close together that the intervening spaces appear like fine dark lines. This part of the spectrum is very bright, but not uniform, being brighter towards the violet end.
5·51	6000
5·69	5948
5·87	5896
6·04	5846
6·20	5802
6·43	5741
6·96	5607	Group of faint but at the same time very broad lines. The last is the brightest.
7·13	5567
7·28	5532
7·55	5470	The dark intervening spaces are somewhat broader, the bright lines somewhat more intense than in the preceding group, and all of almost equal brilliancy.
7·74	5428
7·92	5389
8·09	5353
8·32	5306	Very faint fine lines.
8·50	5272
8·69	5237
9·01	5178	Very bright broad misty line.
9·67	5066	Very bright line.	The bright lines are sharply defined towards the red end of the spectrum, fading away towards the other end of the spectrum.
10·25	4975	””
10·66	4913	””
11·03	4862	Very faint line.
11·41	4811	Bright line.
12·11	4721	””
12·57	4666	Faint line.
12·57	4644	Bright, broad, misty line.
13·42	4570	Very bright line.
14·24	4487	””
15·02	4417	Bright line.
15·66	4363	Bright lines.	Bright lines sharply defined towards red end, indistinct towards other end of spectrum.
15·72	4357	”
15·87	4345	Bright line.
16·72	4273	”

Here follow several lines.

b. In the wide part of the tube.

Screw.	Wave-length.	Remarks.
6·20	5802	Faint, indistinct, broad line.
7·72	5433	Dull stripe.
8·20	5330	Faint line.
8·94	5191	Very faint line.
9·03	5175	Broad band of light.
9·90	5029	Dull band of light.
10·68	4911	Moderately bright line.
11·42	4809	Faint line.
12·59	4663	Bright line.
13·43	4569	”
14·07	4504	Moderately bright line.
14·25	4486	Very bright line.
15·85	4347	””
16·76	4273	Moderately bright line.

c. At the aura of the negative pole.

Screw.	Wave-length.	Remarks.
5·18	6100	Broad, moderately bright stripe, indistinct at the edges.
5·70	5945	Broad, moderately bright stripe, indistinct at the edges.
7·60	5159	Broad, moderately bright stripe.
8·41	5289	Broad, moderately bright stripe.
8·76	5224	Very bright line, somewhat indistinct towards the violet.
9·19	5147	Faint line.
10·00	5004	Bright line, indistinct towards the red.
10·67	4912	Somewhat fainter than the last, indistinct towards the red.
11·43	4808	Very faint line.
12·25	4704	Very intense, broad, indistinct towards the violet.
12·73	4646	Very faint line.
13·43	4569	Moderately bright, indistinct towards the violet.
14·25	4486	Like the last.
15·03	4417	Quite a faint line.
15·86	4346	Moderately bright line.
16·76	4275	Very bright line.

Here follow several other lines.

The observations in the different parts of the tube show plainly the dependence of the spectrum on the temperature. The aura of the negative pole gives the line near 10·07 so characteristic of the air-spectrum. This is the same line which is met with in the spectra of most of the nebulæ. The very striking groups of lines in the red and yellow in the spectrum of the narrow part of the tube disappear entirely in the wide part. If we compare the spectra with those above quoted, of oxygen and hydrogen, we find line Hβ very faint in the spectrum of the narrow part of the tube near 11·03; on the other hand, oxygen-lines appear in the broad part near 8·20, 8·94, and 14·07. Thence I would conjecture that the tube was not filled with pure nitrogen, the appearance of which is precise, but with dry rarefied air, since Wüllner’s researches have proved that dry air yields the same spectrum as nitrogen gas. Perhaps the air in the tube examined by me had not been thoroughly dried, and thus the appearance of some lines of the elements before named is to be explained.

I must further mention that the electrodes of the tubes consisted of aluminium; yet a comparison of the spectra observed and the aluminium spectrum has shown no connexion between them.

IV. Atmospheric Air.

Screw.		Wave-length.	Remarks.
	5·88	5892	Very bright double line (Na).
	6·67	5680	Very bright line.
	7·20	5550	Faint line.
	9·00	5180	Very bright line.
	9·79	5047	Fine faint line.
	10·03	5008	Very bright double line.
	10·07	5002	Very bright double line.
	11·43	4803	Faint confused line.
	12·69	4651	Faint line not sharply defined.
	12·84	4633
	13·04	4612
From	14·61	4453	Confused band of light, which ends with a broad washy line.
to	15·88	4444	Confused band of light, which ends with a broad washy line.

Here follow several other lines.

Rarefied air saturated with aqueous vapour.

Screw.		Wave-length.	Remarks.
	3·97	6562	Moderately bright line.
	(5·88)	5892	Bright double line (Na).
	(6·25)	5789	Bright line (H).
From	7·03	5591	Broad dull band of light; near 7·03 a somewhat brighter line.
to	7·55	5470
	7·59	5461	Bright line (H).
	8·72	5231	Dull stripe.
	8·96	5187	Broad misty stripe.	On a dull steady ground.
	10·07	5002	Faint line.	On a dull steady ground.
	11·05	4859	Very bright line.
	12·21	4709	Moderately bright line.	On dimly lighted ground, becoming fainter towards the violet.
	12·75	4644	Line fainter than the preceding.
	13·28	5585	Very faint line (H).
	(15·71)	4358	Very bright line (H).
	15·90	4341	” ”

Here follow several more lines.

In the first observations, the electric spark, about 1 centim. in length, was allowed to pass between platinum points in ordinary air.

The sodium-line near 5·88 appeared continually. The bright double line at 10·03 and 10·07, with a weaker current or longer spark, was no longer to be recognized as a double line, but appeared as a broad somewhat confused line, of which the brightest part was near 10·05. No lines belonging to the platinum spectrum appeared. Ordinary rarefied air, under a pressure of 25 to 30 millims., and which was enclosed by mercury in a tube 8 millims. wide, showed exactly the same lines as Plücker’s nitrogen-tube (b), except that some lines belonging to the spectrum of mercury also appeared.

This observation may be regarded as a confirmation of the conjecture above expressed as to the condition of Plücker’s tube III. (nitrogen). In the observations described under b, the air saturated with aqueous vapour was under a pressure of 22 millims. Besides the sodium-lines, lines of the mercury-spectrum appeared at 6·25, 7·59, and 15·71. The spectrum of rarefied air under similar pressure was found to accord completely with the spectrum of the light in the broad part of Plücker’s tube.

III. (Nitrogen b.)—A comparison of the spectrum of rarefied air saturated with aqueous vapour with the former shows the striking alterations in the spectrum which are brought about by the presence of the aqueous vapour.

3. Comparison of the Aurora-Spectrum with the Spectra of Atmospheric Gases and of Inorganic Substances.

In the next place, I turn to the comparison of the observed spectra of different gases and of the air with the spectrum of the Aurora. The first band of light in the red part of the Aurora-spectrum most probably coincides with the first system of lines in the spectrum of nitrogen (a). Probably only the bright part of this group of lines is perceptible, on account of the extreme faintness of the Aurora; and as in nitrogen the increase of the brilliancy of the spectrum takes place towards the violet end, the absence of the intermediate spectrum towards this direction would find its explanation. The most intense line of the Aurora-spectrum at 7·12 is to be also found in the spectrum of nitrogen (a)—as a very faint line, however. That this line appears in the Aurora by itself, and with intensity relatively great, need not appear strange, considering the great alteration of the gas-spectra under different conditions of pressure and temperature. The third line of the Aurora-spectrum, very vaguely defined on account of its great faintness, coincides in the same way with a nitrogen-line.

The line at 8·71 is met with in the nitrogen-spectrum (c), as well as in the air-spectrum (b). The third line of the oxygen-spectrum at 8·95, which seems to appear under very different conditions, is found again, as the fifth line in the spectrum of the Aurora. Moreover, the sixth line in the Aurora at 10·06 coincides very exactly with the known nitrogen-line appearing in the spectra of some of the nebulæ. Lastly, as to the broad band of light in the Aurora-spectrum from 12·33 to 12·88, several lines are found in this place in the spectrum of nitrogen as well as the air-spectrum (a, b); so that here, too, a coincidence between the spectra may be regarded as probable.

The observations show with some certainty that at least one line at 10·06 of the Aurora-spectrum coincides with the maximum brilliancy of the air-spectrum, and that the other lines appear with great probability in the spectra of atmospheric gases.

In the very great difference of the gas-spectra under varying conditions of pressure and temperature, it would indeed be difficult to succeed in producing artificially a spectrum which should resemble that of the Aurora in all parts. Moreover, it must be admitted, under the hypothesis that the Auroræ are electric discharges in rarefied air-strata, that these strata, qualified for the transmitting of electricity, will have a very considerable thickness.

In this case the conditions of pressure on these air-strata are themselves so different that, within certain limits, each will yield its own peculiar spectrum; but we shall see the sum of collective spectra, so to speak, spread out behind each other; and therefore the impossibility of attaining a perfect agreement between the Aurora-spectrum and the artificially exhibited spectra of mixed gases is evident.

A comparison of the Aurora-spectrum with the spectra of inorganic substances may be easily worked out by the help of the above-quoted wave-lengths of the single lines of the former, with due regard to probable errors, and with the aid of Ångström’s Atlas of the Solar Spectrum. Here the perfect harmony of the brightest Aurora-line (which was fixed with an exactitude of about one seventh of the separation of the sodium-lines) with the lines of the iron-spectrum is especially striking. The wave-lengths in the above-cited observations of the bright Aurora-line vary between 556·9 and 557·3, whilst, according to Ångström, two lines of the iron-spectrum are situated at 556·85 and 557·17.

Iron-lines corresponding to the other Aurora-lines, within certain limits of accuracy, are also to be found, as will be seen from the following comparison:—

Aurora-lines.		Lines of the iron-spectrum.	Remarks.
	629·7	630·08	Moderately bright.
	629·7	629·85	Moderately bright.
	539·0	539·60	Mostly very faint.
		539·92
		539·05
		538·85
	523·3	523·43	Very faint.
		523·21	Moderately bright.
		522·90	Very faint.
	518·9	519·79	”
		519·40	”
		519·16	Moderately bright.
		519·06	” ”
		518·51	Very faint.
	500·4	500·65	Very faint.
		500·52
		500·49
		500·30
		500·20
From	469·4	3 stronger and 4 very faint iron-lines.
to	462·9	3 stronger and 4 very faint iron-lines.

Yet this agreement, though remarkable, can only be considered as complete proof of the presence of iron-vapour in the atmosphere when we shall have succeeded in showing by observation analogous modifications of the relative conditions of brilliancy in the iron-spectrum by alterations of temperature and density; and in this way explain the appearance of relatively very faint iron-lines in the Aurora-spectrum, or, on the other hand, the absence of the most intense lines.

It will meanwhile remain far more in accordance with probability to regard the Aurora-spectrum as a modification of the air-spectrum; since we are already aware, in the case of gases, of the alteration of the spectra by conditions of temperature and pressure; and an agreement, at any rate, quite as certain between the spectrum in question and the spectra of atmospheric gases has been proved above.

[I am indebted to Miss Annie Ludlam for a translation from the German of the above Memoir.—J. R. C.]