J 297 



amplitiei'. Following the charactei-islics of lig. 3 we can easily check 

 these functions. When the instantaneons coiidition of the audion is 

 given b}- the |)oint B on one of the characteristics, we see that 

 small fluctuations of the grid-potential involve relatixely large altera- 

 tions of the anode-current. Owing to the linearity of the characteri- 

 stics in the neighbourhood of B, these current-fluctuations are 

 proportional to ihe variations of the giid-potential. Here it is of 

 great importance, that the grid itself receives but little current "it 

 reacts upon tensions" ; \ ery snuxU energies arc therefore suflicient to 

 cause the alterations of the grid-[>otentiaI. At the points A and C 

 the audion acts at the same time as rectifier: to ecpial potential- 

 variations in positive and negative sense correspond different current- 

 variations. 



Of late years it has been found that the audion can perform yet 

 a third function. By a suitable arrangement we can form an 

 unstable system, which gives rise to alternating currents of definite 

 frequency, as in the so-called musical arc. Furthermore it has been 

 found that the unstable connections increase to a high degree the 

 amplifying action, so-called "back-coupling". For a good insight into 

 the use of the audion in wiieless telegraphy it is of importance to 

 understand its generative action. 



The question of the instability 



\t\/\\j 1 — T^~Z^ *^^ electrodynamical systems mecha- 



T %v ^*n^^ i 3 nically at i-est has been studied 



£* 'S i ^72 among others by Simon and his 



il j —j-C pupils^). These investigatoi's have suc- 

 1— 1 1 ceeded in establishing some general 



^\- -i- rules which ai-e easily obtained by 



the aid of a simple diagram (see fig. 4). 



£ is a constant electromotive force, II' is a resistance which is 

 so large, that compared with it the variable resistance of the arc B 

 is negligible, L is a sel (induction, whose resistance is R, (J is a 

 capacity. Owing to the assumption with regard to W, the current 

 i, may be considered as a constant. We assume that the arc-tension 

 is only a function of the arc-cuirent ?'„ -\- i, which is lineai- for 

 small values of /. For this problem we therefore an-ive at linear 

 diff'erential equations, whose general solution consists in a "continuous- 

 current-solution" a)id an "alternating-cui-rent-solution", which may 

 be considered independently of each other. Thus the tensicm e may 



1) Phys. Zeitschr. (1902) HI p. 282. 



(1903) IV p. 366 and p. 737. 

 See also Jahrb. f. Dr. Tel. (1918) 1 p. 16. 



