634 



18 



PIPPAKD AND CHITTY ON EXPERIMENTS ON THE 



When the load was removed and then replaced as at P, Q, etc., the 

 looping was negligible. In the region PQ the load was applied in five 

 increments, but the regions QC and CD each represent a single increment. 

 The displacement at D was smaller than was expected and suggested some 

 unaccountable recovery of the cylinder between C and D. To investigate 

 the reason for that, the cylinder was turned over in the sand bed and a 

 second test was made on the undamaged surface. The load-displacement 

 curve for that test (OB) is shown in Fig. 26, and it consists of a number of 



Fig. 26. 



iOOO 



Tl To ^T" 



RAM DiSPLACCMENT INCHES 



Test B on Cylinder 0. 



distinct sections. OB was a region of linear load-displacement and the 

 strains were practically elastic. At B there was a marked increase in the 

 rate of displacement with load, but that was followed by another region of 

 linear proportionality, BC. The slope in that region, however, was very 

 much flatter than in OB. At C the slope became steeper again and CD was 

 a third region of comparatively linear proportionality. At D the slope 

 changed once more and for the short length DE it was very nearly the 

 same as for BC. At E, linear proportionality finally broke down. The 

 maximum load carried was just under 9,000 lb. and the depth of the 

 permanent dent was about 1-75 cm. 



The explanation of the shape of the curve appears to be as follows. In 

 the region, OB, the central ring and its attached shell resisted the loads 

 without plastic deformation. At B the limit of proportionality of stress 

 was exceeded in the material of the ring and an appreciable dent developed 

 in the shell in the range represented by BC. Load was then transferred 



