SELECTirr. ciKci'irs .ixn sr.cnc ixrnRiERF.Ncr. zt^ 



and 



.'il Jo I Z(Jui) ,- 



W'f now intnxluci- tlu- fmulion /? (w), wliirh will In- ii-rnifd the 

 energ^y spectrum of the random interference, and which is analyticijHy 

 detined !>>■ the equation 



/?(co) = y|F(w)|» (5) 



Dixiding both sides of (H) and (4) l)y T wc get 



/- = I , IT / ^ — -. — ., rfo), (b) 



P=\ T^ I , „, ■ ,,., 1 s(t"a))| • cos a ((i)) ■ rfto. (7) 



^0 lZ(la))|- 



/-. P and /? (oj) become independent of the T provided the epoch is 

 made sufficiently ^reat. /- is the mean square current and P the mean 

 power absorbed by the receiinng branch from the random interference. 



In the applications of the foregoing formulas to the problem under 

 discussion, the mean square current /- of the formula (6) will be 

 taken as the relative measure of interference instead of the mean 

 power P of formula (7). The reason for this is the superior sim- 

 plicity, both as regards interpretation and computation, of formula 

 ((i). The adoption of /- as the criterion of interference ma\' be justified 

 as follows : 



(1) In a great many important cases, including in particular ex- 

 perimental arrangements for the measurement of the static energy 

 spectrum, the receiving device is substantially a pure resistance. In 

 such cases multiplication of I- by a constant gi\'es the actual mean 

 p<iwer P. 



(2) It is often convenient and desirable in comparing selective net- 

 works to have a standard termination and receiving device. A three- 

 element vacuum tube with a pure resistance output impedance sug- 

 gests itself, and for this arrangement formulas ((i) and (7) are equal 

 within a constant. 



(3) We are usually concerned with relative amounts of energy 

 absorlwd from static as compared with that absorbed from signal. 

 X'ariation f)f the receiver impedance from a pure constant resistance 

 would only in the extreme cases affect this ratio to any great extent. 

 In other words, the ratio calculated from formula (6) would not 

 differ greatly from the ratio calculated from (7). 



