146 Scientific Proceedings (127) 



was a continuous one as distinguished from a variation employ- 

 ing discrete frequencies. 



The method of procedure in the determination of the mini- 

 mum audibility was to vary the frequency back and forth over 

 a given range, the intensity being changed by successive incre- 

 ments. For different intensities near the low limit, different por- 

 tions of the frequency range will in general be audible, and by 

 a suitable choice of intensity increments, the desired curve of 

 relationship between frequency and necessary intensity for audi- 

 tion may be determined as accurately as is desired. Any peaks 

 or dips in the curve which extend over a narrow range of fre- 

 quencies, will be easily detected by this method. This would 

 not be possible without a prohibitive amount of time and work 

 if determinations were made with a series of separate distinct 

 frequencies, as is usually done. The intensity was first reduced 

 by the operator until the sound was insufficient to be audible to 

 the subject at any point in the frequency range being used. The 

 intensity was then doubled and the subject, by varying the fre- 

 quency, determined what portion of the range he could hear. 

 This audible frequency range and the corresponding currents 

 being noted, the intensity was again doubled. This process was 

 continued until all the range was heard. A sufficient number 

 of these frequency ranges, each somewhat less than an octave, 

 were used to cover the frequencies from 320 to 3800 p. p. s. 



Results are given in terms of "logarithmic sensitivity," this 

 being the logarithm of the reciprocal of the necessary intensity 

 at the limit of audition, the intensity being expressed in absolute 

 units, ergs per square centimeter per second. Thus the high 



CHART I 



