The short-term effect of windthrow gap size and isolation for saproxylic beetle abundance, species richness and assemblage composition was studied in managed oak-hornbeam forest in France. The main gap effects did not increase with gap area. Nonmetric multi-dimensional scaling (NMDS) and ANOSIM tests showed that all gaps, even small, differed from closed-canopy controls in terms of saproxylic species composition. However, only mid-size gaps were distinct from large and especially small gaps. The highest Bray-Curtis dissimilarity occurred between mid-size gaps and closed-canopy controls, and not between large gaps and controls. Partly in agreement with the resource concentration hypothesis, floricolous beetles were more abundant in mid-size and large gaps than in small ones. However, pioneer xylophagous species did not increase and late-successional xylophagous species did not decrease in abundance with increasing gap area. A peak in cumulative and mean local species richness was measured in midsize gaps, but assemblage composition in mid-size gaps was not intermediate between small and large gaps, which does not fully comply with the intermediate disturbance hypothesis. From IndVal results, no strictly characteristic species were detected in small gaps, but several ones in mid-sized and large gaps. On the whole, scolytids and pioneer xylophagous beetles were favoured in gaps by an increase in felling density in the 78 and 314 ha surrounding landscapes, respectively. Individual species responses in occurrence or abundance were varied. As the faunistic peculiarity in the gaps was mainly confined to mid- and large-sized gaps, retaining uncleared gaps above 0.5 ha in size is recommended to improve saproxylic gamma diversity in forests.