250 PROCEEDINGS OP THE AMERICAN ACADEMY 



inserted in the cii-cuit; but as there was some uncertainty about its 

 constant and law of deflection which we were prevented from deter- 

 mining by want of time, the method ah-eady stated was relied upon, 

 merely using the figures given by the galvanometer as a check upon 

 the results. The current strength as given in this paper is therefore 

 the streno-th of the steady current which gives the same dynamometer 

 reading as the alternating currents of the telephone. No correction 

 for the self-induction of the dynamometer was attempted, as the range 

 of variation in the readings shows that at the present stage of our 

 research no real increase in accuracy would be secured by doing this. 



The absence of any means of obtaining a definite and constant inten- 

 sity of the voice renders it very diflicult to obtain uniformity in results. 

 Care was taken, however, not to continue the experiment very long at 

 any one time, as the voice of the singer or speaker became wearied, 

 and was less under control. 



In telephonic experiments care must be taken, as already intimated, 

 to avoid the logographic disturbance that arises from the strong air 

 pressure developed on beginning the utterance of different sounds, as 

 the magnitude of the temporary deviation produced by this often far 

 exceeds the permanent deflection which is given by the continuous 

 telephone current. This can be done by raising the voice very gradu- 

 ally ; but a better method, which we have employed, is to interpose a 

 key, which keeps the circuit open until the sound is fully developed. 

 Then, on closing the circuit, the deflection quickly becomes steady, and 

 the disturbance in question is avoided. 



Measurements were made with the Hunning, Fitch, Blake, Edison, 

 and Bell magneto transmitters. The Hunning ti-ansmitter uses carbon 

 in a granulated form, and the Fitch employs two hard carbon disks as 

 electrodes. The structure of the other transmitters is too well known 

 to require any description. 



The vowels a, o, u, i (German sounds) were spoken or sung into 

 the transmitter successively, all at the same pitch, viz. the B of 480 

 vibrations, and with the same intensity as nearly as this could be real- 

 ized. The deflections obtained are given below in full, in order to fur- 

 nish some idea of the degree of uniformity obtained. Several separate 

 series are given for each transmitter. In some cases the loudness of 

 the voice varied somewhat in passing from one series to the next. 



The readings are given in centimeters and tenths. 



