Global patterns in the biocontrol efficacy of spiders: A meta‐analysis

AIM: To investigate the overall effect of spiders on pest suppression and crop performance, and to explore the extent to which the biocontrol efficacy of spiders depends on the characteristics of spiders, pests, agroecosystems, climate and geography. LOCATION: Global. TIME PERIOD: 1970–2017. MAJOR TAXA STUDIED: Spiders. METHODS: We performed a meta‐analysis of 58 published studies where we investigated (a) the overall effect of spiders on pest density and crop performance; (b) the extent to which the biocontrol efficacy of spiders depends on the taxonomy of pests (aphids, leafhoppers, beetles, and lepidopteran larvae), the hunting strategy of spiders (hunters, web‐weavers), crop type (vine, cabbage, wheat, rice), climate, and geography. RESULTS: Spiders suppressed agricultural pest insects in 79% of cases. The mean effect size of increased spider density on pest suppression was large (Hedge's d = 0.89; 95% confidence interval (CI₉₅)= 0.66–1.12). Spider pest suppression efficacy slightly increased also with taxonomic diversity (d = 0.33; CI₉₅ = 0.05–0.61). The effects of spiders cascaded down and improved crop performance (d = 2.3, CI₉₅ = 0.70–3.84). The effects of spiders seemed to escalate rather than attenuate down through the agricultural food‐chains (regression slopes > 1). The biocontrol efficacy of spiders was highest in rice followed by grape, cabbage and wheat. The pest suppression efficacy of spiders and the positive effect of spiders on crop yield slightly increased towards the tropics and with mean annual temperature. Spiders suppressed the four pest groups with similar efficacy. MAIN CONCLUSIONS: The meta‐analysis provides strong evidence that spiders are effective in natural pest control and improve crop performance. However, the efficacy of spiders differed among crops. Our study substantiates the few earlier findings that predation pressure and the intensity of trophic cascades in terrestrial ecosystems intensify towards the tropics.