<p>Retention harvesting intentionally retains live trees at harvest time to reconcile timber production and biodiversity conservation. Although retention harvesting has been practiced in forests as an alternative to clearcutting, few studies have assessed canopy arthropod responses to retention harvesting. Thus, we examined the effects of retention harvesting on canopy beetle assemblages in a large-scale field experiment in conifer plantations in Hokkaido, Japan. We collected beetles using modified Malaise traps deployed on forest grounds and in canopies of mature plantations in a pre-harvest survey, and in the canopies of three levels of dispersed retention (10, 50, and 100 trees/ha retention of single broad-leaved trees) and unharvested natural forests in a post-harvest survey one year after harvest. In the pre-harvest survey, species richness and total abundance were higher on forest grounds than in forest canopies. Five species were significantly more abundant in forest canopies than on forest grounds, suggesting that the forest canopy harbors specialized beetle species. In the post-harvest survey, unharvested natural forests had higher species richness than and different species composition from three levels of dispersed retention. Although beetle assemblages differed between the sites, dispersed retention maintained species that had 67% overlaps with those in unharvested natural forests. Considering that clearcutting removes all the canopy from harvested areas, this result demonstrates that dispersed retention can mitigate the impact of clearcutting on canopy beetles. Species richness and composition did not differ among three levels of dispersed retention, suggesting that even 10 trees/ha retention is effective to conserve canopy beetles.</p>