MINIMUM AUDIBLE INTENSITY OF SOUND. 441 



very faint sound near the threshold of audibility. The inevitable 

 result with persons untrained to use their ears is to indicate an end- 

 point long before it is properly reached. This would give the final 

 averages too great a threshold value or too small a sensitivity. The 

 trend of their curve bears this out, as it lies quite close to the values of 

 Rayleigh and of Lane, whose results, as we have seen, are undoubtedly 

 too large owing to other causes. 



To sum up, all other investigations in this subject may be grouped 

 in three classes, those carried on out-of-doors, those confined in a 

 closed room with neglect of boundary conditions, and those resulting 

 from too great an admixture of unskilled observation. Each of these 

 has inherent sources of error, tending in the first and third instances to 

 produce too large results, and in the second, too small. The errors in 

 the first case are difficult to eliminate; not so in the other two. It has 

 been the object of this research to undertake such a determination in 

 a closed room with two observers of long experience, using an entirely 

 new method which takes account of the interference and reverberation. 

 The results seem to justify the expectations. 



Jefferson Physical Laboratory, 

 Harvard University. 

 September, 1922. 



