Soil arthropods are important components of savannas, contributing to nutrient cycling and thus primary productivity. To investigate how fire and mammalian herbivores influence arthropod food webs, we used two long term herbivore exclosures (ca. 20 y) and burning trials (ca. 5-y return) located along rivers in Kruger National Park, South Africa.

Herbivory and fire will usually have negative effects on soil arthropods although this can be variable, and dependent on multiple aspects of habitat structure and nutrient cycling. We hypothesised that in our sites, the more chronic herbivory disturbance would have stronger and more effects than fire, and that both fire and herbivory would decrease arthropod abundance, biomass and diversity via changes to habitat structure and nutrient cycling.

We used a structural equation model to investigate these mechanisms, and to compare these drivers. This model supported our hypothesis that herbivory had more and stronger effects than fire, largely through indirect flow-on effects. We also found evidence to support a ‘tolerance/avoidance’ hypothesis, in that herbivory increased soil arthropod diversity by decreasing soil nutrients. Herbivores also decreased arthropod biomass and abundance in total and in all trophic groups excluding omnivores.

Fire and herbivory are closely linked, careful consideration should be made when making decisions in the management of either. In some areas either driver may be more dominant, as was the case in our research. Further studies should incorporate a range of fire frequencies and intensities, as well as herbivore types, densities and abundances.

土壤节肢动物是稀树草原的重要组成部分，有助于养分循环，从而提高初级生产力。为了研究火和哺乳动物食草动物如何影响节肢动物食物网，我们使用了位于南非克鲁格国家公园沿河的两次长期食草动物排泄物（大约20年）和燃烧试验（大约5年返回）。

食草和火灾通常会对土壤节肢动物产生负面影响，尽管这是可变的，并且取决于生境结构和养分循环的多个方面。我们假设在我们的站点中，与草火相比，更长期的草食干扰将具有更强的影响，并且火和草食都会通过改变栖息地结构和养分循环而减少节肢动物的丰度，生物量和多样性。

我们使用结构方程模型来研究这些机制，并比较这些驱动因素。该模型支持了我们的假设，即草食比火具有更多和更强的作用，这在很大程度上是通过间接流动效应产生的。我们还发现支持“耐受/避免”假说的证据，因为食草动物通过减少土壤养分来增加土壤节肢动物的多样性。食草动物还减少了杂食动物的节肢动物总生物量和丰富度。

火与草食是紧密联系的，在两者的管理决策中应谨慎考虑。在某些领域，正如我们的研究一样，任何一个驱动因素都可能更为主导。进一步的研究应包括一系列的起火频率和强度，以及食草动物的类型，密度和丰度。