In this study, lodgepole regeneration reached the early sapling stage on only 

 three clearcuts, the youngest of which was 14 years old. Herbage yields, however, 

 started to decline on 10- and 11-year-old clearcuts where tree heights averaged only 

 2 feet. Apparently, then, lodgepole pine adversely affects herbaceous plants while still 

 in the seedling stage. Moreover, we believe competition for light was negligible on 

 the sampled sites because the maximum tree canopy cover was only 37 percent and averaged 

 a mere 7 percent. After discounting the time effect, tree canopy and understory yield 

 were not significantly correlated. 



Very likely the roots of seedling trees had developed sufficiently by 10 years of 

 age to gain a competitive advantage over herbaceous plants for available moisture. 

 Horton (1958) reported that open-grown lodgepole exhibited about a 1:1 ratio of maximum 

 lateral root length to stem height for the first 20 years. If the 1:1 ratio applies 

 to trees in this study, in which stem heights averaged 2 feet, roots should have radi- 

 ated as far as 2 feet from the stem. Because lodgepole pine density averaged 5,800 stems 

 per acre, then uniformly distributed trees would be spaced about 33 inches apart. Thus, 

 the roots of a given tree would overlap those of its immediate neighbors by about 1 

 foot. Apparently, then, we could apply to our study sites Morton's statement that "in 

 a fully-stocked condition, all of the available rooting space is presumably utilized so 

 that a complex interweaving network of roots occurs." Also, we could assume that this 

 network of tree roots would compete effectively with understory plants for soil moisture. 



Of course, trees were not uniformly distributed on the clearcut sites, but the 

 added effect of stocking rates on accountable variation in forage yields was suffi- 

 ciently high to suggest considerable overlapping of root systems: 



R2 



yield = f (time) 0.38 

 yield = / (time + % stocking of lodgepole) .47 

 yield = / (time + % stocking of all trees) .55 



Further additions of tree density or of tree canopy cover, or both, failed to raise 

 R2 above 0.56. 



SITE INDEX 



Ideally, a prediction model should be based on variables for which measurements 

 are easily obtained. Accordingly, we examined the suitability of site index for pre- 

 dicting understory yield. For this prediction, we used data from the 19 clearcuts for 

 which site index could be approximated from the immediately adjacent uncut stands. 



After fitting the time effect, site index accounted for 1 percent of the variation 

 in yield. As an additive or interacting term with additional variables in the model, 

 the contribution of site index to accountable variation was still too small to warrant 

 its retention in prediction models. Apparently this measure of site potential for tree 

 growth was not closely related to the production of associated understory vegetation. 



5 



