Soil arthropod communities were surveyed and compared in litter layers of selectively harvested and nonharvested forests of coast redwood,Sequoia sempervirens, to assess impact of a l5-yr selective timber harvest cycle. Abundance of four guilds (microphytophages, panphytophages, macrophytophages, and predators) were estimated from litter samples taken from three forest classes: uncut old growth, mature second growth last harvested before 1920, and selectively harvested forest sampled 14 yr after harvest. Microphytophage and panphytophage abundance did not vary significantly among forest classes. Macrophytophages showed significantly higher abundance in selectively harvested forest, whereas predators showed significantly reduced abundance in selectively harvested forest. Because no significant differences in macrophytophage and predator abundance were found between old growth and mature second growth classes, observed differences are attributed to the selective timber harvest regime. Furthermore, since selectively harvested forest was sampled for arthropods 14 yr after harvest, it appears that the 15-yr cycle is of insufficient duration to allow full recovery of soil arthropod communities from the effect of selective timber harvest. In addition, it is likely that species are at high risk of extinction, based on the striking observation that an entire order, Diplura, was absent from all samples taken from selectively harvested sites. Possible explanations for observed differences in guild abundance and potential implications for long-term maintenance of soil arthropod diversity are discussed.