other separation phenomena may be observed when fluid issues from a nozzle with an 

 abrupt expansion. This was discussed in Reference 3 in connection with the discrepancies 

 betweenthe results of Crump in a nozzle with a gradually expanding diffusor and of Numachi 

 in a nozzle with an abrupt expansion. The trends observed by Numachi were confirmed by 

 Crump in a subsequent TMB report^ although the numerical values differ somewhat. The ob- 

 servations of increasing tensions with increasing time of exposure to low pressures in Crump's 

 first nozzle have not as yet been explained. 



EFFECTS OF SCALE ON THE INCEPTION OF CAVITATION 



Only a few remarks were made in Reference 3 on the question of scaling cavitation 

 phenomena. However, this problem is of much importance in view of the need for reliance on 

 results from model experiments. It is desired to mention here only the problem of scaling of 

 the inception of cavitation and, in particular, to review some very interesting results obtained 

 at the California Institute of Technology during the last two years on the effects of geometri- 

 cal scale on inception of cavitation. Before discussing the latter results, however, it seems 

 worthwhile to review the physical picture of the role of nuclei in cavitation inception and to 

 point out the expected consequences of this concept. 



It is now generally accepted that cavitation in a fluid under reduced pressure or boiling 

 in a heated liquid begins with the growth of microscopic nuclei containing gas phase (air or 

 vapor, etc.) It is well known that the absence of such nuclei requires very large forces for 

 rupture since the surface tension forces become very large. Thus in well degassed liquids 

 one expects rupture forces, of the order of those predicted by kinetic theoretical formulations. 

 Experimental evidence has also been obtained that water saturated with air, but denucleated 

 by application of very high press^ires, exhibits high tensile strength (of the order of several 

 hundred atmospheres).^'* Thus, the presence of nuclei is evidently necessary for the incep- 

 tion of cavitation at pressures of the order of vapor pressure. In many engineering applica- 

 tions, it is usually sufficient to assume that cavitation will occur at the vapor pressure corre- 

 sponding to the temperature of the liquid. This assumes that there are sufficient nuclei of 

 large enough initial size to grow to observable size during the time of application of reduced 

 pressure. In supersaturated liquids, it is easy to account for the presence and stability of 

 such nuclei, but in saturated and undersaturated liquids it is necessary to account for such 

 nuclei on the basis that they are stabilized on particles suspended in the liquid (see, e.g.. 

 Reference 3 and the discussion of Reference 10). As a consequence, depending upon the 

 size and number of these nuclei, cavitation may be expected to begin above as well as below 

 the vapor pressure, as was shown in Crump's experiments with sea water and fresh water. A 

 further consequence of this concept is that inception will depend upon the time of exposure 

 of these nuclei to low pressures. Thus, for nuclei of a given size, the longer the time of ex- 

 posure the higher inception pressures that would be expected. 



In addition to the effect of actual time of exposure to low pressure, another factor which 



