i8o 



Scientific Proceedings (114). 



methods not involving use of a microphone all suffer from a 

 common defect. They are insensitive if the instruments are made 

 of high natural period and if made sensitive they give records so 

 distorted as to be almost valueless except for time-relation studies 

 of the first and second sounds and of very intense murmurs. 



To overcome these difficulties an electromagnetic telephone has 

 been used to convert the sound oscillations into electrical oscilla- 

 tions. This telephone is provided with an air-damped diaphragm 

 of high natural period. A shallow ring fastened to the telephone 

 cap serves as "mouthpiece" and has a small lateral opening to 

 maintain atmospheric pressure within it except for the sudden 

 variations due to sound. This device is applied directly to the 

 chest wall and the air column confined between the diaphragm and 

 the chest wall is very short and has a high natural period. 



The currents produced by the telephone are amplified by a 

 four-stage amplifier used in such a manner as to give practically 

 distortionless amplification. It is necessary to protect the vacuum 

 tubes against extraneous sound and mechanical vibration. It is 

 also necessary to shield all parts of the circuit against electrostatic 

 disturbances and to keep the amplifiers away from sources of 

 magnetic disturbance or else shield against them also. 



The recording instrument is a string galvanometer. The pre- 

 liminary work has been done with the natural period of the gal- 

 vanometer approximately .002 second. This is too slow for best 

 results and an instrument of higher period will be substituted in 

 future work. Between the galvanometer and amplifier a con- 

 denser is placed in series, its capacity being so chosen relative to 

 the other parts of the circuit that the comparatively low-pitched 

 first and second sounds are partially suppressed in the electrical 

 record while pulsations due to the higher pitched murmurs are 

 transmitted undiminished. It is thus possible to produce the 

 necessary amplification of the murmur record without the enor- 

 mous excursions of the galvanometer which the first and second 

 sounds would otherwise produce. A further advantage of this 

 selective sound filter is that the appearance of the record corre- 

 sponds more closely with the impression which the sounds make 

 on the ear. 



The general character of the first, second and third heart sounds 



