tubes, as a sand sample falls through one of the tubes, to determine the 

 fall velocities of the particles. A computer program, SEDANL, then relates 

 fall velocity to hydraulic diameter, using tables which were empirically 

 determined at the Woods Hole Oceanographic Institution (Zeigler and Gill, 

 1959) . The hydraulic diameter of a particle is defined as the diameter 

 of a quartz sphere having the same fall velocity as the fall velocity of 

 the measured particle. 



Figure 10 is an example of a SEDANL printout, using data from sample 

 93. The "cumulative percent" column indicates the percentages of sand 

 grains which are coarser than the indicated size. The "frequency percent" 

 column represents a distribution of particle sizes falling between the 

 indicated size and the next coarser size. Thus, 46.33 percent adjacent to 

 2.0 phi in the frequency percent column indicates that 46.33 percent of 

 the sample was between 2.0 and 1.5 phi. 



Although the SEDANL output (Fig. 10) lists results to the nearest 

 0.01 percent, the actual accuracy is unknown. SEDANL uses the fall veloc- 

 ity of an equivalent quartz sphere to determine the diameter of particles, 

 but possible effects of particle shape, density, and concentration on fall 

 velocity, suggest that the actual physical size of the particles may differ 

 from that indicated. 



4. Sieve Analysis of the Samples . 



Eleven sieves with screen openings, in millimeters (phi in parenthesis), 

 of 2.00 (-1.00), 1.41 (-0.50), 1.00 (0.00), 0.707 (0.50), 0.500 (1.00), 

 0.356 (1.50), 0.250 (2.00), 0.177 (2.50), 0.125 (3.00), 0.088 (3.50), and 

 0.062 (4.00) were used in the sample analysis. The weight of sand collected 

 on each sieve was entered on the sediment analysis form (Fig. 11) and per- 

 centages computed for plotting on cumulative curves (Fig. 12) from which 

 sample mean sizes were graphically computed. For ease in following through 

 this analysis, data from sample 93 have been used as examples on the SEDANL 

 output (Fig. 10), the sieve analysis (Fig. 11), and the computation of the 

 dry-sieve mean (Fig. 12). 



To some extent, size depends on the definition used. The sieve diam- 

 eter of a particle is sometimes defined (Kennedy and Koh, 1961, p. 4233) 

 as the geometric mean of the sieve openings in the last sieve through 

 which the particle passed and the sieve openings of the sieve on which it 

 is retained. Usually, the size of each fraction retained on a sieve is 

 defined as equal to the size of the sieve openings on which it rests, and 

 this retaining sieve definition is used in this report. This is consist- 

 ent with the RSA output, since the SEDANL program interprets settling 

 velocity in size classes equivalent to the retaining sieve definitions, 

 assuming the particles are quartz spheres. This difference between the 

 geometric and retaining sieve definitions of size is equal to one-half 

 the size difference between neighboring sieves. For the 0.5-phi interval 

 sieves in this report, the retaining sieve definition gives sizes 0.25 phi 

 smaller than the geometric mean definition (Fig. 13). Thus, sand estimated 



22 



