396 



NATURE 



[.August 21, 1890 



the great hydro-carbon band seen in the spectrum of 

 every comet that has been examined under favourable 

 circumstances. This identification is strongly supported 

 by the second bright line, 5137, which is also found both 

 in hydro-carbon and cometary spectra. It is, however, 

 on bringing the spectra of the star and of the blue flame 

 of a spirit-lamp at the same time into the field of the 

 spectroscope that their exact agreement becomes most 

 evident. For not only do they agree perfectly in wave- 

 length and in beginning with two plainly distinguishable 

 bright lines, but also in the delicate gradations of light by 

 which they similarly fade away towards the violet, thus 

 forcing the extreme probability of a common origin upon 

 the observer. 



" But the presence of luminous lines does not rest on 

 this single band, for the second cometary and hydro- 

 carbon band which has its bright edge at W.L. 472 9 

 (Hasselberg, ' Ueber die Spectra der Cometen') is also 

 found in the new star's spectrum at W.L. 472*2. 



" Of the three other luminous bands agreeing with the 

 coal-gas spectrum, which were all measured at Dun Echt 

 in that of comet 1881 III.,i two lie beyond the limit 

 to which I have yet traced the spectrum of this star, 

 and the third, falling between W.L. 563 and 534, in a 

 bright and otherwise difficult part of the spectrum, has 

 not made its possible presence evident. 



" This leaves the origin of the bright bands beginning 

 at 542'8 and 494*4 an open question ; but excepting their 

 general appearance, there is no reason why they should 

 be due to the same substance as the great band at 516* 2. 

 On the other hand, the presence of the bright hydro- 

 carbon bands in a spectrum of type IIL removes any 

 difficulty there may be in accepting Secchi's conclusion 

 that they appear in a reversed (dark) form in spectra of 

 type IV." 



Dr. Copeland also made determinations of the position 

 of the bright bands in Nova Andromedas,^ and noted — 



" It seems probable that the three ' bright ' bands, of 

 wave-lengths 546-8, 514-0, and 489-2, are identical with 

 the three brightest bands afterwards measured with the 

 same apparatus in Mr. Gore's Nova Orionis, of which 

 the brightest parts were at wave-lengths 542-8, 5 16-2, and 

 494-4. The trace of a condensation of light at W.L. 

 471-6, seen on September 20, agrees well with the bright 

 Hne in Nova Orionis at W.L. 472-2. ... In conclusion, 

 it seems worthy of remark that the spectrum described 

 above is the same as that given by any ordinary hydro- 

 carbon flame, burning so feebly that the spectrum of the 

 blue base of the flame is just beginning to show through 

 the continuous spectrum afforded by the white part of the 

 flame." 



Vogel made some observations of Nova Orionis,^ and 

 found that the wave-lengths of the absorption bands were 

 the same as those of a Orionis and other stars of that 

 group, the only difference being that the bright spaces 

 were more strongly marked. Dun(fr also noted* very 

 bright parts in the green and blue, which he identified as 

 the bright zones 516-8-503-2 and 495-8-484-3. With re- 

 spect to these bright parts, he thought they may be partly 

 due to the contrast with the very dark and broad bands. 



M. Ch. Tripled observed that the spectrum of Nova 

 Orionis was like a Orionis and /3 Pegasi. He also re- 

 marked : — ^ 



" Le 23 ddcembre, j'ai, pour la premiere fois, soupgonne 

 I'existence de lignes brillantes dans le vert ; mais cette 

 observation est un peu incertaine. On salt combien il est 

 difficile de decider si les apparences de lignes ou de 

 bandes brillantes, dans un spectre faible, sont vraiment 

 celles qui caractdrisent I'dtat d'incandescence d'une 

 matiere gazeuse, ou s'il faut les attribuer k un effet de 

 contraste causd par le voisinage des bandes obscures." 



' Ccpernic2ts, vol. il. p. 227. ^ Monthly Notices, vol. xlvii. p. 54. 



3 Astr. Nachr., No. 2704. "* Astr. Nachr., No. 2707. 



5 Comptes rendus, vol. cii. p.*4i, 1886. 



NO. 1086, VOL. 42] 



M. ThoUon observed the same Nova, and recorded — ^ 

 " Ce qui nous frappa tout d'abord fut I'^clat remarqu- 

 able du rouge et surtout du vert, tandis que le jaune ^tait 

 relativement sombre. Cette particularity nous sugg^ra 

 d'abord I'idde que nous nous trouvions en presence d'un 

 spectre de bands brillantes, analogue k celui des com^tes, 

 mais bien plus compliqud. Les observations compara- 

 tives faites sur a d'Orion nous confirmferent dans cette 

 idde. Cette dtoile, en effet, montre avec une parfaite 

 Evidence un spectre continu conservant partout I'dclat qui 

 lui est propre, et coup^ par des bandes et raies obscures." 

 With the exception of Dr. Copeland, however, no 

 observer confronted the spectrum of the Nova with that of 

 carbon, or the identification of the bright spaces with the 

 carbon flutings would have been evident. 



A short time after Dr. Copeland had published his 

 observations, Mr. Maunder challenged the assertion ^ that 

 in the Nova "the spectrum is not so much a continuous 

 one, interrupted by dark lines and dusky bands, as a not 

 very luminous spectrum upon which a series of bright 

 bands are superposed.^^ The accuracy of the observations 

 was not, however, doubted, nor was the importance of the 

 view denied. 



The main objection urged by Mr. Maunder was that 

 Dr. Copeland's measures of the bright parts in Nova 

 Orionis did not exactly agree with laboratory determina- 

 tions of the wave-lengths of the hydro-carbon bands. He 

 does not, however, make mention of the fact that there 

 are two perfectly distinct sets of bands seen under different 

 conditions. Nor does he refer to the " laboratory work " 

 which has been relied on to show that they are not hydro- 

 carbon bands at all. The mean of Dr. Copeland's mea- 

 sures of the bright line in the green, beginning a band, is 

 516-2. The wave-length of the first carbon fluting of one 

 series is given by Thalen as 516-4, which, therefore, gives 

 a difference of 0002 in the two determinations. The 

 line measured by Dr. Copeland at 513-7 is said by Mr. 

 Maunder scarcely to support his view, since the second 

 maximum of the carbon fluting has a wave-length 512-8, 

 and of the blue carbon fluting it is noted, " The third 

 hydro-carbon band, that in the blue with wave-length for 

 its less refrangible edge 473'7, is indeed not far from the 

 bright space Dr. Copeland has observed at X 472-2, but 

 the correspondence is certainly not very exact." Since 

 this criticism was made, however, it has been shown that 

 at different temperatures the maximum of the blue carbon 

 fluting may shift from 468 to 474, so that Dr. Copeland's 

 measures may represent the exact position of the band in 

 the Nova. 



But it is evident that a vast difference must exist 

 between the accuracy attainable in the observatory and 

 in the laboratory. Dr. Copeland's measures appear to 

 give the smallest probable error in the determination of 

 wave-lengths in such an object as Nova Orionis, yet he 

 measured the brightest band once at 5i7'4 and on the 

 following evening at 5i5"6, a difference of 0018. The 

 difference of the observations inter se exceeds any of the 

 differences between the bright parts measured by Dr. 

 Copeland and the accepted wave-lengths of the carbon 

 bands ; nevertheless Mr. Maunder says the observations 

 are " undoubtedly very accurate," hence it cannot reason- 

 ably be argued that the bright bands are not carbon 

 because of a want of exact coincidence with those 

 measured in the laboratory. 



Mr. Maunder also notes that "the second band in 

 order of brightness in the hydro-carbon spectrum begins 

 at X 563-4. This is certainly non-existent in spectra of 

 the third type; a broad dark band— No. 4 in Duner's 

 nomenclature and my own, wave-length 564-2 to 5592 — 

 occupies the very place." This contention, however, is no 

 longer allowable, since the recent researches show that 

 the carbon fluting of one series at 5634 is masked by the 



' Compies rendus, vol. cii. p. 356, 1886. 

 = Monthly Notices R.A.S., xlvi. p. 284. 



