394 PROCEEDINGS OF THE AMERICAN ACADEMY. 



in the open. Another method, which is perhaps more interesting, 

 would be to study directly the characteristics of the magneto-telephone 

 when used as a generator, by measuring directly the amplitude of 

 vibration of the telephone diaphragm and then measure with the rec- 

 tifier the resulting alternating voltage. 



V. Sensitiveness of the Method. 



The galvanometer employed in the above experiments was not 

 particularly sensitive. Its resistance was decidedly too low and 

 entirely inappreciable in comparison with the resistance of the rec- 

 tifier. A galvanometer of the highest attainable resistance would 

 hardly be appreciable in resistance in comparison with the resistance 

 of the rectifier. Also the transformer employed between the telephone 

 circuit and the rectifier circuit did not have high enough resistance in 

 its secondary. With evident improvements in these respects the sensi- 

 tiveness of the apparatus could be greatly increased, in case one should 

 desire to measure extremely feeble sounds. However, without such 

 improvements the sensitiveness of the apparatus seems to greatly 

 exceed that of any of the physical methods heretofore employed for 

 the measurement of sound. 



For a deflection of .2 millimeters on the galvanometer scale, the 

 power in the galvanometer circuit, calculated from the current-voltage 

 curve of Figure 6, amounted to 1.53 X 10"^ ergs per second, while 

 Lord Rayleigh '^ finds the minimum energy that will affect the human 

 ear to be 1.11 X 10"^ ergs per second, for a pitch of 2730 vibrations 

 per second. That is to say, with the apparatus of the present experi- 

 ments, in order to get .2 mm. deflection it is necessary to develop 

 energy in the galvanometer circuit at about the rate at which energy 

 is received by the human ear at minimum audible intensity. On 

 account of the inefficiency of the magneto-telephone receiver when 

 used as a phono-electric generator, energy at a rate much greater than 

 this is required by the magneto -telephone receiver in order that this 

 amount of power may get into the electric circuits. 



The use of a carbon transmitter in place of the magneto-telephone 

 receiver for the sound receptor, while not so constant as the magneto- 

 telephone, is of course enormously more sensitive. With this arrange- 

 ment the condenser C of Figure 2 was replaced by a battery of four 

 storage cells, and a transformer of lower resistance primary was em- 

 ployed. Preliminary tests showed that the galvanometer would then 



* Lord Rayleigh, Proceedings of the Royal Society, 1877, Vol. 26, p. 248. 



