424 



NA TURE 



[August 29, 1895 



than twenty objects of this class, and no exception has been 

 found of a star having this spcctnira whose light docs not really 

 vary. Of the variables of long period which have lieen dis- 

 covered visually, the hydrogen lines have been photographed as 

 bright in forty-one, the greater portion of the others being too 

 fainter too red to be studied with our present means." 



As said before, it seems imiwssible to imagine how our Sun, 

 as it proceeds along its " order of growth," should change into a 

 body with such characteristics as these. But on this point we 



Flame of carbonic 

 oxide. 



Flanie of cyanogen 

 fed with O.Nygen. 



Fig. 39. — Showing j-lhe various intensities of the lines of magnesium as seen under 

 different condi'ions. 



must wait for more large scale photographic spectra ; in other 

 words, more facts. 



Associated with this change in the order of evolution, Prof. 

 Pickering classes the chief stars in Orion, such a.s Bcll.itrix, 

 characterised by spectra containing hydrogen and a few other 

 dark lines of unknown origin, as early forms. On this point I 

 may also quote the following from Prof. Campbell {Astronomy 

 and Astro- Physics, 1894, p. 475) : — 



" In conclusion, I think we can say, from the foregoing 

 observations, that the spectra of the Wolf-Rayet stars are not 

 closely related to any other known type. They apjx'ar to have 

 several points in common with the nebular an<l Orion type 

 spectra ; but the last two appear to be much more closely re- 

 lated to each other than to the Wolf-Rayet spectra. It istherefore 

 difficult to place these stars Ix-tween the nebulxand Orion stars. 

 They certainly do not come after the Orion stars, and one does 

 not like to place them before the nebula;. We can probably say 

 that the bright lines are chronios|iheric, owing their origin to 

 very extensive and highly-heated atmospheres, but showing very 

 little relation, in constitution .and physical condition, to ih.al of 

 our own Sun. I'or the present, at least, this type 

 of spectrum must be considered as distinct from 

 every other known type, just as the nebular spectrum 

 is distinct, and like the nebular spectrum containing 

 lines whose origin cannot now be assigned." 



.\lthough Ur. Vogcl and others ap|)arently still 

 hold in the main to the classification w hich .ossimies 

 that all stars were created hot, and that nebulx 

 have nothing to do with them ; that, in short, every 

 star began in the highest stage of temperature, so 

 that the whole history of every star in the heavens 

 has been a process of cooling, there are signs of 

 wavering here and there. Some f>f the definitions 

 are being " edited " and rc-cdited to fit the facts 

 which the photographic record is ixmring in upon 

 us. I may take, as an instance, the following state- 

 ment made by I)r, Scheincr with reference to 

 a Cygni, which is classified by Dr. \'i:>gel as a 

 wilar star. 



" These figtircs plainly show that the spectrum of 

 o Cygni, in spile of the large numlx:r of its lines, 

 has no resemblance with that of the sun. While it 

 is p<jssilile to identify most of the lines with solar lines in 

 rcsjicct to their |x)sition, yet the total lack of agreement as to 

 intensity of the lines makes many of these identifications 

 worthlcM." 



The " figures" referred to arc micrometer measures of a photo- 

 graph. My ex|K;ricncc in these matters is that it is a pure waste 

 of time to mc-LSure a photograph until it h.as lieen comimred with 

 others to which it is important to refer it, enlarged up to the 

 umc Kale. In this I think I carry I'rof, Kcelcr with mc 



(Astronomy and Astro-Physics, 1894, p. 4S5). " The coincidence 

 of , , . lines is shown more beautifully by inspection ot 

 . . . photc^raphs than by any process of measurement." 

 Thus a comparison of the spectra of a Cygni and of the Sun 

 which Dr. I 'ogel classes together, shows at once the dissimilarity 

 pointed out above without any measurement whatever. I anv 

 glad, however, to find that Dr. Scheiner now regards the identi- 

 fication as " worthless," l)ecause it is such difterences as these 

 which have compelled me to reject Dr. X'ogel's classification. 

 Dr. Scheiner then goes on : — 

 "The magnesium line at \ 4481 is the strongest 

 in the entire spectrum. The other strong lines 

 coincide for the most |X»rt with the fainter solar 

 lines. The presence of numerous iron lines can be 

 scarcely doubted, but here again we have the 

 peculiar phenomenon that the fainter, instead oj 

 the stronger, lines occur. We may conclude from 

 all these facts that very different conditions as to 

 temperature must prevail in a Cygni from those in 

 the stars of class I<;." (Scheiner's " Astronomical 

 Spectrosco|iy," Krost's translation, p. 247.) 



Much of the work of the future, which eventually 

 must smooth down all diflerences between stellar 

 classifications, must consist of the study of single 

 lines in the spectra of difierent stars, and I anv 

 rejoiced to find that the Potsdam observers are a6 

 length beginning to take this matter up. Dr. 

 Scheiner, one of the Potsdam assistants, has, as seen 

 above, called attention to the behaviour of the line 

 4481 of magnesium, and agrees that the variations 

 in the line observed are due to differences of leinjieratvire, and! 

 that therefore it nmy be useil .as a stellar thermometer.' 



But for this work an acquaintance with the literature of the 

 subject is desirable. Had Ur. Scheiner been acquainted with it, 

 I am certain he woukl have iluiie me the honour to quote, or 

 at all events to refer to, a communication I made to the Ro)'al 

 Society (16 years ago !), pointing out that the line in question 

 was visible only at high temperatures, and that such work would 

 help us in the study of " the atmospheres of the hottest stars." * 

 In the same connection, in the "Chemistry of the Sun," pub- 

 lished in 1887, I gave the diagrams, here reproduced, indicating 

 the lines, visible at various lenipenitures in the laboratory, and in 

 the Sun and prominences. 



Having said so much on the different classifications of stars, 

 and indicated, I trust judicially, that the one suggested by the 

 meteoritic hypothesis is so far holding its own, I now pass oiv 

 to some recent work which was undertaken to test it by a 

 limited photographic survey. In the first instance I had ustd 

 the eye observations of others; a sliuly of spectra, entirely 

 photographic, it was hoped would enable an independent 



Bunscn. 



Flame of cyanoKcn 

 fed with oxygen. 



Flame of cnrlwnic 

 o.\idc. 



(. IN. v.iiiii. -111. ^ -III.- iM.i ■.! iii.(.:ni It II ranged in order of 



increasing tenipcraturc>. The lines marked rt A <• (/ c in the diagrams have 

 the following wavelengths :— 5iO</8, 517 (/'). 47o3'5, 457o'3, 4481. 



estimate to be formed as to the validity of the 

 hypothesis. 



The conclusions I came to in the first instance were necessarily 

 based on observations made by others, for the reason that my 

 own work up to that time had been chiefly directed to the Sun. 



So .s<K)n, however, .as my solar work rendered it necessary to 

 determine the S'.ui's true pl.ace among the st.ars in regard to its 



1 " Astronomical Spectroscopy, " p. viii. 

 ' Key. Soc. t'roc. vol. xxx. p. 3», 1879. 



NO. 1348, VOL. 52] 



