such as those discussed above may not "be observed even %\;ith the 

 smallest model. The reason for this suggestion is that local 

 se-:~-aration can easily occur at such a leading ^dge and cavita- 

 tion can then start easily, the separated fluid moving with 

 the blade and allowing sufficient time for inception to occur o 

 It would be of much interest to perform a comparable set of 

 experiments with propellers having well rounded leading edges 

 so that the minimum pressures occur on the blade surface aft 

 of the leading edge=, 



With regard to Dr^ Lerb's comm.ents on the forces developed 

 on cavitating flows with lift, it might be pointed out that 

 exact theoretical results for two-dimensional flows about 

 polygonal obstacles are available (see, e^g, reference 30 of 

 the paper) o However, these commutations only include the 

 formulation of the solutions, and derivations of forces have 

 not as yet been performed. The work of Tulin on lifting 

 surfaces will be of particular value in answering the ques- 

 tion posed by Dr« Lerbs at least for thin sections o For 

 thick sections, ic.e,, sections on which the cavity starts 

 along the profile, it will be necessary to resort to exact 

 theory and include the effects of surface curvature o In 

 principle 5 this is evidently now possible, but the computa- 

 tions will be extremely involved = The high-speed computing 

 machines will be valuable tools in this problem. 



The results of Betz and Walchner merit additional comment, 

 Betz approximation theory gives for the lift coefficient 

 C-r - S"ff< -h cr- where £?( is the angle of attack. The first 

 term on the right represents the lift coefficient of a flat 

 plate at small angles for Kirchhoff floWo In Walchner's experi- 

 ments the separation points were fixed (sharp leading and 

 trailing edges on thin sections) so that Betz' assumption that 

 the pressure over the suction side is essentially fixed is a 

 reasonable one* 'However, the assumption that the pressure 

 distribution on the pressure side is constant seems to have 

 no foundation* Reichardt's assumption (see reference 3) 

 that the ncn=- dimensional distribution remains the same is 

 much more reasonable. Furthermore, Betz" formula which is 

 essentially a flat plate approximation leads to the anomalous 

 result that lift is developed at zero angle of attack o Thus, 

 the good agreement with Walchner's results at very small 

 angles of attack is rather surprising. These remarks are 

 made only to point out the need for rational theory and addi- 

 tional experiments and, thus.* to endorse Dio Lerbs' comments 

 on this problem^ 



The question posed by >3r-o Mandel appears already to have 

 been answered in his observation that separation occurs earlier 

 on the model than on the full scale rudder n If the different 



11 



