mixed for 15 minutes. The sample next was wet sieved InaU.S. Standard 230 mesh 

 (4 phi) sieve. The size fraction coarser than 4 ft was ovendried and then sieved again 

 for 15 minutes using a nest of sieves {openings: 2 mm, 1 mm, 0.5 mm, 250^, 125^, 

 and 62.5m) in an American Instrument shaker. The size fraction finer than 4 ^ was 

 placed in a cylinder of water or a solution of either sodium hexametaphosphate or 

 sodium metaphosphate (10 g/l), and allowed to stand for several hours. Usually the 

 samples in a pipette analysis, which in general follows the procedure of Krumbein and 

 Pettijohn (1938, p. 166-172), were selected only at settlement times corresponding to 

 4, 6, and 9 phi . A pipette analysis was not made if the sample contained 85 percent 

 or more sand-size material . Results of the grain-size analysis were graphed to show a 

 cumulative frequency distribution by weight, from which the phi median diameter, 

 ^50' was selected. The few grain-size analyses made in the BUDOCKS Soil Labora- 

 tory in general follow ASTM designation D422-54T (ASTM, 1958, p. 1119-1129), 

 except that the sample was not dried prior to testing. Several identical test samples 

 were processed by each laboratory for a check of accuracy and precision; differences 

 In the percentage of sand, silt, clay, and median diameter of the samples were 

 negligible. 



Practically all of the samples analyzed were sufficiently fine grained to make 

 the 75th or 84th percentiles unobtainable by 9 $. Consequently, St was impossible 

 to compute either the Trask (1932, p. 70-72) sorting coefficient or the phi deviation 

 measure of Inman (1952). 8n some instances, even the 50th percentile had to be 

 estimated from an extrapolation of the cumulative curve. Converting phi units to 

 microns was facilitated using a conversion table (Page, 1955). 



Results — Cores from Areas C and D and core F 6 were predominantly composed 

 of clayey silt-size particles with more than 10 percent sand-size material (Fig. 3). 

 Cores from Areas A, 8, E, and F were predominantly composed of both silty clay- 

 and clayey silt-size particles with less than 10 percent sand-size material . Area G 

 and H cores almost entirely consisted of silty clay with less than 5 percent sand-size 

 grains. 



The correlation between porosity and grain median diameter Is shown as a band 

 with an eye-fitted mean line when void ratio is related to median diameter in phi 

 units (Fig. 5). This relationship is much less obvious when porosity instead of void 

 ratio is related to grain median diameter, and when samples from other areas 

 (Shumway, 1960; Sutton and others, 1957; and Trask, 1953) also are plotted (Fig. 6). 

 (In converting water contents to porosity — ■ Table 16 in Trask 1953, a grain density of 

 2.7 was assumed; a density of 2.6 instead of 2.7 will change the porosity less than one 

 percent) . The range of values shown suggests that when more figures(such as shown In a 



3 



Ratio of volume of voids to the volume of solids , 



