and the relation 



2L = 21 x H 



Z 7T Z 



is here essentially variable , principally by z_ being so. 



H 

 If , by simple change of variable, we adopt as abscissas the 

 values of the batter t/z instead of the arbitrary t/2L we obtain the 

 graph of Figure 2', as ordered and acceptable as that of Figure 1' . 



This very interesting result shows once again that the funda- 

 mental variable for the study of the breaking or reflection of the 

 waves must be always the batter of the slope t/z, or its inverse the 

 slope, and not t/2L. The data corresponding to the new experimenta- 

 tion, done for these discontinuous slopes shown in the report 

 presented in Grenoble, are the following: 



2L = 5.7; 2L =4.0; H= 0.316 m.j 

 H 2h 



from which is deduced: 



2L - 0.316 x 5.7 = 1.80 m t ; h = 1.80 = 0.0225 m.j 



2 x 40. 



K = Cth H = 1.166: T = In . L . K = 0.58 sees. 



l VT 



In the experiments the relation z/h had the values: 



z_ = lj z_ = 0.45; z_ = 0.27 

 H H H 



For z/h = 1 we have the case of continuous batters whose experi- 

 mentation has been utilized here again, and the corresponding point 

 P will be approximated much as that obtained previously. 



For; 



z_ = 0.4-6; z = 0.45 x 0.316 =» 0.14-2 m. >2h = 0.04.5 

 H 



and we obtain the limiting slope: 



4 0.0225 - 1 



0.58 V9.81 3 



that is, the batter: 



t = 3 



z 



10 



