140 nature's flfcfracles. 



re-enforced by some one of these air-cavities, 

 thus giving a louder and more resonant effect 

 to the musical notes. 



Here were two types of receiver, one that 

 would receive one sound as well as another, 

 but none of them so loud, while the other was 

 constructed on the principle of selection and 

 re-enforcement, so that a particular note would 

 be sounded by the box having a cavity corre- 

 sponding to the pitch of the tone, and was 

 much louder and of much better quality than 

 I could get from the diaphragm receiver. One 

 of these receivers pointed to the harmonic 

 telegraph and the other to the speaking tele- 

 phone. I knew that I had a receiver that 

 would reproduce articulate speech or anything 

 else that could be transmitted. 



My first conceptions of an articulate speech- 

 transmitter were somewhat complicated. I 

 conceived of a funnel made of thin metal hav- 

 ing a great number of little riders, insulated 

 from the funnel at one end and resting lightly 

 in contact with the funnel at the other end. 

 These riders were to be made of all sizes and 

 weights so as to be responsive to all rates of 

 vibration. In the light of the present day we 

 know that such an arrangement would have 

 transmitted articulate speech, but perhaps not 

 so well as a single point would do when prop- 

 erly adjusted. My mind clung to this idea 

 till in the fall of 1875, when an observation I 



