2^6 



NATURE 



iJan. 5, 



their spectra. Therefore it is easily understood that Secchi was 

 only able to seize gradually the characteristics of the different 

 types (thus it was not until late that he introduced the fourth 

 type) ; and again, he once changed the order so that the second 

 and third types changed numbers. On account of this change, 

 some errors may have found their way into his publications. 

 Some even may be explained without having recourse to this 

 supposition. All the spectra which I have excluded from the 

 third class are, according to Secchi, indeterminable, except two, 

 which I consider intermediate between W.a and \\\.a, and the 

 two stars R.A. = 9h. i8m., Decl. = - 21° 50', andR.A. = iSh. 

 14m. 40s., Decl. = +25°-2, which Secchi found independently of 

 Schjellerup's catalogue of red stars. It might happen then that 

 with a clear sky faint bands might be perceived, ^ and as to the 

 last we may well suspect that there is some gross error in their 

 positions, judging from what Secchi says as to the manner in 

 which he discovered the first of them,^ Such a supposition 

 would not be admissible for the star 249 Schj. This .star is 

 situated, according to Herschel, amongst a mass of stars, and 

 Secchi says of it, " Stella di 9^ rossa con rigoni nello spettro 4° 

 tipo certamente " (" Memoria Seconda," p. 52). I have often 

 examined the cluster in which this star was situated, but with- 

 out perceiving it. We may therefore believe that it is variable 

 of long period. 



I still have to refer to the stars which ought to have been 

 transferred from one sub-c\z.%% to another. In the spectra of 

 these, variability seems to me quite inadmissible, the two sub- 

 classes being, as I shall try to prove soon, o-ord nated, and 

 not successive phases of development which every star must 

 undergo. I suppose that at the commencement of his observa- 

 tions of spectra of the third class, when Secchi met with stars 

 111.(5 not very well marked, he did not think them different 

 fronri 1 1 1, a, and he did not perceive the difference until after 

 having seen several spectra of this class as pronounced as tho'.e 

 of 78 and 152 Schj. On the spectrum of the la^er he still says 

 in 1867, "In conclusione e tipo di a Ercole ma con zone vere 

 mancanti" (" Catalogo," pp. 14, 15). 



However, neither Secchi nor even D'Arrest examined a 

 sufficient number of spectra Ill.i^ to thoroughly understand their 

 characteristics. Both appear to admit that there are funda- 

 mental differences between spectra belon'jing to it. For instance, 

 Secchi says of the spectrum of star 136 Schj., " E difficile dire se 

 sia proprio del 4°tipo"(" Memoria Seconda," p. 42), and of the 

 stars discovered by Wolf and Rayet with bright lines which are 

 not hydrogen, 3 and dark bands in the spectra, and there- 

 fore certainly belonging#ito Class W.h, he says, "Accenneremo 

 qui soltanto che esse appartengono al 4° tipo, ma sono di quelle 

 a zone molto irregolari " ("Sugli spettri prismatici delle stelle 

 fisse," p. 194), and "Ad ogni modo som di 4° tipo, e le righe 

 paiono del carbonio diretto" {I.e. p. 216). D'Arrest speaks as 

 follows of the star 74 Schj., "Irregular spectrum of type IV." 

 {A.M. 2016); and of the star \^K,b Schj., "Very remarkable 

 spectrum, &c." {A.N. 2009). By collecting all my observations on 

 all the spectra of this class it is seen that not one of them really 

 departs from what may be called the normal spectrum. Doubt- 

 less there are in different stars notable differences in the darkness 

 of the flutings, and in the brilliancy of the intervals, but all this 

 does not prevent all the spectra being formed according to one 

 constant type, as happens with Class Ill.a. Besides, Secchi seems 

 to th'.n'c that the aspect of a spectrum may change completely 

 with the kind of spectroscope used. He says of the spectrum of 132 

 Schj., " Tipo 4° ben deciso, . . . oculare cilindrico. Coll'oculare 

 piccolo sferico tutto questo era sparito e si credette tipo 3°." 

 When instruments are used which give so vague an appearance 

 to an exceptionally well-defined spectrum, prc'^enting essential 

 characteristics, it is easy to commit serious errors in judging 

 of the spectra examined. Therefore I cannot see that the dis° 

 crepancies which exist between Secchi's observations and mine 

 are a sign of variations in the stellar spectra, although no 

 doubt it is prudent to occasionally examine the stars conce.-ning 

 which these discrepancies have arisen. 



It is quite a different case with the discrepancies that I have 

 found between my observations and thise of D'Arrest, who was 



M. Vogel has as a matter of fact seen feeble bands in the spectrum of 

 one of them - 60 Schj., while in the case of ten stars his observations c )n- 

 firm my own ; in the spectrum of another star M. Pechiil^ has not seen any 

 bands. 



^ ' Trovata cercando 124 Schj." The position of this star differs by 27m. 

 and 40' from that of the star in question. 



^ In his observations at Vienna with the great refractor, M. Vogel was able 

 to see the lines of hydrogen either C or F in the spectra of a'l three stars ; 

 they were, however, feeble in compwson with the other bright lines. 



supplied with excellent instruments, and was a most careful and 

 skilful observer. It is necessary therefore to examine more 

 closely into the cases in which differences exist. There are three, 

 two of which concern the stars 24034 LI. = Weisse XII." 793 

 and DM. -l- 60° 1461 = A. Oe. 13681. D'Arrest says that the 

 latter has "a bright well-marked spectrum of type III." {A.N. 

 2044) and the former "a clear, fluked spectrum, the flutings 

 being very distinct although pretty fine, III." {A.N. 2009). I 

 found both nearly white, and their spectra II. a, or continuous. 

 It is true that I examined the positions of these stars by the help 

 of the two catalogues in which they are, and obtained the same 

 positions, nevertheless one is tempted to believe that D'Arrest 

 made some error in the identification of these stars, especially as 

 he did not observe them several times. This supposition is quite 

 inadmissible, however, for the star DM", -t- 36° 2772 = LI. 3 '500, 

 for here D'Arrest expressly says {A.N. 2X)9), "8'3 mag. with 

 beautiful column-like spectrum. It is one of the stars accom- 

 panying the great Hercules nebula." I have calculated the 

 position of this star with the help of the catalogue and of LI., 

 and besides that I examined all the stars in the neighbourhood 

 of the great cluster in Hercules without finding one of Class III. a. 

 The star DM. + 36" 2772 is of orange colour, but its spectrum 

 is continuous, or at most II. a very poorly developed. As rej;ards 

 this spectrum I shall not even attempt to explain the difference 

 between what I have seen, and the description given by 

 D'Arrest. A variability of the spectrum seems really probable, 

 and the star is certainly deserving of much attention. Besides 

 this star there are others whose spectra I found very feebly 

 developed, whilst D'Arrest says that they are beautiful or even 

 superb. This also may be regarded as a sign of variability, and 

 a fact which also supports this supposition is that D'Arrest has 

 made his observations under atmospheric conditions generally 

 regarded as similar to those found at Lund with a spectroscope 

 of similar construction to one of ours, and a telescope very little 

 larger than the one which I used. But it is remarkable that 

 whereas I have often found expressions used by D'Arrest to de- 

 scribe spectra stronger than I should have used, the contrary is 

 of very rare occurrence. It is p )ssible, therefore, that the 

 differences are only apparent, and that either D' Arrest's observa- 

 tions were mide under exceptionally favourable circumstances, 

 or mine under very unfavourable ones. The latter supposition 

 is scarcely probable however, for when such differences have 

 occurred I repeated the observation several times ; and besides, 

 Vogel's observations on the stars between - 2° and + 20° ajree 

 almost without exception with mine. It appears, moreover, 

 that very often D'Arrest only mide one observation of the same 

 star ; and, without depreciating his researches, it seems to me 

 more probable that there are small and rare inexactitudes in the 

 observations, than that such great changes hive taken place in the 

 stars themselves in the short period of ten years. 



My researches already contradict the hypothesis that important 

 changes in the stellar spectra take place so rapidly. My observa- 

 tions embrace a period of six years, and a much larger number 

 of objects than either D'Arrest or Secchi examined. But there 

 is no spectrum in which my latest observations have differed 

 sensibly from my first ones. It is true that my first observations 

 on the spectrum of R Crateris are in direct opposition to the 

 last, the former making its spectrum lll./> and the latter III. a. 

 But that is in no way a sign of variation in the spectrum. When 

 the bands in the spectrum III. a of a faint star are exceedingly 

 broad in the green-blue, it is easy to fall into the error of regard- 

 ing it as III.i^. At the time of M. Vogel's first observation he 

 also believed that the spectra of stars DM. + 14° 2525 and 

 DM. + 17° 3940 were Ill.b, and it is true that these two spectra, 

 especially the last, are strikingly like the spectra Ill./^in spectro- 

 scopes of small dispersion. 



It was in order to escape such errors that I determined in the 

 spectra of most of the faint stars of Class III.*^ the approximate 

 wave-lengths of their principal flutings ; the wave-length of 

 band 6, and also that of band 10, being a sure mark that the 

 spectra belonged to this section of Class III. This deceptive 

 appearance generally disappears when the star is examined 

 with spectroscopes of considerable dispersion. I am therefore 

 at present of opinion that, excluding the new stars and perhaps 

 17 Argils, which seems to belong to the same category, we 

 have no reason to believe that great and rapid changes take 

 place in the stellar spectra, although it must be confessed that the 

 observations of certain stars, especially DM. + 36° 2772, are such 

 as to render such changes very probable. 

 (To be continued.) 



