Mr. J. N. Lockyer's Spectroscopic Notes. 321 



similar among themselves, as regards molecular condition, than if 

 the spectra varied in character, not only from element to element, 

 but from one temperature to another in the vapour of the same 

 element. 



5. At the temperature of the sun's reversing layer, the spectra 

 of all the elements known to exist in that layer are apparently 

 similar in character — that is, they are all line-spectra ; hence it is 

 more probable that the vapours there are truly homogeneous, and 

 that they all exist in the same molecular condition, than if the 

 spectrum were a mixed one. 



6. The fact that the order of vapour-densities in the sun's atmo- 

 sphere, which we can in a measure determine by spectroscopic 

 observations, does not agree with the order of the modern atomic 

 weights of the elements, but more closely agrees with the older 

 atomic weights, led me to take up the present research. Thus I 

 may mention that my early observations of the welling-up of Mg 

 vapour all round the sun above the ]Na vapour have lately been fre- 

 quently substantiated by the Italian observers ; so that it is beyond 

 all question, I think, that, at the sun, the vapour-density of Mg is 

 less than that of Na. 



7. The vapour-densities of the following elements have been 

 experimentally determined : — 



S 32 (at 1000°) 



I 127 



Hg 100 



N 14 



O 16 



P 62 



8. To pursue this inquiry the following arrangements have 

 been adopted : — 



The first experiments were made last December upon Zn in a 

 glass tube closed at each end with glass plates ; and I have to 

 express my obligations to Dr. Russell for allowing them to be con- 

 ducted in his laboratory, and for much assistance and counsel con- 

 cerning them. 



A stream of dry H was allowed to pass. The tube was heated 

 in a Hofmann's gas-furnace, pieces of the metal to be studied 

 having previously been introduced. It was found that the glass 

 tube melted ; it was therefore replaced by an iron one. The incon- 

 venience of this plan, however (owing to the necessity for intro- 

 ducing the metal into the end of the hot tube when the first 

 charge had volatilized), and, moreover, the insufficiency of the 

 heat obtainable from the gas-furnace, soon obliged me to replace 

 both tube and furnace by others, which have now been in use for 

 many weeks, and which still continue to work most satisfactorily. 



The iron tube is 4 feet in length, and is provided with a central 

 enlargement, suggested to me by Mr. Dewar, forming a T-piece 

 by the screwing in of a side tube, the end of which is left pro- 

 jecting from the door in the roof of the furnace. Caps are 



H ... 



1 



K ... 



39 



As ... 



150 



Br ... 



80 



Cd ... 



56 



01 ... 



35*5 



