Quantitative effects of spider predation on arthropod populations in the eastern Sierra

A naturally occurring assemblage of spiders was manipulated in situ to determine the predatory effects on a multiple prey complex in an eastern Sierran mixed-conifer forest. Exclosures were used to vary the densities of spiders within treatment plots. Litter samples were then analysed to determine the relative abundance of epigeal collembolan prey resulting at each level of spider density. Probably as a result of cannabalism and predator-predator interference, exclosures with artificially high densities of spiders returned to normal spider densities within several weeks and relative collembolan abundance was not found to be significantly different from that found in controls. The sifting procedure used to reduce spiders in the spider reduction exclosures altered both prey and spider densities significantly, a procedure from which the collembola populations did not recover. The following year, pitfall trapping was used to remove spiders in a less destructive manner. However, sufficient numbers of spiders were not removed by this method during the study period, and no significant differences in relative collembolan abundance were detected between treatments and controls. Microcosms with controlled initial densities of spiders and collembola failed to maintain treatment integrity over the term of investigation and differences in resulting densities were due to chance alone. A critical review of the exclosure methodology used in community level studies of predation drawn from this experiment and the literature is made, and recommendations for a mesocosm-microcosm approach to future studies are presented.