The importance of different fractions of coarse wood debris (cwd) for species diversity of wood-inhabiting fungi was investigated in near-natural Danish beech stands. Species number per tree increased significantly with increasing tree size, pointing out large trees to be most valuable for fungal diversity if single samples are compared. Rarefaction curves, evaluating the importance of different cwd fractions in a cumulative space, revealed a different and more complex picture. Rarefaction curves based on wood volume showed small trees and branches to host more species per volume unit than larger trees and logs, respectively, while snags were the most species-poor fraction. Surface-based curves showed species density to be rather similar among cwd types, though species density still decreased slightly with tree size. These results are interpreted to reflect a combination of two factors: firstly, small diameter cwd represent a larger surface area per volume and hence more space for fungal sporocarps, than large diameter cwd. This ‘surface area factor’ explains the high degree of similarity of the surface-based rarefaction curves. Secondly, a collection of small diameter cwd involves more separate units than an equal volume of large diameter cwd, and represents thereby more individual cases of fungal infection and, most likely, more variation in environmental conditions. The effect of this ‘number of item factor’ is reflected in the slightly increasing species density per surface area with decreasing tree size. Richness patterns of red-listed and non-red-listed species were found to be strikingly similar across cwd types, and a general preference for large logs among red-listed species was hence absent. An individual look at the most frequent encountered red-listed species revealed substrate preference patterns to occur in three of six species, of which one, the heart-rot agent Ischnoderma resinosum , preferred large logs. Based on these results, it would be obvious to conclude that local fungal species diversity is most efficiently increased in managed forests if small diameter cwd is prioritised for natural decay. However, small diameter wood appear to be unable to support heart-rot agents and other species depending on a long and diverse infection history and thus the integrity of saproxylic communities may be seriously undermined if only small diameter cwd is left for decay. Therefore, we strongly recommend that whole, naturally dead trees, representing the full range of cwd habitats, are prioritised for natural decay in managed forests whenever possible.