366 Dissociation of Hydrogen in the Sun's Atmosphere, 



the changes in form observed by Lockyer. If this assump- 

 tion is correct, \= 5315*9 (about) must be a wave-length of 

 the above a-spectrum, which does not occur amongst the 

 numbers found by calculation only because the wave-lengths 

 corresponding to the element a in hydrogen are not com- 

 pletely known. In this case \= 5315*9 x § must be a wave- 

 length of the elementary H-spectrum. It is true that the 

 wave-length (about) § x 5315*9 = 3544-0 (about) does not occur 

 in the list of wave-lengths at present known as belonging 

 to hydrogen; but if we take the extremely rich section of the 

 H 2 0-spectrum corresponding to the second hydrogen-spectrum 

 as given by Dr. Hasselberg (see II. (1)) for the comparison, it 

 satisfies in a very complete manner the criteria given (II. (5) a) 

 for hydrogen-rays which belong to the substance a within the 

 hydrogen, since £§x3544'0 = 2244-5, jx 3544-0 = 2658*0, 

 and ffx 3544*0 = 2646-1 differ only by a few tenths of an 



Angstrom unit (10~ 7 milhm.) from the corresponding wave- 

 lengths of the H 3 0-spectrum: 2244*8, 2658*6, and 2645*4. 

 The ray at about X= 3544*0 is, therefore, an hitherto unknown 

 ray of hydrogen, the existence of which will probably be 

 shown experimentally ; X = 5315*9 is a wave-length of the 

 primary element a (previously overlooked in this calculation 

 for want of sufficient data), which is, therefore, probably 

 identical with coronium. I must not omit to remark that the 

 line 5317-28 of the second hydrogen-spectrum described by 

 Dr. B. Hasselberg in 1883 lies very near to the corona-line 

 5315*9; and from what has been said above might possibly be 

 observed, together with other lines of the latter spectrum, in 

 the comparatively cooler regions of the corona far removed 

 from the photosphere. If these two lines are really different, 

 it must be possible, under favourable circumstances, to observe 

 both lines at the same time in the corona — 5315*9 in the hot 

 portion nearer the photosphere, and the other in the less hot 

 regions, further from the photosphere, for which observation 

 a favourable opportunity offers in the total solar eclipse of the 

 18th August. 



The like holds good of the known chromosphere-line 

 X=4471*2 of Rayet (Lorenzonr's/), which differs by only 

 0*3 from the nearest line of that part of the second hydrogen- 

 spectrum which Dr. B. Hasselberg has described in 1884 

 (Zusatz zu meinen Untersuchungen) , viz. the line X = 4470*9; 

 so also with the chromosphere-line 4712*5, which, like the 

 former line 4471*2, is rather a band than a line, and like this 

 is ascribed to Ce, but differs from the nearest line of the 

 second H-spectrum only by 0*6. 

 The line 4471*2, which I was disposed to take for a helium- 



