shift somewhat. Consequently, similarity values for 

 treated lanes must be compared with controls. 



In the Pachistima and Carex types, floristic compo- 

 sition did not change, even on the most heavily tram- 

 pled lanes; that is, similarity values on treated lanes 

 were not significantly different from controls (fig. 5). 

 In contrast, 75 passes significantly changed composi- 

 tion in the Valeriana type and 200 passes significantly 

 changed the Phyllodoce type. On the Valeriana type, 

 the mean similarity value after 500 passes was only 

 36 percent; the comparable mean similarity value on 

 controls was 88 percent. After the year of recovery, 



however, mean similarity Values on trampled lanes 

 of all intensities were not significantly different fi"om 

 controls. Shifts in species composition increased dur- 

 ing "recovery" in the Phyllodoce type, where values on 

 the 200-pass and 500-pass lanes were significantly 

 lower than on controls after the year. One year after 

 500 passes, the mean similarity value comparing com- 

 position to that before trampling was only 56 percent, 

 compared to 90 percent on controls. 



The vegetation types where species composition 

 shifted the most were the most diverse, physiogno- 

 mically in the case oi Phyllodoce and floristically in 



100—1 



(A) Pachistima myrsinites 



After trampling 



80 — 



60 — 



40 — 



§ 20-J 

 O 



I 



T 



25 75 200 500 



(0 



After 1 year 



It 



25 75 200 500 



(B) Carex nigricans 



After trampling 



75 200 500 700 



After 1 year 



75 200 500 700 



^ (C) Phyllodoce empetriformis 



o 

 u. 



100- 



80 - 



60' 



40 - 



20 - 



After trampling 



25 75 200 500 



After 1 year 



X 



X 



(D) Valeriana sitchensis 



After trampling 

 T.T 



1 



25 75 200 500 25 75 200 500 



Number of Passes 



Figure 5 — Floristic similarity, comparing composition before and after disturbance, in four veg- 

 etation types in Washington's Cascade fy^ountains. Vertical bars represent 1 standard error 

 above the mean. 



After 1 year 



I 



25 75 200 500 



10 



