606 



russell-dickey. POROSITY AND PERMEABILITY 



[Ch. 32 



grain-size distribution on permeability, holding all other factors con- 

 stant. They used glacial outwash sand, screened into 24 different sizes. 

 These sizes were than recombined in such proportions that they fitted 

 normal probability curves of predetermined mean and standard devia- 

 tion. The porosity in all tests was 40 ± 0.5 percent. Figure 10 shows 

 the relation between permeability and mean diameter, and Figure 11 



5,000 



1,000 



in 



1 500 



100 



50 



10 1 0.5 1 5 



Geometric mean diameter 

 in millimeters 



Fig. 10. Effect of mean grain size 



on permeability of unconsolidated 



sand. (After Krumbein and 



Monk, 1943, p. 159.) 



0.2 0.4 0.6 Q.8 

 Phi standard deviation 



Fig. 11. Permeability as a func- 

 tion of the logarithmic standard 

 deviation. (After Krumbein and 

 Monk, 1943, p. 160.) 



the relation between permeability and the standard deviation as a 

 measure of sorting. All the sands plotted on curve A (Fig. 10) had a 

 standard deviation of 0.04 phi units, and the mean size varied between 

 0.273 and 1.30 millimeters. The permeability varied from 57 to 1,195 

 darcys. Those in curve B had a standard deviation of 0.21 phi units. 

 The wider particle-size distribution resulted in a somewhat lower per- 

 meability for the same mean grain size. The points on the vertical line 

 represent sands having the same mean grain size, 1.0 millimeter, but 

 standard deviations ranging from 0.15 to 0.80. The curves show that 

 the permeability is proportional to the square of the geometric mean 

 diameter and the negative exponential of the logarithmic standard 

 deviation. 



Although these experiments form a very important contribution to 



