We investigated the regeneration requirements of the two dominant tree species in a mixed-oak forest of SW Portugal: Quercus suber (cork oak, evergreen) and Q. faginea (Portuguese oak, winter-deciduous). We hypothesized that (1) recruits of diVerent oak species are diVerentially inXuenced by soil and overstory variables and (2) diVerent factors explain the recruitment occurrence and performance of the same species. We sampled the recruits' height and diameter, and several environmental and forest structure variables of their microsites. Both recruitment occurrence and performance were modeled using generalized linear models. Our Wnal models predicted the probability of occurrence of recruits of Q. faginea and Q. suber with 74 and 82% of accuracy, respectively, and explained about 50% of the variance of their recruitment performance. The recruits of Q. faginea tended to occur in microsites with higher canopy height, canopy density and litter cover, and closer to both conspeciWc and heterospeciWc adults, while the opposite was true for recruits of Q. suber. The performance of recruits of Q. suber was favored by the higher litter cover (a good surrogate for N and P availability), but negatively aVected by the higher litter depth. We concluded the following: (1) there were signiWcantly diVerent regeneration niches for each species (Q. faginea and Q. suber); (2) the factors explaining the recruitment occurrence diVered from those explaining the recruitment performance; (3) the overstory plays a complex and important role in the regeneration process; (4) diVerent variables apparently related with the same environmental factor (e.g. litter cover and litter depth) could aVect recruits in an opposite way; (5) sensitive trade-oVs must be considered for delineating management actions, since they could favor the regeneration of Q. suber, but, at the same time, negatively aVect the regeneration of Q. faginea.