In multipredator environments, animals encounter the risk

Behavioral Ecology Vol. 7 No. : 0-4 Microhabitat use and behavior of voles under weasel and raptor predation risk: predator facilitation? Erkki Korpimaki, Vesa Koivunen, and Hani Hakkarainen Laboratory of Ecological Zoology, Department of Biology, University of Turku, FIN-20500 Turku, Finland An example of predator facilitation is that a microhabitat shift in a prey species induced by one predator increases the probability of the prey falling victim to other predators. Least weasels (Mustela nivalis) hunt in dense plant cover, whereas kestrels (Falco tinnunculus) hunt in habitats with sparse plant cover. Field voles (Microtus agrestis), the main food of weasels and kestrels, prefer open country with a high grass layer. We simulated a multipredator environment in an aviary (.0 X 4.8 X 2.2 m) to find out whether predator facilitation plays a role in the interactions between voles, small mustelids, and raptors. In each replicate, we placed a field vole in a pen including sides of high and low grass layers (cover and open). In a predator-free situation, voles preferred cover but shifted to open when a weasel was introduced to cover. In the presence of a kestrel, voles occupied cover and decreased their mobility. In the presence of a weasel plus a kestrel, voles behaved as under the kestrel risk alone. Therefore, in these aviary circumstances, voles perceived the kestrel risk as greater than the weasel risk. Predator facilitation in the assemblage of predators subsisting on rodent prey may contribute to the crash of the four-year vole cycle: microhabitat shift due to an avoidance of weasel jaws may drive voles to raptor talons. Key mords: antipredatory behavior, field vole, four-year vole cycle, kestrel, least weasel, microhabitat shift, predator facilitation. [Behav Ecol 7:0-4 (996)] In multipredator environments, animals encounter the risk of being killed by an assemblage of predators covering the spectrum from extremely hazardous to rather harmless (e.g., Polis and Holt, 992; Polis etal., 989; Taylor, 984). Antipredatory behaviors reducing mortality from one predator type may not be efficient against another type and may even increase the risk of being killed by other predators. Predator facilitation means behaviors (e.g., habitat or microhabitat shift) of prey animals that decrease mortality from one predator but which expose them to a second predator (Charnov et al., 976). For example, in an aviary, barn owls (Tyto alba) caught gerbils (GcrbUlus allenbyi and G. pyramidum) more from open microhabitats, and gerbils responded by shifting to forage in the bush microhabitat (Kotler et al., 99) where they encountered increased risk of snake predation (Kotler et al., 992). Other examples of predator facilitation mainly come from studies in aquatic ecosystems (e.g., Lima, 992; Power, 984; Walls et al., 990 and references therein). Small rodents are the main food of a diverse assemblage of predators consisting of many different-sized carnivorous mammals, diurnal raptors, owls, and snakes, especially, when voles are abundant in the peak phase of the four-year population cycle (e.g., Erlinge et al., 98; Goszczynski, 977; Korpimaki and Norrdahl, 99b; Korpimaki et al., 99). Small mustelids (the stoat Mustela erminea and the least weasel M. nivalis nivalis) and raptors create highly different risks for microtine rodents, and this may induce predator facilitation. Small mustelids mainly use olfaction to discover prey (King, 989) and can dius locate their prey at a distance of several meters. Because of their small body size, least weasels are able to hunt in the runnels of small rodents (Erlinge et al., 974; Korpimaki et al., 99). Their smallness makes them effective hunters in a habitat with high plant cover, which also offers a pro- Received 0 November 994; first revision 22 January 995; second revision March 995; accepted 8 March 995. 045-2249/96/S5.00 C 996 International Society for Behavioral Ecology tection against their own enemies, like raptors (Korpimaki and Norrdahl, 989; Powell, 97). Least weasels actively search for their prey and catch it after a pursuit but can also attack prey in ambush (Erlinge et al., 974). Small mustelids scent-mark their home ranges (King, 989), which reveals their presence to small rodents that can recognize their odor and behave differently than when exposed to the odors of larger carnivores (Jedrzejewski et al., 99). The presence of small mustelids or their scent induces a fast shift in temporal activity patterns of microtine rodents (Jedrzejewska and Jedrzejewski, 990). Diurnal raptors mainly use eyesight and nocturnal owls hearing to discover prey. They hunt from a perch or on wing locating vole prey even 00-200 m away [e.g., hovering Eurasian kestrels (Falco tinnunculus), hereafter kestrel, Village, 990]. Because dense vegetative cover forms a shelter against avian predators, hunting raptors prefer habitats with low vegetative cover (e.g., Janes, 985; Korpimaki, 986; Preston, 990). Also in an aviary, prey discovery and strike success of great-horned owls (Bubo virginianus) were higher in the open than in the cover (Longland and Price, 99). Field voles (Microtus agrestis) are herbivorous rodents that occupy open country with high vegetative cover (Hansson, 987; Myllymaki, 977). In our study area in western Finland, field voles prefer uncultivated farmland areas and pastures (Norrdahl and Korpimaki, 99). In this area, small mustelids, die kestrel, and two owl species (Asio otus and A. flammeus) are the main open-country predators of voles (Korpimaki and Norrdahl, 99a,b; Korpimaki et al., 99). When a small mustelid enters a microhabitat (i.e., a patch within the home range of a vole) preferred by field voles, voles can stay there and reduce their activity to decrease mustelid predation risk, or they can shift to an open patch with high avian predation risk. We conducted an aviary experiment to find out how voles actually respond in multipredator situations and whether predator facilitation plays a role in the interactions between voles, small mustelids, and raptors. We asked four questions: () Do field voles prefer a microhabitat Downloaded from https://academic.oup.com/beheco/article-abstract/7//0/2292 by guest on 0 January 209

Korpimaki et al. Predator facilitation in small rodents with high grass layer (cover) in a predator-free situation? (2) If so, do field voles alter their microhabitat use and behavior when a least weasel is introduced to cover? () Do field voles change their microhabitat use and behavior when a kestrel is present in the aviary? (4) What do field voles do in the presence of both a weasel and a kestrel? MATERIAL AND METHODS We conducted experiments in August 99 in an aviary (bottom.0 X 4.8 m, height 2.2 m) situated at the Satakunta Environmental Research Center of Turku University in Pori, western Finland (6 0' N, 2 0' E). The field voles (n = 70) were caught in July and August 99 in the vicinity of Pori and Turku (60 0' N, 22 0' E) using Swedish Ugglan livetraps. The two male least weasels were captured in early August 99 in the Kauhava region, western Finland (6 N, 2 E) with Ugglan live-traps. The kestrels [7 adult (age +l-yr) females and 6 adult males] were trapped in the same area in late June 99 by using a bal-chatri (Berger and Mueller, 959). Both least weasels and kestrels were trapped and held in captivity with the permission of the Finnish Ministry of the Environment. After the experiment, we released them in the same area where they were trapped. We used two identical uncovered pens (bottom 2. X.6 m, height 0.5 m) located symmetrically around a perch (height.4 m) for the kestrel. Both pens were partitioned off in two equal parts, and these parts were connected with each other by two plastic tubes (diameter 5 cm, length 0 cm) so that voles were free to move between two parts of a pen. We simulated an agricultural environment by placing a 4 cm deep hay layer to one part of both pens (a good microhabitat, hereafter referred to as cover) and a 0.5 cm deep hay layer to a second part of both pens (a poor microhabitat, referred to as open). Each part of the two pens also contained an open feeding place for voles (pellets for laboratory mice and water). We recorded the behavior of voles from a hide located on the wall (.5 m above the bottom) of the aviary. Before each replicate, we measured the body mass and checked the reproductive condition of voles (females: vagina open or closed; males: testis developed or not). In each replicate, we used two similar (in sex and body size) field voles. Each trial lasted for 45 min. We introduced each vole randomly to one part of the pen (one vole per pen). In the first 5 min., voles were allowed to become familiar with the pen. In the next 5 min., we recorded the behavior of voles to confirm that they behaved normally. In the control treatment, hereafter CT, vole behavior was recorded in the last 5 min. of each trial without further experimental intervention. In the three experimental treatments, the first two 5-min. periods were the same as in the CT. In the kestrel treatment (KT), we released a kestrel in the aviary and observed the behavior of field voles for 5 min. After a release, the kestrel usually flew around the aviary for -2 min. and thereafter scanned from the perch at a distance of 2-4 m from the pens. We used fullfed (satiated) falcons to remove the effects of strikes on vole behavior. In the weasel treatment (WT), we introduced a least weasel to cover of both pens (one weasel per pen) in a small cage (0 X 40 X 0 cm). The weasel was placed on the cover side of the pens only because in agricultural fields least weasels prefer to hunt in protective cover (Korpimaki E and Norrdahl K, unpublished radio-tracking data). Thereafter, we recorded vole behavior for 5 min. We used full-fed (satiated) weasels to reduce the effects of movements on vole behavior. In the kestrel plus weasel treatment (KWT), we released a kestrel in the aviary and placed a least weasel in cover of both pens, and recorded vole behavior for 5 min. In the last two treatments, the bottom of the pen and hay of the cover treatment were replaced by a similar bottom and hay to exclude the possibility diat the weasel odor remained on the floor and grass layer. Both kestrels and weasels were well accustomed to the test cages before the experiment so that they did not frantically try to escape. We carried out a total of replicates. In the experiment, we used some voles more than once because the total number of voles needed was 88 (2 pens X 4 treatments X replicates), but we had only 70 voles. However, we introduced each vole only once to one pen and treatment, and the order in which we subjected voles to the different treatments was random. This promotes the independence of replicates. Each kestrel was introduced to only one replicate. Between trials, we kept voles individually in cages (bottom X m, height 0.5 m), where they had unlimited access to food and water. The use of microhabitats (cover/open) and behavior of voles in two trial pens was recorded at -min. intervals during die 5-min. treatment. We classified vole behavior as follows: () moving when a vole was mobile in the pen, (2) staying when a vole was immobile in the pen, () eating when a vole was feeding on pellets and drinking water, and (4) cleaning when a vole was grooming its coat, RESULTS Field voles mostly inhabited cover when no predators were in the aviary (mean proportion of time inhabiting cover during the 5-min. trial: pen, 9% and pen 2, 8%). The experimental treatment significantly affected the microhabitat occupancy of voles, whereas the sex of voles and trial pen did not have obvious effects on microhabitat use (Figure and Table ). We found no significant interactions between treatment, sex, and trial pen. The between-treatment difference mainly resulted from voles avoiding cover in WT, whereas in KT and in KWT voles stayed in cover [pen (Figure, above): Tukey test for the difference in the arcsine-transformed proportion of time of voles inhabiting cover between CT and WT, two-tailed p <.00; between KT and WT, p <.00; between WT and KWT, p =.002] [pen 2 (Figure, below): p <.00, p <.00 and p <.00, respectively]. Also, the experimental treatment strongly altered the behavior of field voles, whereas the sex of voles, trial pen, and interaction between the three independent variables had no obvious influence on vole activity (Figure 2 and Table 2). Voles decreased their mobility as a response to KT (Tukey test for die difference in die arcsine-transformed proportion of time moving between CT and KT: pen, two-tailed p =.06; pen 2, p =.; pooled data from two pens p.005), whereas WT did not result in die similar response (p =.99, p =.89, and p =.94, respectively). In KWT, voles reduced their mobility compared with CT (pen, p =.20; pen 2, p =.04; pooled data from two pens p =.007). Field voles mosdy moved or stayed in the trial pens and, therefore, showed no obvious between-treatment difference in the time of eating and grooming (Figure 2). We also examined whether body mass and reproductive condition of voles influenced dieir microhabitat use and behavior in trial pens. The only effect diat approached statistical significance was that males with developed testis tended to occupy cover less dian those with nondeveloped testis (F = 4.00,p=.052). DISCUSSION Four main findings emerged in our experiment First, field voles preferred cover in a predator-free situation. Second, field voles shifted to open when a least weasel entered cover. Third, field voles preferred cover under the kestrel predation Downloaded from https://academic.oup.com/beheco/article-abstract/7//0/2292 by guest on 0 January 209

2 Behavioral Ecology Vol. 7 No. 0.0 ao P«n P«n 2 Figure (Top) The mean (±SD) proportion of time per replicate (n =, 5 min. each) of field voles occupying cover and open microhabitats in four experimental treatments (CT = control, KT = kestrel present, WT = least weasel present, and KWT = both kestrel and weasel present) in pen. (Bottom) The same but for pen 2. risk and reduced dieir mobility. In accordance with our results, in a large enclosure (44 m!, each pen 4 X 2 m), bank voles (Clethrionomys giarsoku) also avoided the pen visited by the common weasel (Mustela nivalis vulgaris) (Jedrzejewski and Jedrzejewska, 990). Common voles (Microtus arvalis), exposed to a kestrel model, also reduced dieir mobility (Gerkema and Verhulst, 990). Finally, simultaneous presence of a kestrel and a weasel induced a response much like the response to die kestrel alone, suggesting that in diese aviary circumstances field voles perceived the kestrel predation risk as more hazardous than the Table ANOVA-table for the proportion of time of field voles inhabiting cover on the experimental treatment (control, kestrel present, weasel present, both kestrel and weasel present), sex of volei, and trial pen ( or 2) in replicates (5 min. each) Source of variation Treatment (T) TX df 7 Mean square 5.76 0.20 0.05 0.0 0.5 0.0 0.7 0.22 F 26.50 0.90 0.24 0.2 0.68 0.0 0.79 P <.00.5.6.95.57.86.50 The proportion of time was arcsine-transformed because of nonnormal distribution. weasel predation risk. An explanation for this rather unexpected result may be an aviary artifact For example, kestrels might be noisy whereas least weasels are silent, but kestrels in our experiment were also silent. In addition, voles might discover that a caged weasel is unable to pursue and attack whereas a free-flying kestrel can do so. However, this explanation seems unlikely because satiated kestrels in our trials did not strike voles, and the presence of a caged weasel in cover induced a significant avoidance of this pen. The microhabitat shift in the presence of a captive weasel was probably attributable to weasel scent, which even humans can smell. Also, laboratory experiments indicate that microtine rodents may assess weasel odor as a real direat (Heikkila et al., 99; Ylonenetal., 992). Least weasel densities usually peak in the crash phase of the four-year vole cycle (Korpimaki et al., 99; Oksanen and Ok- Table 2 ANOVA-table for the proportion of time of field voles moving (A) and staying (B) in the trial pens on the experimental treatment (control, kestrel present, weasel present, both kestrel and weasel present), sex of voles, and trial pen ( or 2) in replicates (5 min. each) Source of variation df (A) Dependent variable: time Treatment (T) T X S x P 7 (B) Dependent variable: time Treatment (T) T X S X P 7 Mean square of moving 0.72.00 0.00 0. 0.02 0. 0.0 of staying 2.42.00 0.22 0.09 0.09 0.02 0.2 F 9.07.2 0.06.62 0.28.4 0..58 0.46 0.8.04 0.42 0.45 P <.00.27.8.9.84.25.80 <.00.50 4.8.74.5.97 The proportion of time was arcsine-transformed because of nonnormal distribution. Downloaded from https://academic.oup.com/beheco/article-abstract/7//0/2292 by guest on 0 January 209

Korpimaki et al. Predator facilitation in small rodents I I.0 O8 O6 O4 i «0.0. CT 2. KT P«n Pen 2. WT Eating Ctoanlng Staying D Moving 4. KWT Figure 2 (Top) The mean (±SD) proportion of time per replicate (n =, 5 min. each) used by field voles in various behaviors (moving, staying, cleaning the coat, and eating) in four experimental treatments (CT = control, KT = kestrel present, WT = least weasel present, and KWT = both kestrel and weasel present) in pen. (Bottom) The same but for pen 2. sanen, 992). Least weasels are patch-searching predators that stay in good vole patches until they have been depleted (Henttonen, 987; Korpimaki et al., 99 and unpublished radiotracking data). Our results suggest that when least weasels find a field vole patch in cover, their presence or even scent may induce a microhabitat shift of voles, which may increase the exposure of voles to avian predation. Therefore, predator facilitation may contribute to a vole crash, as microhabitat shift due to an avoidance of weasel jaws may drive voles to raptor talons. Much field data show that in the peak and crash phases of the four-year population cycle, microtine rodents (including field voles) inhabit uncommon habitats (e.g., Hansson, 969; Henttonen et al., 977; Myllymaki, 977; Norrdahl and Korpimaki, 99). However, the importance of predator facilitation to these habitat shifts remains undocumented because our experiment was made in the scale of a vole home range. Many factors also can induce these shifts (intra- and interspecific competition for space, lack of high quality food, etc.). Future studies on radio-collared voles and least weasels could reveal the importance of predator facilitation in die field. Natural spatial scale of experiments is important when studying prey behavior in the presence of predators (e.g., Korpimaki et al., 994; Lima and Dill, 990). We are well aware that the abnormal spatial scale of our enclosure experiment may restrict the application of the results to field circumstances. Despite diis, our results suggest that predator facilitation may be potentially important in the assemblage of mammalian and avian predators subsisting on microtine rodent prey. The existence of multipredator environments has been largely neglected in studies on antipredatory behavior of microtine rodents. So far, the behavior of voles has been studied under die risk of only one predator, like small mustelids (e.g., Heikkila et al., 99; Jedrzejewska and Jedrzejewski, 990; Ylonen et al., 992) or raptors (e.g., Gerkema and Verhulst, 990; Hakkarainen et al., 992). We thank the staff of Satakunta Environmental Research Center, especially Jukka Jussila and Mikko Ojanen, for providing good working facilities, and Voitto Haukisalmi, Bogumila Jedrzejewska, Wlodzimierz Jedrzejewski, Kai Norrdahl, Christoph Rohner, and anonymous referees for comments on the manuscript. The study was supported by the Academy of Finland. REFERENCES Berger DD, Mueller HC, 959. The bal-chatri: a trap for the birds of prey. Bird-banding 0:8-26. Charnov EL, Orianj GH, Hyatt K, 976. Ecological implications of resource depression. Am Nat 0:247-259. Erlinge S, Goransson G, Hansson L, Hdgstedt G, Liberg O, Nilsson IN, Nilsson T, von Schantz T, Sylven M, 98. Predation as a regulating factor in small rodent populations in southern Sweden. Oikos 40:6-52. Erlinge S.Johansson B, Willstedt H, 974.Jaktbeteende och bytesval hos smavesslan. Fauna o Flora 69:95-0. Gerkema MP, Verhulst S, 990. Warning against unseen predator: a functional aspect of synchronous feeding in the common vole, Microtus arvalis. Anim Behav 40:69-78. Goszczynski J, 977. Connections between predatory birds and mammals and their prey. Acta Theriol 22:99-40. Hakkarainen H, Korpimaki E, Mappes T, Palokangas, P, 992. Kestrel hunting behaviour towards solitary and grouped Microtus agratis and M. ephoticus a laboratory experiment. Ann Zool Fennici 29: 279-284. Hansson L, 969. Spring populations of small mammals in central Swedish Lapland in 964-68. Oikos 20:4-450. Hansson L, 987. An interpretation of rodent dynamics as due to trophic interactions. Oikos 50:08-8. HeikkilaJ, Kaarsalo K, Mustonen O, Pekkarinen P, 99. Influence of predation risk on early development and maturation in three species of Clethrionomys voles. Ann Zool Fennici 0:5-6. Henttonen H, 987. The impact of spacing behavior in microtine rodents on the dynamics of least weasels MusUta nivalis a hypothesis. Oikos 50:66-70. Henttonen H, Kaikusalo A, Tast J, Viitala J, 977. Interspecific competition between small rodents in subarctic and boreal ecosystems. Oikos 29:58-590. Janes SW, 985. Habitat selection in raptorial birds. In: Habitat selection in birds (Cody ML, ed). London: Academic Press; 59-88. Jedrzejewska B, Jedrzejewski W, 990. Antipredatory behaviour of bank voles and prey choice of weasels enclosure experiments. Ann Zool Fennici 27:2-28. Jedrzejewski W, Jedrzejewska B, 990. Effect of a predator's visit on Downloaded from https://academic.oup.com/beheco/article-abstract/7//0/2292 by guest on 0 January 209

4 Behavioral Ecology Vol. 7 No. the spatial distribution of bank voles: experiments with weasels. Can J Zool 68:660-666. Jedrzejewski W, Rychlik L, Jedrzejewska B, 99. Responses of bank voles to odours of seven species of predators experimental data and their relevance to natural predator-vole relationships. Oikos 68: 25-257. King CM, 989. The natural history of weasels and stoats. London: Christopher Helm. Korpimaki E, 986. Diet variation, hunting habitat and reproductive output of the kestrel FaXco tinnunculus in the light of the optimal diet theory. Ornis Fenn 6:84-90. Korpimaki E, Norrdahl K, 989. Avian predation on mustelids in Europe. I: Occurrence and effects on body size variation and life traits. Oikos 55:205-25. Korpimaki E, Norrdahl K, 99a. Do breeding nomadic avian predators dampen population fluctuations of small mammals? Oikos 62: 95-208. Korpimaki E, Norrdahl K, 99b. Numerical and functional responses of kestrels, short-eared owls, and long-eared owls to vole densities. Ecology 72:84-826. Korpimaki E, Norrdahl K, Rinta-Jaskari T, 99. Responses of stoats and least weasels to fluctuating vole abundances: is the low phase of the vole cycle due to mustelid predation? Oecologia 88:552-56. Korpimaki E, Norrdahl K, ValkamaJ, 994. Reproductive investment under fluctuating predation risk; microtine rodents and small mustelids. Evol Ecol 8:57-68. Kotler BP, Blaustein L, Brown JS, 992. Predator facilitation: the combined effect of snakes and owls on foraging behaviour of gerbils. Ann Zool Fennici 29:99-206. Kotler BP, Brown JL, Hasson O, 99. Factors affecting gerbil foraging behavior and rates of owl predation. Ecology 72:2249-2260. Lima SL, 992. Life in a multi-predator environment: some considerations for anti-predatory vigilance. Ann Zool Fennici 29:27-226. Lima SL, Dill LM, 990. Behavioural decisions made under the risk of predauon: a review and prospectus. CanJ Zool 68:69-640. Longland WS, Price MV, 99. Direct observations of owls and heteromyid rodents: can predation risk explain microhabitat use? Ecology 72:226-227. Myllymaki A, 977. Interactions between the field vole Microtus agratis and its microtine competitors in Central-Scandinavian populations. Oikos 29:570-580. Norrdahl K, Korpimaki E, 99. Predation and interspecific competition in two Microtus voles. Oikos 67:49 58. Oksanen L, Oksanen T, 992. Long-term microtine dynamics in north Fennoscandian tundra: the vole cycle and the lemming chaos. Ecography 5:226-26. Polis GA, Holt RD, 992. Intraguild predation: the dynamics of complex trophic interactions. Trends Ecol Evol 7:5-54. Polis GA, Myers CA, Holt RD, 989. The ecology and evolution of intraguild predation: potential competitors that eat each other. Annu Rev Ecol Syst 20:297-0. Powell RA, 97. A model for raptor predation on weasels. J Mammal 54:259-26. Power ME, 984. Depth distributions of armored catfish: predatorinduced resource avoidance. Ecology 65:52-528. Preston CR, 990. Distribution of raptor foraging in relation to prey biomass and habitat structure. Condor 92:07-2. Taylor RJ, 984. Predation. London: Chapman and Hall. Village A, 990. The kestrel. London: Poyser. Walls M, Kortelainen, SarvalaJ, 990. Prey responses to fish predation in freshwater communities. Ann Zool Fennici 27:8-99. Ylonen H, Jedrzejewska B, Jedrcejewski W, HeikkilaJ, 992. Antipredatory behaviour of Ciethrionomys voles "David and Goliath" arms race. Ann Zool Fennici 29:207-26. Downloaded from https://academic.oup.com/beheco/article-abstract/7//0/2292 by guest on 0 January 209