724 Mr. H. Donaldson on the Spectra of the 



permanganate, and was left in contact with solid potash for 

 some time before it was admitted to the apparatus. The bulb 

 in which the discharge was secured was first exhausted and 

 then washed out six times with oxygen, exhaustion to a 

 pressure below that at which any discharge would pass being 

 carried out after each washing. When the pressure of the 

 gas in the bulb was just low enough for the true ring- 

 discharge to pass, the colour of the ring was brick-red, 

 especially on the inner edge, and the spectrum seen was the 

 compound line spectrum. On reducing the pressure, a fairly 

 bright continuous spectrum appeared in the blue, superposed 

 on the compound line spectrum, and this further developed 

 into bright oxygen bands, those having their centres at 

 wave-lengths 5590 and 5245 being especially distinct. 



At a still lower pressure, the bright lines of the elementary 

 line spectrum began to flash out brightly at certain instants 

 when the discharge was running well, and ultimately they 

 became steadily bright while the intensity of the bands 

 decreased. The violet lines of the elementary line spectrum 

 were especially bright. The only line which persisted 

 throughout the whole range of pressures was the red (6157). 



An attempt was made to produce the changes in the spectra 

 at a medium pressure merely by varying the spark-gap and 

 the frequency of sparking. By increasing these, it was 

 found that the elementary line spectrum appeared at a 

 higher pressure than that at which it appeared with smaller 

 spark-gap and frequency. The band spectrum remained in 

 this case with the larger spark-gap, but rather more faintly. 



Argon : — Certain colour changes occurring in the ring 

 discharge in argon have already been obtained and described * 

 and an attempt was made to repeat those observations and to 

 examine the colour changes spectroscopically. The argon 

 first used contained a small percentage of nitrogen which 

 appeared in the spectrum of the light from the inner part of 

 the bulb during the discharge. 



At no pressure could any trace of the red argon spectrum 

 be seen, but the blue spectrum was very bright, about 30 

 lines being measured. A portion of the argon, mixed with 

 pure oxygen, was then sparked for three hours over potash. 

 The excess of oxygen was removed by phosphorus, and the 

 argon remaining was used for the discharge, after the bulb 

 had been washed out several times with it. In this case, at 

 the highest pressures, the ring was red and of the 12 lines 

 measured every one belonged to the red argon spectrum. 

 At an intermediate pressure the number of bright lines 

 * Strutt, Phil, Mag. xlix. p. 293 (1900). 



