Artificial illumination reduces bait-take by small rainforest mammals

Abstract

Small mammals often moderate their foraging behaviour in response to cues indicating a high local predation risk. We assessed the ability of cues associated with a high predation risk to reduce the consumption of bait by non-target small mammal species in a tropical rainforest, without inhibiting bait-take by feral pigs (Sus scrofa). The illumination of feeding stations with a low power light source caused small mammals to reduce their foraging intensity on sunflower seeds mixed through sand by 25% (P < 0.001) and on unprocessed corn-based feral pig bait by 80% (P < 0.001). Illumination also reduced the intensity with which small mammals fed on commercially manufactured baits (odds ratio = 6.17, P = 0.009). Illumination did not cause pigs to reduce their intake of corn bait (P = 0.43). Neither pig nor dingo (Canis lupus dingo) vocalisations had any detectable effect on the foraging intensity of small mammals (P > 0.05 for all treatments). We conclude that site illumination was an effective method of selectively deterring small mammals from consuming feral pig baits in our study region, but had no effect on consumption of those baits by pigs.

Introduction

Animals must balance a variety of risks and opportunities when deciding where and for how long to forage (Lima et al., 1985, Searle et al., 2008). Animals will often restrict their foraging intensity at a particular site to balance the rate at which they gain energy or other resources against the site-specific risk of predation, energetic costs of foraging, and costs of missed foraging opportunities (Bednekoff, 2007). Desert rodents, for example, have been shown to forgo foraging opportunities in response to natural and artificial illumination that renders them vulnerable to predation by visually oriented predators (Brown et al., 1988, Kotler et al., 1991), and also in response to acoustic or visual cues that indicate the immediate presence of a specific predator (Abramsky et al., 1996, Hendrie et al., 1998). Several other species are known to respond to the vocalisations of specific predators by adopting defensive behaviours at the expense of foraging (e.g. Hauser and Wrangham, 1990, Rainey et al., 2004, Randler, 2006).

Acute behavioural responses to environmental cues that indicate an enhanced risk of predation have been exploited by resource managers attempting to manipulate the foraging behaviour of animals; for example, chemical repellents which emit sulphurous odours similar to secretions or excretions deposited in the environment by predators can reduce browsing by herbivores on forestry seedlings (Conover, 1984, Wagner and Nolte, 2001). It is presumed that these odours are perceived by herbivores as indicators of the potential presence of a predator, and hence a heightened local predation risk. Behavioural responses to these predation risk signals can include avoidance, inhibition of foraging and other non-defensive activities, and reduced activity levels in general (Apfelbach et al., 2005).

Wildlife deterrents, based on predation risk avoidance, generally target broad groups of animals, with the aim of protecting a resource from any foraging activity (e.g. Morgan and Woolhouse, 1997); however, it is sometimes useful to selectively deter some species from feeding without inhibiting the activity of others. Deterrents based on taste and other sensations relating to food sampling have been proposed as selective antifeedants for addition to stock feeds and animal control baits (e.g. Mason et al., 1985); however, the selective action of these deterrents may be most effective when applied to phylogenetically distant groups of animals because they often rely on deeply rooted innate differences, such as those between the trigeminal chemosensory systems of birds and mammals (e.g. Mason et al., 1989). Potential selective deterrents based on olfactory or gustatory sensations may be less effective within broad animal groups (e.g. eutherian omnivores, Campbell and Long, 2009).

To date, no studies have sought to exploit predation risk avoidance behaviour as a foundation for selective deterrents in animals. Selective deterrence might be possible because cues that convey information about predation risk can elicit different responses in different species, even those that are phylogenetically close. For example, kangaroo rats (Dipodomys spp.) are more tolerant of open microhabitats in American deserts than other heteromyid rodents, probably because they are morphologically better equipped to detect and avoid predators (Kotler, 1984, Brown et al., 1988). Furthermore, some species appear to be able to distinguish between vocalisations of different potential predators, and adopt appropriate levels of defensive behaviour based on differences in the perceived intensity of risk (e.g. Hauser and Wrangham, 1990, Rainey et al., 2004).

Here we evaluate the ability of cues associated with heightened predation risk for small rainforest mammals to selectively reduce the consumption of feral pig (Sus scrofa) baits without inhibiting bait-take by pigs. A systematic effort to develop target-specific feral pig baiting protocols for the tropical rainforests of north-eastern Australia hypothesised that small mammals might be selectively deterred from feeding on toxic pig baits by augmenting baiting sites with cues indicative of an enhanced predation risk for small mammals (Bengsen et al., 2008). Specifically, we hypothesised that small rainforest mammals would recognise site illumination and the vocalisations of pigs as indicators of increased predation risk, and would respond by reducing their foraging intensity at sites augmented with these cues, whereas feral pigs would not. Site illumination was expected to render small mammals vulnerable to predation by visual predators, such as owls (Strigiformes), which are not known predators of feral pigs (Schodde and Tidemann, 1986). Predator vocalisations were expected to indicate the immediate presence of a specific predator, and hence an immediate predation risk. We tested the first part of this hypothesis by comparing the proportions of feed left uneaten at sites exposed to different cues. We then tested whether any deterrent effects apparent in the first experiment were effective when applied to two types of feral pig bait, and whether these deterrents had any adverse effects on bait-take by pigs.