Predator odours attract other predators, creating an olfactory web of information

Abstract

Many studies have reported the aversive reactions of prey towards a predator's odour signals (e.g. urine marks), a behaviour widely thought to reduce the risk of predation by the predator. However, because odour signals persist in the environment, they are vulnerable to exploitation and eavesdropping by predators, prey and conspecifics. As such, scent patches created by one species might attract other species interested in information about their enemies. We studied this phenomenon by examining red fox investigation of odours from conspecifics and competing species in order to understand what prey are responding to when avoiding the odours of a predator. Surprisingly, foxes showed limited interest in conspecific odours but were highly interested in the odours of their competitors (wild dogs and feral cats), suggesting that odours are likely to play an important role in mediating competitive interactions. Importantly, our results identify that simple, dyadic interpretations of prey responses to a predator odour (i.e. cat odour = risk of cat encounter = fear of cats) can no longer be assumed in ecological or psychology research. Instead, interactions mediated by olfactory cues are more complex than previously thought and are likely to form a complicated olfactory web of interactions.

1. Introduction

The exploitation of signals (e.g. visual, auditory, olfactory) by predators, prey, competitors and conspecifics is a well-established phenomenon in ecology (reviewed in [1,2]). In particular, numerous studies have investigated the response of prey to predator odours, and prey behaviours such as avoidance and reduction in feeding, activity or reproduction are thought to reduce predation risk from the donor species [3,4]. Predator odours are also widely used in the psychology literature to understand fear, and many such studies also assume that fear responses in laboratory animals after exposure to a predator's odour are owing to fear of that predator [4]. However, anti-predator responses to the odours of a predator species will only decrease predation risk if the odour represents a hotspot of that predator's activity. This tight spatial association has been assumed in many studies, given that scent marks made to demarcate territory ownership will attract passing conspecifics [5], but it is rarely investigated.

Olfactory communication in this way is traditionally considered to be a dyadic interaction, with one signaller (e.g. predator) and one receiver (e.g. prey). This simple approach is unlikely to hold however, given the properties of olfactory cues in terms of longevity, spatial association and the need for close inspection to receive specific information [1]. It is likely that olfactory signals are of interest not only to the intended receiver but also to an array of eavesdroppers in the community who seek to gain beneficial information, including predators, prey, competitors and conspecifics. Olfactory cues in particular are vulnerable to eavesdropping, as unlike acoustic and visual cues, odours typically persist in the environment long after the creator has departed [1]. This creates potential for a complex olfactory web of interactions whereby receivers (intended or otherwise) may not only be responding to the signaller, but also to many other potential receivers.

The idea that olfactory information can have multiple users to create an olfactory web of interactions has received some theoretical consideration, albeit indirectly [6]. For example, infochemical webs and the ‘crying for help’ hypothesis are based on the direct signals produced by plants suffering herbivory (herbivore-induced plant volatiles) to attract their herbivores' enemies (predators or parasitoids) [6]. It is also known that some mammals use olfactory cues produced by competitors [7] as well as conspecifics [8] as cues of interaction with shared predators.

Here we investigated the patterns of visitation by red foxes (Vulpes vulpes) to scent marks of two species of competitors: feral cats (Felis catus) and wild dogs (Canis lupus familiaris), as well as conspecifics, to assess the likely risk of fox encounter associated with these odours. Odours from red foxes, cats and dogs have all induced avoidance/fear responses in a wide range of mammalian prey, both in the field and in the laboratory (see review [4]). All of these predators use urine and faeces to maintain territories and are attracted to the odours of conspecifics within their territory to over-mark and signal their presence [5,9], so their odours should represent hotspots of conspecific activity. Like most predators, foxes live in a multi-predator community. There is dietary overlap among foxes, cats and dogs, who compete for food and possibly other resources [10], and likely use each others' odours to mediate competitive interactions [11]. If foxes eavesdrop on the odour cues of their enemies, we predicted that fox visitation would be high not only at scent marks of unfamiliar foxes, but also at scent marks of cats and dogs, to create an olfactory web of information available for prey to assess their risk of fox encounter.

2. Material and methods

This experiment was conducted in a 225 km² area of the mallee wheatlands near Walpeup, Victoria, Australia where the mammalian predator community consists of foxes (density 0.9–2.0 km⁻²) [12], feral cats and feral dogs [13]. Roaming domestic farm dogs and cats may also interact with other predators in the area.

Cat and dog urine was collected from domestic animals (both genders) during routine surgical patient management at a veterinary hospital. Fox urine was collected from the bladders of foxes (both genders) shot during pest control programmes well outside the study area. Urine was kept frozen at −20°C until use. Predator urine (1 ml; typical volume of a scent mark [5]) was dripped onto 10 × 30 cm pieces of cotton towel, each towel receiving odour from a different individual. Controls consisted of untreated towels.

Forty sand plots (2 m wide and a minimum of 500 m apart (average 650 m) to maintain independence of fox visits [14]; nightly results from adjacent plots were independent; χ² = 3.12, p = 0.08, n = 112) were created across roads using in situ sand and randomly assigned to an odour treatment (cat, dog, fox or control; n = 10). Odour-treated towels were fixed to the ground with a tent peg on road verges and alternated between road sides to avoid bias. A light coating of sand was sprinkled over the towel to reduce visual conspicuousness.

Plots were checked for 6 days for predator tracks, and visitors were identified to species using track characteristics (with reference to [15]). An observer (J.P.B.) who was blind to the treatment recorded the number of fox footprints within 1 m of the towel and the spatial arrangement of footprints. Investigation of the treatment towel was determined from 10 or more footprints within 1 m of the towel, scats or diggings present within 1 m of the towel, clear change in direction of footprints towards the towel or obvious disturbance of the towel. All other fox records were considered to be incidental passings. Treatment towels were replaced daily with the same odour treatment.

(a). Statistical analyses

We performed a χ² test of independence using exact probability values to compare the percentage of investigation events by foxes between treatments. Adjusted standardized residuals (ASRs) were calculated to identify where significance could be attributed; ASRs > |2| indicate a lack of fit of the null hypothesis in that cell [16]. To compare the number of fox investigations at an odour plot among treatments, we conducted a one-way analysis of variance (ANOVA). Data were transformed () to meet the assumptions of ANOVA. All statistics were conducted in IBM SPSS Statistics v. 22. Only plots that received some form of visitation (incidental or investigation) were included.

3. Results

Fox investigation of odour plots was significantly affected by treatment (Pearson's χ² = 9.928, d.f. = 3, p = 0.018). Cat and dog odour plots received a high level of investigation, with 52% and 50% of visits respectively resulting in investigation by foxes (ASR = 1.8 and 1.6, respectively; figure 1a). While control plots received some attention (32%, ASR = −0.6), only 18% of fox-treated plots were investigated (ASR = −2.7; figure 1a). The number of investigations to an odour plot also appeared to be higher for dog- and cat-treated plots; although this was not significant (F_3,20 = 2.918, p = 0.059; figure 1b), it is nonetheless indicative of a large effect size (η² = 0.304). Foxes visited (incidental or investigation) almost all (3940) of the sand plots.

Figure 1.

(a) The percentage of visited odour-treated plots that were investigated by foxes. The number of visits to each treatment is indicated above the bars. An asterisk indicates an ASR value > |2|. (b) The average number of times a visited plot was investigated. Error bars represent standard error of the mean. The sample size is indicated above the bars.

4. Discussion

Foxes investigated cat and dog odour 62% and 56% more than untreated controls, respectively. This suggests that odours have a strong role in mediating competitive interactions among these predators. Foxes are likely to be competitively superior to cats (cat populations increase in the absence of foxes [17]) and occasionally kill them [10]. In Australia, wild dogs/dingoes harass and kill foxes [18], compete with them for food [19] and exclude them from other resources such as water [20]. Foxes exhibit small-scale avoidance of dogs [11], and our results suggest that this avoidance behaviour may be facilitated by olfactory cues, which require close investigation to retrieve specific information about competitors.

Foxes were least interested in conspecific odour, which was surprising giving its purported use in social interactions, including territorial maintenance and mating [5,9]. We presented fox odours from unfamiliar individuals to represent interloping individuals, which we predicted to induce greater interest than would odours from local foxes [21]. However, foxes were as attracted to interloper scent marks as much as they were to the novelty in odours of an untreated towel. Fox investigation of one-third of the untreated controls suggests some neophilic attraction to the novelty of the towel, even without the addition of a biologically significant odour.

Given that foxes were highly attracted to odours of other predator species, what risk does a predator scent actually signal to eavesdropping prey? Cat and dog odours, in fact, had a higher probability of visitation by foxes than did fox odour. It is likely that fox odour would, in turn, attract other predators such as cats and dogs/dingoes. Thus, anti-predator responses induced by fox odour, e.g. [4], may be a tactic to avoid predation by other eavesdropping predators rather than only the foxes themselves. Predator–prey interactions mediated by olfactory cues are clearly more complex than previously thought and are likely to form a complicated web of interactions akin to a food web (figure 2). Decision-making on whether to visit a particular olfactory cue must take into consideration who else may be eavesdropping on that cue and the potential outcome of an encounter with either the cue creator or an eavesdropper.

Figure 2.

An olfactory web of communication among a predator : prey assemblage. Thick lines are direct responses to fox odour cues; thin lines are indirect associations generated by direct responses to foxes by other members of the food web.

The concept of an olfactory web predicts that a suite of competitors and predators should all show varying degrees of attraction to and avoidance of each other's odours, as these social signals are likely to be useful in mediating their interactions with their enemies. This concept undermines the traditional approach used in hundreds of studies examining prey responses to predator odours, which assumes that it is a simple dyadic interaction, with one signaller (predator) and one receiver (prey) of the information. Before making inferences about the information encoded in the odours of a predator or indeed any species, future studies need to first untangle the potential olfactory web by quantifying visitation rates of all members of an interacting community to their own and each other's olfactory cues. This is an essential step to provide clarity in our understanding of how eavesdroppers exploit odours in a landscape of risk signalled by olfactory cues.

Ethics

This study was conducted in accordance with the Australian code of practice for the care and use of animals for scientific purposes. Research was approved by The University of Sydney Animal Ethics Committee (approval no. 2013/5743).

Data accessibility

Data supporting this article are available at the University of Sydney's eScholarship Repository at http://hdl.handle.net/2123/14727.

Authors' contributions

All authors contributed to the study design and drafted the manuscript. J.P.B. and A.D. carried out the fieldwork. All authors gave approval for publication of the final version and agree to be held accountable for the content herein.

Competing interests

We declare we have no competing interests.

Funding

This study was supported by an ARC Discovery grant DP140104413 awarded to P.B.B.