The test setup> is shown in Figure 11 being prepared by Messieurs 

 E. J. Winkler and J. J. Jones, technicians in the Structural Research 

 Department. A paper gasket with sealing compound was placed between the 

 lid of the pressure facility and the model to effect a seaL The model was 

 held to the lid of the pressure facility (Fig. 11) by means of a turnbuckle 

 bar (Figs. 7. and 9) until the initial seal between the lid and model was 

 formed. The bar was then loosened before beginning the tests. 



End blocks on the bar were bonded with epoxy to the shell with the 

 test set-up' of Configuration A. This arrangennent proved unsatisfactory so 

 small indentations were made in the model to hold the bar (Fig. 9) for the 

 test of Configuration B. 



The edge of the port window was coated with sealing compound and 

 taped in place for the tests of Configuration A. A small amount of oil seeped 

 between the window and the head until the initial seal was reached. In order 

 to eliminate leakage, the window was bonded to the shell with Duco cement 

 for the tests of Configuration B. 



A Gilmore Strain Gage Plotter (Fig. 1 1) was used to record the strains 

 in the model. The model was cycled between 100 and 1, 600 pipi external 

 pressure several times in order to stabilize the strain gage installations, 

 and then the strains were recorded at each increment of 300 psi from 100 

 to 1, 600 psi, and finally again at 100 psi. For convenience during the tests 



