TEST A: EXPLOSION IN AIR 



Pressures inside gun turrets and ship compartments 

 were small, seldom exceeding a few pounds per square 

 inch. The data obtained were sufficiently extensive to 

 form a firm basis foi- determining where crew members 

 would have been safe from the ]:)last, and where ad- 

 ditional protection would have been needed to insure 

 safety. 



There were, of course, many failures among the 

 thousands of pressure gages used. Some of those which 

 were to be started by radio were started too late as a 

 result of a timing signal failure.* Others operated 

 badly as a result of imperfect adjustment of the gages 

 themselves or of ''black box" starting mechanisms. 

 Some gages found themselves unexpectedly close to the 

 Zeropoint and gave undecipherable off-scale readings; 

 others were unexpectedly far from the Zeropoint and 

 gave no response whatever. Many gages sank with 

 the five mortally- wounded ships they were mounted on ; 

 few of these gages were recovered. Some of the gages 

 were shielded by ships' superstructures and gave ab- 

 normally low readings; in some instances the inter- 



* Arrangements had heen made to send out a master timing sig- 

 nal, which was to start a number of recording instruments an in- 

 stant before the explosion occurred. Actually, the sigyial was sent 

 out a number of seconds too late. This delay was caused partly by 

 imperfect receptiori of a preliminary radio signal and partly by 

 human error. Plans were such that nothing short of this coinci- 

 dence of ill luck could have thrown off the master signal. The great 

 majority of instriiments were not dependent on this signal, and the 

 multiplicity of instrurnents was such that all principal types of 

 information sought were achieved despite inoperativeness of indi- 

 vidual groups of instruments. 



113 



