the equation for tliat study area is used. Obviously, the uncertainty of the estimate 

 increases as the difference between the area of application and the original study area 

 increases . 



The accuracy of the estimates can be improved by considering some site factors that 

 do not appear in the equations but still may influence amount of erosion. 



On most of the plots in this study, the cover was fairly uniform in distribution. 

 If cover is not uniformly distributed, as on bunchgrass range, estimates of erosion 

 will probably be low; the size of bare openings can affect erosion significantly 

 (Packer 1951) . 



Basal area of vegetation is another factor that should be considered. Although 

 protection from direct rainfall impact is likely the most important single function of 

 vegetation, the amount of cover in direct contact with the ground is also important. At 

 any given areal cover percentage, those species having a larger basal area will retard 

 overland flow and erosion more than those with a smaller basal area. In this respect, 

 grasses are superior to tall, single-stem forbs . 



On finer textured soil, litter weight apparently exercises some restrictive in- 

 fluence on erosion in addition to that attributable to protection from direct raindrop 

 impact. On sandy soils, litter appears to have no favorable influence beyond that of 

 raindrop interception. 



The erodibility of the litter itself must also be considered. It may be eroded if 

 it consists of small, easily detached fragments. Erosion tends to be greater on sites 

 with easily detached litter than on otherwise simi lar sites with firmly anchored litter. 



Bulk density occasionally influences erosion but these effects are complex. There 

 is an inverse relation between bulk density and infiltration because soil porosity is 

 inversely related to bulk density; this means more runoff and, consequently, more ero- 

 sion on denser soils. There is also an inverse relation between bulk density and organ- 

 ic matter content; and, as noted earlier in this paper, organic matter exercises a 

 variable effect on erosion, depending on soil texture. Under some circumstances, and 

 this was noted particularly on sandy soils, cohesiveness and resistance to detachment 

 are positively related to bulk density; some light fluffy soils are highly erodible. 



While it is obvious that soil erodibility depends on many factors, results of this 

 study suggest that reliable estimates of erodibility may be made on the basis of a few 

 of the most important ones. The equations in this paper give reasonable approximations 

 of the amount of sheet erosion that will occur on any particular site under the impact 

 of a half-hour simulated design rainstorm. These approximations, augmented by visual 

 observations in the field, provide bases for estimating potential sheet erosion on sites 

 similar to those studied. 



17 



