- Email: [email protected]
Assessment and defence of solitary kangaroo rats under risk of predation by snakes
ANIMAL BEHAVIOUR, 2001, 61, 579–587 doi:10.1006/anbe.2000.1643, available online at http://www.idealibrary.com on
Assessment and defence of solitary kangaroo rats under risk of predation by snakes JAN A. RANDALL & DENISE K. BOLTAS KING
Department of Biology, San Francisco State University (Received 23 February 2000; initial acceptance 5 June 2000; final acceptance 16 September 2000; MS. number: A8722R)
Prey that spend a large proportion of their time foraging must trade-off between predator defence and feeding. Because predation risk can be a cost in desert granivores, we predicted that kangaroo rats (Dipodomys) should weigh the risk of predation when deciding which behavioural option to pursue. We manipulated apparent predation risk of the giant kangaroo rat, D. ingens, and the desert kangaroo rat, D. deserti, and predicted they should be more vigilant and show greater defensive behaviour to the more dangerous stimulus of a real snake than to an artificial snake decoy. Results showed that kangaroo rats decreased foraging with increased vigilance when the snake and decoy were present. The largest decrease in foraging was in response to the live snake, and both species oriented towards, approached and footdrummed more in the presence of the live snake than the decoy. The species also differed in their antipredator behaviour: D. ingens footdrummed more and spent more time within striking distance of the snake than the decoy. Dipodomys deserti spent more time within 1 m of the snake than the decoy and kicked sand at the snake. Sand kicking was never observed in D. ingens. We conclude that kangaroo rats are able to discriminate predation risk. They decrease foraging and increase vigilance in the presence of live snakes to assess risk and may approach and footdrum as a pursuit deterrent to communicate to the snake its chances of ambush are no longer available to cause the snake to leave. Sand kicking may function to harass the snake to cause it to leave.
own defence (Elgar 1989; Scheel 1993). Dugkatin & Godin (1992) proposed four benefits for why prey approach and interact with predators: (1) to acquire information about the nature of the potential threat, (2) inform others of threat, (3) deter predator attack and (4) advertise one’s quality to potential mates. Of these, solitary animals most likely approach predators to gain information and to deter attack. They also may communicate directly to the predator to cause an attack or pursuit to be abandoned (Caro 1986a, b; Hasson 1991; Caro et al. 1995). Such behaviour deters pursuit by informing the predator that the chances of ambush are thwarted (Woodland et al. 1980), continued pursuit is costly (Caro 1995; Randall & Matocq 1997), or the prey is healthy and cannot be caught (FitzGibbon & Fanshawe 1988; Caro 1995). Although smaller species of kangaroo rats (D. merriami), avoid predators, larger species act in their own defence in response to snakes (Randall et al. 1995). The bannertailed kangaroo rat, D. spectabilis, approaches snakes to within striking distance, jumps back and footdrums. Randall & Matocq (1997) interpreted this behaviour to function in predator deterrence (Caro 1995): the close approach informed the snake it was detected, and
Herbivores must trade-off between watching for predators and feeding (Elgar 1989; Lima & Dill 1990), because vigilant prey have a lower chance of being captured by a predator than those that are not vigilant (FitzGibbon 1989; Illius & FitzGibbon 1994). It is well documented that predation is a cost of foraging in desert granivores such as kangaroo rats (Dipodomys) and gerbils (Gerbillus) (Daly et al. 1990, 1992; Kotler et al. 1991, 1994; Brown et al. 1994; Lima 1998). Three factors make desert rodents especially susceptible to predation risk: they are solitary and forage alone, they often forage in open habitats with little cover, and they have patchy food sources that require movement between patches. Solitary rodents probably have to spend more time watching and listening for predators than those in groups (McNamara & Houston 1992; Roberts 1996). Rodents are often depicted as sedentary victims of predators (Brown et al. 1999), but prey are not always passive players in the predator–prey game (Fishman 1999; Swaisgood et al. 1999a, b). Although it seems counterintuitive, prey sometimes interact with a predator in their Correspondence: J. A. Randall, Department of Biology, San Francisco State University, San Francisco, CA 94132, U.S.A. (email: [email protected]). 0003–3472/01/030579+09 $35.00/0
2001 The Association for the Study of Animal Behaviour
579
2001 The Association for the Study of Animal Behaviour
580
ANIMAL BEHAVIOUR, 61, 3
footdrumming signalled the snake that the kangaroo rat was alert and would not be easy prey. The possibility that kangaroo rats approach snakes to assess and gain information about potential threat was not addressed in this study (Randall & Matocq 1997). Kangaroo rats should weigh the risk of predation when deciding what behavioural option to pursue when they encounter snakes. We predicted that if kangaroo rats can assess the danger posed by snakes, they should display a different set of responses to the more dangerous stimulus of a real snake than to an artificial snake decoy. They should be more vigilant, inspect more and show greater defensive behaviour in the presence of a live snake. We also predicted that species differ in their response to predators based on occupancy of different habitats and differences in social tolerance. The giant kangaroo rat, D. ingens, occupies arid grasslands with a rich food source that does not require them to make long-distance foraging trips (Braun 1985). The desert kangaroo rat, D. deserti, is a sand dune specialist that occupies a drier habitat with a less productive food source than the grassland habitat of D. ingens. Dipodomys deserti makes longer foraging trips, visits the burrow of neighbours to pilfer seeds at higher rates and is generally more aggressive than D. ingens. They aggressively chase neighbours that visit their burrows and engage in rollover fights (Stern 1980; Randall 1997, unpublished data). We predicted that this social intolerance and aggressive behaviour of D. deserti might cause them to be more aggressive towards snakes than D. ingens. METHODS
Study Sites We studied D. ingens on a 1.5-ha study area in the Carrizo Plain Natural Area located in California Valley, San Luis Obispo County, California, U.S.A. from 15 June to 10 August 1997. Vegetation consisted of non-native annual grasses (Vulpia, Bromus), annual native herbs (Lepidium, Erodium) and sparse native shrubs (Atriplex). We studied D. deserti near the Desert Studies Center on a 1.6-ha site in the Mojave National Preserve, 18 km south of Baker, San Bernadino County, California from 3 June to 8 July 1998. The habitat consisted of areas of loose sand dunes interspersed with sparse grasses (Bromus, Bouteloua, Achnatherum) and creosote (Larrea tridentata) hummocks. Because 1998 was an El Nin ˜ o year with record rainfall, vegetation was relatively abundant in the dune habitat.
Animals Dipodomys ingens is the largest species in the genus. Adult males (N=16) averaged 136.92.8 g and females 131.92.9 g (N=13) on our study site. Because of habitat loss, distribution is restricted to less than 2% of the former range and D. ingens is listed as an endangered species by both the U.S. Fish and Wildlife Service and the California Department of Fish and Game. Dipodomys deserti is smaller than D. ingens; males (N=10) averaged
114.83.86 g and females (N=13) averaged 101.54.5 g on our site. Dipodomys deserti is the only species of kangaroo rat that lives exclusively in sand dunes. Both species are solitary, and individuals of both species occupied individual burrows during our study. These desert rodents normally reproduce from late autumn to spring (Randall 1993, 1994), and thus, when this study was conducted breeding was finished and there was no evidence of young in the burrow. We trapped animals with wild bird seed in Sherman live-traps (30.57.69 cm) with a 0.25-cm space in the top of the door to prevent tail severance. We marked each individual with numbered ear tags (Monel 1005-1) covered with colour-coded reflective tape (Scotchlite) for individual identification at night. We also sexed, determined reproductive condition and weighed each animal to estimate age. We determined residency of a kangaroo rat at a burrow by trapping records, scan sampling and observation (Randall 1984). A kangaroo rat had to be trapped and observed at a particular burrow more than twice before we assigned residency.
Behavioural Observations and Footdrumming Recordings We used established techniques for observing kangaroo rats at night (Randall 1995, 1997). We sat in the dark approximately 10 m from a burrow entrance and observed behaviour with either a Generation II night vision scope mounted on a tripod or a Generation III night vision goggle. We noted behaviour by talking softly into a hand-held tape recorder (Table 1). Footdrumming was recorded using geophones connected to either a Uher 4400 reel-to-reel tape recorder or a Marantz PMD cassettetape recorder (Randall 1989a). We scattered a small amount of bird seed around the burrow at the beginning of an observation or test to keep the kangaroo rats active and to facilitate observations (Randall & Stevens 1987). We also scattered a small amount of seed at neighbouring burrows of D. deserti to minimize the number of visits by neighbours during our tests.
Stimuli We chose gopher snakes, Pituophis melanoleucus, as the stimulus animal because they are sympatric with both species of kangaroo rat and feed on them in the wild (Kotler 1984; Williams 1992), and they are present on our study sites. We initially used two snakes with snout–vent lengths (SVL) of 1.23 and 1.10 m and weights of 1106 and 907 g, respectively. Because we detected a tumor in the smaller snake after one test, we used only the larger snake for remaining tests. This caused a problem of pseudoreplication. However, because of the strong and consistent response of kangaroo rats to snakes of all sizes (Randall & Stevens 1987; Randall et al. 1995; Randall & Matocq 1997), we considered the use of this large, active, snake sufficient to elicit a normal range of responses, and the consistent use of stimuli provided consistency among tests.
RANDALL & BOLTAS KING: ANTISNAKE BEHAVIOUR
Table 1. Behaviours of Dipodomys ingens and D. deserti recorded in the presence of live and decoy snakes Behaviour
Approach Footdrum Forage Jump back Kick sand Orient Out of burrow Sniff Within striking Within 1 m
Description
Movement towards snake or decoy that decreases distance between focal animal and stimulus. The kangaroo rat strikes hind feet against the ground. Each strike is a footdrum; a series of drums is a footroll and one or more footrolls is a bout. Slow quadrupedal movement with head towards ground and digging with forepaws. A backward hop in the air that causes the rat to increase distance between itself and the stimulus. Kangaroo rat turns away from the stimulus and causes sand to fly towards it by sharp kicks of the hind feet. Duration of standing in a quadrupedal or bipedal posture holding body at 45–90° with its head and anterior part of its body towards the stimulus. Time in burrow subtracted from total time in test period. An approach to within 3 cm of the stimulus with nose lowered towards it. Kangaroo rat moves within 0.5 m (half snake’s body length) of head of snake or decoy. Distance between striking distance (0.5 m) and 1 m.
The snake was tethered using an established technique (Randall & Stevens 1987; Randall & Matocq 1997). To prevent damage to the snake’s skin, we first wrapped a piece of paper athletic tape around the snake, covering about one-third the length of the snake’s body below the head (33 cm), and then wrapped a strip of adhesive tape (3–4 cm wide) on top. We folded the adhesive tape back on itself and punched two holes in the flap, to which we tied two monofilament lines that were fastened to two large nails or tent stakes driven flush into the ground about 0.6 m apart. The snake could move and strike, but it did not strike at the kangaroo rats during our tests. The rubber snake decoy measured 1 m stretched out and was a replica of a real snake with a black body and yellow stripes. Its coiled silhouette in the dark appeared real.
specific to the stimuli in the test period (Table 1). Data recorded for comparison during the pretest, test and post-test included: (1) time spent foraging, (2) time spent footdrumming and (3) time out of the burrow. Behaviour specific to interactions with the snake and decoy included: (1) time oriented, (2) predator inspection as measured by number of approaches, time spent within 1 m, time within striking distance and time sniffing the snake, (3) frequency of jump backs and (4) number of sand kicks. We also recorded visits by neighbours and whether there were any interactions with focal animals or the stimuli. We recorded behaviour and footdrumming into continuously running recorders so that we could transcribe and time behaviour from the tapes.
Data Analysis Procedures We tested seven males and seven females of each species for their responses to the live gopher snake and rubber snake decoy in a counterbalanced order. We tethered the stimuli 1.5–2.0 m from the focal animal’s most frequently used burrow entrance. We determined burrow use during a 1-h focal observation period, which also served to habituate the kangaroo rat to our presence, on a night 1–12 days prior to the test. Each test lasted 30 min. We began a 10-min pretest observation when the focal animal came aboveground and foraged normally. After the pretest observation, we walked softly to the burrow area and tethered the stimulus. Although the kangaroo rats entered the burrow in response to our approach, they usually exited the burrow within 5–10 min to forage on the scattered seeds. We began the 10-min test only after the kangaroo rat exited the burrow and oriented to the stimulus for more than 5 s. At the end of the test, we quietly removed the stimulus and began a 10-min post-test observation. We tabulated one set of behaviours during all three time periods and another set that involved responses
We analysed data collected in the three test periods in a 223 design (speciesstimulustest period) and in response to the stimuli in 22 design (species stimulus) using repeated measures analyses of variance (ANOVAs) in SYSTAT 7.0 (Wilkinson 1997). (We omitted the sex variable because we found no differences in behaviour of males and females.) When data contained zeros, we performed a normalizing log+1 transformation (Zar 1984), and we used nonparametic statistics (Wilcoxon matched-pairs signed-ranks and Friedman tests) when data could not be normalized and assumptions for parametric tests were not met. We used paired t tests and Wilcoxon matched-pairs signed-ranks tests with Bonferroni corrections as post hoc tests. We performed statistical tests in SYSTAT 7.0, with the exception of Wilcoxon tests with an N less than 14, which were calculated by hand following Snedecor & Cochran (1967), and we report the multivariate F statistic (Wilks’ lambda) when possible because it does not require compound symmetry (Wilkinson 1997). All tests are two-tailed and data are reported as the meanSE of untransformed data.
581
ANIMAL BEHAVIOUR, 61, 3
Table 2. Comparisons of time Dipodomys ingens and D. deserti spent foraging (paired t tests, df=27) and footdrumming (Wilcoxon matched-pairs signed-ranks tests N=14) during pretest, test and post-test in response to a snake and decoy. Multiple comparisons are Bonferroni corrected
Behaviour
Pretest versus test
Pretest versus post-test
8 (a)
Pretest Test Post-test
7 6
Test versus post-test
5 4 3
Foraging Snake Decoy Footdrumming D. ingens Snake Decoy D. deserti Snake Decoy
6.2** 3.1*
4.62** 2.87*
1.75 (NS) 0.37 (NS)
0** 23 (NS)
9** 38 (NS)
45 (NS) 29 (NS)
4** 30 (NS)
33 (NS) 35 (NS)
0** 23 (NS)
Time foraging (min)
582
2 1 0 10 9
(b)
8 7 6
*P<0.05; **P<0.001.
5
We also analysed footdrumming of D. ingens to compare differences in patterns of drumming in the snake context with those used to communicate to conspecifics (Randall 1994, 1997). We were unable to analyse footdrumming of D. deserti because of infrequent drumming in any context except with the live snake. We compared drumming patterns of seven D. ingens that drummed five or more bouts of spontaneous drumming during pretests and habituations and in response to the snake stimuli. We analysed the recordings using established techniques (Randall 1989b, 1995), and compared (1) number of drums in the first footroll, (2) number of footrolls in a footdrumming bout and (3) rate of drumming (drums/s). RESULTS
Foraging The kangaroo rats decreased foraging in response to both the snake and decoy with less time foraging in the test than in the pretest (effect for time: Wilks’ lambda: F2,52 =33.15, P=0.0001; Table 2, Fig. 1). The largest decrease in foraging was in response to the live snake (timestimulus interaction: Wilks’ lambda: F2,52 =3.57, P=0.035). Time foraging decreased an average of 4.210.44 min between the pretest and test when the snake was tethered compared with an average decrease of 2.190.42 min in response to the decoy (Wilcoxon matched-pairs signed-ranks test: Z= 3.78, N=28, P=0.0001). Because time foraging remained lower in the post-test after the snake or decoy was removed, there was no difference in time foraging between the test and post-test, but there was a significant difference between the post-test and pretest (Table 2). Although D. deserti spent more total time foraging than D. ingens (main effect species: F1,52 =19.7, P=0.0001), both species spent significantly less time foraging during the test than pretest (Wilks’ lambda: speciestime interaction: F2,51 =7.48, P=0.001; paired t test: D. ingens: t27 =5.58, P=0.0001; D. deserti: t27 =4.9, P=0.0001; Fig. 1).
4 3 2 1 0
Snake
Decoy Stimulus
Figure 1. Time foraging (mean±SE) by (a) Dipodomys ingens (N=14) and (b) D. deserti (N=14) in response to a live gopher snake and an artificial snake decoy during three 10-min periods: pretest, test and post-test.
Dipodomys ingens also spent significantly less time foraging during the post-test than during the pretest (t27 =4.89, P=0.0001). Dipodomys deserti spent about the same amount of time foraging in the post-test as in the test (t27 =1.17, P=1.0) and during the pretest and post-test (t27 =2.59, P=0.09). Time foraging also did not differ for D. ingens during the post-test and test (t27 =1.1, P=1.0).
Avoidance The kangaroo rats did not enter the burrow in response to either the snake or decoy so that time out of the burrow did not differ significantly between the pretest (8.890.24 min), test (8.370.36 min) and post-test (6.910.50 min) (Friedman test: 22 =3.06, P=0.22), and there was no significant effect by the type of stimulus (21 =1.63, P=0.48). There was a significant difference by species (22 =15.88, P=0.001). Dipodomys ingens spent significantly less time out of the burrow in the post-test (4.20.68 min) than in the pretest (8.210.44 min) (Wilcoxon matched-pairs Z= 3.14, N=28, P=0.02) and test (7.150.63 min) (Z= 3.85, N=28, P=0.001). Dipodomys deserti spent equal time out of the burrow in the pretest (9.70.13 min), test (9.80.21 min) and post-test (9.60.19 min) (all comparisons NS).
RANDALL & BOLTAS KING: ANTISNAKE BEHAVIOUR
Kangaroo rats sometimes jumped back after they approached the snake. The species together averaged 1.320.33 jump backs from the snake and 0.500.37 from the decoy during the 10-min test (Friedman test: 21 =5.14, P=0.03). Taken separately, D. deserti jumped back more from the snake (1.140.29) compared with the decoy (0.290.16) (Wilcoxon matched-pairs signedranks test: T=8, N=14, P<0.05), but the frequencies did not differ significantly for D. ingens (snake: 1.50.57; decoy: 0.710.64; T=22, N=14, NS).
300
(a)
Snake Decoy
*
200
* 100
Vigilance and Predator Inspection
Defence Sand kicking was an important element of predator defence for D. deserti, but not for D. ingens. Thirteen of 14 D. deserti kicked sand directly towards the snake compared with six that kicked sand towards the decoy to yield a significantly higher frequency for the snake (14.13.9) compared with the decoy (5.12.9) (T=14, N=14, P=0.05). We never observed D. ingens kick sand in any test or observation. Both D. ingens and D. deserti footdrummed more in the presence of the snake than the decoy (Friedman test:
0 80
(b)
* 60 Time (s)
There was no significant difference in the amount of time that D. ingens and D. deserti oriented to the stimuli (ANOVA: main effect for species: F1,26 =2.4, P=0.134; Fig. 2a). Time oriented to the stimuli during the test showed differences in vigilance between the snake and decoy (main effect for stimulus: F1,26 =30.4, P=0.0001), and the kangaroo rats spent significantly more time oriented to the snake than to the decoy (paired t test: t27 =5.6, P=0.0001). Other comparisons were not significant. Both D. ingens and D. deserti approached the snake and decoy with no species differences found in the total number of approaches (ANOVA: main effect for species: F1,26 =0.04, P=0.95) but with significantly more approaches to the snake (5.540.93) than to the decoy (3.180.73) (main effect for stimulus: F1,26 =6.69, P=0.016). The interaction was not significant. Once kangaroo rats approached the stimulus, however, the species differed in the amount of time within 1 m (Friedman test: 22 =16.2, P=0.001) and striking distance (22 =8.41, P=0.015) of the snake and decoy (Fig. 2b,c). Although D. deserti spent significantly more time within 1 m of the snake compared with the decoy (Wilcoxon signed-ranks test: T=8.5, N=14, P<0.01), they did not spend significantly more time within striking distance of the snake compared to the decoy (T=25, N=14, NS). Dipodomys ingens, however, spent significantly more time within striking distance of the snake than the decoy (T=0, N=14, P<0.01) but showed similar times within 1 m of the snake and decoy (T=22, N=14, NS) (Fig. 2b,c). Species taken together sniffed the decoy (0.640.19) more frequently than they sniffed the snake (0.360.26) (Friedman test: 22 =5.2, P=0.02). All other comparisons were not significant.
40
20
0 80
(c)
60
40
* 20
0
D. ingens
D. deserti Species
Figure 2. Mean±SE time Dipodomys ingens (N=14) and D. deserti (N=14) spent (a) oriented towards (b) within 1 m of and (c) within striking distance of a live gopher snake and an artificial snake decoy during a 10-min test. *P<0.05.
22 =13.4, P=0.001; 22 =10.8, P=0.005, respectively) (Fig. 3). Dipodomys ingens drummed significantly longer in the test and post-test than in the pretest, but because they continued footdrumming into the post-test, time drumming during the test and post-test did not differ significantly (Table 2). Like D. ingens, D. deserti drummed significantly longer in the snake test than in the pretest, but unlike D. ingens, D. deserti also drummed significantly longer in the test than the post-test. No significant differences were found for drumming in the presence of
583
ANIMAL BEHAVIOUR, 61, 3
60 Pretest Test Post-test
(a) 50 40 30
Time footdrumming (s)
584
20 10 0 60 (b) 50
DISCUSSION
40 30 20 10 0
main effect for time: F1,26 =0.8, P=0.45; timestimulus interaction: F1,26 =0.5, P=0.45). Four focal D. deserti interacted with a snake at the burrow of a neighbour during the snake test of the neighbour (one kangaroo rat interacted with the snake at a neighbour’s burrow twice, so we calculated an average of the two encounters). When we compared the total time interacting with the snake at their home burrow compared with a neighbour’s burrow, we found that the kangaroo rats spent significantly more time interacting with the snake at home (2.10.5 min) than at a neighbour’s burrow (0.250.08 min) (paired t test: t3 =8.0, P=0.004). Visiting rats oriented to the snake, approached it and three kicked sand, but none footdrummed when away from the home burrow.
Snake
Decoy Stimulus
Figure 3. Time footdrumming (mean±SE) by (a) Dipodomys ingens (N=14) and (b) D. deserti (N=14) in response to a live gopher snake and an artificial snake decoy during three 10-min periods: pretest, test and post-test.
the decoy for any of the other time variables for either D. ingens or D. deserti (Table 2).
Footdrumming Patterns Dipodomys ingens did not alter footdrumming patterns when they encountered a live snake, but drummed with the same pattern as during spontaneous drumming, presumably as a signal to conspecifics (Randall 1997). The number of footdrums in the first footroll averaged 51.69.1 and 53.19.1 and the number of footrolls in a bout averaged 1.40.23 and 1.210.05 in the presence of the snake and during spontaneous drumming, respectively. Footdrumming rate was also similar, with 16.10.43 footdrums/s to the snake and 15.960.60 footdrums/s during spontaneous drumming (N=7, NS, all analyses).
Visits Although we tried to minimize interference from neighbours during our experiments, neighbours visited 11 of 14 D. deserti during the test period. In contrast, we observed no visits by D. ingens during the test and only one in a pretest. Visits were just as likely to occur with the snake as with the decoy during any time in the 30-min test (ANOVA: main effect for stimulus: F1,26 =2.2, P=0.60;
The kangaroo rats decreased time foraging and increased vigilance when the snake and decoy were present in what appeared to be a trade-off between feeding and predator defence (Elgar 1989; Lima & Dill 1990). Time foraging showed the largest decrease when a live snake was present compared with the decoy, and both species oriented towards, approached and footdrummed more in the presence of the live snake than the decoy. This snake-directed behaviour showed that the kangaroo rats could discriminate the difference in danger between the two stimuli. The decrease in foraging from the pretest to the test in the presence of the decoy shows a response to the decoy, but visual cues, lack of movement and snake scent probably informed the kangaroo rats that the decoy was less of a threat than a live snake. How much the kangaroo rats actually perceived the decoy as a snake is unknown, and the decoy may have been perceived more as an inanimate, novel object than a snake, especially when approached to within 1 m. The presentation of only one example of the decoy and live snake also may have limited the range of antipredator responses normally expressed by the kangaroo rats. The decline in foraging between the pretest and test could result from two other factors. Because we scattered seeds around the burrow area at the beginning of our observation, foraging behaviour could decline as seeds were harvested. If decreased time foraging had been because seed density decreased, we would expect the decline in foraging from the pretest to the test to be the same for the decoy and snake. We also do not attribute the change in foraging to a decrease in activity, because activity remained high in both the pretest and test. Dipodomys deserti was active throughout the 30-min test, and activity did not decrease for D. ingens until the post-test when they remained in the burrow and footdrummed. The difference in the defensive behaviour of D. ingens and D. deserti during tests suggests different antipredator strategies in the two species. Dipodomys ingens spent more time within striking distance of the snake than the decoy and more time in the burrow footdrumming in the post-test. Dipodomys deserti spent more time within 1 m of the snake than the decoy, but they did not spend significantly more time within striking distance of the snake
RANDALL & BOLTAS KING: ANTISNAKE BEHAVIOUR
than the decoy, and only footdrummed during the test. The greatest difference in defensive behaviour of the two species, however, was in sand kicking, which was a major defence for D. deserti, but was never observed in D. ingens. Dipodomys spectabilis also has never been observed kicking substrate in predator defence (Randall & Stevens 1987; Randall et al. 1995; Randall & Matocq 1997). The behaviour seems specific to D. deserti as an adaptation for living in a habitat with loose sand. The sand flying towards the snake seemed aversive to the tethered snake, which hissed when hit with sand, and it was sufficient to cause an untethered gopher snake to move away when we allowed a kangaroo rat to interact with it (unpublished observation). Dipodomys deserti seemed to assess the snake from about 1 m to determine the direction of sand kicking, and therefore could remain out of striking distance when turning around to kick sand. Another factor could account for the different responses of the two kangaroo rat species to snakes. We were unable to control for experience, and the kangaroo rats possibly responded differently to snakes based on prior interactions with them. We sometimes observed rattlesnakes (Crotalus sp.) and gopher snakes on both study sites, but we did not observe any natural interactions between kangaroo rats and snakes in 1997 or 1998. We assume that experience with snakes was limited, but similar, for D. ingens and D. deserti during the years we did our tests (Boltas 1999). Why do kangaroo rats approach snakes to within striking distance? The two species in this study and D. spectabilis (Randall et al. 1995; Randall & Matocq 1997) spend time within striking distance of live snakes. This is a very dangerous behaviour even when interacting with nonvenomous snakes (D. spectabilis was caught 5% of the time by gopher snakes, P. melanoleucus, in laboratory tests with free-ranging snakes; Randall et al. 1995) and possibly more dangerous with venomous species. Although California ground squirrels, Spermophilus beecheyii, approach rattlesnakes to within striking distance and interact with them for long periods, they are in less danger than kangaroo rats because they can detoxify the venom, and they often interact with snakes in groups (Owings & Coss 1977; Poran & Coss 1990). Cooperative approaches to predators are less risky than individual ones (Dugatkin & Godin 1992; Fishman 1999). Kangaroo rats seem unable to detoxify venom, as they die soon after being struck by rattlesnakes (P. Wasser, personal communication), and they interact with snakes alone. Despite the risk, however, kangaroo rats may be even more vulnerable without the close inspection and information they gain about the type of predator or its readiness to attack (Bouskila & Blumstein 1992; Dugatkin & Godin 1992). Hiding, stalking and ambush predators seem to be the principle subjects of close inspection by prey (FitzGibbon 1994; Fishman 1999), and snakes fall in this category. A close inspection provides information about the size and temperature-dependent ability of the snake to move, which is important information for the rodents to assess personal danger and the chances of the snake entering the burrow (Rowe & Owings 1990; Swaisgood et al. 1999a, b). Approaching the snake may also provide
information about species identity of the snake (Loughry 1989). The close approach to snakes may also be a defensive strategy. Kangaroo rats often approach a snake’s head, which is aversive to the snake (Herzog 1986; Burger 1998). The kangaroo rats may approach to signal they have detected the snake. They continue to approach and footdrum to signal that they are alert and not easy prey so the snake will leave (Hasson 1991; Caro 1995; Randall & Matocq 1997). Swaisgood et al. (1999a) proposed that game theory models for intraspecific competition could be applied to interactions between predator and prey (Maynard Smith 1974). Despite the fact that predators seek to kill and consume their opponents, animals engaging in both interspecific interactions with snakes and intraspecific interactions with conspecifics must first assess the situation to determine whether to escalate the conflict to more dangerous and energetically costly levels. Kangaroo rats assess their interactions with conspecifics and distinguish differences in threat between familiar neighbours and unfamiliar strangers (Randall 1984, 1989b; Perri & Randall 1999). Dipodomys spectabilis and D. ingens usually communicate identity or presence by footdrumming before escalating an interaction with a conspecific into a chase or fight (Randall 1984, 1991, 1997). Dipodomys deserti, however, is more aggressive than either D. ingens or D. spectabilis and readily approaches and chases conspecifics from its burrow (Randall 1997, unpublished data). Dipodomys deserti was extremely active during our tests and chased visitors with some rollover fights and readily interacted with the snake at their home burrow, much more than with a tethered snake on the burrows of neighbours. It seems that this species actively tries to remove any unwanted visitor to its burrow by chasing and fighting with conspecifics and approaching and kicking sand at snakes. Interactions with snakes are not only the prerogative of solitary rodents. Because predation by snakes is widespread, it is expected that approach and confrontation of snakes by rodents would be widespread and not restricted to solitary species. Social rodents do indeed interact with snakes and use some of the same snake-directed behaviour as the solitary kangaroo rats. Prairie dogs, Cynomys ludovicianus, interact aggressively with snakes and kick substrate and footdrum (Owings & Owings 1979; Loughry 1987). Spermophilus beecheyi forcefully throw sand with their forepaws and spend time within striking distance (Owings & Coss 1977; Towers & Coss 1990). They also pounce on snakes and bite them, which are behaviours never observed in kangaroo rats. Great gerbils, Rhombomys opimus, the only social desert rodent known to interact with snakes, approach snakes, footdrum and give alarm calls, but they do not throw substrate (Randall et al. 2000). Results from this study and studies of D. spectabilis (Randall & Stevens 1987; Randall et al. 1995; Randall & Matocq 1997) suggest that kangaroo rats respond to predation risk and interact with snakes (1) to assess risk, (2) as a pursuit deterrent to communicate to the snake its chances of ambush are no longer available to cause the
585
586
ANIMAL BEHAVIOUR, 61, 3
snake to leave, and (3) to harass the snake to cause it to leave. One or all of these functions of snake-directed behaviour may apply to D. ingens and D. deserti and cannot be rejected until more information about the behaviour of the snake is obtained to determine whether the snake leaves the burrow area in response to the behaviour of the kangaroo rats. If snakes respond to D. ingens and D. deserti the same as D. spectabilis, the snake-directed behaviour should cause snakes to decrease stalking and find prey elsewhere (Randall & Matocq 1997). Acknowledgments We thank our field assistants, Shane King and Mary Atchison, the California Fish and Game for housing at the Painted Rock Ranch, and Rob Fulton and the staff at the Desert Studies Center for their help. Chris Moffatt, Jim Loughry and anonymous referees provided helpful comments on the manuscript. D. Boltas King is grateful for her support by a GAANN fellowship. The National Science Foundation supported this research, IBN-9723033 to J. A. Randall. Research on D. ingens was conducted under permit PRT 799486 from the U.S. Fish and Wildlife Service and a MOU, 31 May 1995–31 October 1998, from the California Department of Fish and Game. All procedures were approved on 18 April 1998 by the Institutional Animal Care and Use Committee of San Francisco State University under protocols 97-321 and 98-412. References Boltas, D. K. 1999. Antisnake behavior in two species of kangaroo rat (Dipodomys ingens and D. deserti): a comparative study. M.A. thesis, San Francisco State University. Bouskila, A. & Blumstein, D. T. 1992. Rules of thumb for predation hazard assessment: predictions from a dynamic model. American Naturalist, 139, 161–176. Braun, S. E. 1985. Home range and activity patterns of the giant kangaroo rat, Dipodomys ingens. Journal of Mammalogy, 66, 1–12. Brown, J. S., Kotler, B. P. & Valone, T. J. 1994. Foraging under predation: a comparison of energetic and predation costs in rodent communities of the Negev and Sonoran deserts. Australian Journal of Zoology, 42, 435–448. Brown, J. S., Laundre´, J. W. & Gurung, M. 1999. The ecology of fear: optimal foraging, game theory and trophic interactions. Journal of Mammalogy, 80, 385–399. Burger, J. 1998. Antipredator behaviour of hatchling snakes: effects of incubation temperature and simulated predators. Animal Behaviour, 56, 547–553. Caro, T. M. 1986a. The function of stotting: a review of the hypotheses. Animal Behaviour, 34, 649–662. Caro, T. M. 1986b. The function of stotting in Thomson’s gazelles: some tests of the predictions. Animal Behaviour, 34, 663–664. Caro, T. M. 1995. Pursuit-deterrence revisited. Trends in Ecology and Evolution, 10, 500–503. Caro, T. M., Lombardo, L., Goldizen, A. W. & Kelly, M. 1995. Tail-flagging and other antipredator signals in white-tailed deer: new data and synthesis. Behavioral Ecology, 6, 442–450. Daly, M., Wilson, M., Behrends, P. R. & Jacobs, L. F. 1990. Characteristics of kangaroo rats, Dipodomys merriami, associated with differential predation risk. Animal Behaviour, 40, 380–389.
Daly, M., Jacobs, L. F., Wilson, M. I. & Behrends, P. R. 1992. Scatter hoarding by kangaroo rats (Dipodomys merriami) and pilfering from their caches. Behavioral Ecology, 3, 102–111. Dugatkin, L. A. & Godin, J. G. J. 1992. Prey approaching predators: a cost–benefit perspective. Annals Zoologica Fennici, 29, 233–252. Elgar, M. A. 1989. Predator vigilance and group size in mammals and birds: a critical review of the empirical evidence. Biological Reviews, 64, 13–33. Fishman, M. A. 1999. Predator inspection: closer approach as a way to improve assessment of potential threats. Journal of Theoretical Biology, 196, 225–235. FitzGibbon, C. D. 1989. A cost to individuals with reduced vigilance in groups of Thompson’s gazelles hunted by cheetahs. Animal Behaviour, 37, 508–510. FitzGibbon, C. D. 1994. The cost and benefits of predator inspection behaviour in Thomson’s gazelles. Behavioral Ecology and Sociobiology, 34, 139–366. FitzGibbon, C. D. & Fanshawe, J. H. 1988. Stotting in Thomson’s gazelles: an honest signal of condition. Behavioral Ecology and Sociobiology, 23, 68–74. Hasson, O. 1991. Pursuit-deterrent signals: communication between prey and predator. Trends Ecology and Evolution, 6, 325–329. Herzog, H. A. 1986. Development of antipredator responses in snakes: I. Defensive and open-field behaviors in newborns and adults of three species of garter snakes (Thamnophis melanogaster, T. sirtalis, T. butleri). Journal of Comparative Psychology, 100, 372–379. Illius, A. W. & FitzGibbon, C. 1994. Costs of vigilance in foraging. Animal Behaviour, 47, 481–484. Kotler, B. P. 1984. Risk of predation and the structure of desert communities. Ecology, 65, 689–701. Kotler, B. P., Brown, J. S. & Hasson, O. 1991. The specter of predation: factors affecting gerbil foraging behavior and rates of owl predation. Ecology, 72, 2249–2260. Kotler, B. P., Brown, J. S. & Mitchell, W. A. 1994. The role of predation in shaping the behaviour, morphology and community organisation of desert rodents. Australian Journal of Zoology, 42, 449–466. Lima, S. L. 1998. Stress and decision making under the risk of predation: recent developments from behavioral, reproductive, and ecological perspectives. Advances in the Study of Behavior, 27, 215–290. Lima, S. L. & Dill, L. M. 1990. Behavioral decisions made under the risk of predation: a review and prospectus. Canadian Journal of Zoology, 68, 619–640. Loughry, W. J. 1987. The dynamics of snake harassment by blacktailed prairie dogs. Behaviour, 103, 27–48. Loughry, W. J. 1989. Discrimination of snakes by two populations of black-tailed prairie dogs. Journal of Mammalogy, 70, 627–630. McNamara, J. M. & Houston, A. I. 1992. Evolutionary stable levels of vigilance as a function of group size. Animal Behaviour, 43, 641–658. Maynard Smith, J. 1974. The theory of games and the evolution of animal conflicts. Journal of Theoretical Biology, 47, 209–221. Owings, D. H. & Coss, R. G. 1977. Snake mobbing by California ground squirrels: adaptive variation and ontogeny. Behaviour, 62, 50–69. Owings, D. H. & Owings, S. C. 1979. Snake-directed behavior by black-tailed prairie dogs (Cynomys ludovicianus). Zeitschrift fu¨r Tierpsychologie, 49, 35–54. Perri, L. M. & Randall, J. A. 1999. Behavioral mechanisms of coexistence in sympatric species of desert rodents, Dipodomys ordii and D. merriami. Journal of Mammalogy, 80, 1297–1310. Poran, N. S. & Coss, R. G. 1990. Development of antisnake defenses in California ground squirrels (Spermophilus beecheyi): behavioral and immunological relationships. Behaviour, 112, 222–245.
RANDALL & BOLTAS KING: ANTISNAKE BEHAVIOUR
Randall, J. A. 1984. Territorial defense and advertisement by footdrumming in bannertail kangaroo rats (Dipodomys spectabilis) at high and low population densities. Behavioral Ecology and Sociobiology, 16, 11–20. Randall, J. A. 1989a. Individual footdrumming signatures in bannertailed kangaroo rats, Dipodomys spectabilis. Animal Behaviour, 38, 620–630. Randall, J. A. 1989b. Territorial-defense with neighbors and strangers in banner-tailed kangaroo rats. Journal of Mammalogy, 70, 308–315. Randall, J. A. 1991. Mating strategies of a nocturnal desert rodent. Behavioral Ecology and Sociobiology, 28, 215–220. Randall, J. A. 1993. Behavioral adaptations of desert rodents (Heteromyidae). Animal Behaviour, 45, 263–287. Randall, J. A. 1994. Convergences and divergences in communication and social organization of desert rodents. Australian Journal of Zoology, 42, 404–433. Randall J. A. 1995. Modification of footdrumming signatures: changing territories and gaining new neighbours. Animal Behaviour, 49, 1227–1237. Randall, J. A. 1997. Species-specific footdrumming in kangaroo rats: Dipodomys ingens, D. deserti, D. spectabilis. Animal Behaviour, 54, 1167–1175. Randall, J. A. & Stevens, C. M. 1987. Footdrumming and other anti-predator responses in the bannertail kangaroo rat (Dipodomys spectabilis). Behavioral Ecology and Sociobiology, 20, 187–194. Randall, J. A. & Matocq, M. D. 1997. Why do kangaroo rats (Dipodomys spectabilis) footdrum at snakes? Behavioral Ecology, 8, 404–413. Randall, J. A., Hatch, S. M. & Hekkala, E. R. 1995. Inter-specific variation in anti-predator behavior in sympatric species of kangaroo rat. Behavioral Ecology and Sociobiology, 36, 243–250. Randall, J. A., Rogovin, K. A. & Shier, D. M. 2000. Anti-predator behavior of a social desert rodent: footdrumming and alarm calling in the great gerbil, Rhombomys opimus. Behavioral Ecology and Sociobiology, 48, 110–118. Roberts, G. 1996. Why individual vigilance declines as group size increases. Animal Behaviour, 51, 1077–1086.
Rowe, M. P. & Owings, D. H. 1990. Probing, assessment, and management during interactions between ground squirrels and rattlesnakes. Part 1: risks related to rattlesnake size and body temperature. Ethology, 86, 237–249. Scheel, D. 1993. Watching for lions in the grass: the usefulness of scanning and its effects during hunts. Animal Behaviour, 46, 695–705. Snedecor, G. W. & Cochran, W. G. 1967. Statistical Methods. Ames: Iowa State University Press. Stern, B. 1980. Habitat partitioning in two species of kangaroo rats (Rodentia; Heteromyidae): Dipodomys deserti and Dipodomys merriami. Ph.D. thesis. University of California, Berkeley. Swaisgood, R. R., Owings, D. H. & Rowe, M. P. 1999a. Conflict and assessment in a predator–prey system: ground squirrels versus rattlesnakes. Animal Behaviour, 57, 1033–1044. Swaisgood, R. R., Rowe, M. P. & Owings, D. H. 1999b. Assessment of rattlesnake dangerousness by California ground squirrels: exploitation of cues from rattling sounds. Animal Behaviour, 57, 1301–1310. Towers, S. R. & Coss, R. G. 1990. Confronting snakes in the burrow: snake-species discrimination and antisnake tactics of two California ground squirrel populations. Ethology, 84, 177–192. Wilkinson, L. 1997. SYSTAT. The System for Statistics. Evanston, Illinois: SPSS. Williams, D. F. 1992. Geographic distribution and population status of the giant kangaroo rat, Dipodomys ingens (Rodentia, Heteromyidae). In: Proceedings of the Conference on Endangered and Sensitive Species of the San Joaquin Valley, California (Ed. by D. F. Williams, T. A. Rado & S. Byrne), pp. 301–327. Sacramento: California Energy Commission. Woodland, D. J., Jaafar, Z. & Knight, M. L. 1980. The ‘pursuit deterrent’ function of alarm signals. American Naturalist, 115, 748–753. Zar, J. H. 1984. Biostatistical Analysis. Englewood Cliffs, New Jersey: Simon & Schuster.
587
Assessment and defence of solitary kangaroo rats under risk of predation by snakes JAN A. RANDALL & DENISE K. BOLTAS KING
Department of Biology, San Francisco State University (Received 23 February 2000; initial acceptance 5 June 2000; final acceptance 16 September 2000; MS. number: A8722R)
Prey that spend a large proportion of their time foraging must trade-off between predator defence and feeding. Because predation risk can be a cost in desert granivores, we predicted that kangaroo rats (Dipodomys) should weigh the risk of predation when deciding which behavioural option to pursue. We manipulated apparent predation risk of the giant kangaroo rat, D. ingens, and the desert kangaroo rat, D. deserti, and predicted they should be more vigilant and show greater defensive behaviour to the more dangerous stimulus of a real snake than to an artificial snake decoy. Results showed that kangaroo rats decreased foraging with increased vigilance when the snake and decoy were present. The largest decrease in foraging was in response to the live snake, and both species oriented towards, approached and footdrummed more in the presence of the live snake than the decoy. The species also differed in their antipredator behaviour: D. ingens footdrummed more and spent more time within striking distance of the snake than the decoy. Dipodomys deserti spent more time within 1 m of the snake than the decoy and kicked sand at the snake. Sand kicking was never observed in D. ingens. We conclude that kangaroo rats are able to discriminate predation risk. They decrease foraging and increase vigilance in the presence of live snakes to assess risk and may approach and footdrum as a pursuit deterrent to communicate to the snake its chances of ambush are no longer available to cause the snake to leave. Sand kicking may function to harass the snake to cause it to leave.
own defence (Elgar 1989; Scheel 1993). Dugkatin & Godin (1992) proposed four benefits for why prey approach and interact with predators: (1) to acquire information about the nature of the potential threat, (2) inform others of threat, (3) deter predator attack and (4) advertise one’s quality to potential mates. Of these, solitary animals most likely approach predators to gain information and to deter attack. They also may communicate directly to the predator to cause an attack or pursuit to be abandoned (Caro 1986a, b; Hasson 1991; Caro et al. 1995). Such behaviour deters pursuit by informing the predator that the chances of ambush are thwarted (Woodland et al. 1980), continued pursuit is costly (Caro 1995; Randall & Matocq 1997), or the prey is healthy and cannot be caught (FitzGibbon & Fanshawe 1988; Caro 1995). Although smaller species of kangaroo rats (D. merriami), avoid predators, larger species act in their own defence in response to snakes (Randall et al. 1995). The bannertailed kangaroo rat, D. spectabilis, approaches snakes to within striking distance, jumps back and footdrums. Randall & Matocq (1997) interpreted this behaviour to function in predator deterrence (Caro 1995): the close approach informed the snake it was detected, and
Herbivores must trade-off between watching for predators and feeding (Elgar 1989; Lima & Dill 1990), because vigilant prey have a lower chance of being captured by a predator than those that are not vigilant (FitzGibbon 1989; Illius & FitzGibbon 1994). It is well documented that predation is a cost of foraging in desert granivores such as kangaroo rats (Dipodomys) and gerbils (Gerbillus) (Daly et al. 1990, 1992; Kotler et al. 1991, 1994; Brown et al. 1994; Lima 1998). Three factors make desert rodents especially susceptible to predation risk: they are solitary and forage alone, they often forage in open habitats with little cover, and they have patchy food sources that require movement between patches. Solitary rodents probably have to spend more time watching and listening for predators than those in groups (McNamara & Houston 1992; Roberts 1996). Rodents are often depicted as sedentary victims of predators (Brown et al. 1999), but prey are not always passive players in the predator–prey game (Fishman 1999; Swaisgood et al. 1999a, b). Although it seems counterintuitive, prey sometimes interact with a predator in their Correspondence: J. A. Randall, Department of Biology, San Francisco State University, San Francisco, CA 94132, U.S.A. (email: [email protected]). 0003–3472/01/030579+09 $35.00/0
2001 The Association for the Study of Animal Behaviour
579
2001 The Association for the Study of Animal Behaviour
580
ANIMAL BEHAVIOUR, 61, 3
footdrumming signalled the snake that the kangaroo rat was alert and would not be easy prey. The possibility that kangaroo rats approach snakes to assess and gain information about potential threat was not addressed in this study (Randall & Matocq 1997). Kangaroo rats should weigh the risk of predation when deciding what behavioural option to pursue when they encounter snakes. We predicted that if kangaroo rats can assess the danger posed by snakes, they should display a different set of responses to the more dangerous stimulus of a real snake than to an artificial snake decoy. They should be more vigilant, inspect more and show greater defensive behaviour in the presence of a live snake. We also predicted that species differ in their response to predators based on occupancy of different habitats and differences in social tolerance. The giant kangaroo rat, D. ingens, occupies arid grasslands with a rich food source that does not require them to make long-distance foraging trips (Braun 1985). The desert kangaroo rat, D. deserti, is a sand dune specialist that occupies a drier habitat with a less productive food source than the grassland habitat of D. ingens. Dipodomys deserti makes longer foraging trips, visits the burrow of neighbours to pilfer seeds at higher rates and is generally more aggressive than D. ingens. They aggressively chase neighbours that visit their burrows and engage in rollover fights (Stern 1980; Randall 1997, unpublished data). We predicted that this social intolerance and aggressive behaviour of D. deserti might cause them to be more aggressive towards snakes than D. ingens. METHODS
Study Sites We studied D. ingens on a 1.5-ha study area in the Carrizo Plain Natural Area located in California Valley, San Luis Obispo County, California, U.S.A. from 15 June to 10 August 1997. Vegetation consisted of non-native annual grasses (Vulpia, Bromus), annual native herbs (Lepidium, Erodium) and sparse native shrubs (Atriplex). We studied D. deserti near the Desert Studies Center on a 1.6-ha site in the Mojave National Preserve, 18 km south of Baker, San Bernadino County, California from 3 June to 8 July 1998. The habitat consisted of areas of loose sand dunes interspersed with sparse grasses (Bromus, Bouteloua, Achnatherum) and creosote (Larrea tridentata) hummocks. Because 1998 was an El Nin ˜ o year with record rainfall, vegetation was relatively abundant in the dune habitat.
Animals Dipodomys ingens is the largest species in the genus. Adult males (N=16) averaged 136.92.8 g and females 131.92.9 g (N=13) on our study site. Because of habitat loss, distribution is restricted to less than 2% of the former range and D. ingens is listed as an endangered species by both the U.S. Fish and Wildlife Service and the California Department of Fish and Game. Dipodomys deserti is smaller than D. ingens; males (N=10) averaged
114.83.86 g and females (N=13) averaged 101.54.5 g on our site. Dipodomys deserti is the only species of kangaroo rat that lives exclusively in sand dunes. Both species are solitary, and individuals of both species occupied individual burrows during our study. These desert rodents normally reproduce from late autumn to spring (Randall 1993, 1994), and thus, when this study was conducted breeding was finished and there was no evidence of young in the burrow. We trapped animals with wild bird seed in Sherman live-traps (30.57.69 cm) with a 0.25-cm space in the top of the door to prevent tail severance. We marked each individual with numbered ear tags (Monel 1005-1) covered with colour-coded reflective tape (Scotchlite) for individual identification at night. We also sexed, determined reproductive condition and weighed each animal to estimate age. We determined residency of a kangaroo rat at a burrow by trapping records, scan sampling and observation (Randall 1984). A kangaroo rat had to be trapped and observed at a particular burrow more than twice before we assigned residency.
Behavioural Observations and Footdrumming Recordings We used established techniques for observing kangaroo rats at night (Randall 1995, 1997). We sat in the dark approximately 10 m from a burrow entrance and observed behaviour with either a Generation II night vision scope mounted on a tripod or a Generation III night vision goggle. We noted behaviour by talking softly into a hand-held tape recorder (Table 1). Footdrumming was recorded using geophones connected to either a Uher 4400 reel-to-reel tape recorder or a Marantz PMD cassettetape recorder (Randall 1989a). We scattered a small amount of bird seed around the burrow at the beginning of an observation or test to keep the kangaroo rats active and to facilitate observations (Randall & Stevens 1987). We also scattered a small amount of seed at neighbouring burrows of D. deserti to minimize the number of visits by neighbours during our tests.
Stimuli We chose gopher snakes, Pituophis melanoleucus, as the stimulus animal because they are sympatric with both species of kangaroo rat and feed on them in the wild (Kotler 1984; Williams 1992), and they are present on our study sites. We initially used two snakes with snout–vent lengths (SVL) of 1.23 and 1.10 m and weights of 1106 and 907 g, respectively. Because we detected a tumor in the smaller snake after one test, we used only the larger snake for remaining tests. This caused a problem of pseudoreplication. However, because of the strong and consistent response of kangaroo rats to snakes of all sizes (Randall & Stevens 1987; Randall et al. 1995; Randall & Matocq 1997), we considered the use of this large, active, snake sufficient to elicit a normal range of responses, and the consistent use of stimuli provided consistency among tests.
RANDALL & BOLTAS KING: ANTISNAKE BEHAVIOUR
Table 1. Behaviours of Dipodomys ingens and D. deserti recorded in the presence of live and decoy snakes Behaviour
Approach Footdrum Forage Jump back Kick sand Orient Out of burrow Sniff Within striking Within 1 m
Description
Movement towards snake or decoy that decreases distance between focal animal and stimulus. The kangaroo rat strikes hind feet against the ground. Each strike is a footdrum; a series of drums is a footroll and one or more footrolls is a bout. Slow quadrupedal movement with head towards ground and digging with forepaws. A backward hop in the air that causes the rat to increase distance between itself and the stimulus. Kangaroo rat turns away from the stimulus and causes sand to fly towards it by sharp kicks of the hind feet. Duration of standing in a quadrupedal or bipedal posture holding body at 45–90° with its head and anterior part of its body towards the stimulus. Time in burrow subtracted from total time in test period. An approach to within 3 cm of the stimulus with nose lowered towards it. Kangaroo rat moves within 0.5 m (half snake’s body length) of head of snake or decoy. Distance between striking distance (0.5 m) and 1 m.
The snake was tethered using an established technique (Randall & Stevens 1987; Randall & Matocq 1997). To prevent damage to the snake’s skin, we first wrapped a piece of paper athletic tape around the snake, covering about one-third the length of the snake’s body below the head (33 cm), and then wrapped a strip of adhesive tape (3–4 cm wide) on top. We folded the adhesive tape back on itself and punched two holes in the flap, to which we tied two monofilament lines that were fastened to two large nails or tent stakes driven flush into the ground about 0.6 m apart. The snake could move and strike, but it did not strike at the kangaroo rats during our tests. The rubber snake decoy measured 1 m stretched out and was a replica of a real snake with a black body and yellow stripes. Its coiled silhouette in the dark appeared real.
specific to the stimuli in the test period (Table 1). Data recorded for comparison during the pretest, test and post-test included: (1) time spent foraging, (2) time spent footdrumming and (3) time out of the burrow. Behaviour specific to interactions with the snake and decoy included: (1) time oriented, (2) predator inspection as measured by number of approaches, time spent within 1 m, time within striking distance and time sniffing the snake, (3) frequency of jump backs and (4) number of sand kicks. We also recorded visits by neighbours and whether there were any interactions with focal animals or the stimuli. We recorded behaviour and footdrumming into continuously running recorders so that we could transcribe and time behaviour from the tapes.
Data Analysis Procedures We tested seven males and seven females of each species for their responses to the live gopher snake and rubber snake decoy in a counterbalanced order. We tethered the stimuli 1.5–2.0 m from the focal animal’s most frequently used burrow entrance. We determined burrow use during a 1-h focal observation period, which also served to habituate the kangaroo rat to our presence, on a night 1–12 days prior to the test. Each test lasted 30 min. We began a 10-min pretest observation when the focal animal came aboveground and foraged normally. After the pretest observation, we walked softly to the burrow area and tethered the stimulus. Although the kangaroo rats entered the burrow in response to our approach, they usually exited the burrow within 5–10 min to forage on the scattered seeds. We began the 10-min test only after the kangaroo rat exited the burrow and oriented to the stimulus for more than 5 s. At the end of the test, we quietly removed the stimulus and began a 10-min post-test observation. We tabulated one set of behaviours during all three time periods and another set that involved responses
We analysed data collected in the three test periods in a 223 design (speciesstimulustest period) and in response to the stimuli in 22 design (species stimulus) using repeated measures analyses of variance (ANOVAs) in SYSTAT 7.0 (Wilkinson 1997). (We omitted the sex variable because we found no differences in behaviour of males and females.) When data contained zeros, we performed a normalizing log+1 transformation (Zar 1984), and we used nonparametic statistics (Wilcoxon matched-pairs signed-ranks and Friedman tests) when data could not be normalized and assumptions for parametric tests were not met. We used paired t tests and Wilcoxon matched-pairs signed-ranks tests with Bonferroni corrections as post hoc tests. We performed statistical tests in SYSTAT 7.0, with the exception of Wilcoxon tests with an N less than 14, which were calculated by hand following Snedecor & Cochran (1967), and we report the multivariate F statistic (Wilks’ lambda) when possible because it does not require compound symmetry (Wilkinson 1997). All tests are two-tailed and data are reported as the meanSE of untransformed data.
581
ANIMAL BEHAVIOUR, 61, 3
Table 2. Comparisons of time Dipodomys ingens and D. deserti spent foraging (paired t tests, df=27) and footdrumming (Wilcoxon matched-pairs signed-ranks tests N=14) during pretest, test and post-test in response to a snake and decoy. Multiple comparisons are Bonferroni corrected
Behaviour
Pretest versus test
Pretest versus post-test
8 (a)
Pretest Test Post-test
7 6
Test versus post-test
5 4 3
Foraging Snake Decoy Footdrumming D. ingens Snake Decoy D. deserti Snake Decoy
6.2** 3.1*
4.62** 2.87*
1.75 (NS) 0.37 (NS)
0** 23 (NS)
9** 38 (NS)
45 (NS) 29 (NS)
4** 30 (NS)
33 (NS) 35 (NS)
0** 23 (NS)
Time foraging (min)
582
2 1 0 10 9
(b)
8 7 6
*P<0.05; **P<0.001.
5
We also analysed footdrumming of D. ingens to compare differences in patterns of drumming in the snake context with those used to communicate to conspecifics (Randall 1994, 1997). We were unable to analyse footdrumming of D. deserti because of infrequent drumming in any context except with the live snake. We compared drumming patterns of seven D. ingens that drummed five or more bouts of spontaneous drumming during pretests and habituations and in response to the snake stimuli. We analysed the recordings using established techniques (Randall 1989b, 1995), and compared (1) number of drums in the first footroll, (2) number of footrolls in a footdrumming bout and (3) rate of drumming (drums/s). RESULTS
Foraging The kangaroo rats decreased foraging in response to both the snake and decoy with less time foraging in the test than in the pretest (effect for time: Wilks’ lambda: F2,52 =33.15, P=0.0001; Table 2, Fig. 1). The largest decrease in foraging was in response to the live snake (timestimulus interaction: Wilks’ lambda: F2,52 =3.57, P=0.035). Time foraging decreased an average of 4.210.44 min between the pretest and test when the snake was tethered compared with an average decrease of 2.190.42 min in response to the decoy (Wilcoxon matched-pairs signed-ranks test: Z= 3.78, N=28, P=0.0001). Because time foraging remained lower in the post-test after the snake or decoy was removed, there was no difference in time foraging between the test and post-test, but there was a significant difference between the post-test and pretest (Table 2). Although D. deserti spent more total time foraging than D. ingens (main effect species: F1,52 =19.7, P=0.0001), both species spent significantly less time foraging during the test than pretest (Wilks’ lambda: speciestime interaction: F2,51 =7.48, P=0.001; paired t test: D. ingens: t27 =5.58, P=0.0001; D. deserti: t27 =4.9, P=0.0001; Fig. 1).
4 3 2 1 0
Snake
Decoy Stimulus
Figure 1. Time foraging (mean±SE) by (a) Dipodomys ingens (N=14) and (b) D. deserti (N=14) in response to a live gopher snake and an artificial snake decoy during three 10-min periods: pretest, test and post-test.
Dipodomys ingens also spent significantly less time foraging during the post-test than during the pretest (t27 =4.89, P=0.0001). Dipodomys deserti spent about the same amount of time foraging in the post-test as in the test (t27 =1.17, P=1.0) and during the pretest and post-test (t27 =2.59, P=0.09). Time foraging also did not differ for D. ingens during the post-test and test (t27 =1.1, P=1.0).
Avoidance The kangaroo rats did not enter the burrow in response to either the snake or decoy so that time out of the burrow did not differ significantly between the pretest (8.890.24 min), test (8.370.36 min) and post-test (6.910.50 min) (Friedman test: 22 =3.06, P=0.22), and there was no significant effect by the type of stimulus (21 =1.63, P=0.48). There was a significant difference by species (22 =15.88, P=0.001). Dipodomys ingens spent significantly less time out of the burrow in the post-test (4.20.68 min) than in the pretest (8.210.44 min) (Wilcoxon matched-pairs Z= 3.14, N=28, P=0.02) and test (7.150.63 min) (Z= 3.85, N=28, P=0.001). Dipodomys deserti spent equal time out of the burrow in the pretest (9.70.13 min), test (9.80.21 min) and post-test (9.60.19 min) (all comparisons NS).
RANDALL & BOLTAS KING: ANTISNAKE BEHAVIOUR
Kangaroo rats sometimes jumped back after they approached the snake. The species together averaged 1.320.33 jump backs from the snake and 0.500.37 from the decoy during the 10-min test (Friedman test: 21 =5.14, P=0.03). Taken separately, D. deserti jumped back more from the snake (1.140.29) compared with the decoy (0.290.16) (Wilcoxon matched-pairs signedranks test: T=8, N=14, P<0.05), but the frequencies did not differ significantly for D. ingens (snake: 1.50.57; decoy: 0.710.64; T=22, N=14, NS).
300
(a)
Snake Decoy
*
200
* 100
Vigilance and Predator Inspection
Defence Sand kicking was an important element of predator defence for D. deserti, but not for D. ingens. Thirteen of 14 D. deserti kicked sand directly towards the snake compared with six that kicked sand towards the decoy to yield a significantly higher frequency for the snake (14.13.9) compared with the decoy (5.12.9) (T=14, N=14, P=0.05). We never observed D. ingens kick sand in any test or observation. Both D. ingens and D. deserti footdrummed more in the presence of the snake than the decoy (Friedman test:
0 80
(b)
* 60 Time (s)
There was no significant difference in the amount of time that D. ingens and D. deserti oriented to the stimuli (ANOVA: main effect for species: F1,26 =2.4, P=0.134; Fig. 2a). Time oriented to the stimuli during the test showed differences in vigilance between the snake and decoy (main effect for stimulus: F1,26 =30.4, P=0.0001), and the kangaroo rats spent significantly more time oriented to the snake than to the decoy (paired t test: t27 =5.6, P=0.0001). Other comparisons were not significant. Both D. ingens and D. deserti approached the snake and decoy with no species differences found in the total number of approaches (ANOVA: main effect for species: F1,26 =0.04, P=0.95) but with significantly more approaches to the snake (5.540.93) than to the decoy (3.180.73) (main effect for stimulus: F1,26 =6.69, P=0.016). The interaction was not significant. Once kangaroo rats approached the stimulus, however, the species differed in the amount of time within 1 m (Friedman test: 22 =16.2, P=0.001) and striking distance (22 =8.41, P=0.015) of the snake and decoy (Fig. 2b,c). Although D. deserti spent significantly more time within 1 m of the snake compared with the decoy (Wilcoxon signed-ranks test: T=8.5, N=14, P<0.01), they did not spend significantly more time within striking distance of the snake compared to the decoy (T=25, N=14, NS). Dipodomys ingens, however, spent significantly more time within striking distance of the snake than the decoy (T=0, N=14, P<0.01) but showed similar times within 1 m of the snake and decoy (T=22, N=14, NS) (Fig. 2b,c). Species taken together sniffed the decoy (0.640.19) more frequently than they sniffed the snake (0.360.26) (Friedman test: 22 =5.2, P=0.02). All other comparisons were not significant.
40
20
0 80
(c)
60
40
* 20
0
D. ingens
D. deserti Species
Figure 2. Mean±SE time Dipodomys ingens (N=14) and D. deserti (N=14) spent (a) oriented towards (b) within 1 m of and (c) within striking distance of a live gopher snake and an artificial snake decoy during a 10-min test. *P<0.05.
22 =13.4, P=0.001; 22 =10.8, P=0.005, respectively) (Fig. 3). Dipodomys ingens drummed significantly longer in the test and post-test than in the pretest, but because they continued footdrumming into the post-test, time drumming during the test and post-test did not differ significantly (Table 2). Like D. ingens, D. deserti drummed significantly longer in the snake test than in the pretest, but unlike D. ingens, D. deserti also drummed significantly longer in the test than the post-test. No significant differences were found for drumming in the presence of
583
ANIMAL BEHAVIOUR, 61, 3
60 Pretest Test Post-test
(a) 50 40 30
Time footdrumming (s)
584
20 10 0 60 (b) 50
DISCUSSION
40 30 20 10 0
main effect for time: F1,26 =0.8, P=0.45; timestimulus interaction: F1,26 =0.5, P=0.45). Four focal D. deserti interacted with a snake at the burrow of a neighbour during the snake test of the neighbour (one kangaroo rat interacted with the snake at a neighbour’s burrow twice, so we calculated an average of the two encounters). When we compared the total time interacting with the snake at their home burrow compared with a neighbour’s burrow, we found that the kangaroo rats spent significantly more time interacting with the snake at home (2.10.5 min) than at a neighbour’s burrow (0.250.08 min) (paired t test: t3 =8.0, P=0.004). Visiting rats oriented to the snake, approached it and three kicked sand, but none footdrummed when away from the home burrow.
Snake
Decoy Stimulus
Figure 3. Time footdrumming (mean±SE) by (a) Dipodomys ingens (N=14) and (b) D. deserti (N=14) in response to a live gopher snake and an artificial snake decoy during three 10-min periods: pretest, test and post-test.
the decoy for any of the other time variables for either D. ingens or D. deserti (Table 2).
Footdrumming Patterns Dipodomys ingens did not alter footdrumming patterns when they encountered a live snake, but drummed with the same pattern as during spontaneous drumming, presumably as a signal to conspecifics (Randall 1997). The number of footdrums in the first footroll averaged 51.69.1 and 53.19.1 and the number of footrolls in a bout averaged 1.40.23 and 1.210.05 in the presence of the snake and during spontaneous drumming, respectively. Footdrumming rate was also similar, with 16.10.43 footdrums/s to the snake and 15.960.60 footdrums/s during spontaneous drumming (N=7, NS, all analyses).
Visits Although we tried to minimize interference from neighbours during our experiments, neighbours visited 11 of 14 D. deserti during the test period. In contrast, we observed no visits by D. ingens during the test and only one in a pretest. Visits were just as likely to occur with the snake as with the decoy during any time in the 30-min test (ANOVA: main effect for stimulus: F1,26 =2.2, P=0.60;
The kangaroo rats decreased time foraging and increased vigilance when the snake and decoy were present in what appeared to be a trade-off between feeding and predator defence (Elgar 1989; Lima & Dill 1990). Time foraging showed the largest decrease when a live snake was present compared with the decoy, and both species oriented towards, approached and footdrummed more in the presence of the live snake than the decoy. This snake-directed behaviour showed that the kangaroo rats could discriminate the difference in danger between the two stimuli. The decrease in foraging from the pretest to the test in the presence of the decoy shows a response to the decoy, but visual cues, lack of movement and snake scent probably informed the kangaroo rats that the decoy was less of a threat than a live snake. How much the kangaroo rats actually perceived the decoy as a snake is unknown, and the decoy may have been perceived more as an inanimate, novel object than a snake, especially when approached to within 1 m. The presentation of only one example of the decoy and live snake also may have limited the range of antipredator responses normally expressed by the kangaroo rats. The decline in foraging between the pretest and test could result from two other factors. Because we scattered seeds around the burrow area at the beginning of our observation, foraging behaviour could decline as seeds were harvested. If decreased time foraging had been because seed density decreased, we would expect the decline in foraging from the pretest to the test to be the same for the decoy and snake. We also do not attribute the change in foraging to a decrease in activity, because activity remained high in both the pretest and test. Dipodomys deserti was active throughout the 30-min test, and activity did not decrease for D. ingens until the post-test when they remained in the burrow and footdrummed. The difference in the defensive behaviour of D. ingens and D. deserti during tests suggests different antipredator strategies in the two species. Dipodomys ingens spent more time within striking distance of the snake than the decoy and more time in the burrow footdrumming in the post-test. Dipodomys deserti spent more time within 1 m of the snake than the decoy, but they did not spend significantly more time within striking distance of the snake
RANDALL & BOLTAS KING: ANTISNAKE BEHAVIOUR
than the decoy, and only footdrummed during the test. The greatest difference in defensive behaviour of the two species, however, was in sand kicking, which was a major defence for D. deserti, but was never observed in D. ingens. Dipodomys spectabilis also has never been observed kicking substrate in predator defence (Randall & Stevens 1987; Randall et al. 1995; Randall & Matocq 1997). The behaviour seems specific to D. deserti as an adaptation for living in a habitat with loose sand. The sand flying towards the snake seemed aversive to the tethered snake, which hissed when hit with sand, and it was sufficient to cause an untethered gopher snake to move away when we allowed a kangaroo rat to interact with it (unpublished observation). Dipodomys deserti seemed to assess the snake from about 1 m to determine the direction of sand kicking, and therefore could remain out of striking distance when turning around to kick sand. Another factor could account for the different responses of the two kangaroo rat species to snakes. We were unable to control for experience, and the kangaroo rats possibly responded differently to snakes based on prior interactions with them. We sometimes observed rattlesnakes (Crotalus sp.) and gopher snakes on both study sites, but we did not observe any natural interactions between kangaroo rats and snakes in 1997 or 1998. We assume that experience with snakes was limited, but similar, for D. ingens and D. deserti during the years we did our tests (Boltas 1999). Why do kangaroo rats approach snakes to within striking distance? The two species in this study and D. spectabilis (Randall et al. 1995; Randall & Matocq 1997) spend time within striking distance of live snakes. This is a very dangerous behaviour even when interacting with nonvenomous snakes (D. spectabilis was caught 5% of the time by gopher snakes, P. melanoleucus, in laboratory tests with free-ranging snakes; Randall et al. 1995) and possibly more dangerous with venomous species. Although California ground squirrels, Spermophilus beecheyii, approach rattlesnakes to within striking distance and interact with them for long periods, they are in less danger than kangaroo rats because they can detoxify the venom, and they often interact with snakes in groups (Owings & Coss 1977; Poran & Coss 1990). Cooperative approaches to predators are less risky than individual ones (Dugatkin & Godin 1992; Fishman 1999). Kangaroo rats seem unable to detoxify venom, as they die soon after being struck by rattlesnakes (P. Wasser, personal communication), and they interact with snakes alone. Despite the risk, however, kangaroo rats may be even more vulnerable without the close inspection and information they gain about the type of predator or its readiness to attack (Bouskila & Blumstein 1992; Dugatkin & Godin 1992). Hiding, stalking and ambush predators seem to be the principle subjects of close inspection by prey (FitzGibbon 1994; Fishman 1999), and snakes fall in this category. A close inspection provides information about the size and temperature-dependent ability of the snake to move, which is important information for the rodents to assess personal danger and the chances of the snake entering the burrow (Rowe & Owings 1990; Swaisgood et al. 1999a, b). Approaching the snake may also provide
information about species identity of the snake (Loughry 1989). The close approach to snakes may also be a defensive strategy. Kangaroo rats often approach a snake’s head, which is aversive to the snake (Herzog 1986; Burger 1998). The kangaroo rats may approach to signal they have detected the snake. They continue to approach and footdrum to signal that they are alert and not easy prey so the snake will leave (Hasson 1991; Caro 1995; Randall & Matocq 1997). Swaisgood et al. (1999a) proposed that game theory models for intraspecific competition could be applied to interactions between predator and prey (Maynard Smith 1974). Despite the fact that predators seek to kill and consume their opponents, animals engaging in both interspecific interactions with snakes and intraspecific interactions with conspecifics must first assess the situation to determine whether to escalate the conflict to more dangerous and energetically costly levels. Kangaroo rats assess their interactions with conspecifics and distinguish differences in threat between familiar neighbours and unfamiliar strangers (Randall 1984, 1989b; Perri & Randall 1999). Dipodomys spectabilis and D. ingens usually communicate identity or presence by footdrumming before escalating an interaction with a conspecific into a chase or fight (Randall 1984, 1991, 1997). Dipodomys deserti, however, is more aggressive than either D. ingens or D. spectabilis and readily approaches and chases conspecifics from its burrow (Randall 1997, unpublished data). Dipodomys deserti was extremely active during our tests and chased visitors with some rollover fights and readily interacted with the snake at their home burrow, much more than with a tethered snake on the burrows of neighbours. It seems that this species actively tries to remove any unwanted visitor to its burrow by chasing and fighting with conspecifics and approaching and kicking sand at snakes. Interactions with snakes are not only the prerogative of solitary rodents. Because predation by snakes is widespread, it is expected that approach and confrontation of snakes by rodents would be widespread and not restricted to solitary species. Social rodents do indeed interact with snakes and use some of the same snake-directed behaviour as the solitary kangaroo rats. Prairie dogs, Cynomys ludovicianus, interact aggressively with snakes and kick substrate and footdrum (Owings & Owings 1979; Loughry 1987). Spermophilus beecheyi forcefully throw sand with their forepaws and spend time within striking distance (Owings & Coss 1977; Towers & Coss 1990). They also pounce on snakes and bite them, which are behaviours never observed in kangaroo rats. Great gerbils, Rhombomys opimus, the only social desert rodent known to interact with snakes, approach snakes, footdrum and give alarm calls, but they do not throw substrate (Randall et al. 2000). Results from this study and studies of D. spectabilis (Randall & Stevens 1987; Randall et al. 1995; Randall & Matocq 1997) suggest that kangaroo rats respond to predation risk and interact with snakes (1) to assess risk, (2) as a pursuit deterrent to communicate to the snake its chances of ambush are no longer available to cause the
585
586
ANIMAL BEHAVIOUR, 61, 3
snake to leave, and (3) to harass the snake to cause it to leave. One or all of these functions of snake-directed behaviour may apply to D. ingens and D. deserti and cannot be rejected until more information about the behaviour of the snake is obtained to determine whether the snake leaves the burrow area in response to the behaviour of the kangaroo rats. If snakes respond to D. ingens and D. deserti the same as D. spectabilis, the snake-directed behaviour should cause snakes to decrease stalking and find prey elsewhere (Randall & Matocq 1997). Acknowledgments We thank our field assistants, Shane King and Mary Atchison, the California Fish and Game for housing at the Painted Rock Ranch, and Rob Fulton and the staff at the Desert Studies Center for their help. Chris Moffatt, Jim Loughry and anonymous referees provided helpful comments on the manuscript. D. Boltas King is grateful for her support by a GAANN fellowship. The National Science Foundation supported this research, IBN-9723033 to J. A. Randall. Research on D. ingens was conducted under permit PRT 799486 from the U.S. Fish and Wildlife Service and a MOU, 31 May 1995–31 October 1998, from the California Department of Fish and Game. All procedures were approved on 18 April 1998 by the Institutional Animal Care and Use Committee of San Francisco State University under protocols 97-321 and 98-412. References Boltas, D. K. 1999. Antisnake behavior in two species of kangaroo rat (Dipodomys ingens and D. deserti): a comparative study. M.A. thesis, San Francisco State University. Bouskila, A. & Blumstein, D. T. 1992. Rules of thumb for predation hazard assessment: predictions from a dynamic model. American Naturalist, 139, 161–176. Braun, S. E. 1985. Home range and activity patterns of the giant kangaroo rat, Dipodomys ingens. Journal of Mammalogy, 66, 1–12. Brown, J. S., Kotler, B. P. & Valone, T. J. 1994. Foraging under predation: a comparison of energetic and predation costs in rodent communities of the Negev and Sonoran deserts. Australian Journal of Zoology, 42, 435–448. Brown, J. S., Laundre´, J. W. & Gurung, M. 1999. The ecology of fear: optimal foraging, game theory and trophic interactions. Journal of Mammalogy, 80, 385–399. Burger, J. 1998. Antipredator behaviour of hatchling snakes: effects of incubation temperature and simulated predators. Animal Behaviour, 56, 547–553. Caro, T. M. 1986a. The function of stotting: a review of the hypotheses. Animal Behaviour, 34, 649–662. Caro, T. M. 1986b. The function of stotting in Thomson’s gazelles: some tests of the predictions. Animal Behaviour, 34, 663–664. Caro, T. M. 1995. Pursuit-deterrence revisited. Trends in Ecology and Evolution, 10, 500–503. Caro, T. M., Lombardo, L., Goldizen, A. W. & Kelly, M. 1995. Tail-flagging and other antipredator signals in white-tailed deer: new data and synthesis. Behavioral Ecology, 6, 442–450. Daly, M., Wilson, M., Behrends, P. R. & Jacobs, L. F. 1990. Characteristics of kangaroo rats, Dipodomys merriami, associated with differential predation risk. Animal Behaviour, 40, 380–389.
Daly, M., Jacobs, L. F., Wilson, M. I. & Behrends, P. R. 1992. Scatter hoarding by kangaroo rats (Dipodomys merriami) and pilfering from their caches. Behavioral Ecology, 3, 102–111. Dugatkin, L. A. & Godin, J. G. J. 1992. Prey approaching predators: a cost–benefit perspective. Annals Zoologica Fennici, 29, 233–252. Elgar, M. A. 1989. Predator vigilance and group size in mammals and birds: a critical review of the empirical evidence. Biological Reviews, 64, 13–33. Fishman, M. A. 1999. Predator inspection: closer approach as a way to improve assessment of potential threats. Journal of Theoretical Biology, 196, 225–235. FitzGibbon, C. D. 1989. A cost to individuals with reduced vigilance in groups of Thompson’s gazelles hunted by cheetahs. Animal Behaviour, 37, 508–510. FitzGibbon, C. D. 1994. The cost and benefits of predator inspection behaviour in Thomson’s gazelles. Behavioral Ecology and Sociobiology, 34, 139–366. FitzGibbon, C. D. & Fanshawe, J. H. 1988. Stotting in Thomson’s gazelles: an honest signal of condition. Behavioral Ecology and Sociobiology, 23, 68–74. Hasson, O. 1991. Pursuit-deterrent signals: communication between prey and predator. Trends Ecology and Evolution, 6, 325–329. Herzog, H. A. 1986. Development of antipredator responses in snakes: I. Defensive and open-field behaviors in newborns and adults of three species of garter snakes (Thamnophis melanogaster, T. sirtalis, T. butleri). Journal of Comparative Psychology, 100, 372–379. Illius, A. W. & FitzGibbon, C. 1994. Costs of vigilance in foraging. Animal Behaviour, 47, 481–484. Kotler, B. P. 1984. Risk of predation and the structure of desert communities. Ecology, 65, 689–701. Kotler, B. P., Brown, J. S. & Hasson, O. 1991. The specter of predation: factors affecting gerbil foraging behavior and rates of owl predation. Ecology, 72, 2249–2260. Kotler, B. P., Brown, J. S. & Mitchell, W. A. 1994. The role of predation in shaping the behaviour, morphology and community organisation of desert rodents. Australian Journal of Zoology, 42, 449–466. Lima, S. L. 1998. Stress and decision making under the risk of predation: recent developments from behavioral, reproductive, and ecological perspectives. Advances in the Study of Behavior, 27, 215–290. Lima, S. L. & Dill, L. M. 1990. Behavioral decisions made under the risk of predation: a review and prospectus. Canadian Journal of Zoology, 68, 619–640. Loughry, W. J. 1987. The dynamics of snake harassment by blacktailed prairie dogs. Behaviour, 103, 27–48. Loughry, W. J. 1989. Discrimination of snakes by two populations of black-tailed prairie dogs. Journal of Mammalogy, 70, 627–630. McNamara, J. M. & Houston, A. I. 1992. Evolutionary stable levels of vigilance as a function of group size. Animal Behaviour, 43, 641–658. Maynard Smith, J. 1974. The theory of games and the evolution of animal conflicts. Journal of Theoretical Biology, 47, 209–221. Owings, D. H. & Coss, R. G. 1977. Snake mobbing by California ground squirrels: adaptive variation and ontogeny. Behaviour, 62, 50–69. Owings, D. H. & Owings, S. C. 1979. Snake-directed behavior by black-tailed prairie dogs (Cynomys ludovicianus). Zeitschrift fu¨r Tierpsychologie, 49, 35–54. Perri, L. M. & Randall, J. A. 1999. Behavioral mechanisms of coexistence in sympatric species of desert rodents, Dipodomys ordii and D. merriami. Journal of Mammalogy, 80, 1297–1310. Poran, N. S. & Coss, R. G. 1990. Development of antisnake defenses in California ground squirrels (Spermophilus beecheyi): behavioral and immunological relationships. Behaviour, 112, 222–245.
RANDALL & BOLTAS KING: ANTISNAKE BEHAVIOUR
Randall, J. A. 1984. Territorial defense and advertisement by footdrumming in bannertail kangaroo rats (Dipodomys spectabilis) at high and low population densities. Behavioral Ecology and Sociobiology, 16, 11–20. Randall, J. A. 1989a. Individual footdrumming signatures in bannertailed kangaroo rats, Dipodomys spectabilis. Animal Behaviour, 38, 620–630. Randall, J. A. 1989b. Territorial-defense with neighbors and strangers in banner-tailed kangaroo rats. Journal of Mammalogy, 70, 308–315. Randall, J. A. 1991. Mating strategies of a nocturnal desert rodent. Behavioral Ecology and Sociobiology, 28, 215–220. Randall, J. A. 1993. Behavioral adaptations of desert rodents (Heteromyidae). Animal Behaviour, 45, 263–287. Randall, J. A. 1994. Convergences and divergences in communication and social organization of desert rodents. Australian Journal of Zoology, 42, 404–433. Randall J. A. 1995. Modification of footdrumming signatures: changing territories and gaining new neighbours. Animal Behaviour, 49, 1227–1237. Randall, J. A. 1997. Species-specific footdrumming in kangaroo rats: Dipodomys ingens, D. deserti, D. spectabilis. Animal Behaviour, 54, 1167–1175. Randall, J. A. & Stevens, C. M. 1987. Footdrumming and other anti-predator responses in the bannertail kangaroo rat (Dipodomys spectabilis). Behavioral Ecology and Sociobiology, 20, 187–194. Randall, J. A. & Matocq, M. D. 1997. Why do kangaroo rats (Dipodomys spectabilis) footdrum at snakes? Behavioral Ecology, 8, 404–413. Randall, J. A., Hatch, S. M. & Hekkala, E. R. 1995. Inter-specific variation in anti-predator behavior in sympatric species of kangaroo rat. Behavioral Ecology and Sociobiology, 36, 243–250. Randall, J. A., Rogovin, K. A. & Shier, D. M. 2000. Anti-predator behavior of a social desert rodent: footdrumming and alarm calling in the great gerbil, Rhombomys opimus. Behavioral Ecology and Sociobiology, 48, 110–118. Roberts, G. 1996. Why individual vigilance declines as group size increases. Animal Behaviour, 51, 1077–1086.
Rowe, M. P. & Owings, D. H. 1990. Probing, assessment, and management during interactions between ground squirrels and rattlesnakes. Part 1: risks related to rattlesnake size and body temperature. Ethology, 86, 237–249. Scheel, D. 1993. Watching for lions in the grass: the usefulness of scanning and its effects during hunts. Animal Behaviour, 46, 695–705. Snedecor, G. W. & Cochran, W. G. 1967. Statistical Methods. Ames: Iowa State University Press. Stern, B. 1980. Habitat partitioning in two species of kangaroo rats (Rodentia; Heteromyidae): Dipodomys deserti and Dipodomys merriami. Ph.D. thesis. University of California, Berkeley. Swaisgood, R. R., Owings, D. H. & Rowe, M. P. 1999a. Conflict and assessment in a predator–prey system: ground squirrels versus rattlesnakes. Animal Behaviour, 57, 1033–1044. Swaisgood, R. R., Rowe, M. P. & Owings, D. H. 1999b. Assessment of rattlesnake dangerousness by California ground squirrels: exploitation of cues from rattling sounds. Animal Behaviour, 57, 1301–1310. Towers, S. R. & Coss, R. G. 1990. Confronting snakes in the burrow: snake-species discrimination and antisnake tactics of two California ground squirrel populations. Ethology, 84, 177–192. Wilkinson, L. 1997. SYSTAT. The System for Statistics. Evanston, Illinois: SPSS. Williams, D. F. 1992. Geographic distribution and population status of the giant kangaroo rat, Dipodomys ingens (Rodentia, Heteromyidae). In: Proceedings of the Conference on Endangered and Sensitive Species of the San Joaquin Valley, California (Ed. by D. F. Williams, T. A. Rado & S. Byrne), pp. 301–327. Sacramento: California Energy Commission. Woodland, D. J., Jaafar, Z. & Knight, M. L. 1980. The ‘pursuit deterrent’ function of alarm signals. American Naturalist, 115, 748–753. Zar, J. H. 1984. Biostatistical Analysis. Englewood Cliffs, New Jersey: Simon & Schuster.
587