392 THE CHEMISTRY OF THE SUN. [CHAP. 



be compounds, they have in other experiments seen certain lines 

 in their crucibles which require two metals to produce them. 1 



The observations of Messrs. Liveing and Dewar are so im- 

 portant, that I have asked Mr. Taylor, the Demonstrator of 

 Astronomy at the Science Schools, to be good enough to make 

 an abstract of them to print here. Before I give this abstract 

 however, I propose to show what should happen on my view 

 when each method is employed. The reader will then be able 

 to judge whether or no the results in question are in harmony 

 with it. 



Chlorine since it destroys the finer molecules by uniting with 

 them, forming chlorides, should prevent reversals ; whereas sub- 

 stances which prevent destruction of the finer molecules enable 

 them to form coarser molecules which reverse. 



Nitrous Oxide should produce no marked effects, because 

 although it contains 50 per cent, more oxygen than air, the 

 oxygen in the air is free, while that in the oxide is chemically 

 combined with nitrogen. The chemical attraction has to be 

 overcome before the oxygen can oxidise the finer molecules. 



Aluminium at the temperature of the electric arc is known 

 to be one of the most volatile of the metals, and all the 

 photographs taken at Kensington since 1873 confirm this. The 

 addition of aluminium should therefore bring out more reversals 

 of the lines of less volatile and less oxidisable metals by shield- 

 ing the finer molecules and allowing them to combine. 



Magnesium and Potassium are easily volatilised and oxidised, 

 hence they should be very efficient agents in giving reversals, 

 since they are oxidised instead of the finer molecules of the 

 other metals. 



Ferrocyanides of the metals, since they contain no oxygen and 

 yield cyanogen when heated, should give more reversals than 

 the metals themselves. 



1 Loc. cit. p. 97. 



