also within the 95 percent confidence interval and, therefore, is considered 

 within the natural range of variability. 



96. The second interesting but complicating aspect of the experiment 

 was evaluation of a potential difference between basin-predicted sand trans- 

 port rates in the central 24. 8 -cm- wide portion of the tank as compared to the 

 entire width of the tank. A decrease in the sand transport rate might occur 

 near the walls of the tank due to side wall friction; an increase in sand 

 transport could occur in the far side of the tank due to the slight non- 

 uniformity of the flow field established in the hydraulic efficiency experi- 

 ment (see Part III). The differences (Diff.) between center- and full-width 

 basin transport rates were evaluated for two 2.5-min and four 7.5-min basin 

 tests (Table 9). Individual basin collection cloths extending through the 

 first two basins were placed in the center (24.8 cm wide), near, and far 

 (both 21.3 cm wide) "subbasin" sections of the tank. "Center" subbasin fluxes 

 were consistently higher than those calculated for the entire width of the 

 basin, averaging 11.4 percent higher. "Far" subbasin fluxes were also higher 

 than those calculated for the entire width of the basin, in agreement with the 

 trend of the non-uniform lateral flow speed distribution of the tank (Part 

 III). The difference between "center" and total sand fluxes generally 

 increased with midflow speed, as would be expected if the side wall boundary 

 layer increased with flow speed. 



97. The basin threshold power equation was modified to account for the 

 increasing sand flux in the center portion of the tank using Equation 15, with 

 the following additional empirical equation: 



F = F 1 = 2.06(1CT 5 ) (V mid - 28) 3 - 37 



F 2 = 5.03(10' 6 ) (V mid - 28) 378 (18) 



where 



F x = basin flux, V mid < 58 cm/sec (g/cm z /niin) 

 V mid = midflow speed, cm/sec 

 F 2 - basin flux, V mid > 58 cm/sec (g/cm 2 /min) 



69 



