36 On the Spark Discharge from Metallic Poles in Water. [Jan. 31, 



(c.) Variation of Intensities. 



The most prominent lines in the water-spark are not always the 

 chief lines of the air-spark. This is well shown in the spectra of iron 

 and copper. 



Many of the lines in the spark of iron, if their intensities are com- 

 pared under the two conditions of sparking, show distinct inversions. 

 A typical instance of this occurs with the lines at XX 4422 *74 and 

 4427*48. With the spark in air X 4427*48 is quite twice as strong as 

 X 4422*74, whereas in the water-spark there is scarcely any trace of 

 a line at X 4427*48, the 4422*74 line being, however, easily seen. 



Another example, slightly less prominent, is found in the lines at 

 XX 4315*26 and 4337*22. With the spark in air these lines are 

 almost equal in intensity, but in the water-spark X 4315*26 has about 

 three times the intensity of X 4337*22. 



In the case of copper, in the ordinary spark the most prominent 

 lines are those at X 4275*32 and X 4651*31. In the water-spark spec- 

 trum the line at X 4587*19 is almost as strong as either of the lines 

 just mentioned, although in the ordinary spark it is much weaker. 



Application to Stellar Spectra. 



I will next consider the bearing of these results on the explanation 

 of certain features of the spectrum which is characteristic of new stars. 

 It has been seen that in the water-spark the position of the absorption 

 undergoes little if any change of position, while in the case of non- 

 symmetrical reversals, a bright line may be observed greatly displaced 

 towards the red. In the new stars, on the other hand, the absorption 

 lines are greatly displaced, the accompanying blight lines occupying in 

 comparison normal positions. The facts are as follows : — 



In the case of Nova Aurigse the emission lines had practically normal 

 wave-lengths, but the displacements of the dark lines at H e was about 

 10*7 tenth-metres towards the violet, indicating a velocity of approach 

 of about 500 miles per second. 



The recent new star in Perseus exhibited the same normal positions 

 of the bright lines, and indications of even greater displacements of 

 the dark lines, at one time amounting to 15 tenth-metres at H e , repre- 

 senting a velocity of approach of the body producing the dark-line 

 spectrum of over 700 miles per second. 



These values differ enormously from those produced by pressure. 

 The amount of shift produced by subjecting the light source to 

 pressure is given by Humphreys and Mohler, in the paper above re- 

 ferred to, as follows : — 



