492 ANNUAL REPOET SMITHSONIAN INSTITUTION, 1915. 



be standard. In Europe, where the conditions of a purely local or 

 at best a very short-distance transmission seemed to prevail, a more 

 microphonic type of instrument was preferred. In both continents 

 there was already a drift away from the earlier forms of fixed or 

 wall type subscribers' instruments, and the desk stand and hand set 

 were beginning to appear. 



• In the central office the myriads of inventions which had been pro- 

 posed and tried out were gradually giving way to apparatus which 

 embodied the best of all that had been suggested. So-called auto- 

 matic systems for central office operation were beginning to be ex- 

 ploited, but the great bulk of telephone engineering was on the 

 basis of manual switchboards, which, to be sure, involved man}' func- 

 tions not manually performed. In fact, nobody as 3'et had had 

 sufficient experience to say definitely where the ultimate development 

 of central office switchboards was likely to lead. 



Such, in brief, was the state of the art in the more important 

 parts of wire telephony 15 years ago. A comparison of the then 

 existing art with that of to-day and a statement of what has been 

 done in the last 10 years can best be made by a statement of the 

 improvements in each principal line. This survey will also permit 

 of making some sort of a hazard as to the future development which 

 may be expected in the field of wire telephony. 



PHANTOM CIRCUITS. 



As is well known, so-called " phantom '" circuits are those metallic 

 telephone circuits which are obtained by combining two ordinary me- 

 tallic telephone circuits in such a way that a third metallic telephone 

 circuit is secured without producing any mutual interference with 

 either of the component circuits and without causing any mutual 

 interference between these component circuits. Considered from the 

 theoretical standpoint alone, the problem is a simple one, since it 

 consists merely in so arranging the circuits and the terminal ap- 

 paratus that the current in one side of the phantom divides equally 

 between the two wires of one phj'^sical circuit, while the current in 

 the other side of the phantom divides equally between the two wires 

 of the second physical circuit, a further proviso being that the ar- 

 rangement shall be suitably balanced both electromagnetically and 

 electrostatically. Under these conditions, there is no tendencj^ for 

 currents in the phantom circuit to produce circulating currents in 

 the component physical circuits, nor do circulating currents in the 

 latter tend to produce a circulating current in the phantom. 



Technically the realization of this theoretical ideal is extremely 

 difficult. Since in the transmission of speech we are dealing with a ^ 

 band of high-frequency alternating currents, it does not suffice to 



