Indirect interactions and conservation in human-modified environments


Chris R. Dickman, Institute of Wildlife Research, School of Biological Sciences A08, University of Sydney, Sydney NSW 2006, Australia.

Agricultural enterprises usually simplify and fragment native vegetation, rendering the local environment unsuitable for many species that originally occurred there. Some species do not cope with the changed conditions and disappear, whereas others thrive on the elevated productivity of the new monocultures and become pests (Singleton et al., 1999). Such changes in species abundances will often be driven directly by the agricultural onslaught, but Shapira, Sultan & Shanas (2008) show that the indirect effects of agriculture on surviving species may be more pervasive than we have realized.

In essence, Shapira et al. (2008) found that two species of gerbils were less active at sites close to than distant from farms, and also foraged less efficiently near farms on moonlit nights. Ostensibly, these results are not surprising and reflect the expectation that habitat and other resources needed by gerbils are depleted by agricultural activity. However, all sites were located on semi-stable sands with similar shrub cover and access to seeds, rodenticides were not commonly used on the farms and most potential predators were regionally scarce. The major difference between sites was that the red fox Vulpes vulpes was more active and abundant near farmed land. This could be expected to increase the risk of predation for gerbils near farms, thus driving the observed shifts in their behaviour. The authors did not consider the alternative possibility that increased human activity near farms may in itself have affected gerbils (e.g. by increased levels of noise, lighting or movements on access roads), even though their near-farm sites were just 50–200 m from farms. Nonetheless, the effect of foxes is perhaps most plausible; the shifts in gerbil behaviour observed by Shapira et al. (2008) are similar to those exhibited by several other species of rodents when exposed to increased risk of fox predation (Dickman, 1992; Kats & Dill, 1998). If this interpretation is correct, Shapira et al. (2008) have uncovered a fine example of an indirect amensal interaction, in which humans negatively affect gerbils by improving conditions for the fox. It is not clear whether the situation represents an interaction chain or an interaction modification (sensuWootton, 1994), or both, but this could be resolved by further study.

Indirect interactions are being viewed increasingly as critical determinants of distribution, abundance and species composition in many local communities (Bruno, Stachowicz & Bertness, 2003; Glen & Dickman, 2005), especially by conservation managers (Dickman, 2007). Shapira et al. (2008) suggest that agricultural expansion may pose a particularly severe threat in arid environments as the transformation of land is extreme and the secondary and tertiary effects on native biota may be acute. However, their work also suggests that we might look for more such indirect interactions in other human-modified situations.

Firstly, in rangeland environments, top predators are often suppressed to limit their depredatory effects on livestock. Although livestock can damage native vegetation directly by overgrazing, small prey species may be affected less by the loss of vegetation than by meso-predators that increase in abundance in the absence of the top predator. In arid rangelands, the impacts of meso-predators can be exacerbated if they have access to artificial sources of water that have been provided for livestock. The indirect benefits of top predators to small prey have led to calls for reintroductions of wolves Canis lupus to Yellowstone (Smith, Brewster & Bangs, 1999) and of dingoes Canis lupus dingo to the rangelands of Australia (Glen et al., 2007).

Secondly, in urban habitats remnants of native vegetation are often set aside for amenity and contain subsets of the biota that used to be present originally. Large or specialist species may disappear because their spatial or resource requirements are not met in small fragments, but losses of smaller species can be more difficult to explain. However, domestic pets such as cats and dogs are maintained directly by people living near patches of urban vegetation, and these can readily deplete populations of small prey species (Churcher & Lawton, 1987). Similar to the findings of Shapira et al. (2008), humans here are suppressing small prey species indirectly by supporting large populations of predatory pets. Other commensal species including some rodents, marsupials, gulls and corvids have negative effects on taxa as diverse as vascular plants, invertebrates and nesting birds in habitat fragments in towns and rural settings (Belant, 1997; Matthews, Dickman & Major, 1999). These again reflect amensal impacts of humans that are wrought indirectly on these taxa by our improvement of conditions for scavengers and other opportunists.

Indirect interactions are not new. Darwin (1859), for example, famously described how pet cats in English villages kept field mice in check, indirectly reducing their impacts on the combs and nests of bees and in turn facilitating increased pollination and numbers of flowering plants. Shapira et al. (2008) provide us with a valuable and timely reminder that such interactions are likely to be pervasive in human-modified environments where the need for conservation action is often dire. I suspect that our challenge is not to develop more theory – this is already advanced (Dickman, 2006). It is to rediscover our ‘inner naturalist’ and spend the requisite amount of time in the field that is needed to uncover the true range of human impacts on other species.


I thank Guy Cowlishaw for the opportunity to write this commentary.