negative phi sizes. In this case, the mean (center of gravity) is more 

 affected by the long, coarse tail than by the position of the median. 

 Positive skewness arises when the curve tails toward the finer, positive 

 phi sizes. 



Skewness differences are frequently used to compare sediment-size dis- 

 tributions; these comparisons can be quite effective, especially when 

 the parameter is used within some multivariate analysis scheme. However, 

 the skewness parameter is not as stable statistically as the mean and 

 sorting parameters and small deviations from normality can result in 

 fairly large skewness variations. Skewness values are not required for 

 the calculations considered in this study. However, a value of plotting 

 cumulative phi-size distribution curves on probability paper is that strong 

 deviations from normality are easily spotted, preventing unwarranted use 

 of methods to be discussed later. 



2. Sieving versus Settling . 



Grain-size frequency data are usually obtained using either sieving or 

 settling techniques. For sieving, a dried sample of known weight is 

 mechanically shaken through a nest of size-graded, wire-mesh sieves and 

 the data obtained are the weights of sample retained on each sieve. These 

 weights are then usually converted to weight percents for calculation pur- 

 poses. For sand, the dried sample weight should not exceed 50 grams for 

 standard 8-inch-diameter sieves and this amount of sample should be obtained 

 by randomly splitting the original bulk sample. The sieves should be 

 graded in equal phi intervals with a preferred interval of one-half phi 

 (Table 3), and the shaking time should be at least 15 minutes. Weight per- 

 cent loss or gain during analysis should not exceed a few percent. Weights 

 need only be measured to the nearest tenth of a gram. Old sieves with 

 screens that are stretched or have holes and clogged sieves are the major 

 sources of analytical errors associated with sieving. 



With settling, a small amount of sample is allowed to fall through a 

 water column of known length and is either caught and weighed on a micro- 

 balance at the bottom of the column, or the change in pressure at the base 

 of the column is measured as the sediment falls through the fluid. Changes 

 in weight or pressure, with respect to time, can be converted to weight 

 percent data using fall velocity equations for the size range of particles 

 involved. Some common problems associated with this technique are: (a) 

 Failure of the fall velocity equations to account for the effects of 

 varied particle shapes and densities, interference of falling particles 

 with each other, and water turbulence; (b) drag interference between the 

 cylinder walls and the settling particles; (c) the divergent difficulties 

 of accurately timing the rapid fall of larger particles and the longtime 

 periods required to settle fine particles; and (d) various problems asso- 

 ciated with introducing the sediment into the fluid. 



Which size-analysis technique is preferable? The settling method 

 where particles interact with a fluid is probably more analogous to real 

 environmental conditions than sieving. Settling techniques are also 



