High acorn predation limits assisted regeneration in Mediterranean oak woodlands

Regeneration in Mediterranean oak woodlands is increasingly constrained by intense post-dispersal seed predation. Inside an ungulate-exclosure fence in the Castelporziano Presidential Estate (Central Italy), we conducted a factorial field experiment by sowing 1728 acorns of four native oaks ( Quercus cerris, Q. pubescens , Q. ilex , Q. robur ) under two seed-placement treatments (surface vs. buried) across paired microhabitats (clearing vs. understory), while monitoring predators with infrared camera traps. Acorn loss was rapid, with most removal occurring within four days. At the plot level, acorn loss differed among species (Kruskal-Wallis, p = 0.009); Q. ilex showed the highest mean removal, and the only Bonferroni-supported contrast was Q. ilex > Q. robur ( p = 0.010). These two species represent the extremes of the seed-size spectrum, with Q. ilex producing smaller and lighter acorns and Q. robur larger and heavier ones (diameter, length, and mass; ANOVA, all p < 0.0001), suggesting a contribution of acorn morphology to interspecific removal differences. Removal tended to be higher in the understory than in clearings and higher for surface than buried acorns, but plot-paired tests did not support these patterns ( n = 4 plot pairs). Camera trapping revealed strong habitat-diel segregation: wood mice ( Apodemus spp.) were recorded only at night and exclusively in understory plots, whereas Eurasian jays ( Garrulus glandarius ) were strictly diurnal and used both habitats. Behavioural responses shifted with seed placement, with burial increasing exploration and digging and surface acorns eliciting more removal/consumption. Overall, early acorn fate was shaped primarily by predator identity and context-dependent activity, while seed placement modulated accessibility and behaviour rather than consistently reducing losses. These findings indicate that, although ungulate exclosures are essential for preventing browsing and supporting regeneration, they may not be sufficient to reduce early-stage seed predation. Assisted regeneration strategies should therefore combine microhabitat diversification with targeted seed protection during the critical post-dispersal early stages. • Wood mice and jays drove acorn loss within 4 days in Mediterranean oak woodlands. • Q. ilex showed highest removal, Q. robur lowest, along a seed-size gradient. • Buried acorns shifted foraging behaviour, but did not consistently reduce loss. • Predators segregated wood mice nocturnal understory jay diurnal across habitats. • Ungulate fencing alone failed; regeneration needs integrated strategies.