1878.] Dr. A. Schuster on the Spectra of Metalloids. 387 



first to appear in oxygen tubes. Thalen and Angstrom do not give 

 these lines ; Huggins does not give them. ; Salet does not give them. 

 Pliicker and Wiillner are the only observers who experimented under 

 the circumstances under which the lines appear. They come out 

 equally well whatever way the oxygen is prepared, whatever the 

 nature of the electrode, and I have seen them under the large glass 

 receiver already mentioned. The following is the appearance of an 

 oxygen tube as it undergoes gradual exhaustion. 



When the pressure is sufficiently diminished to allow the spark to 

 pass, it shows a yellow colour and the spectrum is perfectly continuous. 

 Almost immediately, however, the four lines are seen in the capillary 

 "part of the tube above the continuous spectrum. The continuous 

 spectrum in the wide part is then stronger than in the narrow part. 

 The four lines seem to have taken away part of the energy of the 

 continuous spectrum. As the exhaustion proceeds the spark spreads 

 out in the wide part, and the continuous spectrum is therefore di- 

 minished and becomes less intense than in the capillary part ; but it 

 gradually loses in intensity also in the narrow part, until the four 

 lines stand out on a perfectly black background. If under these circum- 

 stances the jar and air break is inserted in the circuit, everything will 

 disappear and the elementary line spectrum will come out. We have 

 here as complete a transformation as from the band spectrum of nitrogen 

 to the line spectrum of nitrogen taking place under precisely the same 

 circumstances ; and it is therefore not unlikely that the two phe- 

 nomena are due to the same cause. There are two reasons why the 

 existence of the compound line spectrum of oxygen as a separate 

 spectrum may have escaped previous observers. There is a blue line in 

 the elementary line spectrum which is nearly coincident with the blue 

 line of the compound line spectrum. It requires considerable dis- 

 persion to notice the difference ; the complete disappearance of the 

 compound line spectrum has therefore escaped notice. The two 

 green lines and the red line widen easily at higher pressures, and as 

 has been remarked by Wiillner, even fuse together to a continuous 

 spectrum. If the experiment is therefore made at a pressure at which 

 oxygen has a continuous background, the disappearance of these lines 

 might be taken for their widening and fusing together. No such 

 mistake is possible when the vacuum is good. I have not been able 

 to determine with certainty whether the red line seen at atmospheric 

 pressure is a remnant of the compound line spectrum, or whether it is 

 a line of the elementary line spectrum closely coincident. I am in- 

 clined to the former view, although it often seemed as if the line seen 

 at atmospheric pressure was less refrangible than the red line of the 

 compound line spectrum. I have drawn attention in a letter to 

 " Nature" (December 20), to the fact that the compound line spec- 

 trum of oxygen seems to be reversed in the sun. I have no further 



