626 

 12 



PIPPARD AND CHITTY ON KXPESIMENTS ON THE 



of the ends was sufficient to prevent splitting as shown at (d) in Fig. 15. 

 The load displacement curves for those specimens are shown in Fig. 17. 



The results of tests on specially prepared specimens under hydrostatic 

 pressure, which will be described later, led to the production of a second 

 paper model by Sir Geoffrey Taylor which showed that the type of failure 

 obtained could also occur without stretching of the shell. As it seemed 

 likely that similar distortions could be produced by static loading dis- 

 tributed along a line parallel to the axis of the specimen, several experi- 



Fig. 18. 



ram displacement ; inches 

 Load-Displacement Cubve fob Line Load on 

 Milk Tin with Lid Soldered in Position. 



ments were made with the load applied through bars of different lengths 

 and diameters, the covers of the tins being soldered into position. 



The general result was the same ; the best comparison with the model 

 was obtained when a rod l|-inch in diameter and IJ-inch long was used, 

 the ends of the rod being rounded to prevent local damage to the thin 

 shell. 



The complete load-displacement curve for that case is shown in Fig. 18 

 and the deformed specimen at {d) in Fig. 19 (facing p. 5) which also includes 

 Taylor's second model. 



Dynamic tests. — Tins of both types were subjected to dynamic loading. 

 They were half bedded in sand as in previous static tests and steel balls of 



