Attempts have been made to reconcile the possibility of 

 pressures too small for fatigue damage with the observations 

 that damage does occur. Among these are the suggestions of 

 Petracchi* (mentioned above) of corros:*ve electric currents 

 associated with crystal deformation and of Kyropo.ulos** of 

 initial damage to the intercrystalline cement which may fail 

 at comparatively small pressures. For non- crystalline 

 substances such as lead, however, the hydrodynamic pressures 

 (or jets) are evidently the direct cause of damage. Additional 

 complications arise when cavitation takes place in a corrosive 

 rae'dium, but, in this case, damage is probably accelerated 

 since the action of the corrosive material and of the cavita- 

 tion are mutually assisting. Cavitation in this case acts 

 to assist in exposing corrosive material to the surface while 

 the corrosive material weakens the surface and thus lowers 

 the pressures required for cavitation damage. This action 

 isj in a sense, self- propagating in that as damage proceeds 

 additional surface is exposed to corrosive attack and the 

 roughened surface assists in cavitation lormation and gives 

 rise to crevices where stress concentrations can occur. 



The remarks by Dr. Lerbs are particularly pertinent to 

 the question of propeller scaling discussed above, and he is 

 eminently more qualified by background and experience with 

 the problem of marine propellers to discuss such problems 

 than is the writer. The failure to obtain correlations in 

 Dr. Lerbs'' tests of the type observed by Kermeen is perhaps 

 not too sAirprising since the propeller is a so much more 

 complicated hydrodynamic system than the body of revolution, 

 and detailed control of experiments is a most formidable 

 tssko Furthermore, the smallest propeller size that might be 

 tested at Reyncdds numbers large enough to insure flow 

 similarity may already be too large for effects of the type 

 found by Kermeen to be observed. In Kermeen' s report, it 

 was pointed out that the Re^^molds numbers were large enough so 

 that the form of the pressure distribution under non-cavitating 

 conditions, at least, did not change with changing Reynolds 

 number. This is a difficult condition to attain with model 

 propellers. 



Among many other complications might be mentioned the 

 problem of the blade shapes used in such experiments. For 

 blades with sharp leading edges, it seems likely that effects 



♦Petracchi, G,, "Investigation of Cavitation Corrosion", The 

 Engineering Digest, Vol. X, No. 9, Sept, 19^9, pp. 31^316 



**Kyropoulos, S,, "Cavitation Pressures and Damage", Zelt. fur 

 Angewandte'Math, und Phys , , Vol. II, 1951? PP- ^Oo-^IO. 



10 



