the Elementary Law of Hydrodiffusion. 531 



[centim • I 

 — -t — — was found. 



1 - Q —kt 

 Under these conditions the term ^ e L2 becomes already after 



the lapse of an hour and a half so little (amounting then to 



only about y- 2 Vo) that in comparison with the first term e ^ , 

 it hardly need any longer be taken into consideration. In all 

 the measurements made after this method the observation of 

 the electromotive force began an hour and a half after the in- 

 terruption of the current ; accordingly equation (13) can be 



employed throughout as perfectly exact. The value of ^ k, 



being so considerable, causes the electromotive force to fall 

 very rapidly as the time goes on (from one 10 seconds to an- 

 other the diminution of the electromotive force was distinctly 

 perceptible), and therefore renders a determination of the 

 electromotive force by the compensation method impossible. 

 Nothing remains now but to measure the electromotive force 

 present at a given moment by the first deflection which it pro- 

 duces in a sensitive galvanometer in the circuit of which it is 

 inserted during a half-vibration of the galvanometer-needle. 

 The currents, always lasting 5 seconds only, generated by 

 this electromotive force, never attained intensities exceeding 



Wo o — ~~ m absolute electromagnetic measure. The 



alterations, of the concentrations at the time existing at the 

 electrodes, evoked by these feeble and transient currents turn 

 out therefore so slight that they can be altogether neglected. 

 The first deflection s evoked in a galvanometer by an electro- 

 motive force is proportional to that force ; and when the sen- 

 sitiveness of the galvanometer remains constant and the resist- 

 ance of its circuit invariable, the electromotive forces present 

 at certain moments of time are to one another as the first de- 

 flections which they impart at those moments to the galvano- 

 meter-needle. The connexion between the first deflection s, 

 obtained on a galvanometer when the diffusion-vessel is inserted 

 at the moment t, for the time of half a vibration, in the gal- 

 vanometer-circuit, and the time t has therefore the form 



in which Ci depends on the above-discussed constant C, the 

 resistance of the galvanometer-circuit, and the quality of the 

 galvanometer. 



The quotients of all the pairs of consecutive first deflections 



