208 



On the Chemistry of the Hottest Stars. 



ascribed by Dr. Scheiner to differences of temperature, in accordance 

 with my experimental results of 1879. 



25. The experiments on the spectrum of mercury which have been 

 made by Eder and Valenta have revealed variations which, according 

 to them, favour the dissociation hypothesis. 



26. On various grounds, the view that the differences in stellar 

 spectra represent fundamental differences of chemical composition is 

 untenable. The fact that many stars which are widely separated in 

 space give identical spectra, indicates that they not only contain the 

 same "elements," but that the "elements" exist in the same pro- 

 portions in all. 



27. On the non-dissociation hypothesis, the action of heat on the 

 sun's chromosphere could not produce such a spectrum as that which 

 we know to be associated with hotter stars, since the relative pro- 

 portions of different vapours could not be changed. The only change 

 which can be imagined to take place on this hypothesis is a reduction 

 of intensity of all the lines due to reduced pressure. 



28. On the dissociation hypothesis, increased temperature would 

 bring about fundamental changes in the spectrum due to molecular 

 simplifications, and in this way the effect of an increase of tempera- 

 ture on the sun's chromosphere, as indicated by hotter stars, can be 

 predicted, and receives a simple and sufficient explanation. 



29. The disappearance of the enhanced iron lines in the hottest 

 stars, and the simultaneous intensification of the lines of hydrogen, 

 helium, and gas X, bring us face to face with the fact that iron is a 

 compound, into the ultimate formation of which one or all of these 

 gases enters. 



30. The ultimate molecules of the chemical elements discussed in 

 the present paper may be provisionally arranged in the following 

 order of resistance to the effects of temperature : — 



GasX "I Doubtful which 

 He J 

 H 

 Ca 

 Mg 

 Fe 



31. Each step in advance which has been made since 1873 has 

 demonstrated more and more that there is really such a " celestial 

 dissociation " going on as that which I then suggested. 



In concluding this paper, I am anxious to express my great 

 indebtedness to my assistants. For the photographs of the spectra 

 of stars and the new photographs of the spectra of metals under 



