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



first one L x . (Dimensions of L 2 , axial length 202 cm., inner 

 diameter 4"1 cm., outside diameter 8'0 cm., total number o£ 

 turns 7200.) An iron core consisting of wires (diameter 

 0*9 ram.) was put inside L 2 . Ci = 2 microfarads. The 

 string galvanometer, which was of the Cambridge type, is 

 shown at G, being connected to the mercury cups a and d. 

 The switch abed was provided in order to take oscillo- 

 grams of 



1° anode current : Connexions a-b, c-d, 

 2° grid current : Connexions a-c, b-d, 

 3° momentary sum o£ anode and grid current : 

 Connexions b-c-d. 



The oscillograms of 3° were taken in order to enable us to 

 locate the relative phases of 1° and 2°, the anode and grid 

 currents. 



Several oscillograms were taken of which a few are repro- 

 duced (PL IX. A, B, C, D, E, F). 



A, B, C, D all represent anode currents, obtained with 

 different reaction coupling. 



In A the mutual inductance M was adjusted by sliding 

 the coil L 2 (together with the iron core) a certain distance 

 out of the coil L^, giving a value of M very near bhe critical 

 value, so that the oscillations were just self-maintained. 

 The form of i a is seen here to be very nearly sinusoidal, as 

 one would expect. B was obtained with the reaction 

 coupling slightly stronger. Plere the oscillations are of a 

 flat-topped form, the anode current oscillating between zero 

 and saturation. With the reaction coupling still further in- 

 creased, C was obtained, where again the anode current 

 oscillates between zero and saturation. 



But in the middle of the saturation phase a slight dip is 

 observed which in D (with the reaction coupling as tight as 

 possible) is much more pronounced. This dip, as was pointed 

 out above, is due to the deviation of electrons to the grid. 

 This fact is made clear by E, which represents the grid 

 current corresponding to the anode current D. The rela- 

 tive phases of D and E were obtained with the' aid of F, 

 which represents the momentary sum of anode and grid 

 current. It is clearly seen that the grid current E just fits 

 into the dips of the anode current, the sum F being again 

 of flat-topped form, where the total emission current varies 

 between zero and saturation. 



There is a further interesting point in connexion with the 

 grid oscillogram, Plate IX. E. It is seen that the grid 

 current is zero throughout the negative half of the grid 



