Auroræ: Their Characters and Spectra

Having regard to the near proximity of the phosphoretted-hydrogen band to the bright Aurora-line, to the circumstance of this band brightening by reduction of temperature (a phenomenon probably connected with ozone), to the peculiar brightening of one line in the green in the “Aurora” and “phosphorescent” tubes (the phosphorescent tubes probably containing O), and to the observed circumstance that the electric discharge has a phosphorescent or fluorescent after-glow (isolated, I believe, by Faraday), I feel there is strong evidence in favour of such an origin to the principal Aurora-line, if not to the red line as well.

5th. Professor Ångström opens a wide door to discussion in his proposition of the invariability of gas-spectra, and I do not now attempt to follow in detail this interesting part of the present subject. Suffice it to say, that if the Professor lays down this proposition in its strictest sense (I can hardly suppose he so meant it), there is, so far as I am aware, no one spectrum that can at all claim comparison with the Aurora-spectrum. Giving greater latitude to the Professor’s words, I reply, upon competent authority, that lines vary in number and brilliancy with temperature, and in breadth with pressure. Kirchhoff, too, in speaking of vapour-films as increasing the intensity of lines, states “it may happen that the spectrum appears to be totally changed when the mass of vapour is altered.” We may, too, now add magnetism as capable of effecting a change in certain spectra, not only as to brilliancy, but even as to position of lines. (Chautard’s Researches, ‘Philosophical Magazine,’ 4th series, vol. 1. p. 77, and experiments detailed in Chap. III. of this work.)

CHAPTER XIII.
THE OXYGEN-SPECTRUM IN RELATION TO THE AURORA (PROCTER AND SCHUSTER).

In a communication to ‘Nature,’ Mr. H. R. Procter has pointed out an apparent coincidence in position of several of the Auroral lines with those of a spectrum occasionally obtained from air at low pressure with a feeble discharge. It is, he says, sometimes exhibited in lumière (phosphorescent?) tubes, and he believed it, in part at least, to be the spectrum described by Wüllner (Philosophical Magazine, June 1869) as a new spectrum of oxygen.

He had obtained it very vividly in pure electrolyzed oxygen with a feeble discharge, but some perplexing observations made him doubtful of its origin.

Plate XI. fig. 4 gives a representation of this spectrum as shown by Mr. Procter, except that my drawing is in black for white.

The upper spectrum is that above mentioned, the centre one that of the Aurora, the lower one the lines of Na and H for comparison. The Auroral yellow-green line, in January 1870, was found by Mr. Procter coincident with a bright line or band in the tube (with a spectroscope of a 60° bisulphide prism, and magnifying-power about six). The third and fifth lines in the Aurora seemed also to correspond with tube-lines. As to these Mr. Procter says they were not bright enough to be compared with the same accuracy as the yellow-green line, but that the positions could not be far wrong.

Mr. Procter subsequently (‘Edinburgh Encyclopædia,’ art. “Aurora”) considered he traced the yellow-green tube-line to some form of hydrocarbon. On examination with instruments of greater dispersion, it was found that, though more refrangible than the first band of citron acetylene (candle-flame), it was less so than the Aurora-line. The tube-band, too, was shaded towards the violet, which was not the case with the Aurora-line.

The question as between hydrocarbon and oxygen I did not then consider as disposed of. With the lumière tubes the question might be open, but I did not see how it could be in the case of the electrolyzed oxygen-spectrum.

From a comparison of the tube-spectra, I have shown that although the spectra of the carbon and oxygen tubes are proved to be, photographically, as a whole, distinct, they have, as to position of some of the principal lines in the central part of the spectrum, a very close resemblance.

That oxygen may in some form play a part in the Aurora seems highly probable; how far it is spectroscopically detected seems a different question.

Ångström and Herschel suggest its presence in the Aurora in connexion with phosphorescence or fluorescence. With a spark-discharge in air at ordinary pressure, a mixed spectrum of bright lines of N and O is found; while in the case of Geissler vacuum-tubes (representing a glow-discharge in a much more rarefied atmosphere) the N lines appear mainly to usurp the spectrum.

It must, however, be borne in mind that a Geissler tube, as to temperature at least, in no way represents the conditions of the Aurora; and when we remember the association of oxygen and ozone, and the way in which the latter is affected by heat, it may well be that temperature plays an important part in the matter. In proof of this conduct of oxygen, it may be cited that, in the case of a H₂O tube, the H lines come out sharp and brilliant in the spectrum, while what is seen of the O lines is comparatively weak, misty, and ill-defined. Vogel, it will be remembered, makes 5189 of the Aurora coincident with an O line.

Professor Herschel has pointed out, and I have confirmed, that the residual phosphorescence in Geissler tubes, after the spark has passed, is probably associated with oxygen. He also alludes to the fact that when one of the globes of a “Garland” tube was heated, it did not shine after the spark had passed, apparently because of the destruction of the ozone by heat.

[Some experiments with a tube of this description will be found detailed in Part III. Oxygen was not, I think, the gas it was filled with.]

Subsequently to my examination and comparison of the O and CO₂ spectra before detailed, Dr. Arthur Schuster was good enough to send me three vacuum-tubes of his own preparation, showing an oxygen-spectrum.

One, with large disk-shaped brass electrodes, was unfortunately broken in transit. Dr. Schuster informed me it showed the carbonic-oxide spectrum as well as that of oxygen. The other two tubes had aluminium electrodes. They were similar in shape to ordinary Geissler tubes, but had attached to each a supplemental bulb containing dry oxide of manganese. Illuminated by the larger coil, one of these tubes (which had a slight crack in the manganese bulb) lighted up faintly; the other was fairly bright, and the glow had a somewhat reddish tint.

Plate XVIII. fig. 15 represents as the upper spectrum Vogel’s Aurora, with W.L. numbers, as the middle spectrum the capillary part of Dr. Schuster’s O tube, and as the lower spectrum the negative (violet) pole of the same tube.

The tube-spectra were mapped out with the aid of the diaphragm micrometer before described.

The capillary spectrum was mainly distinguished by four bright sharp lines—one in the red, between the red Aurora-line and D, two in the green, but considerably more refrangible than the yellow-green Aurora-line, while the fourth was found to be hydrogen F. The other lines in the spectrum were considerably fainter, and misty and band-like. The red line, though not brilliant, was fairly bright and sharp.

The place of the less refrangible of the two bright bands in the violet-pole spectrum was occupied in the capillary spectrum by a faint glow only.

The violet-pole spectrum was recognized by two very bright broad bands of light in the green, each including within its limits one of the Aurora-lines. The bright red line in the capillary had a faint representative in the violet-pole spectrum, as also had the two bright lines in the green. Other fainter lines appeared in the blue, and three fairly bright ones towards the violet.

Dr. Schuster remarks that one of these O bright bands is closely coincident with a band in the CO spectrum, but that the CO band is bright towards one edge and fades off gradually thence, while the O band is of pretty uniform strength throughout. Dr. Schuster finds the wave-lengths of the violet-pole O bands to be as follows:—

He also gives as weak bands 5840-5900 and 5969-6010. Dr. Schuster comes to the conclusion that the green line of the Aurora is not due to oxygen, as, under considerable dispersion and with good definition, the oxygen-bands can be broken up into a series of lines, when the brightest part is found to lie at 5586, which is too much towards the red to compare with the Aurora-line. He notices that the more refrangible of the O bands corresponds with a line sometimes seen in the Aurora (Vogel’s 5233). The same remark will, however, apply to this last as to the other coincidence, viz., that a broad band can hardly represent a line—at least, the line can only be said to coincide in a loose and indefinite way. It is evident that Dr. Schuster’s tubes were free from what must now be considered an impurity in those examined by me and by Dr. Vogel, and that Mr. Procter’s suspicions of carbon impurities in these, and the ordinary oxygen-tubes, are thereby quite confirmed.

In some experiments which we made (after receiving Dr. Schuster’s tubes) with an open Geissler tube, so arranged as to connect with an air-pump and gas-receiver, and thus from time to time to wash out the tube and vary its contents, we found the same impure spectrum as in the case of the sealed O tubes; and it seems to require a very large amount of precaution to avoid these impurities.

Dr. Schuster was kind enough to examine the spectra I mapped out, and which are shown in Plate XVIII. fig. 15, with the following results:—The lines Oα, Oβ, Oγ are those he has referred to under that designation in his communications to ‘Nature,’ and undoubtedly belong to oxygen. The bands A, B, and C are the bands characteristic of the negative pole. He finds A divided into two parts by a dark space. The spectrum of the negative pole, under good exhaustion, stretches into the capillary part; hence B appears in the capillary as a faint band. A similar thing happens with nitrogen. I., II., III., and possibly 8 and 9, he thinks, are due to the spark-spectrum of oxygen, obtained when the jar and a break are interposed, the brighter lines of the line-spectrum being always present at the negative pole. These last-mentioned lines I have already referred to, as having been found by me in a tube showing phosphorescence after the spark has passed. (Compare Plate XVIII. fig. 15, O violet pole, with Plate XV. spectrum 5.) Nos. 1 and 2, he thinks, are due to some foreign matter, as they are not in all his tubes.

Dr. Schuster often finds that a spectrum due to the aluminium electrodes is seen in tubes under great exhaustion; and this he considers is the spectrum of aluminium oxide. A drawing of this spectrum is found in Watts’s ‘Index of Spectra,’ plate iii., “Aluminium first Spectrum.” To this, he thinks, are also due the bands, or sets of lines in my aluminium-arc spectrum (‘Photographed Spectra,’ plate ii.), and he believes lines 3, 4, 5, 6, and 7 in the mapped-out spectra are due to it. It would thus appear that the lines due to O are few in number, and do not well compare with the Aurora-spectrum.

PART III.
MAGNETO-ELECTRIC EXPERIMENTS IN CONNEXION WITH THE AURORA.

INTRODUCTION.

The set of experiments detailed in Chapters XIV. to XIX. was mainly conducted for the purpose of testing, in connexion with the Aurora, the action of a magnet upon the electric glow in vacuo and on the spark at ordinary pressure. It also includes some observations on the glow from the violet pole with and without the magnet, and on the glow obtained from one wire only. The apparatus employed was a Ladd’s electro-magnet, with poles 10¼ inches high by 2 inches across, each pole being surrounded by a movable helix, composed of two sets of stout copper wire wound together, so that they could be used either in one length or as independent coils excited at the same time. The latter form of arrangement was employed by us. In most of the experiments conical armatures were employed for the purpose of bringing the action of the poles to bear upon the subjects examined. A contact-maker was added to the magnet, so that it could be put rapidly in or out of action without disturbing the wires. The battery used to excite the magnet was of the form known as that of Dr. Huggins, and consisted of four vulcanite cells in a frame, each holding seven pints of bichromate solution, and containing two carbon and one zinc plate, each 13½ by 6 inches.

A winch and pulley enabled the whole set of plates to be lowered into the liquid and withdrawn at pleasure, and the large quantity of solution gave the battery a considerable amount of constancy. We found it could be used for two evenings’ work, of four hours each, without any material dropping in power. For obtaining the glow in the Geissler and other small tubes, a Ruhmkorff coil, giving a ½-inch spark, excited by one plate of a ½-gallon bichromate (bottle form), was used. For the glow in the larger tubes and the spark in air a larger coil, giving a two-to three-inch spark, and worked by two ½-gallon double-plate bichromates, was employed. Notes were taken of the experiments, and drawings of the effects at the time; and these are reproduced almost literally in the text and Plates comprised in this Part. To ascertain the direction and extent of the magnetic curves, we covered large sheets of cardboard, placed over the poles, with iron filings; excited the magnet so as to obtain the curves, and then obtained permanent prints from the filings by spraying the cardboard with tannin solution. The magnetic effects were thus found to extend to a radius of at least ten inches (see diagram, Plate XVII. fig. 1, showing magnet-poles and curves on a ¼ scale).

In the vacuum-tube experiments we held Mons. J. Chautard’s investigations (on the action of magnets on rarefied gases in capillary tubes rendered luminous by the induced current, Phil. Mag. 4th series, vol. 1. p. 77) in view. We obtained in our experiments plenty of evidence of a change of colour and form in the discharge under the magnetic influence; and both simple and compound spectra were found to be much varied by the exaltation or suppression of some parts of the spectrum, so that apparently new lines sprang up; but we failed to trace actual change of position or wave-length in any given line, though we carefully looked for it. A portion of our researches was directed to the subject of Ångström’s experiment of filling a dry flask with a violet glow, analogous to that from the negative pole. We entirely failed in obtaining the same result while two wires and an uninterrupted circuit were employed. When, however, we attached a negative wire only (the other wire being left free) to an exhausted globular receiver, we obtained an effect very similar to that referred to in Prof. Ångström’s memoir.

The general result of the experiments was to prove, assuming the Aurora to be an electric discharge, the great influence the magnetic forces may exercise on the colours, form, motions, and probably the spectrum also of that phenomenon. It is easy to conceive that the variation in number, and intensity of the lines which has been remarked in Auroral spectra may have its origin in such a cause. The influence of the magnet on the capillary stream was mainly in colour and intensity; but in the bulbs the effects were still more marked and striking, and, in a greater or less degree, different in the case of each gas which we examined. A careful and extended study of these effects, conjointly with the changes in the spectrum, might possibly form a new and valuable mode of analysis of compound gases. This is well illustrated in the case of the iodine and sulphur tubes which we examined.

CHAPTER XIV.
EXAMINATION OF GEISSLER TUBES UNDER ACTION OF THE MAGNET.

Nitrogen-tubes.

(1) A small Geissler tube (No. 1) was lighted up by the small coil. The capillary part showed a very bright, slightly rosy-tinted stream. Negative bulb was filled with rosy-purple light, the violet-pole glow being confined to the extent of the electrode. Positive bulb of the same rosy-purple colour, but stream slightly contracted in volume. Glow throughout quiescent, and no stratification in the tube. A compound-prism spectroscope, taking in the whole of the spectrum, showed in the capillary stream, from yellow to red, a fairly bright wedge, having a dark band in the centre, and six bright columns, with dark lines at intervals, shading off on either side. On the more refrangible side of the yellow, the spectrum was composed of a set of bright bands and lines in the green, blue, and purple, one line only (in the green) standing out very bright. In the yellow and red no bright line stood out alone. The positive bulb gave a fainter spectrum of the same character, mainly confined to the centre, the violet, yellow, and red not being well seen. When the violet-pole glow was examined, the general character of the spectrum was quite changed: a brilliant broad band in the violet, a bright narrower one in the blue, and two bright lines in the green, with intermediate fainter lines throughout, were the main features. The yellow and red part of the spectrum was also changed. The yellow was fairly and evenly distinct up to the dark band; then came a somewhat brighter orange band, and after that the red, but rather obscure and cut off. No absolutely bright line could be traced in the red.

(2) To compare the capillary stream and the violet-glow, a second nitrogen-tube (No. 2) was used. This tube was larger in bulk and bore than No. 1. The glow in the bulbs was considerably fainter and more salmon-coloured; and there was much stratification in both, extending to the capillary bore. (This stratification was considered due to H, as the three principal lines of that gas came out very brightly in the spectrum.) The difference of the spectra of the capillary stream and of the violet-pole glow was extremely well marked—the former consisting of a set of bright lines and bands of fairly uniform intensity, while the latter was split up into a few bright bands with fainter lines between. The yellow and red of the violet-glow were very weak as compared with the same region of the capillary spectrum. No bright line appeared in the red. The tube being properly adjusted for the purpose, the junction of the violet-pole glow and the capillary red-glow was easily observed. The bright bands of the violet-pole were seen to run into the capillary line-spectrum, and then, gradually getting finer and more pointed, to fade out.

(3) The capillary part of tube No. 1 was arranged between the poles of Ladd’s electro-magnet, the conical ends of the armatures almost touching the tube (Plate XVII. fig. 1). With the magnet not excited, the capillary stream was bright and of a slightly rosy-yellow tinge. It varied a little in apparent diameter with the current. As soon as the magnet was excited the capillary stream, as also (in a less degree) that in the bulbs, were seen to contract, and to change from a rosy tint to a distinctly blue-violet. The polished armatures, acting as reflectors, showed this change of tint in a most marked manner each time the magnet was excited. At the same time the capillary stream was seen to run into the negative bulb, as if overflowing, and with an effect resembling the “tailing-over” of a gas-flame. This effect took place each time the magnet was excited, and was not found at the positive-bulb end.

Occasionally, when the magnet was excited, flashes of light were discharged in the negative bulb from the capillary towards the violet-pole. The spectrum was then carefully examined. No change was seen in the actual position of any of the lines or bands when the tube was influenced by the magnet, but those towards the violet end of the spectrum were conspicuously brightened.

(4) The extremity of the negative bulb was now placed between the poles of the magnet. A bright violet-coloured arc, following the magnetic curve, was at once formed, as in the case of the large Plücker tubes; and at the same time a straight stream of not very bright light ran along the bulb. The positive bulb was next placed within the action of the magnet; and immediately a brilliant spiral of flickering light appeared in the bulb, lighting it up, and reminding one in shape of the spiral which water forms on being poured from a lipped jug (see Plate XVII. fig. 9).

This was repeated each time the magnet was excited. The spiral, though flickering in character, was permanent in form, and inclined to the side of the tube which was in contact with the N pole of the magnet.

Oxygen-tubes.

A tube (No. 1) was lighted up and examined with the spectroscope, and found to give the spectrum shown on Plate XIV. spectrum 3, but with a strong set of H lines in addition.

A second tube (No. 2) was then lighted up. The spectrum was a bright one, similar to the foregoing, the principal H lines being present, but not strong.

The red region was indistinct, and showed no prominently bright line. The bulbs were mainly of a slightly blue-grey tint, with a steady glow. Capillary stream quite pale white, with a very slight tinge of red. Violet-glow small and confined to the electrode. Upon the magnet being excited, the capillary stream became intensely brighter, and the glow in both bulbs contracted into a single bright stream, which curved towards the sides of the bulbs at right angles to the magnetic poles, and changed from side to side with the current. This effect was very marked, and was more apparent in the positive than the negative pole. A faint stratification was seen in both bulbs. Upon either bulb being placed between the armatures, the glow left the electrode point and condensed into one bright stream, running along the side of the tube and curving at each end (Plate XVII. fig. 10). No trace whatever of tendency to form a spiral was seen. The spectrum with the magnet on was very conspicuously brightened up throughout. A set of fluted bands with a bright line among them appeared in the red, and several lines or bands appeared in the violet which could not be seen before. The bright red line, upon measurement, proved to be the hydrogen-line C. It thus seemed brighter in proportion than the F line, although, with the magnet off, the latter was well seen, while the C was not. No actual change in position of the spectrum-lines could be detected.

[It is to be noticed that the O tubes employed were those used by me in former experiments, and had the bright lines now attributed to hydrocarbon impurity. Their bulb-effects differed, however, entirely from those of the CO₂ tube. (Compare figs. 10 and 11, Plate XVII.)]

Hydrogen-tubes.

A small H Geissler tube (No. 1) was selected, and lighted up by the small coil. The capillary was a bright white-pink stream, with a tendency to redden at times. The bulbs were both of a faint blue-grey tint, with coarse lenticular stratification. The violet-pole glow was pale and white as compared with that of N.

When the magnet was excited, the whole character of the tube changed. The capillary stream diminished in brightness and in apparent volume, and changed to a deep amber-yellow. The bulbs lost some of their light, and their coarse stratification; being, in lieu, filled with a vertical condensed stream of moderate light, in which a fine stratification only was seen. The stream in the positive bulb had a tendency to the spiral form. The capillary, each time the magnet was excited, “tailed over” into the negative bulb, as in the case of N, looking as if it were squeezed out of the capillary bore. The unexcited spectrum was found to consist of the usual principal lines of H on a continuous glow, with the intermediate bands and finer lines, which are usually suspected to be due to impurity. The sodium-line was also seen. When the magnet was excited, the spectrum grew much fainter—the continuous glow in the red and blue, and the red and blue lines, nearly disappearing, and the line in the green alone shining out conspicuously. No change of place in the lines could be noticed.

A longer H tube (No. 2) was then tried, with similar effects, except that the diminution in brightness was not so conspicuous. When the negative bulb of the tube No. 1 was placed between the poles of the magnet, a stream of light was formed, and the stratification became finer. The same effect took place with the positive bulb, with a tendency to the spiral form.

Water-Gas (H₂O) tube.

A faint purple glow was seen in each bulb, the tube not lighting-up brightly. The capillary showed a slightly rosy-tinted, grey stream of brighter light. With the magnet on, the glow in the bulbs was condensed into a single bright stream. The capillary brightened up, and assumed a yellow tint—this effect being principally confined to that portion which was between the conical ends of the armatures, and gradually diminishing as the distance increased from these. Without the magnet, the principal H lines showed brightly in the spectrum, the O lines being misty and indistinct. With the magnet on, the O lines and spectrum generally brightened up.

Ammonia-tube.

This tube was difficult to light up. Hardly any light was seen in the bulbs, except a very faint purple glow at the electrodes. In the capillary part a fairly bright stream of purple-white light appeared. The spectrum was a faintly shown one of N and H. The effect of the magnet was to reduce the brightness of the glow in the capillary, but with little marked action on the bulbs, except to condense the faint glow into a slightly bright stream running along the side of the tube.

On a subsequent examination the tube and spectrum both brightened up under the influence of the magnet. The N lines, which were faint without the magnet, shone out under its influence distinctly—the red and yellow parts of the spectrum specially showing this effect. The H lines also brightened up, but hardly so much in proportion as the N.

Carbonic-Acid tube.

A Geissler tube marked C A was examined. Capillary stream a brilliant bluish white; bulbs grey-blue, with a slight tint of green; slight stratification in positive bulb; stream diffuse, not quite filling the bulbs, and changing in volume as the coil-break was touched; glow round the violet-pole considerable, but markedly white in tint, rather than violet; stratification strong in capillary. With magnet excited, the capillary stream diminished in volume, but greatly increased in brightness. It “tailed over” into the negative bulb, and the stream through both bulbs curved towards the sides. A slight pattering noise was heard in the tube. In the positive bulb bright, imperfectly formed, saddle-shaped rings of light, with a tendency to spiral formation, were seen, somewhat similar to the effects in the Plücker tube after described (see Plate XVII. fig. 11).

The whole spectrum, under influence of the magnet, became much brightened up. Faint bands in the red came out bright, as also did some in the violet. The violet-glow was examined (without the magnet), and the light was found condensed into four prominent shaded bands, one red, one yellow-green, one green, and one blue, with fainter bands seen between.

Chlorine-tubes.

A chlorine-tube (No. 1) was lighted-up with the small-coil. Capillary stream of a pale green tint. Bulbs with very little glow in them; spectrum pale, and not very distinct. Under action of the magnet this tube brightened up throughout, and the glow became more condensed, and ran to the sides of the tube. The spectrum also brightened, the faint lines becoming stronger, but the general character was preserved.

A second chlorine-tube (No. 2) was then tried. Both bulbs were completely filled with a dense white (very slightly rosy-tinted) opaque light, and capillary the same, but brighter. A very slight violet tinge was seen at the negative pole. When the magnet was put on, both bulbs were at once filled with flickering bright streams of light, running towards the side of the tube, according to the direction of the current.

The capillary stream at the same time changed from white to an intense bright green. The spectrum without the magnet consisted of sets of lines, with two well-marked absorption-spaces between, all seen somewhat faintly, as if through a mist.

When the magnet was put on, the marked character of the absorption spaces was lost. The sets of lines in the yellow-green and green started up intensely bright, while those in the blue only slightly brightened.

The misty appearance was altogether lost, and the bright lines all shone up upon a perfectly dark background, with a strikingly metallic look; we could not, however, trace change of position or actually new lines. It seemed as if lines which had been faint in the yellow-green and green region suddenly increased in intensity, the other parts of the spectrum not being similarly influenced. They quite flashed up when sudden contact was made with the magnet commutator.

Iodine-tubes.

This tube (No. 1) had been used for photographic purposes, and the bulbs were partly obscured by a white deposit.

On lighting it up, both bulbs were filled with a violet-grey diffused light, with much coarse well-marked lenticular stratification. This stratification was mainly lost on changing the direction of the current, but made its reappearance when one conducting-wire was touched with a finger. This effect was still more marked when one finger of each hand was applied to the wire. The capillary stream was of a pale lemon-yellow. On putting on the magnet the light in the whole tube was nearly extinguished, a faint thin stream of condensed light running through the centre of the tube alone remaining.

On placing the bulbs between the magnet-poles, effects were produced similar to those in the case of the tube, after described (p. 144, and marked Si Fl₆), but in a less marked degree.

The iodine-tube was subsequently again tested, and it lighted-up better than on the last occasion, showing nearly the same effects in bulbs and capillary, the former having somewhat of a rosy tint and the latter an amber.

On exciting the magnet, the capillary part of the tube changed from amber to a decided light green. The spectrum, without the magnet, gave one very bright line, and several less bright ones near, in the blue-green. The rest of the spectrum, with the exception of the absorption-spaces, was misty and continuous, with lines showing faintly through. The red and yellow portions of the spectrum were quite bright. When the magnet was excited, the spectrum entirely changed. The red and yellow portions of the spectrum, and the misty continuous light, all quite disappeared; while a set of sharp lines on the yellow-green and green flashed up bright and clear, and stood out alone upon a dark background, in which the absorption-spaces were lost. The effect was very strongly marked, and gave a totally different character to the appearance of the spectrum. The change seemed to arise from the suppression of one part of the spectrum, and the increase in intensity of the lines in the other part.

The principal lines could not be traced to change in actual position.

This tube differing somewhat from a second one we examined (No. 2) in tint of glow and spectrum, it suggested itself to us that there might be a partial mixture of N or H (or both) with the iodine vapour, giving rise to some of the brighter parts of the spectrum which were extinguished under the action of the magnet.

We therefore compared these two tubes, viz. the old one (No. 1) and the new one (No. 2), and also their spectra, by means of a comparison-prism on the slit of the spectroscope. To the eye, the tubes differed much in appearance. No. 1 had a distinct transparent rosy tint throughout, with considerable coarse flickering stratification; and this contrasted strongly with the dense whitish light of tube No. 2, which showed neither movement nor stratification. The spectra were also found different in general look. That of tube No. 1 was strongly tinged in the red and yellow, and showed a bright continuous spectrum, crossed by many sharp lines, with little trace of absorption-spaces. The spectrum of No. 2 was much whiter in tint, showed very little of the red and yellow, and the absorption-spaces were very dark. A few bright lines, mainly in the yellow-green and green, were faintly seen.

The two tubes were then examined separately in detail. No. 2, excited by the magnet, showed curious effects. The glow was rendered weak and intermittent, and the rosy tint almost disappeared. The capillary changed to a decided green colour, and the positive electrode was surrounded by a yellow glow. The changes in the spectrum were no less decided. Without the magnet, the spectrum was found to be a bright continuous one of H (with a full set of principal and intermediate lines) and N—the N spectrum being rather faint and misty, with very slight, if any, traces of the iodine-spectrum. On the magnet being excited, the spectrum changed as if by magic; the H and N spectra disappeared (except hydrogen F, which still faintly remained), and the iodine lines, mostly in the yellow-green and green, shone out wonderfully sharp and bright on quite a dark ground. No. 1, upon examination, showed between the magnet-poles only the same changes as on last occasion. The spectrum seemed to be one of iodine, with the addition of slight traces of the H spectrum.

On excitation of the magnet, the misty continuous part of the spectrum nearly disappeared, and the bright lines shone up sharply upon the dark background as before. The effects in the case of both tubes were strongly marked. The impression as to tube No. 2 was that, without the magnet, the slight iodine-spectrum was overpowered and masked by the N and H spectra; while under the influence of the magnet the N and H spectra were almost altogether suppressed, the iodine-spectrum being at the same time intensified. The disappearance of the continuous spectrum under the action of the magnet in No. 1 (with the supposition it was mainly H) would be accounted for in the same way.

Bromine-tubes.

This tube (No. 1) had been previously worked for photographic purposes. Excited by the small coil, the whole tube was filled with a faint flickering light. The positive bulb contained a faint purple glow, with a yellow-green tinge at the electrode, a curious flickering stream of light flashing from the electrode to the side of the tube. The negative pole showed pretty much the same effect as the positive. The capillary stream expanded at the opening into the positive bulb, but ran in a condensed stream into the negative bulb. In colour it was of a rather bright lilac. Upon putting the magnet on, the light-glow in the tube was at once and permanently extinguished, the coil still working as if the current passed. The same effect happened repeatedly; but now and then the tube lighted-up for a second, showing spiral arrangement in the bulb. We tried another bromine-tube (No. 2): it lighted-up easily; both bulbs were filled with a purple stream of light; capillary stream bright grey. The glass of the tube was strongly fluorescent and of a yellow tinge. When the magnet was excited the stream of light was somewhat condensed in the bulbs, and flew to the side of the tube; while the capillary stream at the same time brightened. The spectrum without the magnet was fairly bright; it increased in brightness under the influence of the magnet, and additional lines appeared; but we considered them to be only faint existing ones brightened up. No change in the position of the principal lines was traced.

Silicic-Fluoride tubes.

(1) A tube marked Si Fl₆ had been worked for photographic purposes; it lighted-up easily. Both bulbs were filled with a brown-pink diffused light, inclined to condense into a stream in the positive bulb. The violet glow was very bright, and nearly filled the space round the electrode. The capillary stream was of a bright violet tint. The effect of the magnet was to decrease the intensity of the light throughout the whole tube.

In the positive bulb the stream broke up into a number of vibrating streamlets, with little bright threads of light intermixed, which flew towards the side of the tube at right angles to the magnetic poles. There was an inclination to spiral arrangement in the streamlets. This stream changed from side to side of the tube coincidently with change in the magnetic poles. At the negative pole the violet glow formed an arc in the direction of the magnetic curves, while a spiral of fainter (positive?) light was formed in the upper part of the bulb. A slight ringing sound was heard in the tube.

(2) We compared two tubes (Si F₄ and the one marked Si Fl₆). The Si Fl₆ tube in general effect, and in its spectrum, when lighted-up, resembled Si F₄. We compared the one tube under the influence of the magnet with the other not so, by means of a comparison-prism on the slit. As the spectroscope and second tube were necessarily removed some distance from the magnet, the spectrum of the tube between the poles was not bright. We could not trace a change of position in any of the principal lines. The tube between the poles was brightened up when the magnet was in action[15].

Sulphuric-Acid (SO₃) tubes.

(1) Excited by the small coil, both bulbs of this tube (No. 1) lighted-up brightly, with a misty light-blue tinted stream of opaque light, a yellow glow appearing at the negative pole. The capillary stream partook of the same blue tint, but was whiter and brighter. Under the magnet’s influence, the glow in the bulbs flew to the side of the tube in flickering streams of light, the capillary at the same time changing to a distinctly green tint. The spectrum without the magnet consisted of four fairly bright bands of light in the yellow, green, blue, and violet, connected by a faint misty continuous spectrum (O or possibly the hydrocarbon spectrum found in O tubes by way of impurity).

When the magnet was excited, this spectrum entirely disappeared; and a set of bright metallic-looking lines upon a dark background (line-spectrum of S) took its place. This effect was produced whenever the magnet was excited, and we tried it several times, to make sure of the complete change. After a time, when the magnet, battery, and the coil-power were all weaker, with the magnet on, we obtained a compound of both spectra, the bright lines being seen upon the continuous spectrum in which the bands appeared. When the magnet was taken off, the bright lines disappeared, and the O spectrum alone remained.