6. Coolwater Ridge, Nezperce National Forest, central Idaho. Vegetation on this 

 deteriorated subalpine range is predominantly low-value forbs . The granite-derived 

 soils are sandy loam and loam. Plots were at about 6,000 feet. 



7. Trinity Mountains Boise National Forest, southern Idaho. Study plots were 

 located in large and small openings in coniferous forest. The granitic soils, typical 

 of much of the Idaho batholith, are sandy loams and loamy sands. Average elevation of 

 plots was about 7,000 feet. 



MEASUREMENTS 



sheet erosion. — Simulated rain was applied to the 20- by 30.5-inch plots at a 

 constant intensity of 5 inches per hour for 30 minutes, using the rainfall simulator 

 described by Dortignac (1951). The raindrops produced by this simulator tend to be 

 larger than those of natural high-intensity storms but possess less impact energy than 

 natural rain because of their lower impact velocity. All water running off each plot 

 was collected and the suspended sediment allowed to settle. This sediment, plus that 

 deposited in the runoff-collecting trough at the bottom of the plot frame, was ovendried 

 and weighed. 



Initial soil moisture content .- -Immediate \y prior to the application of simulated 

 rain, two 240 cc. soil samples were obtained adjacent to each plot in the surface 2 

 inches of soil. These samples were ovendried to determine moisture content. To ob- 

 tain a wider range of initial moisture conditions, half of the plots at each study site 

 were prewet the day before the simulated rain test by applying 0.5 inch of simulated 

 rain during a 15-minute period. 



Protective coyer. - -Density and composition of cover on each plot were measured 

 with a point frame (Levy and Madden 1933), using first strikes of 100 mechanically 

 spaced pins to determine the proportions of the soil surface protected from direct rain- 

 drop impact by plant species, litter, or stone. One or 2 days after the application 

 of simulated rain, all vegetation and litter were removed from the plot, allowed to 

 air-dry at least 2 weeks, and then weighed. 



Soil properties .- -Tvio days after application of the simulated rain, soil core sam- 

 ples were taken at the following depths: to 1 inch, 1 to 2 inches, 2 to 4 inches, and 

 4 to 6 inches. Capillary porosity and bulk density of these soil cores were measured 

 by the tension table method and subsequent ovendrying (Leamer and Shaw 1941). Soil 

 organic matter contents at 0- to 1-inch and 1- to 2-inch depths were determined by 

 the dichromate method (Peech, Alexander, Dean, and Reed 1947). Particle size distribu- 

 tion of the surface 1 inch of soil was measured by the hydrometer method fBouyoucos 

 1962). Soil aggregation in the surface inch was measured by Middleton's (1930) method. 

 In addition, Yoder's (1936) wet-sieving method was used to determine size distribution 

 of water-stable aggregates in the surface inch of soil on the Davis County and Montana 

 plots . 



ANALYSES 



The data of this study were analyzed by multiple regression techniques. In all 

 cases, the dependent variable (y) was the common logarithm of ovendry weight (pounds per 

 milacre) of soil and organic material washed from the plot, including that deposited in 

 the collector trough at the bottom of the plot frame. Logarithms were used because the 

 erosion data were not normally distributed but were skewed to the right, that is, a 

 large majority of the values were less than the mean. The logaritlimical ly transformed 

 data approached a normal distribution. 



For each of the seven study areas, all measured site factors were evaluated for 

 their contribution to explained variance by stepwise multiple regression analyses. 



4 



