QUALITATIVE ANALYSIS. 55 



The apparatus took its origin in a suggestion from Dr. Billings that 

 it would be desirable to have a machine that would test the accuracy 

 of the tracing of a curve by making a gramophone plate from it and com- 

 paring the sound with the original. For tracings like those of my last 

 machine there is no necessity for such a test, as their accuracy can be 

 readily proved in other ways. The test, however, is important for phon- 

 autograph curves (p. 14), which are entitled to little credit until so 

 tested. It is also important for curves that may be obtained where the 

 apparatus has not been carefully tested and cared for. The special impor- 

 tance of the apparatus, however, lies in an utterly new field of research 

 opening up by it, namely, the acoustic analysis of speech sounds. 



