CEKT.ILX r.lCrORS AFFECTING TELEGRAPH SPEED 339 



Radio and Carrier Telegraphy. Stiiiicr has also advoratetl " that 

 the combination of sine wave envelopes, unbroken rc\ersals and a 

 three-current-value code i)e aiiplied to radio and carrier telegrapiu'. • 



The advantages and limitations in applyinj{ codes with more than 

 two current values have been fully discussetl alnne, and do not need 

 to be gone into further here. It will be evident thai the combining 

 with these of sine wave envelopes and unbroken re\ crsals does no good. 



The matter of using sine wave en\ elopes was discussed above, the 

 discussion pointing out that waves with sine-wave envelopes are 

 inferior to waves produced by sending rectangular shaped signals 

 through suitable networks, both from the standpoint of the received 

 signals, anil from the standpoint of interference into other circuits. 



The "unbroken re\ersals" bring in again the use of an inert com- 

 ponent. Due to the fundamental difference between cable telegraphy 

 on the one hand, and radio and carrier as usually practised on the 

 other, the inert component in the latter case is somewhat smaller 

 than in the former. In the code advocated by Squicr, the current 

 which may be subtracted without greatly affecting the intelligcnce- 

 carrjing capacity of the signals, is about one unit in value, which is 

 the current corresponding to a space. When this current has been 

 subtracted, the space current is reduced from one unit to zero, the 

 dot current from two units to one, and the dash current from three 

 units to two. This subtraction having been carried out, it is seen 

 that the maximum intelligence-carrying component is approximately 

 two-thirds of the maximum current actually employed. (This 

 figure of two-thirds compares with the figure of one-third for the 

 submarine cable.) 



In the case of radio, the amount of power which must be radiated 

 from the transmitting station is of particular importance. Since 

 with the system advocated by Scjuier about two-thirds of the maximum 

 voltage which is radiated is effective in transmitting intelligence, it 

 is evident that about twice as much power must be radiated as would 

 be requiretl if the inert component were not transmitted. 



Incorrect Assumptions. Two incorrect assumptions are made in 

 the papers referred to and underlie a considerable portion of the 

 arguments advanced in favor of the systems advocated by Squier. 



One of these is that a wave, whose elements are half-cycle sine 

 waves, lends itself to tuning. It is true that in the case of the "un- 

 broken-re\ersals" code a certain amount of tuning can be secured, 

 but this tuning applies only to the inert unvarying component in the 

 wave, which carries no intelligence. The fact, shown in l-"ig. 2, that 



•Squicr, loc. cit., Franklin Inst., Jl. 



