51 



the breakage depended upon the ability of the liquid to penetrate the specimen. 

 He further conducted cavitation tests with a magnetostriction oscillator and 

 reported similar results — that damage was greater with water and alcohol. 

 There are objections, however, to comparing the damage from the two different 

 experiments. As already pointed out by several writers, the entire life cycle 

 of the transient cavities is well within the minimum time required for pres- 

 sure application before breakage would occur upon pressure release in Poulter's 

 experiments. In connection with Poulter's oscillator experiments, the lesser 

 damage with oils as compared with water might be attributed to the much slower 

 collapse rates which would be associated with the higher viscosity of the oil, 

 and, thus, lower collapse pressures. On the other hand, Poult er showed that 

 a specimen first treated with oil under high pressure and then cavitated in 

 water showed less damage than an untreated specimen. This result might be 

 accounted for on the basis of Petracchi's results- cited below. 



Further evidence that cavitation damage is associated with impact 

 phenomena is the damage produced by small drops impinging on a specimen. 

 Ackeret and de Haller 78 showed that mechanical damage produced by small water 

 drops closely resembles cavitation damage and that the drops need not be trav- 

 elling at exceptionally high speeds to cause damage as long as there were a 

 sufficient number of impacts. 



On the basis of the spherical collapse of cavities, there is suffic- 

 ient theoretical evidence that the pressures developed are high enough for 

 damage. However, Ackeret has pointed out 79 on the basis of Mueller's experi- 

 ments 45 that the collapse is not spherical, but that the "bubbles are col- 

 lapsed by the upstream face being forced against the downstream face," and, 

 further, that such collapse "can be shown to result in a pressure intensity 

 of only a few hundred atmospheres," which are too low to account for most dam- 

 age. Unfortunately, even the much more definitive photographs of Knapp and 

 Hollander 47 do not throw further light on this question since the behavior of 

 the bubbles near minimum radius is somewhat obscured.* Until such work as 

 that of the latter investigators is extended, the following tentative hypothe- 

 sis is proposed as an additional source of damage associated with cavitation: 

 The unsymmetrical collapse of cavitation bubbles in a pressure gradient may 

 give rise to the same type of jet as observed in the steady-state and air- 

 water entry cavities described in the foregoing; in view of Ackeret and de 

 Haller' s experiments, such high speed jets could cause damage. However, the 



*Becent experiments at the David Taylor Model Basin showed clearly, in some cases, collapse of such 

 cavities by contact of upstream and downstream faces and repetition of this manner of collapse over 

 several cycles. 



