Table 2. Criteria for the selection of variables

among the analyzed 54 sites there are 18 sites which are too affected by human impacts and encroachment that the sites cannot be regarded as be representative for closed forest or they contain mainly dense stands of *Erica* trees*.* These stands are omitted from the calculation of total average value and determination of targets, but these sites are of course included in the presentation of the CFEMI score for all sites (Table 5 and Fig. 2). Hermansen et al (2008) gives a detailed description of field work and results.

Some variables are measured by using numerical data (number of tree species and stems, basal area, tree height, crown width, crown width sum and crown depth), and average value is calculated. Cover of epiphytes is variables estimated by using ordinal (categorical) data (covering of climbers and covering of vascular, lichens and bryophytes), and the average value is calculated from the ordinal values.

The score for each site is calculated as the percentage of the average value for all the nine variables for a certain site compared with the target. Hermansen (2008) contains a complete list of calculated values of variables and score for all sites.
