coordinates of the experimental $, Y curve should be multiplied in order 

 to achieve this approximation express the sought-for values of the para- 

 meters A and B . It is evident from the foregoing that 



1_ 



= 30 

 = 4 



1.5 



(5-3) 



We may conclude, therefore, that for a given set of wave characteristics 

 and grain size of bed material it is possible to determine the bed-load 

 rate q g through equations (4-10), (4-27) and (4-26). 



This is an important result, but by no means the final answer to our 

 problem. Because of symmetry the rate in one direction during the first 

 half of the cycle will be exactly equal to the rate in the opposite di- 

 rection during the second half. The net effect of course will be zero 

 steady movement. The question is now as to how can we make a practical 

 use of the results obtained so far. The argument advanced is that al- 

 though the calculated value of q B does not give a direct measure of the 

 amount of sediment that systematically moves in some direction, it can 

 be used to determine the number of solid particles per unit area of bed 

 surface that at any time are exposed to the transportive effort of any 

 incidental flow no matter how weak; this flow is not strong enough to 

 dislocate the particles from their state of rest, but once it finds them 

 in a state of motion produced by the surface wave it is able to move them 

 forward. The rate of sediment transport per unit width of the bed associ- 

 ated with such a secondary flow will be 



2D 



Q D = \ U(y)c(y)dy (5-4) 



Jo 

 where h is the thickness of the layer within which motion occurs, U(y) 

 expresses the velocity distribution of the flow and c(y) the concentration 

 distribution of the solid particles in motion. The magnitude of h may be 

 taken equal to 2D, a customary approximation for flows of this type. 

 Another approximation very common with steady mean flows is that c(y) 

 within the layer may be considered constant. The velocity distribution 

 U(y) of course remains undefined, but for any particular case is presumed 

 known. Equation (5-4) then becomes 



2D 



Q B = 



U(y) dy (5-5) 



Jo 

 which will provice the desired answer to our problem provided of course 

 that c is known. This is the phase of the problem where the results ob- 

 tained in this investigation find a direct and rather important application. 



20 



