elevation sites indicate that ca. 65% of cohorts have died after three years (Lesica and Shelly 

 1995). Thus, the test cannot be considered valid unless comparing counts taken at least three 

 years apart. Number of plants in each of the three size classes can also be compared using the 

 same test. 



RESULTS 

 Changes in vegetation 



Canopy cover of grass has increased in all the exclosure plots since the start of the study 

 but not in control plots (Table 1). Other changes are small and inconsistent and likely due to 

 recording error. 



Arabis fecunda density 



Only the 1991-96 and 1993-96 intervals can be tested for statistical significance. The 

 1991-93 interval is too short because well over 50% of plants alive in 1993 were probably also 

 alive in 1991 (Lesica and Shelly 1995). 



At Thompson Corner density of A. fecunda increased in exclosure plots but not in control 

 plots in 1991-96 and the difference was significant for reproductives and total density (Fig. 2). 

 The difference between exclosure and control plots was greater in 1993 than in 1996. At Dewey 

 Cemetery density of A. fecunda increased in both exclosure and control plots in 1993-96, but the 

 increase was significantly more pronounced in the exclosure for rosettes and total density (Fig. 

 2). At Quartz Hill density of A. fecunda was higher in the exclosure in 1991, but higher in the 

 control in 1993. Unfortunately the significance of these changes cannot be tested statistically. 



DISCUSSION 

 In the first years of the study density of A. fecunda became significantly greater in the 

 exclosures at both Thompson Corner and Dewey Cemetery, and most of this difference was due 

 to increases in the number of rosettes, indicating enhanced recruitment following cessation of 

 grazing. At Thompson Corner, the only site with three recording periods, the difference between 

 exclosure and control plots remained relatively stable between 1993 and 1996. On the other 

 hand, density of A. fecunda appeared to decrease in the exclosure plot relative to the control at 

 Quartz Hill. The reason(s) for this difference in response among the three sites is not known. 

 These results suggest that grazing or trampling have an adverse effect on A. fecunda populations 

 at some but not all sites. Grass cover increased in the exclosures at all sites as expected, 

 suggesting that livestock do use these areas. However, mule deer are also present, and there is no 



