32 BOTANICAL GAZETTE [JANUARY 
The use of mechanical analysis as a basis for determining the 
moisture retentiveness of a soil is further complicated by the fact 
that soils having a high clay content will show great differences 
in the amount of colloidal material, which greatly affects the mois- 
ture retentiveness. Furthermore, the particles constituting a 
given group may lie much nearer one limit of the group than the 
other, so that a given group does not always have the same prop- 
erties. We are then led to conclude that the particles constituting 
a given group in the mechanical analysis do not always have the 
same moisture retentiveness per unit mass, or that their specific 
retentivity when measured alone is modified to some extent by 
admixture with particles from other groups. 
Brices and McLane,” using the method of least squares, have 
established a relationship between the mechanical composition 
and the moisture equivalent, based upon data covering 104 soil 
types. The resulting probable error of the coefficients in the 
relationship established is +1.7 per cent.* In attempting the 
correlation of the mechanical composition with the non-available 
moisture, we have used the same relative values for the sand, silt, 
and clay coefficients that were obtained by Briccs and McLANE in 
their moisture equivalent correlation. The actual values of the 
coefficients were adjusted to give the best calculated values for 
the wilting coefficient, but the same proportion among the coeffi- 
cients was maintained. The formula used was as follows: 
o.o1r sands+o.12 silt+-o0.57 clay=wilting coefficient. 
In this formula the “sands” refer to the percentage of particles 
ranging from 2 too.o5 mm. in diameter, the “‘silt”’ to particles from 
0.05 to 0.005 mm. in diameter, and the “clay” to particles smaller 
than 0.005 mm. in diameter. In the accompanying table” (table 
V) is given the mechanical composition of each of the soil types, 
the computed value of the wilting coefficient as determined by 
the above formula, the observed value of the wilting coefficient, 
2 Brices, L. J., and McLang, J. W., l.c., 18. 
* This value should not be confused with the probable error of a single determina- 
tion, as given by Briccs and McLANE 
22 We are ee to poiak Bureau of Soils for the mechanical analysis. No 
mechanical analyses w samples nos. 6 and 14. 
| 
. 
: 
p: 
ES 
i 
i! 
