﻿220 L. Scliwcndler — On tlie General Theory of Duplex Telegraph ij. [No. 4, 



But if for instance in Station (I) 



^o' + /5' =^f 

 it follows that 



Hence, substituting its value for o-' and reducing, we get more simply, 



-=^'*-^' >»-> 



but as 



and 

 we have 



m 



N' A' n' 





^^^^r__L__i 



Therefore >S' approximates most rapidlyf towards zero if — - — ■ does, or 



m xj/' 



we have 



A' 



m \p 



should be as small as the circumstances will allow of. 



the first movement of the key (up or down stroke) a force is stored up in a spring 

 before the contacts are changed, which force finally causes the change in these contacts ; 

 for this reason the two principal contacts of the key co-exist only for an almost 

 infinitesimal time, the length of which is moreover independent of the signalling speed. 



Thus for this key tv + / ^ j8 is fulfilled in all positions except one, when it is — — — , 



but for such a short time that the error cannot have any disturbing influence what- 

 ever. 



* \p' is the proportion of the total current arriving at point 1 Fig. 1, which passes 

 off through the instrument (/' when the key of Station (I) is at rest. Then \p', being a 

 function of a', h', d', (j\ is also a function of/'. ^' is the proportion of the total 

 current arriving at point 1 Fig. 1, which passes through the instrument g* when the 

 key of Station (I) is sending, thus, besides being a function of (i\ b', d' and ^', it is a 

 function of w* -|- /3' insead of /', and as <p' and ^' are otherwise quite similar functions 

 they become identical if we make 



/' = w;' -I- fi' 

 E' ¥ 



t — p can never become zero, but should on the contrary be as large as possible, 



A' 



and, therefore, S' can only approximate towards zero by -"-j-r^ becoming as small as 



possible. 



