2IO 



NATURE 



\June 28, 1883 



line observed was that named at the head of the column. It is 

 not always possible to identify the lines, and, in some cases, 

 my identifications disagree with J. R. Capron's. 



This table shows eleven distinct and well-separated lines or 

 bands, the existence of all of which may, in my opinion, be 

 considered proved, all but \ having been seen by two or more 

 observers. 



Seven have been seen by numerous observers ; I have myself 

 seen them repeatedly, though none of the other four with cer- 

 tainty ; but these (6, 1, k, and \) have all been seen by trust- 

 worthy observers, and therefore may be accepted, though of 

 course further confirmation would be advantageous. 



I shall now consider each line separately. 



$ is the only line yet seen in the red. It exists in all auroras 

 •that are tinged with red, though the spectroscope does not 

 always show it. I have always seen this line most easily with a 

 single prism ; but often the slit has to be so wide that it appears 

 merely as a red b irder to a. In those reddish auroras where I 

 have not been able to detect it, its invisibility is probably due to 

 moonlight or some such cause. There can be no doubt that this 

 line is the cause of the red colour. It varies greatly in brilliancy 

 with reference to o ; I have seen it as bright as o, but never 

 brighter. 1 It sometimes exists when no redness is perceptible in 

 the aurora, it being overpowered by the other rays. 



a is almost always the brightest line in the spectrum. The 

 only exceptions I have seen were that once $ was as bright, and 

 once e brighter (see below), a probably exists in all true auroras. 

 In a few very faint ones I have not been able to see it, perhaps 

 because it has been overpowered by the diffused light of the 

 spectrum, which certainly varies in brightness relatively to the 

 brightness of the lines. Very often when there is no decided 

 aurora, a luminosity overspreads the sky, uniform in all direc- 

 tions, though fading gradually towards the zenith ; I have several 

 times seen the line a in its spectrum, but at other times it has 

 been invisible, though the light has appeared equally bright. 



a and are slender lines ; it is not clear whether this can be 

 said of any of the other lines or bands. 



8. Wijkander's is the only reliable observation of this line, but 

 Vogel has one not far from the same place, and as he states his 

 position is unreliable, there is no reason to doubt its being the 

 same line. There is more doubt whether Peirce observed the 

 same. 



1 was observed both by Wijkander and Parent, and probably 

 by Peirce and Copeland. 



c consists of two lines, according to Vogel, which I have 

 called e' and e- ; the latter, he says, became very bright when- 

 ever appeared. He is the only observer who describes two 

 lines here (though I have at least once suspected € to be double), 

 and therefore it is difficult to tell which of them other observers 

 have seen, or whether they have seen both combined as one 

 band. In the table I have assumed that the latter has u-ually 

 ■been the case, or at least that a band has been seen in this place ; 

 possibly this band has been different from either. I have there- 

 fore placed the observations which seem to apply to the band, 

 or to the combined lines, in a separate column from those that 

 seem to refer to the individual lines ; but the average of the 

 former includes the observations of the latter. Barker and 

 Procter describe a band here ; also E. B. Kirk, who, in 18S0, 

 August 12, saw it as a band fading towards the violet ; but in 

 1882, August 4, it faded away on both sides, though quite sharp 

 at the edges, and with a pretty narrow slit it was broken up into 

 lines — his impression is there were six or eight ; and that the 

 group was broader than the distance between e' and r\ He 

 used one of Browning's "Maclean star spectroscopes," with an 

 ordinary convex lens instead of the cylindrical one. 



I have carefully examined my observations of this band, to 

 see w hether it has appeared more refrangible when the red line 

 has been visible or bright than when there has been little or 

 no red ; but the result of this examination is inconclusive. 

 However, in 1869, April 16, and 1874, October 3, I noted that 

 « was relatively brighter in the red part of the aurora than else- 

 where, so far confirming Dr. Vogel. 



e exists in nearly all auroras that are bright enough to show 

 any line besides o ; perhaps in all. It varies very much in 

 brightness with respect to the other lines. I have sometimes 

 found it the brightest next to o, and once the brightest of all, 

 viz. in 1882, November 20, between 5.50 and 6.5 a.m. There 



1 In considering ihe relative brightness of the different bands, it must be 

 borne in mind that it varies considerably with the width of the slit, the dis- 

 persion, &c. 



was twilight at the time in addition to the aurora, but I do not see 

 how this could produce the effect. Between 5.40 and 5.50 that 

 morning I estimated a three times as bright as f, which was the 

 second brightest line ; but I see no way of avoiding the conclu- 

 sion that it was the brightest of all a few minutes later. 



C may be >een in most bright auroras. It is sometimes brighter 

 than e. 



7) is much more seldom visible ; but I have several times seen 

 it brighter than either e or ( ; rarely as bright as 7, or brighter 

 than it ; but never so conspicuous as it, as the latter is rendered 

 more visible by its position at the edge of the brighter part of 

 the continuous spectrum. 



7 and S belong, I believe, to all auroras, always being visible 

 when the spectrum is moderately bright. But their brightness 

 varies with respect to each other and to a ; indeed I do not know 

 any two auroral lines that always vary together. 



7 is usually the brightest line next to o, with my mode of 

 observation. Several observers describe it as a band ; Vogel as 

 a double band (if not triple). Capron, on the authority of A. 

 S. Herschel, says it consists of "two lines, the first rather more 

 frequently noted than the second" (the more refrangible). I 

 cannot see that this assertion is borne out by the accompanying 

 table, but if it is correct, the two lines must be about 4700 and 

 4654. I have several times seen 7 as a band fading towards the 

 violet. E. B. Kirk, in 1880, saw it as a band; but in 1882, 

 August 4, resolved it into bright lines — a broader group than «, 

 less distinctly bounded, and with a less bright centre, and con- 

 taining, he thinks, about twice as many lines. 



k appears to have been noted by two observers ; one being 

 Wijkander, who seems very accurate. 



8 is invariably fainter than 7 to me. 



A. Seen only by Lemstrom ; but he says that in 1S71, Novem- 

 ber 22, he " observed it with certainly three separate times." 



The likelihood of the existence of lines in the violet or blue 

 (such as \ or k) in addition to those commonly seen, is manifest 

 to me from the fact that I have twice seen purple in auroras. 

 The first time was at Sunderland, 1869, May 13, at 10.55 p.m., 

 w hen for a minute or two there was a large patch of coloured 

 light — deep crimson, exquisite pink, and most lovely pinkish 

 purple, gradually passing into one another. The crimson was the 

 same colour frequently observed ; the pink was very different, and 

 far more beautiful. The crimson lasted after the other colours 

 faded. The second time was in Skye, 1872, August 3, about 

 10.30 p.m., when for two or three minutes there were large 

 patches of a beautiful, but not deep, pinkish purple. £ had no 

 time to observe the spectrum in either case. It is manifestly 

 improbable that these colours would be caused by any of the 

 ordinary lines of the spectrum ; probably one line in the violet 

 or blue, in combination with the red line, could account for the 

 various tints. 



Colours are closely connected with the spectrum ; but I cannot 

 say I ever saw any in the aurora, except the purples and pink 

 just referred to, that might not be readily accounted for by the 

 ordinary lines of the spectrum with or without the red line ; as 

 the only ether decided colours I have seen are red and the usual 

 greenish colour, varying somewhat in intensity and perhaps tint. 

 I have seen other less decided colours ; but, considering the 

 extent to which the colours of the aurora might be affected by 

 mist, smoke, twilight, moonlight, &c, and one's judgment by 

 the effect of contrast, I could not say that they certainly belonged 

 to the aurora. 



The continuous light of the spectrum always reaches from a 

 to S ; being very faint from 7 to S, rather faint from a to e, and 

 sometimes brighter from e to 7) or to ( than beyond. Sometimes 

 when « is too faint to be detected, the abrupt brightening of the 

 continuous spectrum at that point is plainly visible. Kirk, in 

 1882, on the occasion already mentioned, when the spectroscope 

 was pointed betw een the streamers, saw the spaces Irom a to c 

 and from t to 7 apparently filled with shifting lines, very nume- 

 rous and close. Not that the lines really shifted, but their 

 flickering caused them to appear shifting, and possibly also to 

 appear more numerous than they really were. When the spec- 

 troscope was pointed on the streamers these lines were obscured 

 by the greater brightness "of the rest of the spectium. I have 

 myself often suspected lines between € and 7, besides ( and r. 



It has been suggested that some of the lines may vary some- 

 what in position ; but there is no evidence yet that the apparent 

 variations are due to anything but errors of observation. 



A flickering of the lines has been observed in certain cases ; 

 in all probability this occurs with the whole spectrum w hen the 



