Appendix E. Model statistics for the general linear models with Poisson error term for species richness divided into different groups based on taxonomy and frequency of sexual reproduction.
TABLE E1. Model statistics for the GLMs with poison error term for species richness divided into different groups based on taxonomy and frequency of sexual reproduction. Zvalues and P values are given for the contribution of distance to mature forest edge in a model with site identity as a factor and the covariables (if significantly contributing): frequency of boulders (BOUL), meter of small logs (LOGS), meter of large logs (LOGL), and frequency of dry ground (DRY). Dependent variables were the number of species recorded at different distances from the mature forest edge. N = 23.
Yvariable 
zvalue 
P value 
Residual 
df 
Covariates
included 
All species 
0.58 
0.56 
25 
67 
LOGS 
Mosses 
0.37 
0.71 
9 
68 

Liverworts 
1.2 
0.22 
42 
67 
LOGS 
All species (pine pairs, n = 8) 
0.67 
0.50 
7 
22 

All species (spruce pairs, n = 5) 
0.83 
0.41 
1 
14 

All species (old spruce/young pine pairs, n = 10) 
0.71 
0.48 
18 
29 

Species with significantly positive indicator values for mature forest 
0.63 
0.63 
23 
67 
LOGS 
Mosses with positive indicator values 
0.24 
0.81 
6 
68 

Mosses which seldom produce sporophytes 
0.13 
0.89 
18 
68 

Liverworts with positive indicator values 
1.0 
0.31 
36 
67 
LOGS 
Liverworts which seldom produce sporophytes 
0.75 
0.46 
40 
67 
LOGS 
Liverworts which seldom produce gemmae 
0.91 
0.37 
32 
67 
LOGS 