Vol. Ill, No. 4.] 



The Electrical State of JStascent Gases. 



195 



The table below contains, in a concise form, the first experi- 



mental results. 



Table. 



Substance. 



KMnO 



KCIO. 



3 



Approximate 

 ^Numbers. 



1 = 1-5x106 



ions per c.cm. 



4 



4 



KCIO 



i 



+ 



104 

 110 



12 

 16 or 18 



NagOg 



+ 



HgO 



+ 



37 



IS 



9 



9 



12 

 13 



3 



1 

 1 



The electrometer gave a deflection of 1180 scale divisions for a 

 potential difference of 1 volt; and the capacity of the T-pi 

 electrometer system was found to be 395"5. Calculating with 

 these data in the case of potassium perchloz^ate, we obtain a num* 

 her of the order of l'4x 10''^ positive and a nearly equal number of 

 negative ions per cubic centimetre of gas. This is of the same 

 order of magnitude as was found by Townsend for the number 

 of ions of only one sign in his electrolytic gases. It indicates only 

 one ion among each 10^^ molecules and therefore a quantity far 

 smaller than what is measurable by any of the older chemical 

 methods. 



The smallness of this total number of ions makes it impossible 

 to claim it as direct evidence that the breaking up of each mole- 

 cule of the original substance is accompanied by the production 

 of any definite number of ions. Nor are we at all in a position 

 to assert the absolute accurac;y of these numbers. It will be 

 necessary to determine a (the rate of recombination of the ions) 

 before it can be safe to generalise with confidence. Havinsr, how- 

 ever, made the conditions as closely comparable as possible, and 

 hazarding the assumption that the rate of recombination will be 

 constant for the same gas In each of the cases here considered, 

 and that therefore tlie actual deflections recorded do bear some 

 simple and uniform relation to the number of ions originally 

 generated in the decomposition, one is tempted to notice the 



simple connections that would seem to be indicated by the 



