plate of the same length and wetted area as the model at rest, although occasionally 

 more exact methods taking account of hull curvature and wave profile are tried. C t 

 does of course depend upon F also since both the wave profile and the consequent 

 local velocities near the hull will change with F. The so-called form resistance will 

 be represented in the coefficient C n , as well as the wave resistance, and is strongly 

 dependent on the Reynolds number. Methods are available for estimating this part 

 of C n , but they will not be described here; let us call this estimated coefficient C e . The 

 wave resistance coefficient is then assumed to be given by 



C w = C - C t - C e . (50) 



Experimenters do not, of course, delude themselves concerning the reliability of C w as 

 a coefficient for comparison with a theoretically computed coefficient. However, it is 

 sometimes difficult to know whether a theoretically predicted effect is really absent or 

 is lost in the working-up of experimental data. One can find more comprehensive 

 discussions of these difficulties in various places, for example, Birkhoff, Korvin- 

 Kroukovsky and Kotik [1] and Apukhtin and Voitkunskii [1]. 



In view of the preceding remarks it may seem surprising how good the agree- 

 ment sometimes is between the experimentally estimated and the theoretically computed 

 wave resistance. Figure 1 from a report by Weinblum, Kendrick and M. A. Todd [1] 

 reproduces some values of C w computed from Michell's integral and a faired curve of 

 the "residuary resistance coefficient" C, — C — C t derived from experiments on a 

 towed "friction plane" (here the coefficients are formed by dividing R by V2pc 2 S, S 

 the wetted area; this changes only the vertical scale). One would expect good agree- 

 ment here, if ever, for the towed body approaches in form what one would expect of 

 a "thin ship" (here beam-length ratio = 0.0265), an accurate estimate of C t is possible, 

 and C e should be negligible. And indeed the agreement is good. Comparison of 

 measured and computed wave profiles in the same report shows fairly good agreement 

 at Froude numbers 0.292 and 0.326, but progressively worse as the Froude number 

 increases. Here, however, the comparison is somewhat inconclusive since the calcula- 

 tions were made for a form of infinite draft. 



C r X I0 3 



0.5 



0.0 



0.5 0.6 



FROUDE NUMBER 



I X I0 7 



2 X 10 



3 X I0 7 

 REYNOLDS NUMBER 

 Figure 1. 



117 



4 X I0 7 



