218 Mr. E. V. Appleton and Dr. B. van der Pol on 



resistance in the filament circuit was approximately 

 (40 + R) ohms where K is the filament resistance, so that 

 small variations in R did not affect the filament current 

 appreciably. By altering the position of the anode circuit 

 connexion it was possible to make all the anode current flow 

 in the same direction as the filament current (anode con- 

 nexion at B) or all in the opposite direction (anode connexion 

 at A). 



It was found that the saturation emission current could be 

 very much altered in this way. Fig. 12 shows the actual 

 results obtained. Thus in a typical case, by altering the 

 anode circuit connexion from A to B the emission could be 

 increased by 25 per cent. But an oscillographic experiment 

 carried out with the connexion at B still showed a resultant 

 cooling, although here the Joulean heating effect of the 

 anode current was made as large as it is possible to make it 

 with a triode of the ordinary type. 



Kir. 12. 



Further simple considerations, however, show that with a 

 triode of different design it should be possible to make the 

 Joulean heating of the anode current greater than the cooling 

 effect due to electronic evaporation. Thus in fig. 11, if we 

 deal with a case in which the whole of the thermionic current 

 flows in the same direction as the filament heating current, 

 as was shown possible in the last experiment described, the 

 increase of energy supplied to the filament when the anode 

 circuit is closed can be shown to be approximately i a (2E — <f>) 

 instead of i/O^E — d)) as is found for a short filament with 



