100 Experimental Researches in Electriciiy. 



and the almost total difference existing between them, 

 as to their mode of conducting electricity and becoming 

 charged with it, it might still be contended, although quite 

 hypothetically, that the bounding portions of air were now 

 the surfaces or places of attraction, as the metals had 

 been supposed to be before. In illustration of this and 

 other points, I endeavoured to devise an arrangement, by 

 which I could decompose a body against a surface of water, 

 as well as against air or metal, and succeeded in doing so 

 unexceptionably in the following manner." For the subse- 

 quent details of this interesting and important experiment, 

 we must refer to the researches themselves, as it would 

 occupy too much space to transcribe them here. The result 

 was, however, decisive, and one element of the substance 

 decomposed (Sulphate of Magnesia) made its appearance 

 at the surface of the water employed, instead of the usual 

 metallic pole. " As, therefore, the substances evolved in cases 

 of electro-chemical decomposition may be made to appear 

 against air, which according to the common language is not 

 a conductor, nor is decomposed, or against water which is a 

 conductor, and can be decomposed as well as against the 

 metal poles, which are excellent conductors, but undecom- 

 posable, there appears but little reason to consider the phe- 

 nomena generally, as due to the attraction or attractive powers 

 of the latter, when used in the ordinary way, since similar 

 attractions can barely be imagined in the former instances." 



Electro-chemical decomposition is well known to be an 

 effect essentially dependent upon the current of electricity, 

 and various have been the views taken of the nature of this 

 current. Some, with Franklin, consider it one and undivi- 

 sible, others assume it to be compounded of two distinct 

 fluids, the positive and the negative, distinct in their nature, 

 and also distinct in their effects. Thus MM. Riffaut and 

 Chompre for instance, consider the positive and negative cur- 

 rents as each causing decomposition, and maintain that the 



