Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Austral Ecology (2008) 33, 922–931 Know thy enemy: Behavioural response of a native mammal (Rattus lutreolus velutinus) to predators of different coexistence histories JOANNE MCEVOY, DAVID L. SINN AND ERIK WAPSTRA* School of Zoology, Private Bag 05, University of Tasmania, Hobart, Tas. 7001, Australia (Email: [email protected]) Abstract Predation is recognized as a major selective pressure influencing population dynamics and evolutionary processes. Prey species have developed a variety of predator avoidance strategies, not least of which is olfactory recognition. However, within Australia, European settlement has brought with it a number of introduced predators, perhaps most notably the red fox (Vulpes vulpes) and domestic cat (Felis catus), which native prey species may be unable to recognize and thus avoid due to a lack of coexistence history. This study examined the response of native Tasmanian swamp rats (Rattus lutreolus velutinus) to predators of different coexistence history (native predatorspotted-tail quoll (Dasyurus maculatus), domestic cats and the recently introduced red fox). We used an aggregate behavioural response of R. l. velutinus to predator integumental odour in order to assess an overall behavioural response to predation risk. Rattus lutreolus velutinus recognized the integumental odour of the native quoll (compared with control odours) but did not respond to either cat or fox scent (compared with control odur). In contrast, analyses of singular behaviours resulted in the conclusion that rats did not respond differentially to either native or introduced predators, as other studies have concluded.Therefore, measuring risk assessment behaviours at the level of overall aggregate response may be more beneficial in understanding and analysing complex behavioural patterns such as predator detection and recognition. These results suggest that fox and cat introductions (and their interactive effects) may have detrimental impacts upon small native Tasmanian mammals due to lack of recognition and thus appropriate responses. Key words: behavioural response, introduced predator, olfactory recognition, predator odour, predator-prey interaction. INTRODUCTION Predation is a strong selective force leading to behavioural modifications in prey species and it is well documented that animals take risk of predation into account when making decisions about how to behave in their environment (Kats & Dill 1998; Krupa & Sih 1998). Numerous studies have examined the responses of prey species to possible predation risk in the context of cost–benefit decision making; that is, how animals make decisions regarding trade-offs between predation risk and for example, foraging or mating opportunities (Dickman 1992; Abrams 1993; Jedrzejewski et al. 1993; Jacob & Brown 2000; Sih & McCarthy 2002; Mohr et al. 2003; Powell & Banks 2004; Devereux et al. 2006; Wohlfahrt et al. 2006). Selection is likely to lead to those mechanisms in prey which allow the detection of predators prior to their attack, thereby increasing the probability of escaping or *Corresponding author. Accepted for publication November 2007. © 2008 The Authors Journal compilation © 2008 Ecological Society of Australia avoiding encounters. The major behavioural mechanism by which prey species detect predators is an activity pattern labelled ‘vigilance’ or ‘risk assessment’ (Apfelbach et al. 2005). It involves a number of behaviours (which may be species or taxon-specific) which facilitate the detection, localization and identification of predators through the use of particular sensory modes (Apfelbach et al. 2005). While visual or acoustic cues may provide direct information on the presence of predators, olfactory cues may be especially important because they provide information on predation risk even when the predator is absent at the time of detection or is difficult to see (especially in dense undergrowth and physically complex habitats). Rattus norvegicus and Rattus rattus, for example, have been shown to display innate behavioural responses to the odours of predators such as cats or the red fox (Burwash et al. 1998; Laska et al. 2005). Use of specific predator odour cues may be particularly important for mammals that have a well-developed chemical sense, such as those that are mainly nocturnal or live in physically complex habitats (Monclus et al. 2005). For example, in order for prey species to doi:10.1111/j.1442-9993.2008.01863.x N AT I V E A N D I N T R O D U C E D P R E DATO R R E C O G N I T I O N successfully extract information from predator odours, scent information needs to be a reliable indicator of predator presence and/or activity. Faeces from predators that use latrine sites (such as quolls, Kruuk & Jarman 1995) may be less indicative of their typical movement patterns than those of a predator that deposits its faeces more widely (Dickman 1992; Hayes et al. 2006). On the other hand, scent-marking for the purpose of territory establishment and to convey information about reproductive state or identity is a common practice among all carnivore families (Oakwood 2002; Belcher & Darrant 2004; Rostain et al. 2004; Laska et al. 2005), and this information can be exploited by prey species to provide cues as to possible focal areas of predator activity and hence, risk of predation (Dickman 1992; Powell & Banks 2004; Russell & Banks 2007). Similarly, predator odours derived from integumental odours or scent glands can provide instant information to prey species regarding predator presence, provided prey can determine how long ago predator scent marks were laid down. It has been suggested that prey species should have a generalized predator response when confronted with carnivore scent, as there are similarities in the scent compounds as a result of a carnivorous diet (Nolte et al. 1994; Kats & Dill 1998). However, the ability of prey species to detect and avoid predators should depend in part on the life history, ecology and evolutionary history of both predator and prey, as such, the decision-making process with regards to risk assessment becomes problematic when prey are faced with introduced (and thus, possibly unknown) predators (e.g., Jones et al. 2004; Russell & Banks 2005). Prey species may be better equipped to recognize coevolved and/or sympatric predators than recently introduced and/or allopatric ones (Hayes et al. 2006) and thus native species will be unable to respond appropriately to predation threat of introduced species due to the lack of co-evolutionary/ existence history (Banks 1998, 1999; Blumstein et al. 2002; Russell & Banks 2005). Evolutionarily novel predators, such as the red fox (Vulpes vulpes) and feral cats (Felis catus) in Australia may therefore have devastating impacts upon native prey species because of this lack of appropriate response (Banks 1998, 1999; Jones et al. 2004). Despite these predictions, there are surprisingly few quantitative tests of prey responses to native versus evolutionarily novel predators (but see Banks 1998, 1999; Russell & Banks 2005, 2007), and even fewer where the differences in length of evolutionary association between predator and prey has been explicitly considered (but see Jones et al. 2004; Russell & Banks 2007). The evidence thus far is equivocal on the ability of prey to recognize or respond to predator odour. For example, many species from the northern hemisphere show consistent avoidance of areas tainted with scent marks of co-evolved predator species only (Murray © 2008 The Authors Journal compilation © 2008 Ecological Society of Australia 923 et al. 2004; Apfelbach et al. 2005), while some native Australian mammals show little response to, and avoidance of, the faecal or urine odours of native or introduced predators (Banks 1998; Blumstein et al. 2002; Banks et al. 2003; Russell & Banks 2005). Importantly, a recent paper by Russell and Banks (2007) is the first to demonstrate native Australian mammal avoidance of predator odour. In their study, in which Elliot traps were treated with fox faeces and quoll faeces, or left untreated, native rodents showed avoidance of both native (quoll) and introduced (fox) predators.There are at least two potential reasons for the apparent equivocality in the results of these previous studies.The first of these is the manner in which a ‘predator’ is presented to potential prey.Visual, acoustic, olfactory or physical threat presence may all result in a different prey response, and the response (or lack thereof ) may be ecologically irrelevant depending on the animal’s environment and primary sensory mode. The second reason is the manner in which the prey response to predation risk is assessed. The question here refers to how prey response to predation threat is examined, for example, is a singular behavioural variable sufficient, or is an aggregate score made up of a number of predator avoidance/detection behaviours more appropriate? This study examines the risk assessment response of native rats, Rattus lutreolus velutinus, to the integumental odours of predators of different coexistence history. Spotted-tailed quolls (Dasyurus maculatus) are a native predator of R. l. velutinus (Glen & Dickman 2006; D. Moyle, pers. comm. 2006) and it may be predicted that rats will respond to the cues of the presence of quolls. Feral cats (F. catus) and red foxes (V. vulpes) are two introduced predators to which R. l. velutinus are relatively naïve. Feral cats have been present in Tasmania for the last 200+ years and are a known predator of rodents ( Jones & Coman 1981), and are specifically a threat to the R. l. velutinus population trapped for this study (Wellington Park Manangement Trust 2006). Foxes are a recently introduced species with less than 10 years history in the state (Department of Primary Industries and Water 2006), and pose a major threat to a number of native small mammal species (Banks 1998; Banks et al. 1998; Short et al. 2002; Jones et al. 2004), including R. lutreolus (Department of Primary Industries and Water 2006; Russell & Banks 2007). MATERIALS AND METHODS Study species, capture and maintenance The velvet-furred rat, R. l. velutinus (Rodentia: Muridae) is a moderately sized (ca. 150 g) endemic subspecies found throughout the state of Tasmania, Australia, but is especially prevalent in south-west doi:10.1111/j.1442-9993.2008.01863.x 924 J. M C E VOY ET AL. Tasmania. Although rarely observed, they are one of the most abundant and widely distributed mammal species, occurring at altitudes from sea level to 1600 m (Rounsevell et al. 1991; Hocking & Driessen 2000) and are found in a variety of habitats (wet and dry sclerophyll, coastal heath, button grass sedge and moorlands). Rattus lutreolus velutinus are predominantly herbivorous, with a highly varied diet (Driessen 1998). Rattus lutreolus velutinus are wholly protected under the Tasmanian National Parks and Wildlife Act of 1973, but they are nevertheless under some threat due to range constriction imposed by encroaching farmland and cattle grazing, as well as the recent introduction of the fox (V. vulpes) (Department of Primary Industries and Water 2006). Rattus lutreolus velutinus were trapped in an area of heterogeneous forest near Shoobridge Bend inWellington Park, near Hobart, Tasmania (42°56′S, 147°15′E) and in which quolls and cats are present (foxes have not yet been reported in Wellington Park). Sixteen female and 21 male R. l. velutinus were trapped during late January and early March 2006. Elliot traps (33 ¥ 10 ¥ 9 cm, Elliot Scientific, Upway,Victoria, Australia) were placed near discernable runs in the undergrowth along a 500 m straight line transect. Traps were set before dusk each evening (five nights/week), and checked each morning before 10 . Upon capture, all animals were micro-chipped (Allfelx ISO implants, 8 ¥ 2 mm) for unambiguous identification before being housed in the small mammal rooms in the School of Zoology, University of Tasmania. Rattus lutreolus velutinus were maintained in captivity in individual plastic cages (90 ¥ 40 ¥ 40 cm), with wire mesh lids and paper-pellet substrate. Each cage was provided with shredded paper for nest construction, leaf litter, bark and rocks for environmental enrichment. Water was available ad libitum. A mix of food consisting of standard small mammal pellets (rabbit food), fresh fruit and vegetables, a fruit/nut mix, crickets, meal worms and dog food biscuits was provided once a day. Light was kept constant at 11:13 hours light/dark photoperiod, and the temperature was kept at a relatively constant 10°C (!3°C). We attempted to minimize disturbance of rats while in captivity, therefore human interaction with rats was restricted to daily feeding and fortnightly cleaning of cages. All animals experienced the same conditions and handling experience while in captivity (a total of 3 months prior to these experiments). RESPONSE TO PREDATOR SCENT Integument odours from spotted tailed quolls (D. maculatus), cats (F. catus) and red foxes (V. vulpes) provided a coexistence predator-prey time scale with which to assess behavioural responses of R. l. velutinus. doi:10.1111/j.1442-9993.2008.01863.x A number of studies have found that the greatest behavioural, neurological and endocrinological responses elicited from prey species have been from integument odours (Blanchard et al. 2003; Apfelbach et al. 2005; Masini et al. 2005). It is thought that, especially in highly complex habitat (such as experienced by the population of R. l. velutinus where the individuals in this study were sourced), integument odours should provide vital cues to predator presence. While the chemical stimulus from integument of predator species may be a preferred scent stimulus for use by researchers, it is rarely used. This can be due to a number of reasons, perhaps most notably the difficulty involved in obtaining integument cues as opposed to the relative ease of using faeces and urine.Arguably however, faeces and urine may provide relatively less information to the prey species about the range of the predator, or the temporal nature of predator presence. Quoll (Dasyurus maculatus) scent was obtained from captive quolls at Bonorong Wildlife Park (Brighton, Tasmania, Australia) by rubbing each animal with cotton wool, concentrating on the back of the neck where there is a concentration of sebaceous glands (Oakwood 2002). Fresh samples were sealed in plastic bags and frozen (at -20°C) until use (e.g. Masini et al. 2005; Hayes et al. 2006). Similarly, cotton wool swabs were rubbed under the chin area of male and female domestic cats (F. catus). Cats have a scent region under their chin (sub-mandibular gland) and around the mouth (peri-oral glands) which they use to rub against surfaces to mark their territory (Feldman 1994). Fox (V. vulpes) scent was obtained by swabbing cotton wool on the anal glands (obtained from the Department of Sustainability and Environment in Victoria). Foxes, like other canids, use integument scent (as well as urine) in order to mark territory (Laska et al. 2005; Wood et al. 2005). As with quoll scent, cat and fox swabs were frozen (-20°C) until use. Assessment of behavioural responses In each predator scent test, a control (distilled water) and predator odour test were run concurrently, and individuals were randomly assigned to predator or control on day one, with the opposite test conducted at the same time on the following day (order of presentation was included as a factor in data analysis). A total of six trials were run for each individual: three controls, and one predator odour for quoll, cat and fox (in that order). Due to space limitations and the fact that scent compounds can infiltrate surrounding materials (e.g. wood, paper pellets), two isolated rooms (a control room and a scent room) were used and predator treatments were temporally separated rather than randomized. Both rooms had the same temperatures and lighting conditions over the course of the testing © 2008 The Authors Journal compilation © 2008 Ecological Society of Australia N AT I V E A N D I N T R O D U C E D P R E DATO R R E C O G N I T I O N 925 period. Each different predator/control trial was conducted at least 5 days apart, allowing all rooms to air out and scent to dissipate between trials. In order to standardize individuals’ reaction to conspecific smell, paper pellet from all individual’s home cages were used to line the base of each test arena. Tests were conducted in a 1.2 ¥ 0.6 m wooden arena, under dark conditions (50W red light/test arena) which is ecologically relevant as R. l. velutinus and all three predators are primarily nocturnally active ( Jones & Coman 1981; Jones & Dayan 2000; Glen & Dickman 2006) and each trial lasted for 10 min plus a pre-test acclimation time of 2 min. Each test arena had a familiar shelter at one end (similar to those in home cages and used in previous trials), and a feeding and scent dish at the opposite end. To eliminate the confounding effects of scent and taste being integrally linked (Chabot et al. 1996), scent (four cotton wool swabs) was placed on a separate dish to food ( Jones & Dayan 2000; Monclus et al. 2005). Each trial was videorecorded (B&W Bullet CCD Cameras, each camera connected to a Panasonic video Cassette Recorder, Series NV-FJ630) with tapes analysed at a later date. In order to quantify individuals’ behavioural response to each predator scent, nine discrete behavioural variables were recorded in each test (Table 1). These nine behaviours were chosen on the basis of previous studies (Thor et al. 1988; Burwash et al. 1998; Kats & Dill 1998; Campbell et al. 2003; Takahashi et al. 2005; Engh et al. 2006) and preliminary pilot observations. Frequency and duration of behaviours were measured, and during video analysis, the arena was divided into thirds on the monitor to calculate time spent in each third (shelter end, middle or food/scent end), and number of thirds moved. An a priori 5-s rule was used for all frequency counts, in which a behaviour was scored as a multiple frequency only if there was at least a 5-s break between occurrences (Martin & Bateson 1993; Sinn & Moltschaniwskyj 2005). While it is possible to assess response to predator scent for each specific behaviour (e.g., feeding, grooming, contact events with scent, and this method was also used, see below), the use of an aggregate behavioural score provides an alternative way to assess an overall, integrated and arguably more realistic, behavioural response (Epstein 1983; Sih et al. 2004a,b). Many behaviours within each test were highly inter-correlated. In order to reduce the number of variables used in subsequent analyses and to facilitate use of a reliable single score (see Ray & Hansen 2005) representative of ‘risk assessment behaviour’ (Apfelbach et al. 2005), we subjected the nine behavioural variables to principal components analysis (PCA).We first summed the observed behaviours (total number of times each behaviour was performed) from each rat in each test (three from predator scent, three from control) and subjected these summed behaviours to PCA with orthogonal varimax rotation (Tabachnick & Fidell 1996). The number of components interpreted was based on a scree test (Cattell 1966), and the interpretability of the components themselves (Zwick & Velicer 1986). For component interpretation, behaviours with a loading of at least 0.40 were considered to contribute to the meaning of a component (Tabachnick & Fidell 1996). The majority of the measured behaviours tended to load strongly on a single component in PCA analysis (Table 2) . Because PCA loadings (and therefore PCA scores) are often not replicable across studies, we used PCA to inform our choice of variables for inclusion in aggregate scores (Tabachnick & Fidell 1996).Therefore, for each predator and control test, we computed scale scores by summing the normalized variables measured in each test which loaded highly on PCA1 (activity, groom, food, scent, shelter third, final third and shelter – see Table 1 for behavioural definitions). This method Table 1. Behavioural variables of Rattus lutreolus velutinus recorded in the predator response context.Variables in italics were used to create the behavioural score Table 2. The solution matrix obtained from the principal components analysis ( Varimax rotation) of the nine behaviours recorded from the predator test (n = 32) Behavioural variable Time shelter Activity Groom Apple Food Scent Shelter third Final third Shelter Description Time spent in shelter Thirds crossed; measure of distance moved Number of grooming events, involved five second rule Amount of apple consumed during the test period Number of contact events with food dish Number of contact events with scent dish Proportion of time spent in shelter third Proportion of time spent in final third (where food and scent is) Number of times entered the shelter © 2008 The Authors Journal compilation © 2008 Ecological Society of Australia Time in shelter Thirds moved Groom events Apple eaten Contact food Contact scent Shelter third Final third Shelter % variance explained % variance total Component one Component two -0.802 0.599 0.702 -0.0900 0.704 0.848 -0.720 0.818 0.198 53.315 67.281 0.090 0.700 -0.022 0.625 0.539 0.428 -0.260 0.185 0.804 13.966 The highest factor loadings on each component are indicated by boldface type. doi:10.1111/j.1442-9993.2008.01863.x J. M C E VOY ET AL. resulted in aggregate scores which were highly correlated with PCA scores (r = 0.796, P < 0.001, n = 31). We computed separate risk assessment scores for each rat in each predator test and its corresponding control test, resulting in six unique scores per rat (three predator scent tests, three controls). In each test, higher scores describe an individual which spends less time in the shelter and less time in the shelter third, spends more time in the final third (where the food and scent is located), moves a greater number of thirds, has more groom events and initiates a greater number of contact events with both the scent and food dishes. A high score indicates an increase in the level of risk assessment, while a lower score indicates the reverse (Genaro & Schmidek 2000; Whishaw et al. 2006). It is worth noting here that while groom behaviour is not a risk assessment behaviour per se, it is a displacement behaviour that rodents typically display when they are in unfamiliar environments, are unsure of their surroundings, or are presented with unfamiliar and possibly threatening stimuli (Thor et al. 1988). It is used as a measure of an individual’s distress (Thor et al. 1988; Engh et al. 2006). In this situation, when considered in conjunction with the other behaviours that make up the risk assessment behaviours, it represents an individual’s recognition of new and possibly threatening stimuli in the form of predator scent. We used repeated-measures to determine if R. l. velutinus react to the presence of predator scent (as opposed to a control odour). As predator trials were temporally separated, risk assessment scores of rats were first analysed using three separate one-way repeated-measures ’s (one for each predator scent against its concurrent control, e.g. quoll vs. quoll control, cat vs. cat control and fox vs. fox control). In addition to using aggregate scores, we examined two singular behaviours, to determine whether there was a difference in our interpretation of risk assessment behaviour due to the method of measurement (aggregate score vs. singular behaviour). These two behaviours (amount of apple eaten and number of thirds crossed) were chosen a priori (based on previous studies). Amount of apple eaten was not included in the risk assessment scores because it did not load on PCA1; the other behaviour (number of thirds crossed) was a variable which contributed to aggregate scores. Apple eaten was used because a food variable is often used in assessing prey response to predation threat (e.g. Jones & Dayan 2000; Blumstein et al. 2002), as is an activity measure (i.e. number of thirds crossed) (Kats & Dill 1998; Monclus et al. 2005). The two singular behaviours were subjected to three separate one-way repeated-measures analyses to determine if there was a difference in behavioural response to predator versus control scents. Risk assessment scores and singular behaviours were distributed normally, so no transformations were doi:10.1111/j.1442-9993.2008.01863.x required prior to analyses. Analyses on risk assessment scores and singular behaviours were first conducted with sex and order of presentation as between-subjects factors for all repeated-measures s. In all cases, neither sex nor order of presentation was significant (P > 0.05), and so these are not considered further. All data were analysed with SPSS 14.0 for Windows. RESULTS Rattus lutreolus velutinus differed significantly in their response to quoll scent as opposed to control scent (F(1,29) = 0.425, P = 0.034; Fig. 1A), increasing their a 0.40 0.30 0.20 0.10 0.00 -0.10 -0.20 -0.30 -0.40 Mean Behavioural score 926 -0.50 b Quoll Control - Quoll Cat Control - Cat 0.40 0.30 0.20 0.10 0.00 -0.10 -0.20 -0.30 -0.40 -0.50 c 0.40 0.30 0.20 0.10 0.00 -0.10 -0.20 -0.30 -0.40 -0.50 Fox Control - Fox Scent situation Fig. 1. Mean behavioural scores of Rattus lutreolus velutinus to scent of (a) quoll, (b) cat and (c) fox, associated responses under control conditions are also presented. Higher scores indicate individuals which spend less time in the shelter and less time in the shelter third, spends more time in the final third (where the food and scent is located), moves a greater number of thirds, has more groom events and initiates a greater number of contact events with both the scent and food dishes. Error bars represent standard error of the mean, n = 28. © 2008 The Authors Journal compilation © 2008 Ecological Society of Australia N AT I V E A N D I N T R O D U C E D P R E DATO R R E C O G N I T I O N risk assessment behaviour when in the presence of quoll scent. However, there was no mean level behavioural response to either cat scent or fox scent compared with controls (cat scent to control: F(1,30) = 0.585, P = 0.450; fox scent to control: F(1,28) = 1.067, P = 0.310, Fig. 1B and 1C). We were also initially interested in comparing predator scents directly, however, mean risk assessment scores in controls differed across the 3 weeks of study (repeated-measures : F(2,26) = 3.72; P = 0.03, Fig. 1). Risk assessment scores in each of the predator trials were corrected for the difference in mean response to controls (Dingemanse et al. 2002), on the assumption that predator and control responses co-vary, and were then analysed with repeated-measures . Using the corrected scores to compare directly between predator scents, there was a significant mean-level shift in response to the scent of the different predators (repeated-measures : F(2,27) = 5.533, P = 0.01). Simple contrasts revealed that there was no significant difference in behavioural response of R. l. velutinus to cat and fox scent (F(1,28) = 1.924, P = 0.18), but that R. l. velutinus differed behaviourally in their response to quoll scent compared with both cat and fox scent (quoll to cat scent response: F(1,28) = 4.893, P = 0.035; quoll to fox scent response: F(1,28) = 9.393, P = 0.005; supporting the conclusion that R. l. velutinus respond to quolls differently to control, and that they do not respond to either of the introduced predators). These results confirm our above results that R. l. velutinus do not respond to introduced predators. However, due to the confounding nature of the change in controls, these results are not discussed further. Using the single behavioural variable ‘amount of apple eaten’ there was no significant difference in how R. l. velutinus responded to any of the three predator scents, analysed against their concurrent controls (repeated-measures ; quoll to quoll control: F(1,30) = 0.012, P = 0.913; cat to cat control: F(1,30) = 2.625, P = 0.116; and fox to fox control: F(1,28) = 0.041, P = 0.842). Similarly, using the singular behavioural variable ‘number of thirds crossed’ showed that there was no significant difference in how R. l. velutinus behaved with respect to predator scents (repeatedmeasures ; quoll to quoll control: F(1,30) = 0.975, P = 0.331; cat to cat control: F(1,30) = 1.877, P = 0.181; and fox to fox control: F(1,28) = 0.288, P = 0.596). DISCUSSION Olfactory senses are commonly used by prey species to detect the presence of predators, and prey individuals may display a behavioural recognition of, and response to, predator odour (Banks 1998; Blumstein et al. 2002; Powell & Banks 2004; Russell & Banks 2007). A ‘risk assessment’ behavioural response is frequently used by © 2008 The Authors Journal compilation © 2008 Ecological Society of Australia 927 prey individuals in recognition of predator scent (Apfelbach et al. 2005), and there is a suggestion that recognition of predator scent is an innate response which should exist against the majority of carnivorous species (Nolte et al. 1994; Burwash et al. 1998; Kats & Dill 1998). However, the evidence regarding the response of native prey species to introduced predators is conflicting. In this study, R. l. velutinus responded to quoll scent (as opposed to control) but did not respond to either cat or fox scent (as opposed to control). This indicates that R. l. velutinus is clearly capable of behavioural responses to predator scent, and respond behaviourally to the predator scent cues of its native, sympatric predators, the spotted tailed quoll (D. maculatus). Rattus lutreolus velutinus do not respond to the introduced predator species, cats (F. catus) or red foxes (V. vulpes), indicating behavioural recognition of native predators only. Such lack of responses by prey to evolutionarily novel predators may explain the devastating impacts these predators have had within Australia; prey appear to lack the co-evolutionary history required to have developed the appropriate responses (Coss 1999). However, there is an alternative explanation of our results which relates to R. l. velutinus behaviour in response to predator and control scent as an artefact of the experimental design. We used a fixed order of presentation, rather than a randomized approach (for reasons outlined in the Materials and methods). However, because of the experimental design, rat response to odour may reflect habituation to trials, as is indicated by changes in rat response to control scent over time. The lack of a significant response in overall risk assessment behaviour when R. l. velutinus were confronted with the scent of introduced predators may simply reflect their familiarization with trials.We do not believe that this is the case however, as discussed below. There is evidence from odour recognition trials with rodents to suggest that they do not habituate to scent based stimuli, especially when scent trials are temporally separated by long periods of time (Wallace & Rosen 2000; Blanchard et al. 2003; Burman & Mendl 2006). With R. norvegicus, the odour recognition memory for individuals that had been housed together for 18 days was between 48–96 h (Burman & Mendl 2006). After 96 h, rats investigated unfamiliar odours and the odours of former cage mates equally. Similarly, a study examining the fear response of rats to trimethylthiazoline (TMT), a component of fox faeces, found no within-sessions or between-sessions habituation to TMT, nor did TMT produce contextual conditioning (Wallace & Rosen 2000; Blanchard et al. 2003). Furthermore, Blanchard et al. (1998) found that after 20 days of 60 min visual exposure per day to cats, rats had minimal behavioural habituation to the threat, and had reliably higher basal corticosterone concentrations, indicating no habituation of the endocrine system (i.e., doi:10.1111/j.1442-9993.2008.01863.x 928 J. M C E VOY ET AL. they were still responding with a stress-based endocrine response to cat presence). Similarly, experimentation by File et al. (1993) showed that rats exposed to cat odour showed no behavioural habituation to the threat after repeated exposures; they continued to avoid the odour cloth. Given the above experiments that indicate little or no behavioural habituation to predator threats, the length of time our rats were exposed to each predator threat (<35 min per scent), the length of time between predator presentation (>5 days) and the fact that the effect of order of presentation (within trials) of scents was non-significant, we believe that the results may reasonably be interpreted as R. l. velutinus responding to the odours from its native predators (quolls, D. maculatus) but not to that of either of the introduced predators. Thus, our experimental design, where response to potential predation risk (quoll, cat or fox) was measured against a concurrent control (even though each predator scent trial was separated by a time period), was suitably designed to test recognition of, and response to, native versus introduced predation risk. The fact that R. l. velutinus did not alter their behaviour in response to fox or cat scent (compared with their concurrent controls) strongly suggests that they did not perceive or respond to this risk. Interpretation of our results that R. l. velutinus responded to their native predator and not to either of the introduced predators contrasts with suggestions that prey species should respond to novel predators because of common odour constituents reflecting a generalized carnivore diet (including sulphurous metabolites of protein digestion: Jedrzejewski et al. 1993; Nolte et al. 1994; Kats and Dill 1998; Apfelbach et al. 2005; Hayes et al. 2006). While anti-predator behaviours and predator avoidance strategies are highly beneficial when there is a real and significant predation threat, they may also be costly to the individual. Anti-predator behaviours may include the flight or fight response which is facilitated by an increase in corticosterone (Dufty & Crandall 2005). Corticosterone can be costly over a long-term period (Marquez et al. 2004), and it can thus be detrimental for prey species to respond physiologically to predation threat unnecessarily. Similarly, predator avoidance strategies can involve decreased activity patterns and restricted movement, resulting (in some cases) in a lack of foraging or mating opportunity (Mohr et al. 2003; Devereux et al. 2006); these effects constitute sub-lethal impacts of predation (Powell & Banks 2004).Thus, response to predator odours may be expected to be finely tuned to those risks associated with sympatric predator species due to these detrimental effects (Agrawal et al. 1999; Jones et al. 2004). Our results support these suggestions; in this study, R. l. velutinus displayed a risk assessment behavioural response to their native predators only, perhaps because of the costs associated with using a generalized predator response, or because they doi:10.1111/j.1442-9993.2008.01863.x do not recognize the threat as the constituents of the scent are not common enough. Results on Australian prey species responding to potential predation threats have not been consistent. A recent study (Russell & Banks 2007) demonstrated native rodents (Rattus fuscipes, R. lutreolus and Pseudomys gracilicaudatus) tended to avoid traps scented with both native quoll (D. maculatus) and introduced fox (V. vulpes) odour. These responses clearly differ from our own study (differential responses to quoll and fox/cat) as they do from earlier work by the same authors (Banks 1998; Russell & Banks 2005) where prey showed no response to evolutionarily novel predators using similar methods. The discrepancy between the results of previous studies and our own may lie in how responses were assessed. In our study, we used an aggregate score composed of a number of discrete, observable investigatory, displacement and locomotory behaviours (see Table 1) which was representative of an overall risk assessment response (Apfelbach et al. 2005). Rattus lutreolus velutinus displayed an increase in risk-assessment behaviours in response to quoll scent, characterizing an active search and evaluation of the potential predation risk associated with the scent. This is similar to the response seen to a decrease in habitat complexity (and thus increased perceived predation risk) in which R. l. velutinus increased activity patterns ( J. McEvoy et al. unpubl. data 2006), thus supporting a specific stimuli-based response. The use of an integrated approach to the measurement of behavioural responses of R. l. velutinus to predators, unlike singular measures such as trap avoidance (e.g. Banks 1998) or food consumption (e.g. Jacob & Brown 2000), provides an alternative (and not yet widely used) method to assess prey response to predation. Using two examples of singular behavioural measures in this study (apple consumption and activity, commonly used to assess prey response to predation risk (Kats & Dill 1998; Jones & Dayan 2000; Blumstein et al. 2002; Monclus et al. 2005)) indicated that there was no differential behavioural response of R. l. velutinus to native or introduced predators. The conclusion from these results would be that R. l. velutinus do not respond to olfactory cues from either native or introduced predators; however, the use of an aggregate behavioural score provides a different result, and is (in our opinion) a more accurate and holistic representation of prey response to predation threat. In our study, the use of predator integumental odours (as opposed to urine or faeces) and the response by R. l. velutinus to the odours of native predators suggests that they are used as an indicator of predation risk, perhaps conveying realistic information about predator presence, habitat boundaries and frequently used travelling routes (Blanchard et al. 2003; Apfelbach et al. 2005; Masini et al. 2005). Faecal predator cues may be poor predictors of pos© 2008 The Authors Journal compilation © 2008 Ecological Society of Australia N AT I V E A N D I N T R O D U C E D P R E DATO R R E C O G N I T I O N sible predation risk (Banks et al. 2003; Blanchard et al. 2003), especially for predators that use latrine sites (such as quolls, Kruuk & Jarman 1995) because they convey little information on typical movement patterns compared with predators that deposits its faeces more widely (Dickman 1992; Hayes et al. 2006). The ability of prey to detect predator presence through the use of scent is likely to be related to both predator and prey life history and ecology (Jedrzejewski et al. 1993), and further study is needed to test differential prey responses to alternative scent types (e.g., integumental vs. urinary/faecal) from the same predators. Even though R. l. velutinus has coexisted with cats for over 200 years, our results suggest that they have yet to evolve a risk assessment response to cat integumental odour (see also Griffin et al. 2001; Blumstein et al. 2002). This, coupled with the fact that they also did not respond to a recently introduced predator, the red fox, suggests that native fauna may be highly vulnerable to interactive effects induced by multiple evolutionarily novel predator species. Given the equivocal results on the response of prey to their native and introduced predators, clarifying the response of native species to these predators will provide vital knowledge for assessing the level of risk introduced predators pose to native wildlife. ACKNOWLEDGEMENTS This study was carried out under the University of Tasmania Animal Ethics Permit numbers of A0008756, A0008627, and the Department of Primary Industries and Water permit number FA05257. Thanks to the Behavioural Ecology and Evolutionary Research Group for assistance in rat catching and maintenance, as well as stimulating discussions and feedback on earlier versions of this manuscript. The authors also appreciate the efforts of Dan Purdey,TorraneVergis and Tom Sloane in particular in obtaining predator scent. The authors also wish to thank the anonymous reviewers for valuable feedback and comments on earlier versions of this manuscript. REFERENCES Abrams P. A. (1993) Optimal ‘traits’ when there are several costs – the interaction of mortality and energy costs in determining foraging behavior. Behav. Ecol. 4, 246–53. Agrawal A. A., Laforsch C. & Tollrian R. (1999) Transgenerational induction of defences in animals and plants. Nature 401, 60–3. Apfelbach R., Blanchard C. D., Blanchard R. J., Hayes R. A. & McGregor I. S. (2005) The effects of predator odours in mammalian prey species: a review of field and laboratory studies. Neurosci. Biobehav. Rev. 29, 1123–44. © 2008 The Authors Journal compilation © 2008 Ecological Society of Australia 929 Banks P. B. (1998) Responses of Australian bush rats, Rattus fuscipes, to the odour of introduced Vulpes vulpes. J. Mammal. 79, 1260–4. Banks P. B. (1999) Predation by introduced foxes on native bush rats in Australia: do foxes take the doomed surplus? J. Appl. Ecol. 36, 1063–71. Banks P. B., Dickman C. R. & Newsome A. E. (1998) Ecological costs of feral predator control: foxes and rabbits. J.Wildlife Manage. 62, 762–72. Banks P. B., Hughes N. K. & Rose T. A. (2003) Do Australian native small mammals avoid faeces of domestic dogs? Aust. Zool. 32, 406–9. Belcher C. A. & Darrant J. P. (2004) Home range and spatial organization of the marsupial carnivore, Dasyurus maculatus maculatus (Marsupialia: Dasyuridae) in south-eastern Australia. J. Zool. 262, 271–80. Blanchard D. C., Markham C., Yang M., Hubbard D., Madarang E. & Blanchard R. J. (2003) Failure to produce conditioning with low-dose trimethylthiazoline or cat feces as unconditioned stimuli. Behav. Neurosci. 117, 360–8. Blanchard R. J., Nikulina J. N., Sakai R. R., McKittrick C., McEwen B. & Blanchard D. C. (1998) Behavioral and endocrine change following chronic predatory stress. Physiol. Behav. 63, 561–69. Blumstein D. T., Mari M., Daniel J. C., Ardron J. G., Griffin A. S. & Evans C. S. (2002) Olfactory predator recognition: wallabies may have to learn to be wary. Anim. Conserv. 5, 87–93. Burman O. H. P. & Mendl M. (2006) Long term social memory in the laboratory rat (Rattus norvegicus). Anim Welfare. 15, 379–82. Burwash M. D., Tobin M. E., Woolhouse A. D. & Sullivan T. P. (1998) Laboratory evaluation of predator odors for eliciting an avoidance response in roof rats (Rattus rattus). J. Chem. Ecol. 24, 49–66. Campbell T., Lin S., DeVries C. & Lambert K. (2003) Coping strategies in male and female rats exposed to multiple stressors. Physiol. Behav. 78, 495–504. Cattell R. B. (1966) The scree test for the number of factors. Sociol. Methods Res. 1, 245–76. Chabot D., Gagnon P. & Dixon E. A. (1996) Effect of predator odours on heart rate and metabolic rate of wapiti (Cervus elaphus canadensis). J. Chem. Ecol. 22, 839– 68. Coss R. G. (1999) Effects of relaxed natural selection on the evolution of behaviour. In: GeographicVariation of Behaviour: An Evolutionary Perspective (eds S. A. Foster & J. A. Endler) pp. 180–208. Oxford University Press, Oxford. Department of Primary Industries and Water (2006) Natural Environment:Weeds, Pests and Diseases. [Cited 3 July 2006.] Available from URL: http://www.dpiw.tas.gov.au Devereux C. L., Whittingham M. J., Fernandez-Juricic E., Vickery J. A. & Krebs J. R. (2006) Predator detection and avoidance by starlings under differing scenarios of predation risk. Behav. Ecol. 17, 303–9. Dickman C. R. (1992) Predation and habitat shift in the house mouse, Mus-Domesticus. Ecology. 73, 313–22. Dingemanse N. J., Both C., Drent P. J., van Oers K. & Noordwijk A. J. (2002) Repeatability and heritability of exploratory behaviour of great tits from the wild. Anim. Behav. 64, 929–38. Driessen M. M. (1998) Observations on the diets of the longtailed mouse, Pseudomys higginsi, and the velvet-furred rat, Rattus lutreolus velutinus, in southern Tasmania. Aust. Mammal. 21, 121–30. doi:10.1111/j.1442-9993.2008.01863.x 930 J. M C E VOY ET AL. Dufty A. M. & Crandall M. B. (2005) Corticosterone secretion in response to adult alarm calls in American Kestrels. J. Field Ornithol. 76, 319–25. Engh A. L., Beehner J. C., Bergman T. J. et al. (2006) Behavioural and hormonal responses to predation in female chacma baboons (Papio hamadryas ursinus). Proc. Roy. Soc. B 273, 707–12. Epstein S. (1983) Aggregation and beyond: some basic issues on the prediction of behavior. J. Pers. 51, 360–92. Feldman H. N. (1994) Methods of scent marking in the domestic cat. Can. J. Zool. 72, 1093–9. File S. E., Zangrossi H., Sanders F. L. & Mabbutt P. S. (1993) Dissociation between behavioural and corticosterone responses on repeated exposures to cat odor. Physiol. Behav. 54, 1109–11. Genaro G. & Schmidek W. R. (2000) Exploratory activity of rats in three different environments. Ethology 106, 849– 59. Glen A. S. & Dickman C. R. (2006) Diet of the spotted-tailed quoll (Dasyurus maculatus) in eastern Australia: effects of season, sex and size. J. Zool. 269, 241–8. Griffin A. S., Evans C. S. & Blumstein D. T. (2001) Learning specificity in acquired predator recognition. Anim. Behav. 62, 577–89. Hayes R. A., Nahrung H. F. & Wilson J. C. (2006) The response of native Australian rodents to predator odours varies seasonally: a by-product of life history variation? Anim. Behav. 71, 1307–14. Hocking G. J. & Driessen M. M. (2000) Status and conservation of the rodents of Tasmania. Wildl. Res. 27, 371–7. Jacob J. & Brown J. S. (2000) Microhabitat use, giving-up densities and temporal activity as short- and long-term antipredator behaviors in common voles. Oikos 91, 131–8. Jedrzejewski W., Rychlik L. & Jedrzejewski B. (1993) Responses of bank voles to odours of seven species of predators: experimental data and their relevance to natural predator-vole relationships. Oikos 68, 251–7. Jones E. & Coman B. J. (1981) Ecology of the feral cat, Felis catus (L), in southeastern Australia. 1. Diet. Aust. Wildl. Res. 8, 537–47. Jones M. & Dayan T. (2000) Foraging behaviour and microhabitat use by spiny mice, Acomys cahirinus and A. russatus, in the presence of blanford’s fox (Vulpes cana) odor. J. Chem. Ecol. 26, 455–69. Jones M. E., Smith G. C. & Jones S. M. (2004) Is anti-predator behaviour in tasmanian eastern quolls (Dasyurus viverrinus) effective against introduced predators? Anim. Conserv. 7, 155–60. Kats L. B. & Dill L. M. (1998) The scent of death: chemosensory assessment of predation risk by prey animals. Ecoscience. 5, 361–94. Krupa J. J. & Sih A. (1998) Fishing spiders, green sunfish, and a stream-dwelling water strider: male-female conflict and prey responses to single versus multiple predator environments. Oecologia 117, 258–65. Kruuk H. & Jarman P. J. (1995) Latrine use by the spotted-tailed quoll (Dasyurus maculatus: Dasyuridae, Marsupialia) in its natural habitat. J. Zool. 236, 345–9. Laska M., Fendt M., Wieser A., Endres T., Salazar L. T. H. & Apfelbach R. (2005) Detecting danger-or just another odorant? Olfactory sensitivity for the fox odour component 2,4,5-trimethylthiazoline in four species of mammals. Physiol. Behav. 84, 211–5. Marquez C., Nadal R. & Armario A. (2004) The hypothalamicpituitary-adrenal and glucose responses to daily repeated doi:10.1111/j.1442-9993.2008.01863.x immobilisation stress in rats: individual differences. Neuroscience 123, 601–12. Martin P. & Bateson P. (1993) Measuring Behaviour: An Introductory Guide. Cambridge University Press, Cambridge. Masini C. V., Sauer S. & Campeau S. (2005) Ferret odour as a processive stress model in rats: neurochernical, behavioral, and endocrine evidence. Behav. Neurosci. 119, 280–92. Mohr K., Vibe-Petersen S., Jeppesen L. L., Bildsoe M. & Leirs H. (2003) Foraging of multimammate mice, Mastomys natalensis, under different predation pressure: cover, patchdependent decisions and density-dependent GUDs. Oikos 100, 459–68. Monclus R., Rodel H. G., Von Holst D. & De Miguel J. (2005) Behavioural and physiological responses of naive European rabbits to predator odour. Anim. Behav. 70, 753–61. Murray D. L., Roth J. D. & Wirsing A. J. (2004) Predation risk avoidance by terrestrial amphibians: the role of prey experience and vulnerability to native and exotic predators. Ethology 110, 635–47. Nolte D. L., Mason J. R., Epple G., Aronov E. & Campbell D. L. (1994) Why are predator urines aversive to prey. J. Chem. Ecol. 20, 1505–16. Oakwood M. (2002) Spatial and social organization of a carnivorous marsupial Dasyurus hallucatus (Marsupialia: Dasyuridae). J. Zool. 257, 237–48. Powell F. & Banks P. B. (2004) Do house mice modify their foraging behaviour in response to predator odours and habitat? Anim. Behav. 67, 753–9. Ray J. & Hansen S. (2005) Temperamental development in the rat: the first year. Dev. Psychobiol. 47, 136–44. Rostain R. R., Ben-David M., Groves P. & Randall J. A. (2004) Why do river otters scent-mark? An experimental test of several hypotheses. Anim. Behav. 68, 703–11. Rounsevell D. E., Taylor R. J. & Hocking G. J. (1991) Distribution records of native terrestrial mammals in Tasmania. Wildl. Res. 18, 699–717. Russell B. G. & Banks P. B. (2005) Response of four Critical Weight Range (CWR) marsupials to the odours of native and introduced predators. Aust. Zool. 33, 217–22. Russell B. G. & Banks P. B. (2007) Do Australian small mammals respond to native and introduced predator odours? Austral Ecol. 32, 277–86. Short J., Turner B. & Risbey D. (2002) Control of feral cats for nature conservation III. Trapping. Wildl. Res. 29, 475–87. Sih A. & McCarthy T. M. (2002) Prey responses to pulses of risk and safety: testing the risk allocation hypothesis. Anim. Behav. 63, 437–43. Sih A., Bell A. & Johnson J. C. (2004a) Behavioural syndromes: and ecological and evolutionary overview. Trends Ecol. E. 19, 372–8. Sih A., Bell A. M., Johnson J. C. & Ziemba R. E. (2004b) Behavioural syndromes: an integrative overview. Q. Rev. Biol. 79, 241–77. Sinn D. L. & Moltschaniwskyj N. A. (2005) Personality traits in dumpling squid (Euprymna tasmanica): context-specific traits and their correlation with biological characteristics. J. Comp. Psych. 119, 99–110. Tabachnick G. & Fidell L. S. (1996) Using Multivariate Statistics. HarperCollins, New York. Takahashi L. K., Nakashima B. R., Hong H. C. & Watanabe K. (2005) The smell of danger: a behavioral and neural analysis of predator odour-induced fear. Neurosci. Behav. Rev. 29, 1157–67. Thor D. H., Harrison R. J., Schneider S. R. & Carr W. J. (1988) Sex-differences in investigatory and grooming behaviors of © 2008 The Authors Journal compilation © 2008 Ecological Society of Australia N AT I V E A N D I N T R O D U C E D P R E DATO R R E C O G N I T I O N laboratory rats (Rattus norvegicus) following exposure to novelty. J. Comp. Psych. 102, 188–92. Wallace K. J. & Rosen J. B. (2000) Predator odor as an unconditioned fear stimulus in rats: elicitation of freezing by trimethylthiazoline, a component of fox feces. Behav. Neurosci. 114, 912–22. Wellington Park Manangement Trust (2006) Environment. [Cited 3 July 2006.] Available form URL: http://www. wellingtonpark.tas.gov.au/environmentnatureindex.php Whishaw I. Q., Gharbawie O. A., Clark B. J. & Lehmann H. (2006) The exploratory behaviour of rats in an open © 2008 The Authors Journal compilation © 2008 Ecological Society of Australia 931 environment optimizes security. Behav. Brain. Res. 171, 230– 39. Wohlfahrt B., Mikolajewski D. J., Joop G. & Suhling F. (2006) Are behavioural traits in prey sensitive to the risk imposed by predatory fish? Freshwater Biol. 51, 76–84. Wood W. F., Terwilliger M. N. & Copeland J. P. (2005) Volatile compounds from anal glands of the wolverine, Gulo gulo. J Chem Ecol. 31, 2111–17. Zwick W. R. & Velicer W. F. (1986) Comparison of five rules for determining the number of components to retain. Psychol. Bull. 99, 432–42. doi:10.1111/j.1442-9993.2008.01863.x