624 - u - 



values of n to De asssciatea with clumes which fork ana break out strongly on either side of the 

 centre *hi1j ncjative v-ilujs cf n zrc founo when the clume develocs as a central column. 



A mor- soliO t)5sis for th; commrison th.in the jeneral outline of the olume will 00 fcunfl 

 By observing in esch eximole the Jiroctions in which the Dlume velocity is highest, sarticularly 

 in the e.'.rly stajes if Jevelocment. It will oe seen then that oositive values of n are associateO 

 with oluTCS which thrust out most strongly to either side and are relatively weak at the centre. 

 Values of n near zoro ar- found with rljnes which develoo evenly in all directions, negative values 

 of n with cluraes which thrust uowards most strcngly at the centre. 



Some exctctions ionn to occur out these T-.ay De sxolaineo Dy the fact that the surface 

 Dhenomena are three-o imensional and the clune outlines arawn in Figure u may not be in one olane. 

 Two views at right angles are really required to olace the olume contours relative to the dome. 



It has been remarkeo al reaoy that oositivo values of n are found with domes which are 

 relatively high at the rim and negstive values for domes which are relatively high near the centre. 

 It seems cU.?r then that the initial olume is very closely related to the initial dome shape. 

 The 'Jome velocity tenas to be increase-J in the areas where the olume, which aooears later, is 

 most ccncentratsc. It is not intenoe.i to sujigest by this, that the come velocity is simply 

 reinforce:: by the aa-'ec olume velocity. Tne dome velocity is imoartec tc the scray within a 

 few milliseconds :f ^6t;naticn ana the Jistribution cf velocity which Is usee in the deoth 

 calculation is determined from photograbhs taken within the next one or two tenths of a second. 

 Generally thr plume does not aocear until more tnan a secona later. 



To exDl--.in tne relation oetwein the dome and clume it seems necessary to find that a 

 pressure Dulse associ.itod with the early aevelooment cf the plume horns is superimposed on the 

 general dptcnatlcn oul-se. If this is fcund to be true it will indicate that the gas bubble may 

 oe far fr:m soht-^rical in its early stages and that its initial irregularities tend tc persist right 

 uc tj the time ,.f the az:ezr-,r\cs :f the clume. It seems :;robable that the irregularities in the 

 gas bubble will be mucn more marked above the charge; any instability in this direction wculd be 

 :nccur;gsJ ty the hyarjstatic :ressure grajient. As a ccrcllary to this it arrears tc follow 

 that the destructive effect of an explosion may be most marked above the charge. 



Conclusions . 



Cine records obtained in tests with charges exploded at known depths have been analysed ano 

 have shown that the- original dome analysis method of deoth determination will give the deoth fairly 

 closely in most instances, but tnat examoles when the metnoo is 25? in error may occur. 



It has been fcuno, however, tnat the errors- intrccuceo in the Gome analysis methoa are 

 relate; to the snaoe jf the come ana to the early behaviour of the plume. By introducing factors 

 into the dome analysis to allow for dome shape and plume behaviour the depth of detonation can be 

 determined with confidence to within ± 104 anc will in gener^il De much closer than this. 



Some difficulty in applying the carrection factor may arise in practice with bombs which 

 are detonatrd at small depths while the cavity forn-,ed at entry is still open. Special tests 

 would be required to investigate tnis condition but at tne shallow depths involved the uncorrected 

 dome analysis method may De expected to give the aeoths suff Icifntly accurately for most purposes. 



The broad relation of dome and olume velocity and the relation between dome and plume 

 shape which nave been found in the analysis suggest that variations must occur in the pressure 

 pulse which Influence both the oome and the plume. 



