MECHANICS OF THE HUMAN VOICE 245 



methods of computation employed by other investigators were first 

 tested and found to be not fully adequate ; they were therefore 

 somewhat altered and developed and the approved methods were then 

 applied to the plates on hand. 



A large amount of material had been collected during two 

 years' work of the tracing machine. This could not be studied 

 because the measurements and computations required so much time. 

 It was decided to concentrate the eSort at the start on this material. 

 It was found quite impracticable to carry this out at New Haven, 

 one reason being the difficulty of securing intelligent labor at small 

 rates. I therefore located in Munich, hired work rooms, and ob- 

 tained labor by advertisements. Any number of doctors of philoso- 

 phy, university students, and retired army officers could be found 

 at 10 to 15 cents an hour, graduates of the Realgymnasium at 7>^ 

 to 10 cents, scholars from the Realschulen at 7^^ cents. The great 

 number at disposal made it possible to select specially reliable ones. 

 As many as fifteen persons were emploj^ed at one time. The control 

 of these persons, the systematization of the work, and the checking 

 of results w^ere placed in the hands of a retired Prussian major, the 

 Baron von Hagen, at 25 cents per hour. 



The organization was developed with German minuteness. Each 

 piece of work — for example, the analysis of a wave — was kept in a 

 separate little book ; each w^orker had his particular task, and the 

 books were passed in order from one to another. Every item of 

 work was signed by the worker ; when mistakes were found the 

 person at fault was discharged. The separate books — nearly 500 in 

 number — were classified and inserted in larger holders, from which 

 they could be assigned systematically for working up any problem. 

 A card index showed the progress and the material on hand. 



The analysis of a single wave of a vowel usually required a mini- 

 mum of 5 hours' measurement and computation. The study of each 

 hundred waves from various vowels thus required at least 500 hours. 

 The measurement for the melody of a single four line stanza of verse 

 required the time of one person for 6 hours a day for three weeks. 



The work of this measuring and computing bureau was carried on 

 at Munich until October i, 1903. 



Problems Attacked. 



The great amount of material accumulated comprised records of 

 Continuous prose speech and verse in English and German ; they are 



