377 



tetryl (table II and figure 2), the def lecticais c alculated from the two- 

 mode Bessel flinction treatment are on the average 40^ higher than the ex- 

 perimental deflecticxis, and those calculated from the parabolic treatment 

 are 3»7^ lower, the theoretical pressure-time cxirves being used in both 

 cases. It must be admitted that the excellent agreement between the para- 

 bolic theory and experiment for small charges is due to an accidental can- 

 cellation of errors. The tv/onnode Bessel-function treatment is in a sense 

 in better agreement with e xperiment in that it inyolves errors which are 

 rather insensitive to variations of the charge weight, whereas additional 

 error of opposite sign which rapidly increases in absolute magnitude as the 

 charge weight becomes quite small causing the cancellation mentioned above. 

 The nature of the additional error peculiar to the parabolic theory will 

 be discussed later. 



The greatest contribution to the discrepancy between theory and 

 experiment is probably due to the idealization of the stress-strain curve 

 and to the use of the value of the yield stress quoted by the supplier in 

 lieu of better information. Another contribution of importance maybe due 

 to the yielding of the diaphragm movuiting which in the calculations was as- 

 sumed to be rigid. The theoretical pressure-time curves, of course, contrib- 

 ute error to the cases in which they are used. The rectification of any of 

 the above-menticned errors would be to tend to lower the calculated values 

 of the deflection, possibly bringing them nearer to the experimental values. 

 There are numerous minor sources of error, among vftiich may be mentioned the 

 neglect of the effect of diaphragm curvature on the diffracted wave, the 

 neglect of all other nonlinear terms in the deflection, and the incompressive 

 approximation in treating the diffracted wave. 



60 



