CHAPTER X.

SPECTROSCOPE ADAPTED FOR THE AURORA.

[Sidenote: Must be of moderate dispersion, with ready mode of measuring line-positions.]

Any form of spectroscope of moderate dispersion will suffice for observations of the spectrum of the Aurora; but, for sake of convenience, a hand or direct-vision spectroscope is to be preferred, and it is desirable also to have some quick and ready mode of measuring the position of the lines while the Aurora lasts.

[Sidenote: Mr. Browning’s instrument described.]

Mr. John Browning arranged for me a form of instrument which I have found very convenient for observations by hand of the Aurora-lines, and also, when fixed on a stand, for tube and chemical investigations. A representation of this instrument is given on Plate X. fig. 1. A brass tube carries a large compound (5) direct-vision prism (shown dark in the drawing). An arrangement is made so that a second prism can at will be slipped into the tube (shown in outline in the drawing). With one prism and a fine slit the D lines are widely separated, and the field of view extends at one glance from near C to near G. When the second prism is inserted and used in combination, the nickel line can be seen between the two D lines, and the instrument may be used for solar work. A photograph of the sun’s spectrum, taken with one prism only, shows a great number of the dark solar lines and many of the bright ones, ascribed by Prof. Draper to oxygen and nitrogen.

[Sidenote: Diaphragm micrometer described. Mode of use of the micrometer.]

The collimator and observing telescope are respectively 6 inches in length, and carry achromatic lenses of one inch aperture. The telescope traverses the field so that the extremities of the spectrum may be observed. The dispersion of the instrument was ascertained by a set of observations of the principal solar and some metallic lines, made with an excellent filar micrometer. For the Auroral observations, Dr. Vogel has described an instrument (see Appendix E) in which the usual spider’s-web wires are replaced by a needle-point, as being easily seen upon a faint spectrum. Illuminated wires may also be used; but I was led ultimately to employ, in preference, a diaphragm micrometer which the spectrum itself illuminates, as being adapted for speedy, yet fairly accurate, observations. It was made in this manner:—A card was first of all prepared (Plate X. fig. 2), and within a circle described on this, a scale was drawn of moderately wide white spaces, with black divisions between, short and long, so as to read off easily by eye. The upper half of the circle was then entirely filled in with black; and from the card as thus prepared a reduced negative photograph was made. In this the spaces and lower half of the circle were opaque, and the upper half of the circle and the lines between the spaces were transparent (Plate X. fig. 4). This photograph was about the size of a shilling (fig. 3, same Plate). It was mounted carefully in Canada balsam, with a thin glass cover, and then placed in the focus of the eyepiece. In use, the spectrum is brought upon the scale so that the upper half shows above the scale without any interruption at all; while the lower half illuminates the scale and renders the divisions visible, showing the spectrum-lines falling either upon them or the spaces between. The photographed scale was next enlarged to a considerable size and printed upon faintly ruled paper; and the enlargement was so arranged as to comprise five of the faint ruled lines between each division of the scale. Each of these faint lines in turn represented a certain portion of the spectrum as read off with the filar micrometer; so that the scales as constructed with the filar micrometer and with the photographed micrometer corresponded for all parts of the spectrum included in the field of the eyepiece.

[Sidenote: Advantage of the method.]

One of the photographed enlargements being laid on the table under the spectroscope, the observed lines were marked off with ease and accuracy upon it; and as the photograph was an exact copy of the scale, any want of exactitude in the divisions was of no moment.

One great advantage of this method was, that all the lines seen could be recorded at one time and with all in view, and without the risk of slight shift in the instrument, which frequently happens when lines are read off seriatim.

I found this plan most effective for the rapid and correct recording of a faint and evanescent spectrum, and it gave close results when compared with traversing-micrometer measured spectra. The records, too, admitted of subsequent examination at leisure.

[Sidenote: Double-slit plate arrangement.]

Mr. Browning subsequently constructed for me a double-slit plate (lately in the Scientific Loan Collection at South Kensington) for the same instrument (Plate X. fig. 5). The lower half of the plate is fixed. The upper half traverses the lower by the aid of a micrometer-screw. The slit is widened or closed at pleasure by loosening the small screws by which the jaw-plates are attached. A scale is engraved on the fixed lower half of the plate for an approximate measurement, while the division of the micrometer-screw-head completes it.

In use, one half of the spectrum slides along the other, and a bright line in the upper spectrum is selected as an index. The distances between the lines of the lower half of the spectrum are read off by means of the bright line above. This form of micrometer was suggested by Mr. Procter (in ‘Nature’) many years ago as a substitute for a more complicated apparatus by Zöllner. Other instruments on a similar principle have been lately introduced, but for Aurora purposes I prefer a fixed scale.

[Sidenote: Photographed spectrum suggested.]

In ‘Photographed Spectra’ I have pointed out that we shall probably obtain no spectrum of the Aurora to be absolutely depended upon for comparison with other spectra until we succeed in a photographed one. From experiments made with a special prism of the Rutherfurd form, constructed for me by Mr. Browning (with which many gas-spectra have been already photographed), I see no reason, should an unusually bright Aurora favour us with a visit, why its spectrum may not be recorded in a permanent form, and with lines sufficiently well marked to be compared with other spectra. Rapid dry plates would be especially useful for such a purpose, and some Auroræ, if wanting in brilliancy, would doubtless compensate by their period of endurance.

[Sidenote: Mr. Hilger’s half-prism spectroscope.]

Mr. Adam Hilger has also made for me one of his “half-prism” spectroscopes, in which considerable dispersion is obtained with but very little loss of light. This instrument has a simple and rapid micrometer arrangement, with a bright line as an index. I have (for want of Auroræ) had no opportunity of trying it, but I doubt not it is well adapted for such a purpose.

_Spectrum of the Aurora described._

[Sidenote: Lines or bands and continuous spectrum.]

The spectrum of the Aurora consists of a set of lines or bands upon a dark ground at each extremity of the spectrum, but with more or less of faint continuous spectrum towards the centre. The extreme range of the spectrum, as observed up to the present time, is from “_a_” (between C and D) in the red to “_h_” (hydrogen) in the violet.

[Sidenote: Lines nine in number.]

The lines have been classified and arranged by Lemström and others as nine in number, but I believe not more than seven have ever been seen simultaneously.

The author of the article “Aurora Polaris,” in the ‘Encyclopædia Britannica,’ classes the lines as nine, and gives a table with the following results (to these I have added Herr Vogel’s lines, for the purpose of identification and comparison):—

[Sidenote: Table from Encyc. Brit.]

+------+-------------+------+--------+-------+ |No. of| Number of | Mean |Probable|Vogel’s| |line. |observations.| W.L. | error. |lines. | +------+-------------+------+--------+-------+ | 1. | 5 | 6303 | ± 8·1 | 6297 | | 2. | 10 | 5569 | ± 2·9 | 5569 | | 3. | 4 | 5342 | ±16 | 5390 | | 4. | 6 | 5214 | ± 5·4 | 5233 | | 5. | 4 | 5161 | ± 9·7 | 5189 | | 6. | 6 | 4984 | ±11 | 5004 | | 7. | 4 | 4823 | ± 9·3 | | | 8. | 8 | 4667 | ± 9·8 | 4663 | | 9. | 8 | 4299 | ± 9·3 | | +------+-------------+------+--------+-------+

The probable errors are large, and it is a question whether any thing is gained by thus endeavouring to average the lines.

[Sidenote: Ångström’s line. Zöllner’s line in the red. Other Lines of the spectrum.]

The principal and brightest line, in the yellow-green, is generally called “Ångström’s,” and his (probably the first) measurement of its position at 5567 adopted. This was in the winter of 1867-68, and he saw in addition, by widening the slit, traces of three very feeble bands situated near to F. Zöllner is credited with the first observation of the line in the red. These two lines are generally described as with similar characteristics, and in about the same respective positions, by all observers, and have never been remarked to spread into bands. The other lines in the spectrum are difficult to position, owing to the many discordant observations of them. They seem also variable in intensity as well as in number (sometimes even in the same Aurora), and are not unfrequently observed to have their places supplied by bands.

[Sidenote: Second German expedition observations. Austro-Hungarian.]

The spectroscope was used in the second German expedition, but only the one brightest line seen—Dr. Börgen stating he had never seen a trace of the weak lines in the blue and red, which were observed so distinctly with the same spectroscope on 25th October, 1870, after the return of the expedition. Lieutenant Weyprecht used a small spectroscope during the Austro-Hungarian Expedition, and saw only the well-known yellow-green line.

[Sidenote: Swedish expedition, 1868. Lemström’s observations.]

In the Swedish Expedition, 1868, Lemström mentions that in the Aurora spectrum there are nine lines (he does not say he saw them simultaneously), which he considers to agree with lines belonging to the air-gases. He also thinks the Aurora could be referred to three distinct types, depending on the character of the discharge.

[Sidenote: Spectrum or Aurora seen at Tronsa.]

At Tronsa an Aurora was seen October 21st, 1868, which commenced in the north and became very brilliant. The spectroscope showed:—

1. A yellow line at 74·9.

2. A very clear line in the blue at 65·90.

3. Two lines of hair’s breadth, with very pronounced (horizontal?) striæ on the side of the yellow, one at 125 and the other about 105.

[I presume the striæ were really vertical, and that the explanation intended to convey that these lines shaded off towards the yellow. From a comparison of the figures they must have been in the red, and are the only instance recorded of two auroral lines in that region. They are subsequently spoken of as “shaded rays.”—J. R. C.]

[Sidenote: MM. Wijkander and Parent’s observations.]

M. Auguste Wijkander and Lieut. Parent, of the Swedish Expedition in 1872-73, under Professor Nordenskiöld, used a direct-vision spectroscope, with a micrometer-screw movement of the prisms, the reading being afterwards reduced to wave-lengths upon Ångström’s line-values.

The following Table gives the results, with Dr. Vogel’s lines added for the sake of comparison:—

+------+------------------------+-----------------------+-------+------+ | |Observations, Wijkander.| Observations, Parent. | | | | |-------+-----+----------+-------+-----+---------|Mean of| | |Lines.| | | Probable | | |Probable | both. |Vogel.| | |Number.|W.L. | error. |Number.|W.L. | error. | | | +------+-------+-----+----------+-------+-----+---------+-------+------+ | .. | .. | .. | .. | .. | .. | .. | .. | 6297 | | .. | .. | .. | .. | .. | .. | .. | .. | 5569 | | (1) | 5 | 5359| ±3 | .. | .. | .. | 5359 | 5390 | | (2) | 6 | 5289| ±5 | 3 | 5280| ± 1 | 5286 | .. | | (3) | 6 | 5239| ±4 | 2 | 5207| ±11 | 5231 | 5233 | | .. | .. | .. | .. | .. | .. | .. | .. | 5189 | | (4) | 5 | 4996| ±9 | .. | .. | .. | 4996 | 5004 | | (5) | 1 | 4871| .. | 1 | 4873| .. | 4872 | .. | | (6) | 8 | 4692| ±2 | 10 | 4708| ± 5 | 4701 | 4663 | | (7) | 1 | 4366| .. | .. | .. | .. | 4366 | .. | | (8) | 4 | 4280| ±3 | 3 | 4286| ±16 | 4284 | .. | +------+-------+-----+----------+-------+-----+---------+-------+------+

The brightest line in all Auroræ, 5567, was intentionally not included in the Tables. The red line was not seen. Nos. 5 and 7 were only seen once, and not in the same Aurora.

[Sidenote: Spectrum of Aurora of October 24th, 1870.]

The Aurora of October 24th, 1870, came at a time when spectroscopes of a direct-vision form were being introduced, and a number of observations were communicated at the time to ‘Nature.’

[Sidenote: T. F.’s observations. W. B. Gibbs’s observation. Elger’s observation.]

A correspondent, T. F., writing from Torquay, saw, with a direct-vision spectroscope, one strong red line near C, one strong pale yellow line near D, one paler near F, and a still paler one beyond, with a faint continuous spectrum from about D to beyond F. The C line was very conspicuous and the brightest of the whole. It was intermediate in position and colour to the red lines of the lithium and calcium spectra. Plainly there were two spectra superposed, for while the red portions of the Aurora showed the four lines with a faint continuous spectrum, the greenish portions showed only one line near D on a faint ground. W. B. Gibbs saw, in London, only two bright lines, one a greenish grey, situate about the middle of the spectrum, and the other a red line very much like C (hydrogen). Thomas G. Elger, at Bedford, on the 24th and 25th, saw:—(1) a broad and well-defined red band near C; (2) a bright white band near D (same as Ångström’s W.L. 5567), on 25th visible in every part of the sky; (3) a faint and rather nebulous line, roughly estimated to be near F; (4) a very faint line about halfway between 2 and 3. The red band was absent from the spectrum of the white rays of the Aurora, but the other lines were seen.

[Sidenote: J. R. Capron’s observation.]

With a small Browning direct-vision spectroscope on the 24th, I found no continuous spectrum, but two bright lines, one in the green (like that from the nebulæ, but more intense, and considerably flickering), the other in the red (like the lithium line, but rather duskier: Plate V. fig. 6). The latter was only well seen when the display was at its height; it could, however, be faintly traced wherever the rose tint of the Aurora extended. The line in the green was well seen in all parts of the sky, but was specially bright in the Auroral patches of white light.

[Sidenote: Mr. Browning’s observation. Alvan Clarke’s, jun., observations.]

Mr. Browning also saw the red line, but found comparison difficult. On the evening of the 24th October, Mr. Alvan Clarke, jun., at Boston, used a chemical spectroscope of the ordinary form, with one prism and a photographed scale illuminated with a lamp. Four Auroral lines were seen at points of his scale numbered 61, 68, 80, and 98. These were reduced to wave-lengths by Professor Pickering, with the following results:—

+-----+---------+--------+-------+-------------------+--------+ |Line.| Reading | Wave- |Assumed| Comments. |Probable| | |on scale.|lengths.| line. | | error. | |-----+---------+--------+-------+-------------------+--------| | 1. | 61 | 5690 | 5570 |Common Aurora-line.| -20 | | 2. | 68 | 5320 | 5316 | Corona line. | + 1 | | 3. | 80 | 4850 | 4860 | F, hydrogen. | - 3 | | 4. | 98 | 4350 | 4340 | G, hydrogen. | + 6 | +-----+---------+--------+-------+-------------------+--------+

[61 is evidently wrong, and was probably a mistake for 63.]

[Sidenote: G. F. Barker’s observations.]

George F. Barker, observing at New Haven (U.S.A.), saw, on November 9th, a crimson and white Aurora, which he examined with a single glass-prism spectroscope, by Duboscq, of Paris. The line positions were obtained by an illuminated millimetre scale. In the white Aurora were four lines (the red one being absent); in the red Aurora five. The wave-lengths of the Aurora-lines were run out as follows:—

(1.) Between C and D, 6230 (Zöllner’s 6270). (2.) ” D and E, 5620 (Ångström’s 5570). (3.) ” E and _b_, 5170 (Winlock’s 5200). (4.) ” _b_ and F, 5020. (5.) ” F and G, 4820 (Alvan Clarke’s, jun., 4850).

[Sidenote: Spectrum of Aurora of Feb. 4, 1872. Prof. Piazzi Smyth’s observations.]

Mr. Procter’s Aurora-lines will be found noticed in connexion with the spectrum of oxygen; and Lord Lindsay’s lines, with a comparison scale drawing, are separately described further on in this Chapter. The Aurora of February 4th, 1872, had many observers; some of whom communicated at the time spectroscopic notes. Professor Piazzi Smyth minutely describes the display as seen in Edinburgh, and saw “Ångström’s green Aurora-line perpetually over citron acetylene[13] at W.L. 5579, and the red Aurora-line between lithium _a_ and sodium _a_, but nearer to the latter, say at W.L. 6370.” Extremely faint greenish and bluish lines also appeared at W.L. 5300, 5100, and 4900 nearly.

[Sidenote: Rev. T. W. Webb’s observations.]

The Rev. T. W. Webb, with a very fine slit, saw the green Auroral line even in the light reflected from white paper. With a wider slit he saw a crimson band in the brighter patches of that hue, and beyond an extent of greenish or bluish light, which he suspected to be composed of contiguous bands.

[Sidenote: R. J. Friswell’s observations.]

R. J. Friswell, coming up the Channel at 9.40, with a Hoffman’s direct-vision spectroscope (the observing telescope removed), saw the green line, a crimson line near C, and faint traces of structure in the blue and violet.

[Sidenote: The Rev. S. J. Perry’s observations.]

The Rev. S. J. Perry observed at Stonyhurst four lines, and, on examining one of the curved streamers, found the red line even more strongly marked than the green. A magnetic storm was observed to be at its height from 4 to 9 P.M. of the same day.

[Sidenote: J. R. Capron’s observations.]

With a Browning 7-prism direct-vision spectroscope I saw the green line in all parts of the Aurora, attended with a peculiar flickering movement. I did not see the other lines.

[Sidenote: His catalogue of lines up to Nov. 9, 1872.]

In a letter to ‘Nature,’ dated November 9th, 1872, I catalogued the lines observed up to that date as follows:—

1. A line in the red between C and D. W.L., Ångström, 6279.

2. A line (the principal one of the Aurora) in the yellow-green, between D and E. W.L., Ångström, 5567.

3. A line in the green, near E (corona line?). W.L., Alvan Clarke, jun., and Backhouse, 5320.

4. A faint line in the green, at or near _b_. W.L., Barker, 5170.

5. A faint line or band in the green, between _b_ and F. W.L., Barker, 5020 (chromospheric?).

6. A line in the green-blue, at or near F. W.L., Alvan Clarke, jun., 4850.

7. A line in the indigo, at or near G. W.L., Alvan Clarke, jun., 4350.

8. The continuous spectrum from about D to beyond F.

[Sidenote: Dr. H.C. Vogel’s observations of Auroral lines. Spectrum described.]

Dr. H. C. Vogel, formerly of the Bothkamp Observatory, near Kiel, and since of the Astrophysical Observatory, Potsdam, made several observations of the Auroral lines, October 25th, 1870. Besides the bright line between D and E, he found several other fainter lines stretching towards the blue end of the spectrum on a dimly-lighted ground. February 11th, 1871, he observed the same set of lines, and an average of six readings gave 5572 as the W.L. of the Ångström line. February 12th gave 5576 as Dr. Vogel’s reading, and 5569 as Dr. Lohse’s. April 9th gave 5569, and April 14th 5569. The Aurora of April 9th, 1871, was exceedingly brilliant, so that micrometer measurements of the lines were taken. The spectrum consisted of one line in the red, five in the green, and a somewhat indistinct broad line or band in the blue. The lines are thus described:—

[Sidenote: Table of lines.]

Table of Dr. Vogel’s lines. Aurora, April 9th, 1871.

+-----+--------+------------------------------------------------+ |W.L. |Probable| Remarks. | | | error. | | +-----+--------+------------------------------------------------+ |6297 | 14 |Very bright stripe. } | | | | } | |5569 | 2 |Brightest line of the spectrum, } | | | | became noticeably fainter at } On a faintly | | | | appearance of the red line. } lighted | | | | } ground. | |5390 | .. |Extremely faint line; } | | | | unreliable observation. } | | | | } | |5233 | 4 |Moderately bright. } | | | | | |5189 | 9 |This line was very bright when the red line | | | | appeared at the same time; otherwise equal | | | | in brilliancy with the preceding one. | | | | | |5004 | 3 |Very bright line. | | | | | |4694}| |{Broad band of light, somewhat less brilliant | |4663}| 3 |{ in the middle; very faint in those parts of | |4629}| |{ the Aurora in which the red line appeared. | +-----+--------+------------------------------------------------+

A translation of Dr. Vogel’s interesting paper will be found printed _in extenso_ in Appendix E, and his lithographed drawings of the spectrum in the green and red portions of the Aurora respectively on Plate VI. figs. 2 and 3. The observations of April 9th by Dr. Vogel are probably, up to the present time, the most exact of any one Aurora, and I have therefore in most cases used them for comparison.

[Sidenote: Mr. Backhouse’s catalogue of lines.]

Mr. Backhouse, in a letter to ‘Nature,’ commenting upon my catalogue of lines, gave the following as the latest determinations from his own observations:—

No. 1. Wave-length 6060 2. ” 5660 3. ” 5165 4. ” 5015 6. ” 4625 7. ” 4305

(6060 must be a mistake for 6260, and 5660 for 5560.—J. R. C.) Mr. Backhouse never saw a line at 5320 again. He found the continuous spectrum to reach from No. 2 to No. 7, being brightest from a little beyond No. 2 to No. 6. This part of the spectrum did not give him so much the idea of a true “continuous spectrum” as of a series of bright bands too close to be distinguished.

[Sidenote: Subsequent full catalogue of Auroral lines.]

I have subsequently, in another section of this Chapter, added a full catalogue of the Auroral lines, prepared by myself from the foregoing and other sources and observations; and I also append to it a Plate [Plate XII.], in which these lines are positioned and the wave-lengths and names of observers are given. The numbers of the lines on the Plate correspond with those in the catalogue. The solar spectrum and the spectrum of the blue base of a candle-flame are added for purposes of comparison. [The telluric bands in the solar spectrum are shown more distinctly than they actually appear, and do not profess to give details.]

_Flickering of the Green Line._

[Sidenote: Flickering of the green line. Herschel’s observation. J. R. Capron’s observation.]

A. S. Herschel noticed this, April 9th (1871?). He says:—“A remarkable circumstance connected with the appearance of the single line observed on this occasion was the flickering and frequent changes with which it rose and fell in brightness; apparently even more rapidly than the swiftly travelling waves, or pulsations of light, that repeatedly passed over the streamers, near the northern horizon, towards which the spectroscope was directed.” In the spectrum of the Aurora of 20th October, 1870, I saw and noted the green line as “considerably flickering;” and in the Aurora of 4th February, 1872, I again saw and noted “the peculiar flickering” I had remarked in 1870. I have not seen the peculiarity noted by other observers.

_Mr. Backhouse’s graphical Spectra of four Auroræ._

[Sidenote: Mr. Backhouse’s graphical spectra of Auroræ.]

Mr. Backhouse has been good enough to supply me with some details of four several Auroræ seen by him at Sunderland, accompanied by drawings, showing in a graphical way the spectrum of each display as seen with a spectroscope with rather a wide slit and as drawn by eye. I have reduced the four drawings to the same scale, and in this way they are extremely interesting for comparison (Plate V. fig. 4). The line on the left in each spectrum is Ångström’s bright Auroral line, and is supposed to be considerably prolonged. The height of the lines denotes intensity.

[Sidenote: April 18, 1873.]

April 18th, 1873, was a bright Aurora. No. 3 is a faint band, which Mr. Backhouse had not perceived before. No. 5 had not been visible lately, and Mr. Backhouse thought it must belong to Auroræ of a different type from those which had appeared latterly.

[Sidenote: Feb. 4, 1874.]

February 4th, 1874. In the spectrum of this Aurora Mr. Backhouse saw seven lines, all that he had ever seen. (The red line, not shown in the diagram, makes the seventh.)

The spectrum is represented as seen between 6.50 and 7.5 P.M. Mr. Backhouse had only once before seen No. 4, and it became quite invisible between 7.45 and 7.55, though the other lines were as bright as before and the red line had appeared.

[Sidenote: Oct. 3, 1874.]

October 3rd, 1874. This spectrum was examined, and diagram made between 10 and 10.25 P.M. Five lines only are indicated.

It is mainly distinguished from the two preceding spectra by the brightness of the continuous spectrum on which the lines 2, 3, and 4 lie, and by the weakness of No. 5.

[Sidenote: Oct. 4, 1874.]

October 4th, 1874. Taken between 11.10 and 11.20 P.M.; distinguished, like the last, by a considerable amount of continuous spectrum and by a faint line (No. 3), not seen in the last spectrum, while No. 3 in the last is missing in this spectrum.

[Sidenote: Mr. Backhouse’s remarks as to comparative frequency of some of the Auroral lines.]

Mr. Backhouse, as to both these last spectra, remarks that the lines were very variable in intensity, and sometimes some were visible and sometimes others. They varied also in relative brightness in different parts of the sky at the same time. Mr. Backhouse, in a communication to ‘Nature,’ referring to a statement of Mr. Procter’s, that the bands of the Auroral spectrum are seldom visible, except the bright line at 5570, says that he always found two bands, “doubtless Winlock’s 4640 and 4310,” to be invariably visible when the Aurora was bright enough to show them. Of thirty-four Auroræ examined by Mr. Backhouse, fourteen showed the lines 4640 and 4310, and three others at least one of these, while eight showed the red line. (Ångström only once saw this line.) In five Auroræ, all more or less red, he saw a faint band, the wave-length of which he placed at 5000 or 5100. He never saw the line 5320 (also Winlock’s coronal line), unless it were once, probably from want of instrumental power. With regard to these observations, I may say that with a Browning’s miniature spectroscope I saw only two lines (the red and the green) in the grand display of the 24th October, 1870; and with an instrument of larger aperture the green line only on the 4th February, 1872; while I saw the green line and three others towards the violet with the same instrument during the Aurora of 4th February, 1874. (See description of this Aurora, _antè_ p. 21, and drawing of spectrum, Plate VI. fig. 1 _a_.)

_Lord Lindsay’s Aurora-Spectrum, 21st October, 1870._

[Sidenote: Lord Lindsay’s Aurora of 21st Oct., 1870.]

Lord Lindsay observed a fine Aurora at the Observatory at Dun Echt on the night of the 21st October, 1870. It commenced about 9.30, reached its maximum about 11, and faded away suddenly about 11.30 P.M.

[Sidenote: Spectrum described.]

A spectrum obtained in the north-west gave five bright lines with a Browning’s direct-vision spectroscope—two strong, one medium, two very faint. A tallow candle was used to obtain a comparison spectrum of sodium and carburetted hydrogen.

A drawing of the spectrum obtained is given on Plate XI. fig. 2. No. 1 is a sharp well-formed line visible with a narrow slit.

No. 2, a line very slightly more refrangible than F. The side towards D is sharp and well defined, while on the other side it is nebulous.

No. 3, slightly less refrangible than G, is a broad ill-defined band, seen only with a wide slit.

No. 4, a line near E, woolly at the edges, but rather sharp in the centre. This, says Lord Lindsay, should be at or near the position of the line 1474 of the solar corona.

No. 5, a faint band, coincident with _b_, extending equally on both sides of it.

The lines are numbered in order of intensity. It is questionable, from observations with instruments carrying a scale, whether the line-positions are exact; but the description of their characters is valuable.

[Sidenote: Candle-spectrum.]

As a candle blue-base spectrum is at times a ready and handy mode of reference in Auroral observations (as was found in this instance), I have, on Plate XI. fig. 5, given a representation of it as seen with my Auroral spectroscope. Dr. Watts’s corresponding carbon-spectrum is added on the lower margin. The numbers on the upper margin refer to my scale.

_Spectrum of the Aurora Australis._

[Sidenote: Captain Maclear’s spectra of Aurora Australis.]

Captain Maclear, on examining the streamer seen by him Feb. 9th, 1874 (_antè_, p. 27), with the spectroscope, found three prominent lines in the yellow-green, green, and blue or purple, but not the red line. In the Aurora of March 3rd, 1874 (p. 27), he could trace four lines, three bright and one rather faint. They must have been exceedingly bright to show so plainly in full-moon light.

[Sidenote: Instrument used, and mode of registering lines.]

The spectroscope used was a Grubb single-prism with long collimator. A needle-point in the eyepiece marked the position of the lines; and a corresponding needle-point, carried on a frame by a screw movement in concord with the point of the eyepiece, scratched the lines on a plate of blackened glass. Two plates were taken. On the first were scratched the auroral lines and the solar lines as seen in the moonlight; on the second plate were scratched the auroral lines, the Solar lines from the moon, and the carbon lines in a spirit-lamp.

[Sidenote: Copies of the two spectra obtained. Discrepancy in the spectra. Remarks on the spectra.]

The next morning the solar lines were verified in sunlight. I subjoin (Plate XI. fig. 3) copies of the two spectra as printed in ‘Nature,’ the auroral lines being marked A, the solar lines by the usual designating letters, and the carbon by _Car_. To these spectra I have added for comparison Dr. Vogel’s spectrum of the Aurora Borealis. Captain Maclear could not account for the different positions of the auroral lines in the two plates; for the prism, as far as he was aware, was not moved during the observations. As the solar lines are indicated in the same place in both spectra, the case would seem one of actual change of position of the auroral lines during observation. A comparison of the two spectra gives the impression that the lower one is the same as the upper, except that the dispersion is greater, the lines remaining relatively in position. One does not, however, see how the dispersion could have so varied in a single-prism instrument, and the position of the solar lines is adverse to such an explanation.

There is a suspicion that Auroræ are not always identical in position of some of the lines; but the line in the green (considerably out of place in the lower Australis spectrum) has always, within small limits, the same position. It will be noticed how much further the Australis spectrum runs into the violet than Vogel’s Borealis, the latter having no lines much beyond F.

The faint line (No. 2) mentioned by Captain Maclear possibly corresponds with Dr. Vogel’s band. The absence of the four lines of the Aurora Borealis in the green part of the spectrum of the Australis is peculiar; and in this respect, too, the two Australis spectra agree.

[Sidenote: Comparison of the lines.]

The nearest approaches to Captain Maclear’s lines (of the upper spectrum) which I can find are:—

Line Corresponding line.

1. 5567, Ångström. 2. (The faint line.) Vogel’s band, 4694-4629. 3. I find no approximately corresponding line. 4. 4350, Alvan Clarke.

But the comparisons are not by any means close. Further observations of the Australis spectrum are very desirable.

_Prof. Piazzi Smyth’s Aurora-Spectra._

[Sidenote: Prof. Piazzi Smyth’s chemical and auroral spectra.]

Prof. Piazzi Smyth, in volume xiv. of the ‘Edinburgh Astronomical Observations,’ 1870-77, has compared simultaneously the Aurora-spectrum with the sets of bright lines seen in the blue base of flame—the lines of potassium, lithium, sodium, thallium, and indium being also introduced for comparison. The spectra are drawn as seen under small dispersion, and will prove most useful in cases where an Aurora is not bright enough to admit of the lines being measured by micrometer, and the eye and comparison spectrum are obliged to be resorted to.

_Author’s Catalogue of the Auroral Lines._

(See Plate XII.)

1. W.L. 6297, Vogel. Very bright stripe; first noticed by Zöllner. Seen only in red Auroræ; stands out on a dark ground, without other lines near it. Character of line sharp and well defined; varies in colour from dusky red to bright crimson. Intensity, Herschel, 0 to 4 or 8. According to same, position coincident with atmospheric absorption-group “_a_” in solar spectrum (between C and D). I confirm this position according to my scale of solar lines, and a drawing of the coincidence (in which, and in Plate XII., the absorption-lines are drawn too dark) is given on Plate XIII. fig. 2. Herschel says this line coincides with a red band in the negative glow-discharge, but its identity is doubtful. Its isolation and want of adjacent lines seem to separate it from the air-spectrum and gas-spectra in general. At the appearance of this line, 5569 (No. 2) becomes noticeably fainter. When this line is bright, 5189 (No. 5) is bright also (Vogel).

If we propose to assign to this line, as well as to 5569, a phosphorescent origin, it would be strongly confirmatory of such a theory (in connexion with the phosphoretted-hydrogen spectrum) to find it brighten at low temperatures.

_Note._—Sir John Franklin says, in his ‘Polar Expeditions,’ that a low state of temperature is favourable for the production of brilliant coruscations. It was seldom witnessed that the Auroræ were much agitated, or that the prismatic tints were very apparent, when the temperature was above zero.

2. Line in the yellow-green. Brightest of all lines in the Aurora-spectrum. W.L. 5567, Ångström; 5569, Vogel. Intensity 25, Herschel. To me more pale green than yellow, sometimes flickering and changing in brightness (Herschel and Capron). Seen in all Auroræ usually sharp and bright, but Procter has once recorded it nebulous. Its character as to width, sharpness, and intensity, if carefully observed, might indicate height and structure of Aurora. Becomes noticeably fainter at appearance of red line (Vogel). Found by me to correspond in position with a faint atmospheric absorption-band (see Plate XIII. fig. 2). According to Ångström and Herschel, arising from a phosphorescent and fluorescent light, emitted when air is subjected to the action of electrical discharge.

3. Line in green near last. W.L. 5390. An extremely faint and unreliable observation (Vogel). Seen only by him, unless Alvan Clarke’s 5320 (coronal?) be the same.

4. Line in green-blue. W.L. 5233, moderately bright (Vogel); 5200, Winlock. Intensity, 2 or 0? to 6, Herschel. Coincides with line in the negative glow according to same. Frequently observed.

5. Line in green-blue. W.L. 5189. This line is very bright when the red line appears at the same time; otherwise equal in brilliancy with No. 3 (Vogel); Winlock, 5200. Not so frequently observed as No. 3. Barker gives a band extending from 5330 to 5200. Intensity of 5189, 0 to 8, Herschel, who considers it coincident with a constant strong line in the spark-discharge.

6. Line in blue. W.L. 5004. Very bright line, Vogel; 5020, Barker (coronal?). Intensity 2 or 0? to 8, Herschel. Coincides with line of nitrogen in the nebulæ according to same. Barker gives a band extending from 5050 to 4990.

7. Line in the blue not found by Vogel in Aurora, April 9th, 1871. W.L. 4850, Alvan Clarke; 4820, Backhouse and Barker. Intensity, Herschel, of 4820-4870, 0 to 4? Herschel suspects this and No. 4 to be seen only in Auroral streamers of low elevation. Barker gives a band extending from 4930 to 4850.

8. 4694, 4663, 4629. Broad band of light, somewhat less bright in the middle; very faint in those parts of the Aurora in which the red line appears (Vogel). Intensity 3-6 (Herschel). A double band, consisting of two lines, the first rather more frequently noted than the second in Auroral spectra, agrees well in position with the principal band in the negative glow-spectrum (same). Barker gives a band extending from 4740 to 4670; Backhouse and Winlock give a line at 4640, situate within the same.

9. There seems a good deal of confusion about a fairly bright line (intensity 0-6, Herschel) seen in most Auroræ (not, however, by Vogel, April 9th, 1871), and situate somewhere near G in the solar spectrum. Alvan Clarke places it at 4350, on the less refrangible side of G; Backhouse and Barker at or very near to G; while Lemström and others position it on the more refrangible side of G. Accurate observations, for which a quartz spectroscope might be useful, are much wanted. Herschel makes this line, at 4285, correspond with a strong band in the violet in the negative glow-spectrum.

Herschel also refers to an apparently additional line near the hydrogen-line, or between G and H₁, in the solar spectrum, as mentioned once by Lemström at Helsingfors. I am not aware of any other observation of this line, which must be considerably beyond that at or near G, and would probably be difficult to detect, except in instruments specially adapted for examination of the violet end of the spectrum.

_Theories in relation to the Aurora and its Spectrum._

[Sidenote: Lemström’s.]

Lemström (1):—That the Polar light is caused by an electric current passing from the upper rarefied layers of the air to the earth, producing light-phenomena that do not arise in the denser layers of the air. (2) That there are nine rays (lines or bands) in the Aurora-spectrum, which in all probability agree with lines which belong to the gases of the air. (3) That the Aurora-spectrum can be referred to three distinct types, which depend on the character of the discharge.

[Sidenote: Vogel’s.]

Dr. Vogel:—(1) That the Auroræ are electric discharges in rarefied-air strata of very considerable thickness. (2) That the Aurora-spectrum is a modification of the air-spectrum, involving the question of alteration of the spectrum by conditions of temperature and pressure.

[Sidenote: Ångström’s.]

Ångström:—(1) seems to adopt the hypothesis that the Aurora has its final cause in electrical discharges in the upper strata of the atmosphere, and that these, whether disruptional or continuous, take place sometimes on the outer boundary of the atmosphere, and sometimes near the surface of the earth.

(2) That the Aurora has two different spectra.

(3) That the green line is due to fluorescence or phosphorescence, and that there is no need to resort to Dr. Vogel’s variability of gas-spectra according to circumstances of pressure and temperature.

(4) That an agreement exists between the lines of the Aurora (except the red and green before mentioned) and the lines or bands of the violet light which proceed from the negative pole in dry air.

[Sidenote: Zöllner’s remark as to temperature of Aurora and character of spectrum.]

Zöllner has pointed out that the temperature of the incandescent gas of the Aurora must be exceedingly low, comparatively, and concludes that the spectrum does not correspond with any known spectrum of the atmospheric gases—only because, though a spectrum of our atmosphere, it is one of another order, and one which we cannot produce artificially.