junegrass, and Idaho fescue (north aspect) on non- 

 harvested sites. But nitrogen levels of all species re- 

 mained below recommended levels for the nutritional 

 needs of deer (16 percent protein = 2.56 percent N: Halls 

 1970; Verme and UUrey 1972). 



Percentage phosphorus (P) was numerically greater on 

 tree-harvested plots for all species (except squirreltail) 

 and significantly so (p = 0.1) for Sandberg bluegrass 

 and junegrass. Minimum phosphorus requirement for 

 lactating cows of 1,100 lb (500 kg) is 0.28 percent P 

 (National Research Council 1976). This value would be 

 marginally adequate for deer nutritional needs as well 

 (Verme and Ulb-ey 1972). 



Soil Microsite Impact on Species 



We were unable to determine differences (p — 0.1) in 

 percentage nitrogen or percentage phosphorus of grass 

 species growing on different microsites in tree-harvested 

 or nonharvested plots. Our results are at variance with 

 other reports of increased percentage nitrogen levels in 

 grasses under semiarid shrubs (Rickard and others 1973) 

 or mesic tree cover (Holecheck and others 1981). We 

 speculate that on our tree- harvested plots, the increased 

 grass yields (table 1) of the tree-associated microsites 

 diluted nutrient concentrations. On our nonharvested 

 plots, uniform moisture stress (Everett and Sharrow, un- 

 pubhshed) may have limited nutrient uptake and plant 

 growth equally among microsites. 



Composite Forage Response by Soil 

 Microsite 



We found no yield differences among soil microsites on 

 any of the nonharvested plots, and grass yield was not 

 different (p = 0.1) for individual microsites among 

 aspects. Tree dominance was sufficiently intense to mask 

 inherent microsite differences that emerged following 

 tree removal. 



Grass yield was greater on tree-associated microsites 

 (duff and transition) than in interspace on west and 

 north tree-harvested plots. Grass yield was not different 

 among microsites on the south aspect (table 4). Yields of 

 interspace microsites on tree-harvested plots were consis- 

 tently similar to interspace yields on nonharvested plots. 



Composite Forage Response by Aspect 

 and Harvest Treatment 



We caution that because aspect plots were not repU- 

 cated, statistical results apply only to these specific 

 plots. These plots are, however, characteristic of the 

 population of pinyon-jimiper communities from which 

 they were drawn. 



Grass cover increased for 2 years (1979 and 1980) fol- 

 lowing tree harvest on north and west aspects, but the 

 rate of increase declined the next 2 years (fig. 1). Cover 

 on nonharvested plots increased to a lesser extent from 

 1979 to 1983 and may reflect the effect of livestock ex- 

 clusion on the site. The large peak in cover on the west 

 aspect in 1981 reflects the rapid dominance and decline 

 of squirreltail following tree harvest. 



Table 4.— Total grass yield (lb/acre) by soil microsite on 



tree- harvested and nonharvested plots on south, 

 west, and north aspects 



Nonharvest 



Harvest 



Aspect Qi 



I 



South 41. 

 West 31.2bc 

 North 31. 2C 



Lb/acre 



41. ia 31.23 26.ia 39.3a 

 41.1 be i4.3bc 574.7a 334.63 

 43.7bc 33.9c 205.2ab 223.13^ 



41.9a 

 97.3b 

 169.5abc 



= duff, T = transition, I = interspace microsites. 

 ^Superscripts {a,b.c) denote significant (p = 0.05) differences be- 

 tw^een jnicrosites on the same aspect for harvested and nonharvested 

 plots. 



•79 '81 '83 



'79 '81 '83 



SOUTH 



J L 



'79 '81 '83 



YEAR 



Figure 1 .—Percentage grass cover on tree-harvested (H) and 

 nonharvested (N) plots on north, west, and south aspects 

 over time. (*) denotes significant (p = 0.05) differences be- 

 tween harvested and nonharvested plots in the same year. (O) 

 denotes significant differences in cover from the preceding 

 year on the same plot. (S) refers to cover of squirreltail CSitan- 

 ion hystrixj. 



4 



