162 The Ohio Journal of Science [Vol. XVIII, No. 5, 



Blake and Sheard showed that the inconstancy of Ko/ k 

 for larger values of y was due to the effect of the phase changes 

 introduced by the factor which they called "0 (t/) " due to the 

 fact that a portion of the wires was at right angles to their main 

 lengths and perhaps to the fact that at points like B and K, 

 Figure 1, there is a change of diameter of the wire where it 

 enters the small rod to which the plate is attached. It is 

 conceivable that the reflection coefficient at such points is not 

 necessarily negligible for the higher harmonics whereas it might 

 be for the lower tones. The greater y is the larger the phase 

 changes for the higher harmonics in comparison with the lower 

 harmonics. These phase changes act in the same direction as 

 the phase change y due to the end capacities, hence the apparent 

 Ko/ K would be larger for the higher tones. 



If we plot the values of Ko/k given in column 23 against y we 

 get Curve I of Figure 9. For the sake of comparison the values 

 Blake and Sheard obtained for plates of the same diameter, 

 but of thickness 4.5 mm. instead of 1 mm. are shown in Curve II. 

 Thus, the extra capacity of the thicker plates for the coupling 

 11 cm. is about 10 per cent. 



Summary. 



We have determined experimentally the relation between 

 the tone intensity and the distance between the plates edge-on 

 and have found the optimum distance z/s for a tone of a given 



frequency number .j to fulfill the empirical relation z/s = Ce — ^ s 

 A theory to account for the tone-intensity as a function of the 

 phase change y and the wave-length X is given in the next 

 paper. 



We are indebted to Mr. B. F. Salisbury for assistance in 

 taking the observations and to the University for further 

 financial assistance. 



Physical Laboratory, Ohio State University. 



