Dissociation of Water Vapor. 5 



of oxygen. In the spectrum one sees two reversed bands 

 which coincide with the great H. H. lines in the solar spec- 

 trum ; one also sees a reversed line at wave length 4227, closely 

 coincident with a strong solar line ; one has also a reversed 

 band coinciding with a region of reversals in the sun between 

 wave lengths 4315-4285. These regions in the sun are doubt- 

 less composite photographs of many reversals of different ele- 

 ments. I believe, however, that the basis of the reversals is 

 the presence of dissociated water vapor. 



The nomenclature, too, of the stars in relation to their spec- 

 tra I believe should be changed to dissociation spectra ; an 

 excess of water vapor produces what is termed the hydrogen 

 type, and the dissociation of this water vapor in the presence 

 of other gases, nitrogen for instance, together with metallic 

 vapors, may account for other types. 



The intense light due to the dissociation of water vapor 

 may, in some cases, mask the fainter light of the metallic lines 

 in the stars which show only gaseous spectra, especially when 

 we consider the varying distances of the stars. I have em- 

 ployed electrodes of platinum, copper, silver, aluminum, iron, 

 and aluminum, and no trace of the lines of their vapor can be 

 perceived in the photograph of the gaseous dissociation spectra; 

 moreover, when sodium is present in the spectrum tubes, 

 although it fills the tube with a brilliant yellow light at low 

 voltage and strong currents, and exhibits the sodium lines very 

 strongly, no trace is seen of it, with very powerful discharges. 

 The tubes then show only the brilliant white light due to the 

 dissociation of water vapor : a light which is the nearest ap- 

 proach to sunlight which I have been able to produce. Its 

 actinic effect is greater than that of any of the metals, such as 

 magnesium, aluminum or zinc. Possibly the varying amount 

 of water vapor may be a factor in the variability of certain 

 stars ; and one is led to conjecture whether the light of the 

 sun's atmosphere is not due to an electrical dissociation due to 

 discharges of very high period. 



The selective reversibility of the silver salts seems to me of 

 momentous importance in the subject of astrophysics ; for we 

 have reversible actions on the photographic plate which are 

 not due to the radiations of a gas passing through colder layers 

 of the gas. In other words, we have actions recorded which 

 are photochemical and are not in the heavens. It may be that 

 certain reversals observed in the spectrum of J^ova Persei may 

 be of this nature. This selective reversibility serves to reveal 

 certain lines which escape observation. Thus we see that 

 there is a gaseous line at wave length 4227, and a line not 

 shown on the reproductions but clearly seen on the negatives 

 between the great H. H. lines, having a wave length of 3953. 



