METHODS 



Fortv-seven soil samples were collected for laboratory analysis from 28 sites in 

 the southwestern portion of the Idaho Batholith in seven forested experimental water- 

 sheds, that are tributaries of the Middle Fork of tlie Payette River. Elevations range 

 from 1,200 to 2,200 m. on these watersheds, and a large variety of soil families, typi- 

 cal of the southern Batholith, can be sampled here. Great Soil Groups include both 

 Lithic and Entic Cryoborolls and Cryopsamments , and Typic Cryorthents. Families are 

 generally shallow, sandy-skeletal, mixed or coarse-loamy mixed. A-C profiles over 

 weathered quartz monzonite bedrock predominate. 



The 47 soil samples used in deriving the original models include All, A12, AC, and 

 C horizons. N'o B horizons were present at the 28 sampling sites. Samples were oven- 

 dried at 105° C. and then put through a 2-mm. sieve. Textures were run on the less 

 than 2 mm. fraction by the hydrometer method of Day (1965) . Percent fines (silt plus 

 clay) is presented as percent by weight of the <2 mm. fraction. Organic matter analyses 

 were made in a carbon combustion furnace in an oxygen stream, and determined by weight 

 loss. Although this method overestimates organic matter (OM) slightly by dehydration 

 of clays it was found to be more reproducible than wet oxidation methods, probably be- 

 cause of the coarse nature of organic material in the surface horizons. Bulk density 

 samples were taken in the field by use of a 5.71 cm. diameter soil core sampler. Water 

 retention at 15 bars was run on a pressure membrane apparatus, and at 1/3 bar in a 

 porous-plate pressure device. Samples run were from the <2 mm. fraction. Results are 

 expressed on a percent by volume basis. Experience has shown that no attempt at re- 

 packing these loose, weakly aggregated soils best simulates the original bulk density 

 (personal communication with Delon Hampton, Howard Univ. 1971). The change in pore size 

 distribution might be quite drastic using disturbed samples, so we attempted pressure 

 plate analyses of water retention on undisturbed core samples. However, reproducibil- 

 ity was poor, probably because of inadequate contact between the soil and pressure 

 plate. For this reason, we resorted to the use of disturbed samples. This objection 

 is probably small at 15 bars because most water is held by adsorption on particle 

 surfaces. Tliis may be the case for much of the water at 1/3 bar in very coarse soils. 



The soil information thus derived for the samples was then subjected to analysis. 

 Preliminary scanning of alternative two- to four-dimensional data plots resulted in the 

 isolation of three independent variables deemed worthy of inclusion in the final models: 

 percent organic matter, percent fines (F) which is equivalent to 100 minus the percent 

 sand, and bulk density (Pb) . 



Expectation and Data Trends 



Expected curves for water retention represent our attempts at describing the 

 theoretical relationships between water retention and independent variables, the latter 

 being allowed to vary within prescribed limits. These curves are presented here 

 ^gs. 1, 2, and 3) for the three independent variables included in the final models. 



3 



