172 MM. Elster and G-eitel on 



so that the platinum wire, introduced into the flame, burns for 

 a moment like potassium. As soon as the colouring of the 

 flame is over, the same wire, used as the electrode in air, again 

 gives the usual (mostly somewhat higher) values. 



The apparent diminution of the electromotive force is con- 

 sequently caused by the coating of the electrode which is in 

 air with potassium chloride, corresponding to the experiment 

 recorded in § 6. 



In the following Table, E denotes the electromotive force 

 of an absolutely pure non-luminous gas-flame; E /f , the electro- 

 motive force of the same flame when a bead of potassium 

 chloride is introduced; E', the electromotive force of the flame 

 after removal of the bead, but with the electrodes B and S 

 (fig. 2 a&e) covered with Ka CI. Accordingly E^.— W repre- 

 sents the increase of electromotive force produced by the intro- 

 duction of the Ka CI. 



Series. 



E. 



E . 



E'. 



•o _ -p, Position of the 

 k ' electrodes as in 



I. 



150 



60 



26 



31 [ Fig.2«. 



II. 



171 



111 



80 



III. 



182 



142 



30 



*%} ■*«• 



IV. 



174 



132 



75 



reason 



that the 



values of E 



!*-» 



show so little accord- 



ance lies in the impossibility of making two series of experi- 

 ments under exactly the same conditions. Besides depending 

 on the position of the electrodes, E*— W depends, in a more 

 complicated manner, on this — into which part of the flame 

 the potassium-chloride bead is introduced. Nevertheless the 

 above numbers prove that an increase of the electromotive 

 force is produced by the vaporization of the potassium chloride 

 in the flame. This can also be verified on the employment of 

 liquid electrodes. 



Different flames being employed, the following values were 

 obtained for the electromotive force when platinum electrodes 

 were introduced in the position fig. 2 a: — 



Flame. E. 



(1) That of a Bunsen burner . . 180-200 



(2) A luminous gas-flame . . . 180-200 



(3) Stearine candle-flame .... 180-200 



(4) Spirit-flame 180-200 



(5) Magnesium-flame 20- 30 



(6) Bisulphide-of-carbon flame . . 85-100 

 With magnesium the experimental difficulties are very 



great. Here it could not be ascertained with certainty whether 

 the air-electrode was excited positively, as was at other times 

 the case with all flames, or negatively. 



