Clearcut Herpetofauna 35 



The lack of replication allowed only a rough comparison of 

 capture frequencies, not a statistical test of differences in abundance 

 (Hurlbert 1984). The original intent was to estimate abundances of 

 anurans using mark-recapture methods, but a lack of sufficient recapture 

 prevented this. To make direct comparisons of capture frequencies, 

 we assumed that capture probabilities for each species were the same 

 across habitats. This assumption could not be tested, so habitat 

 preferences should be interpreted with caution. 



Habitat Characteristics 



Sampling — In an attempt to relate herpetofaunal diversity to 

 microhabitat variability, the percent cover of midstory and overstory 

 trees was measured with a spherical densiometer at five points around 

 each drift fence array. One reading was taken in the center and one at 

 a distance of 15 m in each cardinal direction. Percent cover of veg- 

 etation less than 2 m tall was also measured, using a line intercept 

 method (Barbour et al. 1987). There were four 15-m transects at each 

 drift fence array, originating at the middle of the array and extending 

 in each cardinal direction. A complete description of the vegetation 

 of the site is given by Pavel (1993). 



The cross-sectional area of coarse woody debris was measured 

 using the same transects. In this case, the transects were thought of 

 as vertical planes that extended from the ground to the highest piece 

 of downed wood. Each piece of wood greater than 8-cm diameter (at 

 the point of intersection) was measured to the nearest centimeter and 

 classified as "sound" or "rotten" (Brown 1974). 



Surface soil temperatures and soil densities were measured by 

 Perison (In Press) in the control and clearcut areas. All habitat data 

 were analyzed using Mests at the 0.05 alpha level. Each measurement 

 was considered to be independent, and means and standard errors 

 were calculated for each variable. In the clearcut, soil temperature 

 and density were measured in areas with skidder traffic, but not in 

 skidder ruts. These variables were not measured at the edge. 



RESULTS 



Herpetofauna 



Shannon-Weaver Diversity Indices calculated for the drift fence 

 and coverboard captures are given in Table 1. There was no significant 

 difference between clearcut and control diversity indices for either of 

 the two trapping methods. The diversity of the edge drift fences was 

 lower than either the control (t = 8.70, d.f. = 1,538) or clearcut drift 

 fences (t = 9.70, d.f. = 2,209). The reason for this difference was the 



