212 



BELL SYSTEM TECHNICAL JOURNAL 



except that the corners of the Hps were rounded. The third, Fig. 3>e, 

 had square corners as before, but the sUt was about .1 mm. wider 

 in the middle than at the ends. 



Fig. 3— Glottis Models. 



The resistance R is given as the ratio of the product of pressure 

 and slit area to the linear velocity of flow. Measurements were made 

 in each case through a range of pressures such as to give fluxes through 

 the slit through a range of 50 to 200 cm.^/sec. (Stanley and Sheldon 

 values, see Sci. Am., Dec. 1924) and through a range of slit width W 

 of .01 to .10 cm. The data can be represented approximately in 

 this range for the three slits by the following formula?: 



i? = 3.6 PW'-' X 10-«, 

 i? = 6.1 PW- X 10-«, 



R = 800 /-n^--3 X io-«. 



In these expressions / is the velocity of flow of air through the slit. 

 More careful data taken through a wider range of / and W would 

 undoubtedly have given i? in a power series. 



These formula^ are taken to indicate that the resistance of the 

 actual glottis increases faster than a linear function of / and W due 

 to turbulence and may be represented as a single valued function of 

 either displacement of the vocal cords q-i (or glottis width) or of air 

 velocity as expressed by a Taylor's series as follows: 



R = R, 



, d,R 

 dq-i 



, d,R . 



+ 1 



d^R ., , 2d,fR . 

 dqi" dqodti 



+ 



do'R 

 ~d7r 



ir + etc. (7) 



