ANALYSIS 



Successive, full-screen, multiple regression sets enabled us to identify combina- 

 tions of environmental factors that substantially affect forage production. Final 

 acceptance of such combinations was conditioned by agreement of the algebraic signs of 

 estimated effects with those of expected^ effects, under least squares fit. In these 

 analyses, average dry weight in pounds (or yield) per acre was the dependent variable, 

 and 37 environmental factors constituted the complete set of independent variables 

 (listed below) : 



A. Time (Number of Years Sinoe CZearautting) 



B. Tree Factors (Regeneration) : 



Density of Lodgepole Pine, determined by actual count 

 Density of All Trees, determined by actual count 

 Percent Canopy Cover of Lodgepole Pine 

 Percent Canopy Cover of All Trees 



Stocking Rate of Lodgepole Pine, expressed as a percent of milacre plots 

 in which a tree occurs 



Stocking Rate of All Tree Species 



Average Height of Lodgepole Pine Reproduction, estimated to nearest 

 1/2 foot 



Average Height of All Trees 



C. Site Index 



D. Topographic Factors: 



Percent Slope 

 Elevation 



Cosine of Azimuth +1, expressed in 2 it radians 



Cosine of Azimuth +sine azimuth +2, expressed in 2 it radians 



E. Soil Factors: 



Each of the following represents two variables, one for the A horizon and one 

 for the B horizon: 



Field Capacity, estimated by moisture retention at 1/3-atmosphere of pres- 

 sure, using standard pressure membrane apparatus and procedures outlined 

 by Richards (1954). 



^The dependent variable is fitted as a function of each of all combinations of 

 specified groups of independent variables via computer. R^, the Coefficient of Deter- 

 mination, is printed out for each regression. 



^"Expected" from prior knowledge. 



2 



