820 Mr. H. Smith on the Spectroscopy of the Elect fie 



current is needed in solutions where the potential is low, to 

 produce them in a definite intensity, than is required in 

 distilled water where the potential is high (PL XI. fig. la, c). 

 I have compared these bands with the ultra-violet bands 

 occurring in the cone of a Bunsen flame and witli the bands 

 occurring in damp air in a Geissler tube with a quartz 

 window. 



Electrode, 



When a platinum wire is used as the electrode in the pro- 

 duction of the brush, the arc spectrum appears when the 

 brush is negative. Even with very long exposures, there is 

 no trace of the spectrum in the positive brush (fig. 9 h, c) 

 The lines of platinum are most numerous and strongest in 

 the ultra-violet. Traces of the visible lines, however, some- 

 times appear, and these are concentrated close to the point. 

 The lines in the ultra-violet, being out of focus on the slit, 

 do not show this. The relative intensity of these lines to 

 the lines in the water-vapour bands is practically constant. 

 There is, however, a certain amount of fluctuation, and in a 

 few cases the platinum spectrum is very weak. In the 

 condensed brush the spark spectrum of platinum appears 

 (PI. XL fig. 8). 



When copper wire is used for the electrode in the produc- 

 tion of the brush in distilled water, the lines \ = 3247 and 

 A, = 3274 appear faintly in the positive brush and a little 

 more strongly in the negative. In acid solutions the lines 

 are very bright and other lines also appear. This, no doubt, 

 is due to the copper passing into solution. 



Nitrogen, 



The bands of the nitrogen spectrum appear in both the 

 negative and positive brush in distilled water. In both cases 

 the positive and negative bands occur. The negative bands 

 are stronger relatively to the positive in the positive brush 

 than in the negative. In the positive brush the edge of the 

 negative band at \ = 3914 is about equal in intensity with 

 the edge of the positive band \~3801, and this relation 

 is maintained throughout all solutions, as well as in distilled 

 water. 



The presence of the nitrogen spectrum in solutions depends 

 upon the point being more or less cracked. Small cracks do 

 not affect the spectrum in any way except to cause the 

 nitrogen band spectrum to appear. Its presence, however, 

 unless it is very weak, is a serious trouble, as it complicates 

 a spectrum already very crowded with lines. Every care 



