56 Mr. E. P. Perman. [May 8, 



says ('Roy. Soc. Proc.,' vol. 42, 1887, p. 344), "The appearance of 

 the dissociated iodine is not greatly different from that of the 

 unsparked ; its colour, however, is, I think, a little lighter, and not so 

 Tiniform. I was not able to detect any change in the absorption- 

 spectrum produced by the sparking. The electric strength of the 

 sparked gas was, however-, less than that of the unsparked." The 

 .same apparatus was used, platinum wires being fused into the glass 

 (see fig. 4). The jacket used was methyl salicylate vapour, tempera- 

 ture 223'5 ; the iodine was boiled out at the atmospheric pressure, 

 and the tube then sealed at D. The mean distance between the heaps 

 without sparking was 14'24 mm., and after sparking for fifteen 

 minutes, and cooling for three minutes (to allow the vapour to regain 

 the temperature of the jacket), the mean distance was 14'3 mm. 

 Another experiment was then made, so that the heaps were formed 

 during the sparking ; the mean distance between the heaps was 

 14'2 mm. on the side of the piston near the vibrator, and 14'4 mm. 

 on the other side ; the difference is probably due to the heating effect 

 of the induction sparks. 



The results may be tabulated thus : 



Mean distance between heaps. 



Before sparking. After sparking. During sparking. 

 14-24 14-3 14-2 



14-4 



Suppose that complete dissociation of the iodine vapour took place 

 on sparking, then the pressure would be doubled, and therefore the 

 elasticity would be doubled ; but the density (taking no account of 

 the pressure) would remain unaltered. 



Let Vj = velocity of sound in the undissociated iodine vapour. 

 v 2 = completely dissociated ditto. 



e = elasticity of the undissociated vapour. 

 d = density of the iodine vapour (same in both cases). 



Then 



= \J %> 



/2e 

 and v 2 = /y -T- 



But the distance between the heaps of silica (half a wave-length) 

 is directly proportional to the velocity of sound in the vapour ; there- 

 fore, if any dissociation took place, there would be a marked increase 

 in the distance between the heaps on sparking. 



The results given above agree within the limits of the errors of 



