38 
to small abscissae in fig. 18 it is seen that in case 3 the energy 
communicated to the plate is about three times or more than the 
corresponding communicated energy in case 2 More damage is thus to be 
expected in case 3 than in case 2, 
99 The upper ourve of fig, 18 is based on the neglect of the energy lost 
by impac as the cavitated water catches up the plate. A more rigorous 
analysis, allowing for such impact losses, suggests that for practical 
cases of largeE , the energy finally transferred to the target is about 
two-thirds of the incident energy. This comm ated energy is still, 
however, considerably greater than that given by the lower ourve (case 2) 
in fige 18 for small abscissae corresponding to lerge€. 
100. The actual mechanian of cavitation as indicated by underwater 
photographs of air-backed plates subjected to small-scale explosions is 
the formation of bubbles in the water which at first grow in size and 
later collapse as the water piles up on the decelerating plate. Whilst 
the preceding simpler picture of the bombardment of ths plate by successive 
layers is thus not strictly correct, the main conclusions already given 
are not invalidated. The essential feature fran the damage aspect is that 
cavitation enables the water to follow-up, the plate and transfer appreciable 
extra energy from the water to the target. It appears to be relatively 
unimportant whether such follow-up takes place by the water splitting into 
layers which bombard the plate or whether it occurs by the water "stretching" 
due to the formation of bubbles which subsequently collapse as the water 
piles up on the plate and moves forward with it 
101. The effect of a plane pulse arriving at an oblique angle instead of at 
normal inoidence oan be taken into account quite simply If % be the 
angle between the pulse front and the plate (mK = 0 for normal incideme), 
then first, equations 30 to 35 can be simply gemralised by writing 
Pp, cosxfor p, andm cosXform In partioular, equation 32 becomes 
Ea ae eee eee eee eee eee eee eee (40) 
Secondly, the energy incident on unit area of the plate is no longer the 
same as the energy per unit area of the pulse front, the two unit areas 
making anglei with one anothers It is the former energy which is relevant 
and equation 38 is replaced by 
P. 008% 
OF ‘ 2pon 
Thus, if € and(lyare defined by equations 40 and 44, the ourves of fig, 18 
hold, in general, for oblique incidence as well as normal incidence, 
Taking into account impact losses, the general result that about two-thirds 
of the incident energy is transferred to the target still holds for 
oblique incidence, the only effect of incidence being the decrease of (2; ,. 
the energy incident per unit area of plate, with increasing obliquity® 
HORmOG- = Goo . HGememoco! .d6om (G4) 
‘a For glanoing inoidence, it is possible that anomalous results ocour 
but such cases are usually unimportant since\ 2; is then so small 
