On definite Proportions. 40* 



of proceeding is required. The greater the surface of the 

 plates, the greater must be the extent of the decomposing 

 surface. Each point of it possesses indeed a less intensity 

 of the electrochemical operation than in a smaller surface, 

 but the sum of all the decompositions is greater. There is 

 however for evcrv given magnitude of the plates a maximum 

 of the extent of the decomposing surface, beyond wliich 

 the effect will not be increased. If the two surfaces are 

 not parallel, the intensity of the discharge is increased in 

 those points which are nearest to each other, and at the 

 same time the sum of all the effects is diminished. If the 

 battery acts powerfullv, a vegetation of p(Uassium shoots 

 up from the nc2,ative conductor opposite to the<e points, 

 but does not come into contact with the positive conductor, 

 before the battery is weakened, and the evoluiioii of gas 

 from the positive wire, which has kept it at a distance, is 

 diminished. If now the potassium touches the positive 

 conductor, it discharges the column without any further 

 decomposition, until the potassium is again oxidated, and 

 converted into potass. 



The quicksiU'er acts, in all tjiese experiments, a very re- 

 markable part. Its affinity to the base of the alkali has so 

 great a share in the decomposition, that with the column 

 which I have mentioned, taking every possible precaution, 

 T have never been able to separate the component parts of 

 potass without its assistance. I was very much surprised 

 to find that even in Davy's battery, which was nearly thirty 

 times as powerful as mine, the alkaline earths only afforded 

 their bases distinctly when quicksilver was employed. I 

 was first induced to employ this method in order to collect 

 the very minute portions of metal which are set at liberty 

 at the negative conductor, and dissipated by the evolution 

 of gas: it was some time after this, that I observed, that 

 the quicksilver operated also by its affinity, as will hereafter 

 be shown by a direct experiment. 'When the negative wire 

 is taken out of the quicksilver, the decomposition of the 

 potassium ceases together with the negative state of the 

 quicksilver, and the wire emits, as long as it continues in 

 the alkaline solution, only hydrogen, without the slighteist 

 trace of a separation of potassium. Here then the affinity 

 of hydrogen to oxygen appears to be the weaker, the de- 

 composing o'peraiion of the electricity on the fluid being no 

 longer strengthened by the afliinity of quicksilver for potas- 

 ei u m . 



While the quicksilver contains no more than -j-J-y of 



potassium, it remains fluid; but afterwards, that part of it 



C c 2 which 



