NATURE OF SPEECH AND ITS INTERPRETATION 



137 



scale is used, all of the eur\es below the mean are crowded together 

 in the small space between zero and one, and all those above the 

 mean are stretched out from one to infinity. By using a logarithmic 

 plot a symmetrical distribution is obtained. The method of obtain- 



ing these ear sensitivity curves was fully described in a recent paper^ 

 given before the Natural Academy of Sciences. 



It is interesting to note that they indicate that each individual 

 has a hearing characteristic which is quite difTerent from other in- 

 dividuals. Consequently speech sounds differently to difTerent 

 persons. Any distortions of the speech sounds will necessarily 

 affect some persons differently from others. It is evident then that 

 in discussing speech and hearing we must deal w^ith statistical averages. 



Experimental articulation tests showed that the ear interpreted 

 the speech which was transmitted over this high quality system 

 practically as well as that transmitted through the air. Some may 

 wonder why such good quality is not furnished telephone users in 

 commercial practice: Scientifically speaking, it is possible to furnish 

 such quality, but it is evident that the equipment involved is so com- 



2 Fletcher and Wegel, Proc. Nat. Acad. Science, Vol. 8, No. 1, pp. 5-6, Jan., 1922. 



