cover explained by number of passes increased from 52 

 percent (eta- = 0.52) to 72 percent (r^ = 0.72) when 

 pretreatment differences in cover of two resistant plants, 

 Xerophyllum tenax and moss, were statistically con- 

 trolled. In other words, local variation in the distribution 

 of just these two plants accounted for about 40 percent 

 of the variation in cover not explained by trampling 

 intensity. Differences in the distribution of other species 

 probably accounted for much of the rest. 



If this assumption about the cause of erratic relative 

 cover values is accepted, then local variations in species 

 distributions— on a scale of a few meters or less— can 

 account for differences in relative cover of as much as 25 



percent in the ABLA/VACA, ABLA/XETE, and PSME/ 

 SYAL types (fig. 10). Xerophyllum tenax and/or mosses 

 are resistant and patchily distributed in these habitat 

 types. The relative importance of these two plants in the 

 treatment lane can often have a greater influence on 

 cover loss than differences in trampling intensity on the 

 order of many hundred passes. For the ABLA/CLUN, 

 ABLA/CLUN-VACA, and FESC-FEID types, local vari- 

 ation in species distribution is not very important. In 

 the first two types, no resistant species are abundant 

 enough to dampen the effect of a given number of 

 passes. In the grassland, resistant species are so abun- 

 dant and evenly distributed that no lane was inherently 

 more resistant than any other. 



19 



