The canopy gaps and the expanded gaps covered 12.8 % and 28.5 % of the study site, respectively. The frequency distribution of the canopy gap sizes corresponded to the negative exponential distribution, with most of the gaps (60 % ) smaller than 100 m2, 34 % between 100 and 300 m2 and only 2 % larger than 500 m2. Canopy gaps smaller than 300 m2 were responsible for 71 % of the total gap area, suggesting a dominance of small and intermediate gaps in this forest. The pattern of the canopy gaps is characterized by a pronounced soft-core effect, and in one part of the study area a tendency towards regularity. Most of the gaps (84 % ) were caused by more than one gapmaker and seemed to be created in more than one disturbance event since, in 72 % of the gaps, gapmakers of at least two different decay classes were found. The disturbance regime was driven by the mortality of sessile oaks, the main gapmaker species, caused mostly by uprooting. The other main canopy tree species was European beech, which died often by snapping. It was less common as gapmaker (20 % ), but was the main gapfiller (91 % ). In contrast sessile oak was almost absent among the gapfillers.
These results suggest that the current small-scale disturbance pattern dominating this old-growth forest is more suitable for shade-tolerant species such as European beech, accentuating the already steady decline of oaks in mixed sessile oak-European beech stands.