Population differences in how black-tailed prairie dogs deal with snakes

Behav Ecol Sociobiol (1988) 22:61-67 Behavioral Ecology and Sociobiology Ii) Springer-Verlag 1988 Population differences in how black-tailed prairie dogs deal with snakes W.J. Loughry* Animal Behavior Graduate Group, Department of Psychology, University of California, Davis, CA 95616, USA Received April 1O, 1987/ Accepted July 27, 1987 Summary. This paper examines the antisnake behavior of a snake-experienced Texas (TX) population and a snake-naive South Dakota (SO) population of black-tailed prairie dogs (Cynomys ludovicianus). Animals in both populations were presented with tethered rattlesnakes and nonvenomous bullsnakes before and after the first emergence of pups from their natal burrows. SD and TX adults were more likely to actively harass snakes than were pups. There were no differences in the behavior of SO adult males vs females or fathers vs non-fathers, contrary to what was found in TX. However, SO mothers did call more and stayed closer to snakes than did SD non-mothers. All SD adult sex-parental classes spent less time dealing with snakes after the emergence of pups, which was not observed in TX. The primary difference between the two populations was the behavior of TX males who spent a great deal of time harassing snakes. SD adults behaved most like TX females in that these groups spent little time actively dealing with snakes. Pups in both populations behaved similarly. These results are interpreted in terms of the relationship between potential predator and potential prey. Introduction Several recent papers have argued that antipredator behavior might best be understood in terms of the relationship between potential predator and potential prey (Hennessy 1982, 1987; Owings and Loughry 1985; Loughry 1987a). One possible descriptor of this relationship (from the prey's perspective) is vulnerability. Vulnerability has at least * Present address: Department of Zoology, University of Tennessee, Knoxville, TN 37996, USA two components: personal and reproductive vulnerability (Loughry 1987a). Reproductive vulnerability refers to the value at stake in dealing with a potential predator (cf Robertson and Biermann 1979; Shields 1984; Regelmann and Curio 1986). Personal vulnerability refers to the cost to an individual of being the target of a particular predator (cf Hennessy 1987). This relationship may be influenced by the identity of the potential prey, the identity of the potential predator, and the context of the encounter. For example, the behavior of black-tailed prairie dogs (Cynomys ludovicianus) dealing with potentially dangerous snakes has been shown to vary as a function of the type of snake being confronted (Owings and Loughry 1985; Loughry in prep.); the age, sex and parental status of the prairie dogs, the time of year (Loughry 1987a); and the type of encounter (experimental or natural; Loughry 1987b). Another possible influence on the relationship between predator and prey is a population's history with a predator. Several studies indicate that ground squirrels from populations with little or no exposure to rattlesnakes behave differently than individuals from rattlesnake-experienced populations (Owings and Coss 1977; Owings and Owings 1979; Coss and Owings 1985; Coss 1985; Rowe et al. 1986; Poran 1987). Thus, the historical context within which the predator-prey relationship is embedded may also influence antipredator behavior. To test this idea, two populations of blacktailed prairie dogs that differed in their recent history of exposure to snake predation were examined. The results from the snake-experienced population have been published (Loughry 1987a, b). In this paper I examine the behavior of snake-naive prairie dogs and then compare th.ebehavior of each age/sex class of prairie dogs in the two populations. \

62 Methods Black-tailed prairie dogs are large, diurnal, colonial ground squirrels whose social behavior has been described in detail (King 1955; Smith et al. 1973; Hoogland 1981, 1982, 1985). Black-tails live in groups called coteries (King 1955) usually containing one adult male, several closely related adult females, yearling males and females, and the young of the year (pups). I obtained data for two populations of prairie dogs that differed in their history of exposure to snakes. Details of methodology were identical at both sites and have already been published for the Texas (TX) site (Loughry 1987 a). Here, except where necessary, I present information only on the South Dakota (SD) site. Study sites. The snake-experienced population was located on the Muleshoe National Wildlife Refuge in Muleshoe, Texas and has been described previously (Loughry 1987 a). The snakenaive population was located on Lower Rankin Ridge in Wind Cave National Park, South Dakota and has been described by Garrett et al. (1982). Loughry (1987a, b) observed 21 natural encounters between TX prairie dogs and snakes in less than 6 months. Only 4 such encounters have been observed in SD in 14 years of intensive study (Hoogland pers. comm.), indicating that these two populations differ at least in their recent history of exposure to snakes. Observations. Observations were conducted in SD from 26 March through 10 June 1983 and in TX from 22 January to 28 June 1984. Snake experiments. Four different snakes were presented to prairie dogs at particular sites within each colony (TX prairie dogs also encountered snakes naturally, but these encounters are not considered here). Each site was presented with each snake type once before pup emergence and once after, for a total of 8 presentations at each site. The order of presentation of each snake type was random for each site, and no site was tested more than once per day. The snakes used were a large (126 cm, 1.4 kg) and small (106.5 cm, 625 g) prairie rattlesnake (Crotalus viridis viridis), and a large (140 cm, 685 g) and small (114 cm, 355 g) bullsnake (Pituophis melanoleucus sari). These snakes were not the same individuals used in the TX experiments. However, as in TX, the differences between large and small snakes were primarily due to weight rather than length. Three coteries within the colony were chosen for snake experiments. In two of these coteries at least one female weaned a litter. In the third coterie, no females produced pups and it is unlikely any female had pups at the time of the pre-pup experiments. Experimental procedure. Snakes were tethered (see Hennessy et al. 1981) at each site within 3-5 m of an active burrow. In coteries with pups, tether sites were near a burrow containing pups. The ensuing encounter was recorded on videotape for later analysis. The location and behavior of members of the coterie who were out of camera view were narrated onto both audio and video records of the encounter. Data presented here include these off-camera individuals. Statistical analysis. Data were obtained from the videotapes at 10 s intervals for the duration of the encounter. The variables decoded from the video tapes are described in detail elsewhere (Loughry 1987 a; see Table 1). Measures for each variable were obtained for each individual during a snake presentation as long as that individual was above ground. Average values for each individual for each encounter were then calculated and used in subsequent analysis. Variables using percentages were arcsin transformed prior to analysis. Because some of the variables were correlated with one another (Loughry 1987 a) a factor analysis was run (SPSS Inc. 1983) to identify those variables for use in MANOV A analyses. Different classes of prairie dogs (e.g., males, females) might have had different variables that were most important, so the factor analyses were run separately for each group. Although individuals contributed unequally to sample sizes, this did not affect the outcome of the analyses. The analyses were also performed with each individual represented only once, and the same univariate differences among groups were found (using independent samples t-tests; Bruning and Kintz 1977). Consequently, only the results of initial analyses are reported here. Results SD adults vs pups SD adults and yearlings did not differ in how they dealt with snakes (MANOV A F=O.l1, P<O.98), so the results for these two age classes were pooled for comparison with the behavior of pups. Fourteen pups and 23 adults were thus compared (for adults, only post-pup experiments are considered here). In TX, pups behaved as if they were more vulnerable to snakes than were adults and the same was true in SD (Fig. 1). These differences are borne out by the multivariate analysis as well (MAN- OVA F= 7.46, P<O.OOl, df=4,98). SD males vsfemales Contrary to what was found in TX, where males spent more time dealing with snakes, there were no differences between 8 adult male and 17 female SD prairie dogs for any of the 15 variables measured (Table 1). Multivariate analyses confirm the lack of sex differel),cesin snake-directed behavior by SD prairie dogs (MANOV A F=O.63, P<O.70, df = 4,99). SD fathers vs non-fathers In TX, fathers were more active in dealing with snakes. This was not the case in SD. Two fathers and 6 non-fathers did not differ in a multivariate comparison (MANOVA F=O.50, P<O.90, df= 5,27). However, unlike TX prairie dogs, SD nonfathers spent more time within 2 m of snakes and called closer to snakes than did fathers (Table 1). SD mothers vs non-mothers There were only minor differences due to maternity in TX prairie dogs. There were slightly larger dif-

2.5 PUPS. ADULTS!WI tiates the difference (MANOV A F= 2.36, P < 0.05, df = 5,65). 63 2 Pup emergence 1.5 0.5 0 18 16 B 0 14 D Y 12 L 10 E 8 N G 6 T H 4 S 2 0 60 P 50 E R C 40 E N T 30 T 20 I M E 10 0 ORIENTATION PORTION JY/MIN % S.D. DISTANCE JY DIST %2m % FEED Fig. 1. Age differences in the antisnake behavior of SD prairie dogs. Variables shown were significantly different in an ANO- VA comparison with df=1,101 [except ly dist (jump-yip distance, Table 1) where df = 1,46]. For orientation, values only range from 0-1 (proportion of time positively oriented at snakes) and for portion, values closer to 3 reflect more time spent near a snake's head and values around 1 of more time spent near a snake's tail. ly/min (jump-yips/min, Table 1) is self-explanatory. % S.D. refers to the time prairie dogs spent "snake-directed", see Table 1 ferences between 4 mothers and 13 non-mothers in SD. Mothers barked more at snakes and spent more time within 2 m of snakes than did nonmothers (Table 1). Multivariate analysis substan- Table 2 shows that SD prairie dogs became apparently less concerned with snakes after pup emergence. It could be argued that these animals habituated to the snakes, but there is an alternative explanation as well. The number of prairie dogs involved in snake encounters increased after pup emergence (pre-pup = 3.33 adults/encounter, postpup = 5.33 adults/encounter, two-tailed t-test for related measures; Bruning and Kintz 1977; t = 3.25, P<0.01, df= 12). Thus, the per capita involvement of an individual adult might have decreased after pup emergence due to the increased participation of other adults. This did not occur in TX where coterie sizes were smaller and no change in group size occurred after pup emergence (1.5 vs 2.0 adults/encounter, t= 1.32, P< 0.20, df= 12). Between population comparisons Pups. Fourteen SD pups and 7 TX pups were compared. Univariate comparisons revealed only that TX pups were further from snakes (45.44 vs 17.54 body lengths, Table 1). However, the overall similarity between the two populations is demonstrated by the multivariate analysis, which shows no significant difference between the two groups (MANOVA F=0.82, P<0.60, df=5,51). Fathers. Three TX fathers spent far more time actively dealing with snakes than did 2 SD fathers (Table 1). These considerable differences between the two populations are confirmed by the results of the MANOVA (F=-17.49, P<0.0001, df=5,17). Non-fathers. Six SD non-fathers were compared with two TX non-fathers. SD non-fathers were less active in dealing with snakes than were TX nonfathers (Table 1). A MANOV A confirms these results (F=4.66, P<0.003, df= 5,33). Mothers. Four SD and three TX mothers behaved similarly towards snakes (MAN 0 VA F = 2.19, P < 0.09, df=5,25). However, SD mothers did spend more time snake-directed but less time above ground than did TX mothers (Table 1). These differences indicate that, if present, SD mothers were slightly more willing to deal with snakes than were their TX counterparts.

Table1. Differencesin antisnakebehaviorwithinand betweenpopulationsof black-tailedprairiedogsa '".j:>. Within SD comparisons TX vs SD comparisons Adults (A) Males Fathers (F) Mothers (M) vs vs vs vs Variable pups (P) females non-fathers (NF) non-mothers (NM) Pups Fathers Non-fathers Mothers Non-mothers Strikes by snakes ns ns ns ns ns * SD<TX ns ns Swats by prairie dogs ns ns ns ns ns ns SD<TX ns ns Bouts of barking by prairie dogs ns ns ns M>NM ns ns ns ns ns Barks/min ns ns ns M>NM ns ns ns ns ns Distance from snake A<P ns ns ns SD>TX SD > TX SD>TX SD<TX SD<TX Orientation to snake b A>P ns ns ns ns ns SD>TX ns ns Portion of snake closest toc A>P ns ns ns ns ns ns ns SD> TX % time snake-directed A>P ns ns ns ns SD < TX ns ns ns % time within 2 m of snake A>P ns F<NF M>NM ns SD<TX SD < TX ns ns % time feeding A<P ns ns ns ns SD>TX SD<TX ns ns % time above ground during encounter ns ns ns ns ns SO<TX SO<TX SO<TX ns Jump-Ylps/min A>P ns ns ns ns SD<TX ns ns ns Jump yip qualityd ns ns ns ns ns ns ns ns SD>TX Jump-yip distance from snake A<P ns F<NF ns ns SD > TX ns ns ns Number of jump-yips per bout ns ns ns ns ns ns ns ns ns ~. Results of univariate ANOVAs for each comparison. For variables in which a significant difference (P<O.O5)was found, the direction of the difference is indicated b Either positive or negative (see Fig. 1) C Proportion of time spent near a snake's head, mid-body, or tail (see Fig. 1) d "Completeness" of the jump component of a jump-yip (see Smith et al. 1976; Loughry in press a for a description of jump-yipping) * P<O.O8

r=i 65 TX Fathers TX Non-Fathers SD Mothers TX Mothers SD Non-Fathers SD Fathers SD Non-Mothers TX Non-Mothers Fig. 2. Visual depiction of the relationships among all adult sex/parental classes in both populations. The bars indicate groups which are not significantly different from one another Table 2. Differences in the antisnake behavior of adult SO prairie dogs as a function of pup emergence Variable Before After pa Distance 14.60 10.67 0.02 Orientation 0.73 0.62 0.004 Portion 2.45 2.01 0.007 ly /min 1.32 0.63 0.009 % s.d. 37.09 16.92 0.001 % feed 21.82 45.29 0.001 a Results of an ANOVA run for each variable, in all cases df = 1,96; n = 40 and 64 respectively Non-mothers. Five TX and 13 SD females were non-mothers. SD non-mothers were closer to snakes, spent more time closest to the snake's head, and jump-yipped more conspicously than did TX non-mothers (Table 1). These differences are confirmed by multivariate analysis (MANOY A F= 2.46, P < 0.05, df = 5,55). Comparison among all groups. So far I have only examined differences between corresponding groups of prairie dogs in the two populations. Such an analysis does not tell us everything of interest. For instance, SD fathers are markedly different from TX fathers in how they behave in the presence of snakes but are SD fathers similar to some other TX sex/parental class? To answer this question, I compared all possible pairs of adult sex/ parental classes in a MANOY A. The results are depicted in Fig. 2. Figure 2 can be used as very gross indicator of the quality of snake-elicited behavior by each group. Of all groups, TX fathers were most active in directly confronting snakes. Proceeding along the line, active participation in snake encounters declines so that TX non-mothers were the adults least likely to deal with snakes. Discussion Although in some cases sample sizes were small, these results show considerable population differences in how black-tailed prairie dogs deal with snakes. These differences are largely due to the fact that TX males (fathers and non-fathers) spent considerably more time near snakes actively dealing with these potential predators than did any other individuals from either population. In general, adult SD prairie dogs behaved like TX adult females with the exception of SD mothers, who behaved more like TX non-fathers. Pups from both populations behaved similarly and rarely interacted with snakes, which suggests that some kind of ontogenetic change may lead to the greater involvement by TX males as adults. It would be interesting to know if TX male pups are more likely to approach encounters with snakes and to participate in harassing snakes than are female pups. One would also predict that this difference would become more pronounced as the TX pups got older. It is unclear from these results how SO prairie dogs perceive their vulnerability to snakes. Pups acted as though they were more vulnerable than adults, but adults did not differ much according to sex or parental status (reproductive vulnerability). I have already noted that SD adults seem similar in their behavior to that of TX females. Both groups can be described as organizing their behavior so as to monitor the encounter without actively participating (Le. "snake-indirected", see Hennessy et al. 1981; Loughry 1987a). However, the fact that some SD and TX adults behaved similarly need not imply that their behavior have the same functional basis. TX. females may have stayed away from snakes in order to prevent pups from approaching the encounter. SD adults may have behaved this way because they perceived themselves as very vulnerable to snakes. It seems more probable that, as Owings and Owings (1979) found, snake-naive prairie dogs are not as sensitive to the presence of snakes, unless the snakes are moving. While I did not distinguish snake-directed behavior as a function of snake movement, it was apparent that prairie dogs in both populations were much more attentive to a moving snake. Movement by tethered snakes was not regular throughout a trial, often being instigated by the approach of a prairie dog. As I have shown, this was far more likely to occur in TX, where some prairie dogs got very

r-r 66 close to the snakes. Thus, by staying further away, SD prairie dogs may not have generated the context necessary to evoke consistent snake-directed behavior (cf. Hennessy and Owings 1987). The results presented here confirm previous reports in documenting substantial differences between sciurid populations that differ in their history of exposure to snakes (Owings and Coss 1977; Owings and Owings 1979; Coss and Owings 1985; Coss 1985; Poran 1987). This influence of history is a higher-order influence on prairie dog-snake relationships, in which lower-level influences are embedded. For example, the influences of sex and parental status seem to be insignificant in SD due to the overriding importance of the population's history with snakes, while they are much more important in TX, operating within a differnt historical context. Population history may then be viewed as constraining the organization of antipredator behavior. It is important to note that snake-naive SD prairie dogs did not seem incompetent in dealing with snakes. There were no discernible differences in the motor patterns employed by SD prairie dogs in confronting the snakes. With the exception of swats (Table 1), SD prairie dogs" did" the same things as their TX counterparts. Rather, what differed was the way in which these motor patterns were organized in the two populations. As Hennessy (1982, 1987) has argued, differences in organization seem best understood in terms of the relationship between potential predator and potential prey. I have argued that the relationship between prairie dogs and snakes can be partly described in terms of a prairie dog's vulnerability (both personal and reproductive) to a particular snake. Influences on this relationship include the age, sex, and parental status of the prairie dog (Loughry 1987a), the type of snake being confronted (Owings and Loughry 1985), and the immediate context of the encounter (e.g. Loughry 1987b). I have shown here that the historical context within which the encounter is embedded also influences this relationship. Further understanding of prairie dog antisnake behavior will come from continued description of this relationship. Acknowledgements. In Texas, I thank the staff of the Muleshoe National Wildlife Refuge for all their invaluable help. Doug Stein performed vital duties as a field assistant. In South Dakota, I am especiallygrateful to John Hoogland for the opportunity to work with his study animals and for allowing me to parasitize his field assistants, particularly H. Spanglet and S. Birnbaum. The staff of Wind Cave National Park, especially R. Klukas, were also very helpful. B. Texal of the Rapid City Reptile Garden and R.O. Heatley of the Mangum Oklahoma Rattlesnake Roundup kindly provided the snakes used in this study. J. Williams, D. Turk, M. Clark, E. Ramones, and S. Chiabotti worked endlessly decoding videotapes. D.H. Owings, D.F. Hennessy, D.F. Lott and c.m. McDonough zealously critiqued the various drafts of this manuscript. This study was funded by Sigma Xi, The Theodore Roosevelt Memorial Fund, a U.c.D. Graduate Research Award, the U.C.D. Psychology Department, U.c.D. Faculty Research Grants to D.H. Owings, and William R. Loughry. References Bruning JL, Kintz BL (1977) Computational handbook ofstatistics. Foresman, Glenview Coss RG (1985) Evolutionary restraints on learning: phylogenetic and synaptic interpretations. In: Weinberger NM, McGaugh JL, Lynch G (eds) Memory systems of the brain: animal and human cognitive processes. Guilford, New York, pp 253-278 Coss RG, Owings DH (1985) Restraints on ground squirrel antipredator behavior: adjustments over multiple time scales. In: Johnston TD, Pietrewicz AT (eds) Issues in the ecological study oflearning. Erlbaum, Hillsdale, pp 167-200 Garrett MG, Hoogland JL, Franklin WL (1982) Demographic differences between an old and a new colony of black-tailed prairie dogs (Cynomys ludovicianus). Am Midi Nat 108:51-59 Hennessy DF (1982) Paradigmatic constraints on explaining predator harassment: risky versus conservative snake harassment by California ground squirrels. In: Functional significance of variation in predator harassment. Unpubl PhD diss, Univ California, Davis Hennessy DF (1987) A format for describing the generative structures of behavioral patterning. In: Greenberg G, Tobach E (eds) Levels of social behavior: evolutionary and genetic aspects. Gordian Press, New York (in press) Hennessy DF, Owings DH (1987) Rattlesnakes create a context for localizing their search for potential prey. Ethology (in press) Hennessy DF, Owings DH, Rowe MP, Coss RG, Leger DW (1981) The information afforded by a variable signal: constraints on snake-elicited tail flagging by California ground squirrels. Behaviour 78: 188-226 Hoogland JL (1981) Nepotism and cooperative breeding in the black-tailed prairie dog (Sciuridae: Cynomys ludovicianus). In: Alexander RD, Tinkle DW (eds) Natural selection and social behavior: recent research and theory. Chiron Press, New York, pp 283-310 Hoogland JL (1982) Prairie dogs avoid extreme inbreeding. Science 215:1639-1641 Hoogland JL (1985) Infanticide in prairie dogs: lactating females kill offspring ofciose kin. Science 230:1037-1040 King JA (1955) Social behavior, social organization, and population dynamics in a black-tailed prairie dog town in the Black Hills of South Dakota. Contr Lab Vert Bioi, Univ Mich 67: 1-123 Loughry WJ (1987a) The dynamics of snake harassment by black-tailed prairie dogs. Behaviour (in press) Loughry WJ (1987b) Differences in.natural and experimental encounters of black-tailed prairie dogs with snakes. Anim Behav (in press) Owings DH, Coss RG (1977) Snake mobbing by California ground squirrels: adaptive variation and ontogeny. Behavior 62: 50-69 Owings DH, Loughry WJ (1985) Variation in snake-elicited jump-yipping by black-tailed prairie dogs: ontogeny and snake specificity. Z Tierpsychol 70: 177-200 Owings DH, Owings SC (1979) Snake-directed behavior by

-.-.. 67 black-tailed prairie dogs (Cynomys ludovicianus). Z Tierpsychol 49: 35-54 Poran NS (1987) Adaptive variation and developmental aspects of the innate physiological and behavioral anti snake defense in California ground squirrels (Spermophilus beecheyi). Unpubl PhD diss, Univ California, Davis Regelmann K, Curio E (1986) Why do great tit (Parus major) males defend their brood more than females do? Anim Behav 34: 1206-1214 Robertson RJ, Biermann GC (1979) Parental investment strategies determined by expected benefits. Z Tierpsychol 50:124-128 Rowe MP, Coss RG, Owings DH (1986) Rattlesnake rattles and burrowing owl hisses: a case of acoustic Batesian mimicry. Ethology 72:53-71 Shields WM (1984) Barn swallow mobbing: self-defence, collateral kin defence, group defence, or parental care? Anim Behav 32: 132-148 Smith WJ, Smith SL, DeVilla JG, Oppenheimer EC (1976) The jump-yip display of the black-tailed prairie dog, Cynomys ludovicianus. Anim Behav 24:609-621 Smith WJ, Smith SL, Oppenheimer EC, DeVilla JG, Ulmer FA (1973) Behavior of a captive population of black-tailed prairie dogs: annual cycle of social behavior. Behaviour 46:189-220 SPSS Inc (1983) SPSSX user's guide. McGraw-Hill, New York