46 Prof. E. Wiedemann on the 



time, and at still higher exhaustion only in a. When this was 

 the case, if the mercury reservoir of the pump was raised, 

 and the gas present in it thus driven again into the discharge- 

 tubes, then the discharge took place again in b and not in a. This 

 behaviour at high pressures, and at medium pressures, follows 

 at once from the views ordinarily held : at high pressures the 

 induction exerted by the electrodes upon each other, and on 

 the layer of gas between them, is of importance ; at medium 

 pressures the resistance of the positive column of light disap- 

 pears in comparison with that of the negative, light ; in fact, 

 at medium pressures the two tubes differ only by having a 

 positive column of different length, which exerts no percep- 

 tible influence upon the potential which is necessary to the 

 discharge, so that the discharge can take place in both tubes 

 at the same time. At the third pressure mentioned above, the 

 dark space surrounding the kathode has, in the tube in which 

 the electrodes are close together, spread out so far that it 

 completely surrounds the anode. This opposes so great a re- 

 sistance to the passage of the electricity, that the discharge 

 traverses only the tube in which the electrodes are far apart, 

 in which the dark kathode-space does not yet reach the posi- 

 tive electrode. Upon still further exhaustion the dark kathode- 

 space surrounds the anode also in the tube b ; and now the 

 discharge must of course take place more easily in the tube 

 in which the electrodes are near each other, since here there 

 is a smaller part of the dark space to be traversed. 



From the experiments described above we draw the follow- 

 ing two important conclusions : — 



(1) The dark kathode-space opposes a very great resistance 

 to the positive discharge. 



(2) The union of the positive and negative electricities takes 

 place in the "glow-rays." 



Hittorf * had already concluded, from experiments with 

 electrodes near together, that resistances existed in the neigh- 

 bourhood of the kathode; but since he did not vary the distance 

 of the electrodes from each other, he was not able to determine 

 the position of the resistance, so as to explain the phenomena. 



3. Heating of the Gas at different Distances from the 

 Kathode. — A whole series of experiments was undertaken, in 

 order to investigate the decrease of heating-effect in passing 

 from the negative pole towards the positive. 



For this purpose a method was employed which has already 

 been used by Gr. Wiedemann and R. Riihlmann | for another 

 purpose. The junction of a thermo-electric element, consisting 



* Pogg. Ann. cxxxvi. pp. 1, 197 (1869). 

 t Pogg. Ann. xlv. p. 35 (1872). 



