Brush Discharge in Weak Acids and Solutions. 821 



therefore was taken to keep it out o£ the spectrum. Conse- 

 quently there are not a very great number of examples of 

 the bands appearing under varying conditions, and most 

 of these are rather faint. They are, however, sufficiently 

 numerous for one to draw some conclusions with some 

 security. The variation which occurs is a change in rela- 

 tive intensity of the negative bands to the positive in the 

 negative brush. In acid solutions (HC1, H 2 S0 4 , HN0 3 , H 2 S ; 

 other acids do not show the bands in the negative brush) the 

 negative bands disappear from the negative brush. In two 

 cases in nitric acid with a copper point the negative bands 

 appear ; in other two cases they are entirely absent even 

 with the positive bands very strong (see fig. 5 b). In 

 LiOH, and K 2 S0 4 solutions, although the positive bands 

 were weak, the negative band X = 3914 appeared distinctly. 

 Better examples occurred in Li 2 S0 4 , NaOH, and NaS 2 S0 4 

 solutions. These photographs revealed the interesting fact 

 that the negative bands increased relatively to the positive 

 with an increase in concentration of solution and constant 

 current. In sodium sulphate, the spectrum goes through 

 nearly the whole series of spectrograms, and it is found that 

 the maximum relative intensity of the negative to the posi- 

 tive bands occurs at a concentration when the sodium and 

 oxygen spectra have reached their maximum intensity. 

 Consequently, the behaviour of the negative bands is 

 strikingly similar to that of the oxygen elementary line- 

 spectrum. The same conditions are not required for each. 

 In sodium sulphate it was found that the nitrogen spectrum 

 was produced only in a region remote from the point, while 

 the oxygen spectrum is produced only in a region close to 

 the point. In general, the nitrogen band spectra are produced 

 remote from the point in the negative brush and close to the 

 point in the positive brush. There are one or two exceptions 

 in both cases. In distilled water, the bands stretch out from 

 the point with diminishing intensity. 



No bands occur in zinc sulphate. In magnesium sulphate, 

 the positive bands occur in the negative brush faintly in one 

 case ; the negative bands are absent. 



No trace of the nitrogen line-spectrum appears in any of 

 the photographs. 



Impurities. 



Besides nitrogen, the impurities which yield spectra are 

 calcium, sodium, silicon, and lead. They appear in the 

 Phil. Mag. S. 6. Vol. 27. No. 161. May 1914. 3 I 



