How deterministic or stochastic is tree regeneration in a temperate forest? To answer this question, we analyzed the relative effects of abiotic and biotic factors on seed and seedling demography of tree species in an old-growth temperate forest, Ogawa Forest Reserve, Japan. The working hypothesis was that seed and seedling demography of trees would be controlled more deterministically in comparison with tropical forests and that the magnitude of deterministic effects of the abiotic and biotic factors would depend on regeneration traits relating to seed dispersal ability. We examined 5 abiotic factors (litter and humus cover, vegetation cover, light level, soil moisture, and relative elevation) and 2 species-specific biotic factors (distance from the nearest conspecific adult and density of conspecific seeds or seedlings) to quantify their contribution to the spatial variation in seedling emergence and survival for 18 principal tree species in the community. The results showed that conspecific seed density had a negative effect on seedling emergence for almost all species, as suggested by the Janzen—Connell model. On the other hand, various factors were detected for seedling survival of each species. We also found that regeneration traits relating to seed dispersal ability were significantly correlated with some of the abiotic factors. However, the 7 factors tested were less important than expected; seed and seedling demography appeared to be determined rather stochastically in this stable forest community. Even in a temperate forest, seed and seedling demography appeared to be rather stochastic, and niche partitioning among species during these stages was of limited importance.