Page images

VII. On the Spectrum of the Aurora. (Extract from a Letter from Professor A. S. HERSCHEL to R. H. Scott, Esq., F.R.S.*)

THE paper by Ångström, noticed in ‘Nature' (vol. x.

No. 246, July 16, 1874), was a rare legacy of his last days to spectroscopists of auroras. Nothing more conclusive and satisfactory, that I know of, has been written on the subject; and little more will be done now, I expect, by future observers than to verify his conclusions and to extend the research in the direction that he points out, in its more numerous details.

The spectrum of the aurora is no doubt in the main the same as that of the pale blue light round the negative pole in an air or nitrogen vacuum-tube, with the induction-spark passing through it. There are so many well marked lines in this spectrum that, looking at Angström's representation of them, it is probably owing to the insignificant appearance of that part of the vacuum-spark that its proper spectrum has not been more frequently studied with reference to the aurora, as Angström seems to have done by an experiment specially adapted for the purpose.

There are several forms or modes (apparently four or five) of electrical discharge through rarefied gases f. When very much rarefied, air transmits the electricity so as to discharge the Rubmkorff poles without a spark. In that state there is still a glow of heated air round both poles, which increases in size and length along the tube as the air-pressure is increased, faster round the positive than round the negative pole. This has been accounted for by showing that the air offers far greater resistance to the passage of electricity when it surrounds a cathode or negative, than when it surrounds an anode or positively electrified pole. The difference becomes more obvious as the pressure and density of the gas are increased. The negative glow shrinks into a very small space, while the positive brush extends through nearly the whole length of the tube, abolishing at last the dark space that at very low tensions separates the two lights from each other. At pressures not exceeding one or two millimetres the positive glow is stratified; but if the pressure is increased it becomes continuous; and if the air-pressure amounts to that of

or an inch of mercury, or upwards, it again gathers into

* Communicated by R. H. Scott, Esq.

† An examination of these with revolving mirrors, by A. Wüllner, at Aix la Chapelle, appeared inothe “ Jubelband” of Poggendorff's Annalen this year at the same time as Angström's paper in that volume, which also contains some other tracts (by A. de la Rive and others) tracing the effects of magnets and of metallic vapours in augmenting the discharge through air.

Phil. Mag. S. 4. Vol. 49. No. 322. Jan. 1875. F

somewhat larger light clouds; and at about 1 or 2 inches of barometric pressure a spark passes between the poles. This spark is red; it scarcely diminishes the strength of the concomitant glow discharge; and it is far less luminous than the white spark which begins to appear at 5 or 6 inches of pressure, and may often be seen at first broken up along its length into parts which are alternately white and red. The spectra, like the general appearances, of these two forms of the spark are quite distinct.

I do not know if these several phases of the positive part of the discharge have all been examined spectroscopically. They pass into each other according to the shape and size of the tube or flask, as well as the air-pressure; and it is difficult to say how much of each is concerned in those observations which have been made of air-spark spectra in comparisons with the aurora.

No one, so far as I know,, has compared with it the negative-glow spectrum so fully as Angström has now done; and it seems very probable that its peculiar fitness for the comparison has been overlooked-the feat of filling a bottle with the negative glow discharge being certainly a novelty; if it is really true that he succeeded in obliterating the positive brush entirely in its favour.

The next remarkable novelty in the paper is the way in which be proposes to account for the “citron” line of auroras ; for there is evidently nothing of the kind in the negative glow, however well that answers to all the secondary facts of faint blue, red, and greenish lines. If oxygen and its compounds are (as has, I believe, been lately shown) strongly fluorescent, Tait and Dewar have also proved, as shown by some of their experiments this year, that they also possess powers of phosphorescence

-Geissler tubes shining for some time after the spark has passed through them, from the production of ozone during the discharge. When one of the globes of a phosphorescent "garland” tube was heated over a Bunsen flame, that globe which was heated did not shine after the spark had passed, apparently because, as we know, a very little heat is sufficient to destroy ozone. Whatever the way may be in which the ozone or otherwise electrified gas remains self-luminous after the discharge, it seems very reasonable to suppose some action of the same kind (perhaps, as Angström says, simply fluorescence) as common in all auroras, and that this produces the well-known auroral line.

Pocket spectroscopes can, of course, do nothing further to fix the position of the citron line; nor can they alone fix very exactly the places of any of the fainter ones. But as every aurora shows this strong monochromatic light, it might be used to bring out a row of punctures transverse to the slit, as a divided scale in the field of view whereby to map the fainter lines, or at any rate

to recognize those which appear most frequently. For this purpose they should be made large, and the slit should be a wide one.

For ordinary miniature spectroscopes, two holes on the red and five on the blue side of the slit tor of an inch apart, would suffice for recognitions and even for very useful measurements. The jaws of the slit can be cut with a fine saw across the middle about to or ģ of an inch deep each way; and a piece of copper foil, provided with the row of holes and a sufficiently wide slit across it, can be fastened to one of them inside, opposite to the crosscut and adjoining the edges of the spectroscope jaws.

Some other means may be found of piercing the jaws of a pocket spectroscope at regular intervals; but as a simple plan I have found this very efficient, in finer divisions, for laboratory use. The holes are pierced at tto of an inch apart; and thirty of them include the whole visible spectrum. Sodium-light, which is common in laboratory flames, exhibits the punctures with admirable distinctness; and each fifth hole being punched double, the scale is very easily read off. There are ten holes on the red, and twenty on the blue side of the slit. If the mechanical difficulty of perforated jaws could be overcome, nothing perhaps could be better suited for examining auroras than a pocket spectroscope so prepared with a few close but clear and tolerably open holes on each side of the slit.

The secondary auroral lines can only be seen in small spectroscopes) with a pretty broad slit; and the strength of the yellow line might then prove embarrassing. I would abolish it, if so, by a blue* glass nearly covering one half, and a red glass the other half of the slit-the blue and red parts of the spectrum respectively, not in its immediate neighbourhood, being freely transmitted. The slit might also be made longer than usual for auroral study.

I have been bere supposing that special spectroscopes would be provided for Arctic observers. But it is quite certain that much may be done with common pocket spectroscopes without any such provision. They should have adjustable slits and good dispersion, as the secondary lines are faint; and though abundant enough in the blue to make the spectrum there pretty luminous, they can only be individualized by varying the slit-aperture. On the ouly occasion when I have seen this spectrum (in Pebruary 1872) they seemed to run into each other, and presented a light so nearly continuous in the blue part that, although the slit of the Browning's pocket spectroscope which I was using was extremely fine, and was focused on the

* Some care would be necessary in selecting the blue glass, as these generally transmit a yellow ray closely corresponding with the auroral line.

[ocr errors]
[ocr errors]

yellow line, no interruption or appearance of lines could be made out. It was probably also through not opening the slit that I missed seeing a red line which another observer, using a similar instrument and looking with me at the aurora, saw very plainly. Although its red colour was intensely brilliant, I failed to see the slightest trace of light on the red side of the yellow line. Had I opened the slit, or perhaps opened and closed it alternately (as the yellow line, though tine, was still very bright), the result would probably have been different.

I send with this a copy of the Plate of spectra accompanying Ångströni's original paper (see fig. p. 71); and I can confirm the appearance of the negative or "cathode" spectrum which he gives, from the results of some examinations of it which I have lately made. On projecting the recorded lines in wave-lengths, there is a very exact agreement with the chief lines and shadings as figured in the Plate. Some fainter lines, however, are visible, which Ångström has perhaps omitted purposely, to avoid encumbering the drawing. As regards the aurora spectrum, considering the numbers of lines which have been mapped and measured, I am not so well prepared to say how complete Angström's selection of its principal ones may be. It may, I think, be taken for granted that his map includes every line of whose existence, approximate place, and brightness there is really no doubt, and that it does represent the normal spectrum of the aurora when sufficiently bright. During the years 1871 and 1872 there were several résumés of the subject, accompanied with new measurements, in Poggendorff's Annalen and the American Journal of Science' (by Vogel, Barker, and others), with which, from his

his comments on them, Angström was evidently familiar. Adopting his selection, therefore, and the descriptions which he gives, from his own and other observers' notes, of relative brightness and appearance, I have endeavoured to condense the information in the annexed Table as a guide for further observation.

Professor Piazzi Smyth, who has given much attention to the auroral spectrum, bas published, in the introduction to vol. xiii. of the Edinburgh Astronomical Observations,' a set of simple comparison spectra, with notes of desiderata, which would be of great service to observers well furnished with instrumental means and applying them to measurements of the aurora ; and something similar would very much assist observers using directvision spectroscopes to map their spectra, where auroras frequent and of great brightness, in high latitudes.

Hoping that these notes may be of use for the purpose of comparison with occasional observations of auroras,

I remain, &c., Newcastle-on-Tyne, July 24, 1874.

(Signed) A. S. H.


[ocr errors]
[ocr errors]


Positions and General Characters of Principal Lines in the Auroral

Spectrum, according to Ångström.
Red 1. Brightness 0 to 4 or 8. Wave-length 6300.

Relative position. Atmospheric absorptiou-line

a near C in the solar spectrum. General description of source and frequency. Seen

chiefly, if not only, in red auroras; a clearly defined line, sometimes intense ; no other red

line visible. Identification with lines of electrical air-spectra.

Coincides with a red band in the negativeYellow glow discharge. or citron- a. 'Brightness 25. Wave-length 5570. green.

Position. Second separable line in the first or

citron band of blue gas-flame spectrum. Description &c. The inost characteristic auroral

line; constant and conspicuous in all auroras ;

not divisible; sharp and bright. Identification 8c. Not identified; possibly a

phosphorescent or fluorescent light emitted when air is subjected to the action of electrical


3. Brightness 2 or 0? to 6.... Wave-length5225. Greenish

0-8 (increases
with red line)

5170–5190. blue. 3b. 2 or 0? to 8....

5000. 0 to 4 ?

4820–1870. Position. 3 and 3 a closely adjoin the solar line

b and the second or green band of the blue gasflame spectrum. 36 is at (6,F); and the

line or lines 3 c are near F. Description &c. The first three are distinct lines ;

the first most frequently observed; the second and third less commonly; lines in the fourth place (3 c) noted by Alvan Clark, jun., Barker,

and Angströin. Identification &c. 3, 3 b coincide with lines in

the negative glow, 3 b that of nitrogen in the nebulæ ; 3a with a constant strong line in the spark-discharge. The latter and 3 c, it may be, are only seen in auroral streamers of low


4. Brightness 3–6 (fainter Full blue. with red line)..

Wave-lenyth 4663-4740. 4a.



blue or


3 c.

[ocr errors]
« PreviousContinue »