noting that cracking noises were heard during tlie tests of Cylinders 1 and 

 6 as well as during the tests of several other cylinders. In each instance, 

 the pressure at which these noises occurred was well below the maximum 

 pressure applied. It was observed after tests that, in many cases, 

 the cylinder ends had punched through the lead gaskets and were in direct 

 contact with the steel and plates. 



DISCUSSION AND CONCLUSIONS 



Cylinders 7 through 20 were of sufficient length to behave as semi- 

 infinite tubes. Figure 1 compares the collapse pressure of these 



4 5 

 cylinders with the classical theory of Bresse and Bryan for the 



elastic buckling strength of long, unstiffened cylinders under external 

 hydrostatic pressure. The agreement between experiment and theory was 

 extremely good for al.l cylinders except Cylinder 18 when based on min- 

 imum measured thicknesses; see Figure la. UTien the mean measured 

 thickness was used in theoretical calculations the experimental values 

 were on an average about 15 percent below the calculated values; 

 see Figure lb. 



There is no apparent explanation of the relatively low collapse pres- 

 sure of Cylinder 18. After collapse, it was observed that the ends of the 

 cylinder had punched through the lead gasket during the pressure test. 

 However, each of the 1 l/2-in. -diameter cylinders punched through the 

 lead gaskets, and yet Cylinder 18 was the only one of the group which 

 collapsed at a pressure appreciably below that pressure predicted by 

 theory. 



The theory of Bresse and Bryan was modified for the pressures 

 presented in Figure 1 to account for the difference in the radius 

 to the neutraJ- axis of the tube and the outside radius where the 

 pressure load is acting. 



