CHEMICAL AFFINITY 163 



angles to it. What a chain of molecules does here, there 

 can be no doubt all the molecules entering into combination 

 would produce in ordinary chemical actions ; but in such 

 cases, the direction of the lines of combination being irregu 

 lar and confused, there is no general resultant by which the 

 magnet can be affected. 



What the exact nature of the transference of chemical 

 power across an electrolyte is, we at present know not, nor 

 can we form any more definite idea of it than that given by 

 the theory of Grotthus. We have no knowledge as to the 

 exact nature of any mode of chemical action, and, for the 

 present must leave it as an obscure action of force, of which 

 future researches may simplify our apprehension. 



We have seen that an equivalent or proportionate electri 

 cal effect is produced by a given amount of chemical action ; 

 if we, in turn, produce heat and magnetism and motion by 

 the electricity resulting from chemical action, we shall be able 

 to measure these forces far more accurately than when they 

 are directly produced, and thus to deduce their equivalent re 

 lation to the initial chemical action. Thus M. Favre, after 

 ascertaining the quantity of heat produced by the oxidation 

 of a quantity of zinc, and finding, as have others, that the 

 heat is the same when evolved from a voltaic battery by the 

 same consumption of zinc forming its positive element, makes 

 the following experiment. 



A voltaic battery and electro-magnet are immersed in cal 

 orimeters, and the heat produced when the connection with 

 the magnet is effected is noted. 



The electro-magnet is then made to raise a weight, and 

 thus perform mechanical work, and the heat produced is 

 again noted. It is found in the latter case that less heat is 

 evolved than in the former, a certain quantity of heat has 

 therefore been replaced by the mechanical work ; and by esti 

 mating the amount of heat subtracted, and the amount of 

 work produced, he deduces the relative equivalent of work to 



