INTRODUCTION It has been empirically and theoretically demon- strated that the number of dispersed seeds decreases with increasing distance from a parent plant in wind- and bird-dispersed species (Greene & Johnson 1989; Okubo & Levin 1989; Willson 1992; Debussche & Isenmann 1994; Masaki et al. 1994; Wada & Ribbens 1997). This results in the emergence of many seedlings in the vicinity of parent plants (Ribbens et al. 1994; Wada & Ribbens 1997) if seeds do not require particular conditions for germination. For instance, Christian (1991) showed that the germination rate of Euro- pean oak Quercus robur L. acorns is not significantly different in open and shaded sites, and the seed- lings are more abundant under conspecific trees than elsewhere, although the spatial pattern of plants is supposed to change with progressing growth stages due to inter- and intraspecific com- petition, light and nutrient heterogeneity and density- or distance-dependent mortality. However, there is not always a high negative correlation between the number of seedlings and the distance from the parent tree. In a tropical forest, for example, Janzen (1970) and Connell (1971) suggested that density- or distance- dependent predator effects may play a major role in the spacing of trees. That is, enemies seem to be more frequently attracted to the aggrega- tions of seedlings around adult trees (Clark & Clark 1984) and juveniles are more dispersed than expected from the initial seed dispersion (Augspurger 1983, 1984). In a temperate forest, Fox (1977) and Woods (1979) found a reciprocal replacement of sugar maple (Acer saccharum Marsh) and American beech (Fagus grandifolia Ehr.), where the saplings of one species tended to establish themselves under the canopy of the other species, suggesting that conspecific adults have a depres- sive effect on the growth and survival of their offspring (Cypher & Boucher 1982; Woods 1984; Maetô & Fukuyama 1997). Nevertheless, the mechanisms of the enemy effects and reciprocal replacement have not been fully investigated so far. Ecological Research (2000) 15, 219–227 Effects of herbivore-bearing adult trees of the oak Quercus crispula on the survival of their seedlings Naoya Wada, 1 * Masashi Murakami 2 and Kunikichi Yoshida 3 1 Department of Biosphere Science, Faculty of Science, Toyama University, Toyama 930-8555, Japan, 2 Tomakomai Research Station, Hokkaido University Forests, Tomakomai 053-0035, Japan and 3 Tomakomai City Museum, Tomakomai 053-0011, Japan The acorn-producing oak Quercus crispula Blume is shade-tolerant and its seedlings were found to be underdispersed on the forest floor, forming seedling aggregations near fertile canopy trees. However, 84.4% of the seedlings were dead within 2 years, with the highest mortality near the adult canopy trees of Q. crispula. This was mostly due to the high levels of herbivory by lepidopteran larvae. The most serious damage was caused by spring consumers, which fed on the leaves of canopy trees in June and, when the food quality of the leaves declined, fell down to the forest floor to consume the juveniles, including the seedlings. Contrary to the expectation of the herbivore–host arms race theory, most of the herbivores were polyphagous. The most dominant spring consumer Telorta edentata (Leech) was also polyphagous, and its distribution on the forest floor showed a high level of association with the distribution of dead oak seedlings. Key words: herbivory; lepidopteran larvae; oak; polyphagy; Quercus; reciprocal replacement; seedling survival. *Author to whom correspondence should be addressed. Email: wada@sci.toyama-u.ac.jp Received 3 February 2000. Accepted 10 February 2000.