ANIMAL BEHAVIOUR, 1999, 58, 1085–1093 Article No. anbe.1999.1224, available online at http://www.idealibrary.com on
Differential behavioural and hormonal responses of voles and spiny mice to owl calls DAVID EILAM*, TAMAR DAYAN*, SHAMGAR BEN-ELIYAHU†, IFAT SCHULMAN*, GABI SHEFER* & COLIN A. HENDRIE‡
*Department of Zoology, Tel-Aviv University †Department of Psychology, Tel-Aviv University ‡Department of Psychology, University of Leeds (Received 11 August 1998; initial acceptance 2 November 1998; final acceptance 7 July 1999; MS. number: 5966R)
Rodents usually respond to the presence of owls by reducing overall activity, in particular foraging. In this study, a playback of recorded tawny owl, Strix aluco, calls was sufficient to induce a marked effect in the social (Gunther’s) vole, Microtus socialis. Some of the voles exposed to owl calls reduced their activity (‘freeze’ pattern) unlike control voles exposed to a human voice. Other voles, however, dashed around the cage (‘flee’ pattern). Owl calls also increased corticosterone levels in the voles, showing that the calls induced stress. We suggest that the behavioural dichotomy to freeze or flee in voles is a result of differences in individual normal behaviour and/or in stimulus interpretation. In the common spiny mouse, Acomys cahirinus, no behavioural changes were detected after exposure to owl calls, despite increased cortisol levels which are indicative of stress. Differences in the habitats of voles and spiny mice may explain the apparent lack of behavioural response in the latter. They are rock-dwelling rodents preferentially foraging between boulders and in rock crevices, where they are relatively protected from aerial predation, whereas voles forage in relatively open spaces.
We also used an owl call in the present study to determine any behavioural responses and increase in blood glucocorticosteriod level. Previous studies indicate that the threat of predation tends to decrease locomotor behaviour in the prey (Desy et al. 1990; Whishaw & Dringenberg 1991; Ronkainen & Ylonen 1994; Hendrie et al. 1998). In field studies, rodents exposed to owl calls reduced their foraging behaviour (Kotler et al. 1994; Otter 1994; Abramsky et al. 1996) and similar responses were evident in studies carried out in experimental enclosures (Kotler et al. 1992). In a laboratory setting (Hendrie et al. 1998), rodents exposed to owl calls became less active; this may account for the decrease in foraging behaviour observed in the field and in seminatural environments. Freezing and reduced activity, however, are clearly not the only behavioural responses to threat. Rather, many species have developed antipredator defence strategies that also involve fleeing or defensive aggression (fighting). Under an apparent predation risk, rodents freeze or flee (Kieffer 1991; Randall et al. 1995; Weary & Kramer 1995). For instance, woodmice, Apodemus mystacinus, freeze or leap when exposed to stoats, Mustela ermina (Erlinge et al. 1974) but usually bolt for a hole or ‘scamper away’ when exposed to other predators (Bolles 1970; King 1985). Voles, Microtus agrestis and Clethrionomys britannicus, exposed to a silhouette of a
Owl predation poses a significant threat for small mammals, especially rodents, which are a major food source for many owl species (Mikkola 1983; King 1985; Martin 1990; Belm et al. 1993; Selaas 1993; Tome 1994; Jedrzejewski et al. 1993, 1996). Owls are efficient predators and, once they have initiated the final attack sequence, their success rate rarely falls below 90% (Curio 1976). It is thus crucial for small mammals, which are highly susceptible to owl predation, to develop a perception of their risk. The territorial calls of owls are the most obvious indication of their presence. Other stimuli appear minor and less obvious: owls fly silently, perch without motion and are well camouflaged by their feathers. Owl odour alone is probably not considered a threat by rodents, since bank voles, Clethrionomys glareolus, do not respond to the odour of the tawny owl, Strix aluco, their main predator, while they do respond to odours of terrestrial predators (Jedrzejewski et al. 1993). Thus, owl calls may be the major source of information indicating that owls are active within a given area. Consequently, owl calls have been applied in many studies of owl–rodent interactions in order to simulate the presence of an owl. Correspondence: D. Eilam, Department of Zoology, Tel-Aviv University, Ramat-Aviv 69 978, Israel (email:
[email protected]). C. A. Hendrie is at the Department of Psychology, University of Leeds, Leeds LS2 9JT, U.K. 0003–3472/99/111085+09 $30.00/0
1999 The Association for the Study of Animal Behaviour
1085
1999 The Association for the Study of Animal Behaviour
1086
ANIMAL BEHAVIOUR, 58, 5
hawk either freeze or increase locomotion as if trying to escape (Fentress 1968). One possible explanation for the diversity of responses to predatory risk relates the response to the proximity of the potential predator (Gallup 1974; Ratner 1977), referring to defensive behaviour as a ‘distance-dependent defence hierarchy’ (see Hendrie et al. 1996 for a review). According to this hierarchy, a distant predator evokes freezing, a closer predator evokes flight and an even closer one evokes defensive fighting. However, additional factors may also affect the behavioural response of the prey. These include the hunting strategies of the particular predator species involved, the prey species’ characteristics (e.g. Bolles 1970; Erlinge et al. 1974; King 1985; Kieffer 1991; Randall et al. 1995; Weary & Kramer 1995) and individual variability. In the present study, we examined the different behaviour patterns of two species of rodents induced by playback of tape-recorded owl calls. A perception of owl calls as a threat should not be restricted to the behavioural domain but should also involve concomitant physiological changes (Selye 1950; Blanchard et al. 1998; Levy et al. 1998). An increase in corticosteroid level is a reliable indicator of stress (Blanchard et al. 1998). In the present study, we measured corticosteroid levels alongside the behavioural response of rodents exposed to owl calls, to determine a possible relationship between hormonal changes and behaviour. We chose to expose rodents to calls of the tawny owl, which are more effective than calls of other predators at inducing defensive behaviours (Hendrie & Weiss 1994). We studied the response of two rodent species: the social (Gunther’s) vole, Microtus socialis, and the common spiny mouse, Acomys cahirinus. We selected the voles because, among rodents, they are heavily preyed upon by owls and comprise 40–70% (sometimes over 90%) of the diet of barn owls, Tyto alba, and tawny owls (Mikkola 1983; Mendelssohn & Yom-Tov 1987; Martin 1990; Selaas 1993; Tome 1994). We selected the common spiny mouse because, surprisingly, it did not appear to react to recorded tawny owl calls in a previous study (Hendrie et al. 1998). There are no published data on the part of common spiny mice in the diet of owls; however, two out of six Hume’s tawny owl, Strix butleri, pellets included spiny mice (Mandelik et al., in press). We therefore aimed to elucidate the relationship between hormonal change and behaviour in order to gain insight into the apparent lack of response of spiny mice to owl calls, in comparison to the conspicuous response of voles (Hendrie et al. 1998). METHODS
Study Animals The social (Gunther’s) vole weighs 37–50 g and is 11 cm long, plus a 2-cm tail. It is a burrow-dwelling rodent that feeds on seeds and green vegetation. We obtained 49 male voles, bred in captivity, from the Agricultural Research Organization and from the Vulcani Institute, Israel. They were housed in groups of 5–10, in metal cages measuring 4070 cm and 25 cm high, which were
located outdoors in our zoo yard under natural (uncontrolled) temperature and light conditions. Overturned ceramic pots and wooden boxes were placed within each cage to provide shelter. Seeds and diced fresh vegetables were provided daily. Based on years of experience in maintaining colonies of voles in our zoo, provision of water was unnecessary once sufficient fresh vegetables were provided. The common spiny mouse weighs 38–44 g and is 11 cm long, plus a 10-cm tail. It is an agile, omnivorous rodent (Kronfeld & Dayan, in press), common on rocky mountains, where it shelters in rock crevices (Shkolnik 1971; Kronfeld et al. 1994; Shargal 1997). We obtained 27 male spiny mice from colonies in the Meir Segals Zoological Garden at Tel Aviv University. Each colony comprised several dozen males and females kept together in a cage measuring 14070 cm and 70 cm high. After their removal from the colony, they were housed in groups of three to six in metal cages measuring 4070 cm and 25 cm high which were located outdoors in the zoo yard, under natural (uncontrolled) temperature and light conditions. Overturned ceramic pots and wooden boxes were placed within each cage to provide shelter. Diced vegetables, seeds and live fly larvae were provided daily. As for the voles, provision of water was unnecessary. Animal maintenance and treatment were in accordance with the guidelines of the ‘Helsinki Committee of the Faculty of Life-Sciences and Medical School in Tel-Aviv University’. Based on our experience in maintaining these species for many years, the number of animals in each cage was reasonable. We also placed animals of the same size and age together and provided plenty of shelters to minimize aggression. All the animals used in the study were in good physical condition and seemed to live peacefully in this group size. However, we did not know the dominance hierarchy of individuals; we assumed that the predatory threat we used was strong enough to affect any social rank. At the end of the experiments, the animals were returned to the exhibition and breeding colonies that are currently maintained in our research zoo.
The Call Box This apparatus was a box measuring 3040 cm and 30 cm high with a transparent front wall and three opaque walls. A small ‘burrow’ (820 cm and 6 cm high) in the floor of the call box could be entered via an oval hole (35 cm). When we tested the spiny mice, we added a hollowed rock (ca. 111511 cm) to the far corner of the box to provide additional shelter. This setting was designed to mimic the natural habitat of these rodents, since voles are burrow dwellers and spiny mice live among rocks and stones. In the call box the animals were exposed to playbacks of tape-recorded stimuli by means of a small loudspeaker (2.5 inch (6.35 cm), 8 , 0.2 W max) placed on top of the rear wall, 25 cm above floor level. The sound level transmitted through the loudspeaker was standardized to 60–70 dB. The front wall of the box and the burrow consisted of transparent
EILAM ET AL.: RESPONSE OF RODENTS TO OWL CALLS
Plexiglas, to allow videotaping of the animals. The apparatus was placed in an empty room illuminated with red light (300 W), facing a video camera (Sony CCD-v700E). We wiped clean the call box with a cloth before testing each animal.
Procedure We introduced the animals to the test conditions 5 days before testing. We took them out of their cages and transferred them individually to the test room, into a wooden box (3040 cm and 30 cm high; no burrow, hollowed rock or transparent walls) for 5 min of habituation to the test procedure. This acclimation procedure was repeated once daily for 3 consecutive days in order to reduce stress caused by the test conditions. After these 3 days of pre-exposure, the animals were not touched for 1 day (the day before testing). The next day, we placed the animals in a corridor next to the test room for 1 h in order to habituate them to the laboratory environment, and then they underwent the test in the call box. We made our observations at dusk (1700–2000 hours), when these rodents show peak activity (Mendelssohn & Yom-Tov 1987). We gently trapped each animal in a jar and placed it in the shelter (burrow or the rock hollow) in the apparatus, exposed it to the sound stimuli and videotaped it for 10 min. Each animal was tested only once, in either the test group or one of the control groups. Owl call and control stimuli were used in counterbalanced order. Immediately after the 10-min test, we took a blood sample from each animal and measured the glucocorticosteroid level (see below). In the test groups (N=19 for voles and N=8 for spiny mice), the 10-min behavioural test session comprised three phases. (1) The precall period was 3 min of playback of a recorded human male voice, monotonically reading a story in English. We used this sound as a ‘neutral stimulus’, and the monotone guaranteed that the animals were not distracted by sudden changes in sound. This period provided the animal with an opportunity to habituate to the test set-up and separated in time the effect of owl calls and the initial exploration of a novel environment. Preliminary observation revealed that the animals were still active after 3 min, whereas with a longer precall period they sometimes rested inside the burrow so that the effect of the owl calls was not discernible. (2) The call period was 2 min of playback of continuous territorial tawny owl calls (obtained from the British Library National Sound Archive, London, U.K.), providing the test animal with as much information concerning the presence of an owl as would be available to it in the wild. All the test animals heard the same 2-min recording. The possibility that the response is to sound novelty (the change from human voice to owl calls), and not to owl calls, has been disproved in previous studies, which revealed that stimuli such as calls of other birds, sounds of terrestrial predators and jumbled owl calls did not affect the behaviour of rodents in the same way as owl calls (Hendrie & Weiss 1994; Hendrie et al. 1998).
(3) The postcall period was 5 min of silence, simulating the hunting pattern of the tawny owl, which, after a period of territorial calling, flies to a hunting perch where it waits in complete silence to pounce on its prey (Mikkola 1983). Results in previous studies in the laboratory (Hendrie & Weiss 1994) and on wild rodents (Hendrie et al. 1996, 1998) revealed that these three phases of testing simulate well the predatory threat and hunting pattern of tawny owls. We tested two control groups in parallel with the experimental groups. (1) The human voice of the precall period continued throughout the call period (this control thus comprised 5 min of human voice followed by 5 min of silence; N=12 for voles and N=9 for spiny mice). This control was the same as the experimental procedure except for the owl call and therefore reflects the effect of test procedure. (2) To obtain a baseline, we took animals from the home cage and bled them directly to determine stress hormone levels (N=18 for the voles and N=10 for the spiny mice). This control group provides a reference for the effect of the experimental procedure without the owl call (the ‘human voice’ group), and for any additional effect of the test stimuli (‘owl group’).
Bleeding At the end of the 10-min test (5 min after the owl calls ceased), each vole and spiny mouse was removed from the call box, quickly anaesthetized with Halothane, and bled (100–500 l of blood) from the eye sinus with a heparin-coated capillary. The bleeding procedure was completed within 3 min. We then washed the eye with tetrazoline ophthalmic solution, and returned the animal to its the home cage. Within minutes the animals seemed to behave normally, and on the next day we could not distinguish which of the eyes was bled. We used the serum to evaluate the level of glucocorticosteroids. In the vole, the major glucocorticosteroid is corticosterone, which we measured using an 125I double-antibody RIA kit (ICN Biomedicals, Inc., Costa Mesa, CA, U.S.A.). In the spiny mice, the major glucocorticosteroid is cortisol (Lamers et al. 1986), which we measured using a Tritum RIA Kit (ICN Biomedicals, Inc.).
Behavioural Analysis A VITC timecode was recorded on the videotapes, allowing an accurate identification of each frame on screen. The VITC was read to a computer by a special device (EasyReader II, Telcom Research, Burlington, Ontario, Canada). A custom-designed computer program allowed us to analyse the frequency and duration of the following criteria and locomotor patterns. (1) Distance travelled is the distance that the animals travelled during the 10-min test. This was further divided into edge distance (8-cm-wide zone along the walls of the apparatus) and centre distance (the remainder of the apparatus).
1087
Distance travelled (m)
ANIMAL BEHAVIOUR, 58, 5
90 80 70 60 50 40 30 20 10 0
90
(a)
60 45 30 15 10
20
30
40
0
50 1200 (c) 1000
800
800
600
600
400
400
200
200 10
20 30 Rank order
40
50
an eze flee um l fre wl H O w O
(d)
1000
0
(b)
75
1200 Corticosterone (ng/ml)
1088
0
e e e n le ez in ma el re wl f u f l H O w O
s Ba
Figure 1. (a) The distances travelled by each vole from both the group exposed to a human voice and the group exposed to owl calls. The individual distances were pooled and then ranked from low to high along the X axis. As shown, voles exposed to the human voice (C) were in the middle ranks, whereas voles exposed to owl calls dichotomized to a ‘freeze’ group () and a ‘flee’ group (). (b) The distance travelled (X±SE) by each group (overall effect of exposure to owl calls: F2,28 =19.01, P
