262 



NATURE 



[January 12, 



gen lines is about the same as in the solar spectrum. These also 

 are arranged in two sub-divisions. 



In all the stars included in the third table there is a very con- 

 siderable amount of continuous absorption in the violet, extend- 

 ing to about G, and it is a matter of great difficulty to photograph 

 these spectra, as most of the stars of this class are below the 

 third magnitude. The hydrogen lines are very thin. One sub- 

 division incluHes the spectra which show flutings shading away 

 towards the less refrangible end of the spectrum. The other 

 comprises stars without flutings in their spectra. The brightest 

 star in this table, o Orionis, is discussed in detail, the result 

 tending to show that the temperature of the absorbing iron 

 vapours is not much greater than that of the oxy-hydrogen flame. 



The relations of the various sub-divisions to which reference 

 has been made are then traced. 



One important fact comes out very clearly, namely, that 

 whether we take the varying thicknesses of the hydrogen lines 

 or of the lines ot other substances as the basis for the arrange- 

 ment of the spectra, it is not possible to place all the stars in 

 one line of temperature. Thus, there are stars in which the 

 hydrogen lines are of the same average thickness, while the 

 remaining lines are almost entirely different. These spectra 

 cannot, therefore, be placed in juxtaposition, and it is necessary 

 to arrange the stars in two series. 



The next part of the paper consists of a discussion of the 

 photographic results in relation to the meteoritic hypothesis. 

 In the Bakerian Lecture for 1888, I brought together the 

 various observations of the spectra of stars, comets, and nebulas, 

 and the discussion suggested the hypothesis that all celestial 

 bodies are, or have been, swarms of meteorites, the difference 

 between them being due to different stages of condensation. 

 The new classification rendered necessary by this hypothesis 

 differed from previous ones, inasmuch as the line of evolution 

 followeri, instead of locating the highe-t temperature at its com- 

 mencement, as demanded by Laplace's hypothesis, placed it 

 much later. Hence bodies of increasing temperature were 

 demanded as well as bodies of d crea-ing temperature. 



The question how far this condition is satisfied by the new 

 facts revealed by the photographs is next discussed. 



This involves the consideration of some points in connection 

 with the hypothesis to which brief re'erence alone has been 

 made in previous communications. The phenomena to be ex- 

 pected on the hypothesis, and the actual facts, are given side by 

 side below : — 



'Nebulce. 



The bright lines seen in 

 nebulas should have three 



origins : — 



(1) The lines of those 

 substances which occupy the 

 interspaces between the 

 meteorites. Chief among 

 these, from laboratory experi- 

 ments, we should expect 

 hydrogen and gaseous com- 

 pounds of carbon. 



(2) The most numerous 

 collisions between the meteor- 

 ites will be partial ones — 

 mere grazes — sufficient only 

 to produce comparatively 

 slight rises of temperature, 



(3) There will, no doubt, 

 be a small number of end-on 

 collisions, producing very 

 high temperatu es, and there 

 should he evidence of some 

 high-temperature lines. 



(i) Lines at wave-lengths 

 approximately very closely to 

 the lines of hydrogen, and to 

 some of the carbon flutings, 

 appear in the spectra of 

 nebulas. 



(2) There is a fluting most 

 probat)ly due to magnesium at 

 A. 500, and the longest flame 

 lines of iron, calcium, and 

 magnesium are seen. 



(3) The chromospheric line 

 Dj and anoiher line at \ 4471 

 (which is always assoclnted 

 ^ ith D3 in the chromosphere) 

 have been recorded in the 

 spectrum of the Orion 

 Nebula. 



The lines seen 

 spectra of bright-line stars 

 should, in the main, resem 'le 

 those which appear m iie ulse. 

 They will differ, however, for 

 two reasons given in the 

 paper. 



Bright-Line Stars. 

 in the I Prof. 



Pickering has 

 sh iwn that the Draper 

 Memorial photographs prove 

 that brifiht-line >tars are inti- 

 mately connect d wnh the 

 planetary nebulae, the lines in 

 I he specira being almost 

 identical. 



Stars of Increasing Temperature. 



NO. 12 I I, VOL. 47] 



Stage I. — Immediately fol- 

 lowing the stage of condensa- 

 tion giving bright-line stars, 

 the bright lines from the in- 

 terspaces will be masked by 

 corresponding dark ones, due 

 to absorption of the same 

 vapours surrounding the in- 

 candescent meteOrites, and 

 these lines will therefore 

 vanish from the spectrum. 



Owing to the interspaces 

 being restricted, absorption 

 phenomena will be in excess, 

 and low-temperature metallic 

 fluting absorption will first 

 appear. The radiation spec- 

 trum of the interspaces will 

 now consist chiefly of carbon. 



Under these conditions the 

 amount of continuous absorp- 

 tion at the blue end will be at 

 a maximum. 



Stage 2. — With further con- 

 densation, the radiation spec- 

 trum of the interspaces will 

 gradually disappear, and dark 

 lines replace the fluting ab- 

 sorption owing to increase of 

 temperature, though this line 

 absorption need not necessarily 

 resemble that in the solar spec- 

 trum. 



Stige 3.— (i) The line ab- 

 sorption and the continuous 

 spectrum at the blue end will 

 diminish as the condensations 

 are reduced in number, as 

 only those vapours high up in 

 the atmospheres surrounding 

 the condensations will be 

 competent to show absorption 

 phenomena in consequence of 

 the bright continuous spec- 

 trum of the still disturbed 

 lower levels of those atmo- 

 spheres. 



{Z) Lines of iron and other 

 substances will disappear at 

 this stage, because the bright 

 lines from the interspaces wdl 

 counteract the lines in the 

 same positions due to absorp- 

 tion of surrounding vapours. 



(3) The chances ot violent 

 collisions being now enor- 

 mously increased, we should 

 expect the absorption of very 

 high - temperature vapours. 

 The solar chromospheric lines 

 may be taken as examples of 

 lines produced at such tem- 

 peratures. 



The spectra of stars given 

 in the third table answer these 

 requirements. They show no 

 bright lines under normal con- 

 ditions. 



The dark flutings in the 

 visual spectrum agree very 

 closely in position with the 

 flutings seen in the flame spec- 

 tra of manganese, lead, and 

 iron. The evidence afforded 

 by the photographs proves the 

 actual presence of carbon 

 radiation. 



The photographs show a 

 considerable amount of con 

 tinuous absorption in the ultra 

 violet and violet. 



The spectra consist ofnume 

 rous dark metallic lines, bu: 

 they do not exactly resemble 

 the solar spectrum, a Taur: 

 and 7 Cygni are types of stars 

 at this stage. 



(i) These conditions are 

 satisfied by such stars as o 

 Cygni, Rigel, Bellalrix, S 

 Orionis, and a Virginis. In 

 these there is no continuous 

 absorption at the blue end, the 

 spectra consisting of simple 

 line absorption. 



(2) In the spectrum of o 

 Cygni, which represents the 

 earliest example of this stage, 

 there are a few of the longest 

 lines of iron, but in other 

 stars of this class the iron 

 lines disappear. 



(3) The new lines which 

 appear include the chromo- 

 spheric line at A 4471, and 

 possibly a few others. 



The Hottest Stars. 



The order of the absorb- 

 ing layers should follow the 

 original order of the extension 

 of the vapours round the 

 meteorites in the first condi- 

 tion of the swarm, and the 

 lines .seen bright in nebulae, 

 whatever their origins may 

 be, should therefore appear 

 almost alone as dark lines. 



In stars like a Andromedas 

 we have absorption lines 

 agreeing in position with 

 some of the bright lines 

 which appear in nebulae. 



