TRANSATLANTIC TELEPHONING. 305 



fectly uniform speed of the generator, in order to preserve a uniform 



frequency of alternations. (This of itself was a very ingenious device, 

 and is worth a brief description. It is shown upon the table in fig! 

 7 (PI. II.) A stretched piano wire was arranged to be kept in vibra- 

 tion by magnetic impulses. It emitted a continuous note which was 

 tuned to unison with a standard tuning fork, seen on the right. A 

 telephone connected in circuit with the generator emitted a note cor- 

 responding to the frequency of the alternating current which the 

 generator produced. Stretched along the front edge of the table was 

 a resistance wire connected with the electric motor which drove the 

 generator. By varying the resistance of this wire, by sliding along 

 it the weight seen resting upon it, the speed of the generator and the 

 number of alterations per second could be varied. An assistant sat 

 at this table with the telephone at his ear, and, by varying this resist- 

 ance, endeavored to keep the telephone note in unison with that of the 

 piano wire.) With all these refinements the measured wave-length 

 still came out 18 miles. All these remeasurements and recomputa- 

 tions took about three weeks, and it began to look as though the 

 reconciliation of theory with experiment was hopeless. At last, after 

 going over the computations many times, it was discovered that a 

 factor, the square root of two, had been overlooked in the denomina- 

 tor of one of the fractions. Dividing the 20 miles by this gave for 

 the theoretical wave length about 18 miles, agreeing with the meas- 

 ured wave length to about one-tenth of 1 per cent. This wave length 

 of 18 miles corresponds to a velocity of transmission of these electric 

 waves of 10,800 miles per second. 



On the table in fig. 8 (PI. 11), at the left, are shown some coils similar 

 in size and shape to those which will be used on land lines and under- 

 ground cables, but wound with different wire. On land lines they 

 would be placed on top of poles at intervals of one to two miles. On 

 ocean cables the coils would be much smaller, and placed only about 

 one-eighth of a mile apart. They would be inclosed in the protecting 

 sheath, and would appear as swellings on the cable. They add but a 

 small fraction of one per cent to the weight, and will not interfere 

 with the laying of the cables. These coils consist of an iron core 

 made of rings punched from sheet iron two one-thousandths of an inch 

 in thickness, packed up to form a hollow cylinder of the proper length. 

 This is wound with the conducting wire by threading it through the 

 center as the figure shows. 



Not only does Dr. Pupin's line serve for telephone transmission. 

 There are systems of multiplex telegraphy that depend upoi the trans- 

 mission of electric waves. They work beautifully in the laboratory 

 over short lines, giving a record on paper of dots and dashes at the 

 rate of 300 words per minute for each instrument, But as soon as 

 the line becomes of any length, the waves begin to lose their character 

 sm 1901 20 



