Schmitz Oswald J, Hambäck Peter A, Beckerman Andrew P
Am Nat. 2000 Feb;155(2):141-153. doi: 10.1086/303311.
We present a quantitative synthesis of trophic cascades in terrestrial systems using data from 41 studies, reporting 60 independent tests. The studies covered a wide range of taxa in various terrestrial systems with varying degrees of species diversity. We quantified the average magnitude of direct effects of carnivores on herbivore prey and indirect effects of carnivores on plants. We examined how the effect magnitudes varied with type of carnivores in the study system, food web diversity, and experimental protocol. A metaanalysis of the data revealed that trophic cascades were common among the studies. Exceptions to this general trend did arise. In some cases, trophic cascades were expected not to occur, and they did not. In other cases, the direct effects of carnivores on herbivores were stronger than the indirect effects of carnivores on plants, indicating that top-down effects attenuated. Top-down effects usually attenuated whenever plants contained antiherbivore defenses or when herbivore species diversity was high. Conclusions about the strength of top-down effects of carnivores varied with the type of carnivore and with the plant-response variable measured. Vertebrate carnivores generally had stronger effects than invertebrate carnivores. Carnivores, in general, had stronger effects when the response was measured as plant damage rather than as plant biomass or plant reproductive output. We caution, therefore, that conclusions about the strength of top-down effects could be an artifact of the plant-response variable measured. We also found that mesocosm experiments generally had weaker effect magnitudes than open-plot field experiments or observational experiments. Trophic cascades in terrestrial systems, although not a universal phenomenon, are a consistent response throughout the published studies reviewed here. Our analysis thus suggests that they occur more frequently in terrestrial systems than currently believed. Moreover, the mechanisms and strengths of top-down effects of carnivores are equivalent to those found in other types of systems (e.g., aquatic environments).
我们利用41项研究的数据对陆地系统中的营养级联进行了定量综合分析，报告了60项独立测试。这些研究涵盖了各种陆地系统中广泛的分类群，物种多样性程度各异。我们量化了食肉动物对食草动物猎物的直接影响以及食肉动物对植物的间接影响的平均强度。我们研究了效应强度如何随研究系统中的食肉动物类型、食物网多样性和实验方案而变化。对数据的荟萃分析表明，营养级联在这些研究中很常见。当然也有偏离这一普遍趋势的情况。在某些情况下，预计不会出现营养级联现象，实际也确实未出现。在其他情况下，食肉动物对食草动物的直接影响强于食肉动物对植物的间接影响，这表明自上而下的效应减弱了。只要植物含有抗食草动物防御机制或者食草动物物种多样性较高，自上而下的效应通常就会减弱。关于食肉动物自上而下效应强度的结论因食肉动物类型以及所测量的植物反应变量而异。脊椎动物食肉动物的影响通常比无脊椎动物食肉动物更强。一般来说，当以植物损害而非植物生物量或植物繁殖产量来衡量反应时，食肉动物的影响更强。因此，我们提醒，关于自上而下效应强度的结论可能是所测量的植物反应变量造成的假象。我们还发现，中型生态系统实验通常比开放式田间实验或观测实验的效应强度更弱。陆地系统中的营养级联虽然不是普遍现象，但在所审查的已发表研究中是一种一致的反应。因此，我们的分析表明，它们在陆地系统中的发生频率比目前认为的更高。此外，食肉动物自上而下效应的机制和强度与在其他类型系统（如水生环境）中发现的相当。
