Brush Discharge in Weak Acids and Solutions. 809 

 Method of Work. 



Spectrograms of the positive- and the negative brush were 

 taken alternately, exposures o£ a quarter of an hour being 

 given to each. By varying the current in the transformer 

 primary, the current through the brush was kept as nearly 

 constant as possible during the exposure, and for each couple 

 of spectrograms the same current was used. Three readings 

 of current and potential difference were taken during each 

 exposure. Very careful use must be made of these readings 

 to interpret the spectrograms, as they do not take into 

 account the rapid fluctuations which occur both in the 

 positive and negative brush, and which are probably different 

 in each. It will be shown later that changes occur in the 

 spectra which can only be accounted for by supposing changes 

 in the electrical conditions which are not taken into account 

 by the method of measurement adopted. To keep even the 

 mean potential and current constant was quite impossible, 

 though occasionally readings would remain quite steady 

 during the whole exposure. 



The liquids examined were distilled water, tap-water, 

 solutions of H 2 S0 4 , HC1, HN0 3 , H 2 O0 4 , H 2 Mo0 4 , H 3 P0 4 , 

 S0 2 , H 2 S, H,0 2 , NaOH, Na 2 S0 4 , Na 2 C0 3 , NaCl, KOH, 

 K 2 S0 4 , K 2 C0 3 , KBr, LiOH, Li 2 S0 4 , Li 2 C0 3 , MgS0 4 , 

 ZnS0 4 . 



Various concentrations were used (*001 per cent, to "5 per 

 cent, by weight in the case of salts, and from '03 per cent, 

 to 2 per cent, by vol. of concentrated acid in the case of 

 acids), and in a large number the current also was varied. 

 Altogether some 370 photographs were taken. 



Spectrum of the Bubble. 



In discussing the spectra produced I shall first give the 

 results obtained in the case of the " bubble discharge." 

 As was stated above, in the case of dilute sulphuric acid, 

 when a bubble forms underneath the point it becomes 

 suffused with a beautiful violet colour. Though the image 

 thrown on the slit by the quartz lens is not achromatic, it is 

 possible to notice differences in the spectrum near to and 

 remote from the point for a region a little longer than the 

 visible part. The spectrum of the bubble discharge shows 

 not only the usual spectrum produced when there is no 

 bubble, but also a series of bands degraded towards the red 

 and stretching from about H^ to X = 3100. They are most 

 intense in the region of the H and K lines. While the 

 ordinary brush spectrum is confined quite close to the point, 



