2-week intervals throughout the summer. The twice-per- 

 week (16X) treatments never occurred on consecutive 

 days. Four people, ranging in weight from about 130 lb 

 (60 kg) to 190 lb (85 kg), did the trampling; they all wore 

 lug-soled boots. FoUowup measurements were taken by 

 the end of August. These were identical to those taken 

 initially. Penetration resistance was measured beneath 

 the unconsohdated surface organic litter layer with a 

 pocket soil penetrometer on each subplot. On the FESC- 

 FEID habitat type, it was so difficult to identify indivi- 

 dual grass species after trampling that only the cover of 

 all graminoids combined and of Festuca scabrella were 

 recorded. 



EFFECTS OF TRAMPLING ON 

 VEGETATION COVER 



Change in vegetation cover as a result of trampling 

 was expressed as relative cover. This measure, first sug- 

 gested by Bayfield (1979), utilizes a correction factor to 

 compensate for inherent variability in cover between 

 treatment lanes and account for changes in cover which 

 occur over the trampling period but which are not 

 caused by trampling. Relative cover is the percentage of 

 original cover that survives trampling, adjusted for any 

 changes occurring on control lanes. It is calculated as 

 follows: 



Relative cover = 

 surviving cover on each trampled subplot ^ ]^oo% 

 initial cover on each trampled subplot 



where 



_ mean initial cover on four control subplots 

 ~ mean surviving cover on four control subplots. 

 The effects on relative cover of the two major con- 

 trolled variables— amount of trampling and habitat 

 type— were analyzed in a two-way analysis of variance 

 using a factorial design (Steel and Torrie 1960). It was 

 necessary to utilize an inverse sine transformation of the 

 relative cover data to reduce the heterogeneity of vari- 

 ances. With four subplots in each lane and one replica- 

 tion for most habitat types, there were usually eight 

 observations for each treatment in each habitat type. 



Both of these main effects were significant (p < 0.001), 

 as was the interaction between these two variables. 



Effect of Trampling Frequency 



As reviewed at the beginning of this report, some 

 previous studies have suggested that the effect of a 

 given number of passes varies with trampling frequency. 

 This possibility was examined with an analysis of covari- 

 ance that tested whether or not relative cover differed 

 significantly with trampling frequency, after statistically 

 controlling differences in the total number of passes. 

 Only in the ABLA/VACA type did frequency of tram- 

 pling make a significant difference at the 0.05 level 

 (table 2). Here the most concentrated trampling did the 

 least amount of damage. This would corroborate the con- 

 clusions of Singer (1971) and Hylgaard and Liddle 

 (1981). Nevertheless, the effects are not pronounced. 

 Even for the ABLA/VACA type, the percentage of vari- 

 ation in relative cover (r') explained by both frequency 

 and total number of passes. (41 percent) is only 4 percent 

 greater than the r^ value for just total number of passes. 



Another line of evidence supports the conclusion that 

 the effect of trampling frequency is both inconsistent 

 and not a profound factor. Lanes trampled 300 times, 

 once per season (abbreviated 1—300) had the same total 

 number of passes— 300— as lanes trampled 100 times, 

 three times per season (3 — 100). Table 3 presents relative 

 cover after each of these treatments and two other 

 comparisons. 



For all habitat types combined, there is a slight, but 

 insignificant (t-test, a = 0.05), tendency for concentrated 

 trampling to be more damaging. This tendency, the 

 opposite of what the literature reports, is significant, in 

 all three comparisons, only on the PSME/SYAL type. 

 Most of this difference, however, appears to result from 

 a coincidental greater abundance of resistant species on 

 the lanes that received more frequent trampling. For ex- 

 ample, mean moss cover before trampling was 23 per- 

 cent on the 3 — 100 lane and 2 percent on the 1 — 300 

 lane. Most of the surviving vegetation on the 3 — 100 

 lane was moss. Similar differences in species composition 



Table 2. — Mean relative cover on lanes treated 1, 3, 8, and 16 tinnes during the season. Means 

 were adjusted, through analysis of covariance, to compensate for differences in the 

 total number of passes they received 



Trampling Level 

 frequency of 



Habitat type 



1X 



3X 



8X 



16X 



F 



significance 







Adjusted mean 

 relative cover 

 (percent) 









Abies lasiocarpa/Clintonia uniflora 



43 



34 



27 



28 



0.75 



0.53 



Abies lasiocarpa/Clintonia uniflora- 















Vaccinium caespitosum phase 



49 



46 



37 



58 



1.59 



.20 



Abies lasiocarpa/Vaccinium caespitosum 



73 



50 



41 



50 



4.77 



.004 



Abies lasiocarpa/Xerophyllum tenax 



73 



81 



76 



59 



2.09 



.11 



Pseudotsuga menziesii/Symphoricarpos albus 



36 



47 



32 



48 



2.30 



.08 



Festuca scabrella-F. idahoensis 



99 



100 



100 



100 



.08 



.97 



All habitat types 



70 



68 



60 



64 



1.41 



.24 



12 



