932 SHORT COMMUNICATIONS

The Condor89~932-935 0 The CooperOrnithological Society 1987

DOMINANCE-SPECIFIC VIGILANCE IN THE TUFTED TITMOUSE: EFFECTS OF SOCIAL CONTEXT ’

THOMAS A. WAITE Department of Zoology, The Ohio State University,Columbus, OH 43210

Key words: Tufted Titmouse; Parus bicolor; vigi- METHODS lance; dominance;social foraging. Four pairs of wild-caught Tufted Titmice were housed The hypothetical benefits accruingto animals partici- in a large indoor aviary 4.8 m x 4.2 m x 2.7 m high pating in foraginggroups are of two classes,improved between 19 December 1984 and 5 March 1985. To predator avoidance and enhanced foraging efficiency permit instant recognition of individuals, each bird’s (reviewedby Krebs and Davies 1987).These two classes cheek patcheswere painted a unique color with water- of benefitsoften are not independent;for example,there proof felt tip markers. As Tufted Titmice are sexually is a growing literature demonstrating that the size of monomorphic and difficult to sex reliably on the basis the foraging group influences how individuals’ time of external features,and as the birds in this study were budgetsare apportioned among such antagonisticac- neither sacrificed nor laparotomized, the sex of the tivities as vigilance for predators and foraging (e.g., birds was inferred on the basis of wing cord length, Pulliam 1973,Powell 1974,Caraco 1979, Barnard 1980, body mass, and whether singing occurred.The domi- Bertram 1980, Caraco et al. 1980a, Elgar and Catterall nant individual in three replicateswas a putative male, 1981, Lendrem 1983, Studd et al. 1983, Elgar et al. and in the fourth replicate a putative female. The sub- 1984). Several studies on how birds make this trade- ordinate individual in two replicates was a putative off between foragingefficiently and avoiding predators female. The sex of the subordinate bird in the other have demonstratedthat as flock size increasesthe pro- two replicatescould not be determined usingthe above portion of time allocated by individuals to foraging criteria. To ensure that the titmice had been familiar activities is increased,while the flock’s aggregatevig- with each other in the wild, the birds of eachdvad were ilance level is maintained or even increased(e.g., Pow- captured contemporaneouslyat a single trapping sta- ell 1974, Siegfried and Underhill 1975, Caraco 1979, tion. The birds were held at 18.6 +- 1.X (X + SE) on Jenningsand Evans 1980, Sullivan 1984a). natural photoperiod, and were maintained on an ad Most previous studies have failed to consider that libitum diet of sunflower seeds (Heiianthus sp.) and individuals differing in dominance statusprobably ex- mealworms (Tenebrio sp.). In order to be able to cal- perience different benefits and costs associatedwith culate rates of feeding in kcal hr-I, the energeticcon- flock membershin (but seeMoore 1972. cited in Caraco tents of sunflower seeds(0.227 kcal seed-l) and meal- 1979; Caraco 1979; Ekman and Askenmo 1984; Ek- worms (0.086 kcal mealworm-l) were determined man 1987). In particular, the time available for (Lancaster Labs., Lancaster, Pennsylvania; details in foraging should be more constrainedfor subordinates Waite 1986). Each dvad of titmice had been used in if higher-ranking conspecificsinterfere with their for- other experiments (Waite 1986), and had been part of aging. The present study addressesexperimentally the a captive mixed-speciesflock that was comprised of a possibility that subordinates are more vigilant than male and a female Downy Woodpecker (Picoidespu- dominants becausethey must keep higher-rankingcon- bescens),a male and a female White-breastedNuthatch specificsunder surveillance to avoid aggressiveinter- (Sitta carolinensis),two Carolina Chickadees (Parus actions while also remaining vigilant for predators(cf. carolinensis),and a third Tufted Titmouse. Each tit- Robinson 1981, Waite 1986). Using captive Tufted mouse had been exposedto four playbacksof a Tufted Titmice (Parus bicolor),I examined the nonexclusive Titmouse alarm call on a single day within the 4-day hypothesesthat socially foraging animals are vigilant period prior to this study (Waite 1986). In addition, for (1) predatorsand (2) other foraginggroup members while held as a member of a captive mixed-species and, thus, vigilance is subjectto control by at least two flock, each titmouse had been exposedto an overflight proximate factors, group size and dominance status, of a one-half life-size model of a Sharn-shinnedHawk respectively.One predictioncan be generatedfrom each (Accipiterstriatus, unpubl. results).Thus, it is assumed of these hypotheses,respectively: (1) a dominant tit- that the titmice perceived some risk of predation dur- mouse shouldbe more vigilant when foragingsolitarily ing this study. Details concerning the aviary and the than when foraging as a member of a dyad, and (2) a conditionsunder which the birds were held are in Waite subordinate titmouse should be more vigilant than a (1986). dominant titmouse when they forage together. After the three Tufted Titmice had spent at least 3 days acclimatingto the aviary as members of a mixed- species flock, dominance relationships among them ’ Received 9 January 1987. Final acceptance8 April were determined. This was accomplishedby recording, 1987. during 20 15-min observationsessions, all interactions SHORT COMMUNICATIONS 933 in which one bird used a supplanting attack or aerial chaseagainst another titmouse. 51 The general experimental procedurewas as follows. (0.57) On at least the 10th day of captivity for each mixed- 1 speciesflock, the Downv Woodneckers.White-breast- r * ed Nuthatches, Carolina Chickadees, and the Tufted 4 0.011 Titmice intermediate in dominance status were re- moved from the aviary, leaving only the dominant and subordinatetitmice. I then observed the titmice each i r --I I (0.6 of the next 3 days beginningbetween 09:45 and 11:25 in one of the following contexts: subordinate solitary, 3- dominant solitary, and foraging socially (subordinate with dominant presentand dominant with subordinate present). For each dyad of titmice, the order of these three contextswas generatedfrom a random-numbers table. Observationsof birds of each dominance status foraging socially were made on the same day (order determined by a coin toss). Irrespective of dominance statusand social context, trials lasted 72.8 min (SD = 17.6 min) on average.All supplantingattacks and chas- 1 es occurringduring these trials were recorded. During the observationsof solitary individuals, the birds were in both visual and acousticisolation from one another. In each trial, the focal bird was observedas it arrived 0 at a 0.6 m x 0.6 m feedingtray positioned 1.5 m above sot SOL sot SOL the aviary floor. The time lag between the beginning of a trial and the focal bird’s first visit to the feeder DOMINANTS SUBORDINATES averaged6.6 min (SD = 5.0) and was statistically un- FIGURE 1. Time spent scanning at the feeder by related to either dominance status or social context dominant and subordinateTufted Titmice in two con- (paired t-tests, all Bonferroni’s Ps > 0.77; Snedecor texts, social (i.e., as a member of a dyad) and solitary. and Cochran 1967). No food was consumed during The bars represent the mean of means for four repli- this interim; however, the titmice presumably were cates, and vertical lines indicate f 1 SE. P-values are relatively well-fed at the onset of all trials as they had for paired t-tests. Parenthesesindicate two-tailed tests; been housed in the test arena and allowed continuous others are one-tailed. Asterisksindicate statistical sig- free accessto the feeder since dawn. As an index of nificanceafter applying Bonferroni’s probabilitiestech- vigilance, I recorded the time spent scanning at the nique for multiple comparisonsto achieve a 0.05 ex- feeder (Lendrem 1983). After arriving at the feeding perimentwise error rate (Snedecorand Cochran 1967). table, birds often would remain immobile for 1 to 5 set, and sometimes for as long as 16 sec. During this brief “freeze,” they habitually assumed a very erect any clear-cut pattern. Thus, the assumption that 1112 posture while moving the head from side to side, ap- perceived risk of attack by a predator or dominant parently scanning the environment. For each of 15 conspecific was constant throughout the experiment consecutivevisits to the feeder by the focal bird that appearsjustified. resulted in the consumption of a food item (always Multiple comparisons among results from the two fewer than 19 and 23 total visits by dominants and socialcontexts and two dominance ranks were accom- subordinates,respectively), I recordedthe elapsedtime, plished by repeated paired t-tests (ns = 4) using Bon- to the nearest second,between alighting on the feeder ferroni’s probabilitieswith an experimentwiseerror rate and either the first perch change or the first lowering of 0.05 (Snedecorand Cochran 1967). Asymmetries in of the head suchthat the axis of the bill dropped below dominancewere testedby the binomial probability test the horizontal. Either of these behaviors was followed (Snedecorand Cochran 1967). almost invariably by the graspingof a food item with the bill. Any subsequentscanning during a single visit RESULTS to the feeder was not recorded. Scanningat the feeder The dominance relationship between the two birds of could be measuredunambiguously, and appearsa rea- each replicate was unequivocal as the dominant won sonable index of the extent to which individuals are every dominance interaction both when the titmice sensitive to the risk of imminent attack by a predator were housedtogether during observationsof vigilance (Lendrem 1983) or supplantingattack by a conspecific. behavior (X = 9.5 interactions hrrl, SD = 6.6) and In only one of the 16 observation sessionswas there a when they were members of the mixed-species flocks significantcorrelation between time spent scanningat (X = 3.2 interactions hrl, SE = 1.6; one-tailed bino- the feeder and the order of visits (solitary subordinate mial Ps < 0.032 for the probability of suchan extreme in replicate 4; I = -0.608, P = 0.016, two-tailed). In outcome occurringby chance). During trials in which addition, a visual inspectionof scanningdurations over the subordinate was foraging in the presence of the the three days of trials (_z + SE: 2.1 + 0.6 set, 1.5 ? dominant, an averageof 15.8% of the supplantingat- 0.6 set, and 2.9 ? 1.2 set, respectively), independent tacks occurredat the feeder. Moreover, dominant tit- of dominance statusand social context, did not reveal mice (K = 2.08, SD = 0.35) had a significantlyhigher 934 SHORT COMMUNICATIONS feeding rate (kcal hr-I) than did subordinates(1.62 + birds respondedby foraginghigher in the canopy,where 0.40) when they foraged together (t = 4.31, Bonfer- dominants usually forage. In addition, Ekman et al. roni’s P = 0.046, one-tailed). In contrast, the compar- (198 1) documentedthat subordinate,first-year Willow ison between dominants (1.80 * 0.40 kcal hr-I) and Tits were more likely to fall victim to predators than subordinates(1.37 & 0.46 kcal hrr I) when they foraged were older, dominant birds. My resultssuggest another solitarily was nonsignificant(t = 1.63, P = 0.80, two- benefit of social superiority in titmice, that is, domi- tailed). Similarly, no statisticaldifferences emerged in nants may derive a greater benefit from foraging so- feeding rates while solitary compared to while feeding cially in terms of reduced vigilance than may lower- socially for either dominants or subordinates.Taken rankingconspecific flock-mates. Any suchbias in favor together, these results sustain the assumption that of dominants in a vigilance-reductionbenefit from for- priority of accessto a contested resource,namely the aging socially may allow dominants to allocate more food supply at the feeder, was biased in favor of dom- time to foraging. inants. Moreover, my results prompt the speculation that Dominant titmice scannedat the feeder significantly in addition to any microhabitat-specificdifferences in longer when tested as an isolate than when allowed to vulnerability to predation, subordinates could be at feed with a subordinatetitmouse present(Fig. 1). This greater risk of predation (or experience reduced for- result supportsthe hypothesisthat vigilance is directed agingefficiency) owing to the extra constraint on their toward predators and, thus, is influenced by the size time budgets of vigilance directed toward dominant of the foraging group. animals. Admittedly, however, vigilance for predators When the dominant and subordinate titmice were and vigilance for dominant group members may be housedtogether in the aviary, the subordinatescanned complementary to some degree; subordinates some- at the feeder significantlylonger than did the dominant times at least may be able to scan the environment (Fig. 1). This result supportsthe hypothesisthat vigi- simultaneouslyfor approachingpredators and socially lance is directed toward other members of the social dominant animals. Alternatively, to the extent that group and, thus, is influenced by an individual’s dom- keeping dominant flock-mates under surveillanceand inance status. being vigilant for predators are synchronous tasks, Whereas dominants increasedtheir scanningdura- dominants might improve their safety by forcing so- tions when solitary, the durations of scanningby sub- cially inferior individuals, via supplantingattacks and ordinates in the two social contexts were statistically chases,to maintain high levels of vigilance. Such in- indistinguishable (Fig. 1). Moreover, the time spent terpretations are somewhat confounded because, al- scanningby solitarydominant and solitary subordinate though it is common for free-ranging dominant and titmice did not differ statistically(Fig. 1). These results subordinate Tufted Titmice to forage within several further supportthe assertionthat socialdominance was meters of each other (pers. observ.), subordinatesin the principal proximate causeof the differencein scan- this study were forced to remain within some maxi- ning durations by dominant and subordinate titmice mum distanceof dominant. Thus, the vigilance of sub- when they were tested together. ordinates when foraging socially might have been ex- aggerated. DISCUSSION My results differ from Hegner’s (1985) results on My resultssupport the hypothesesthat vigilance is (1) Blue Tits (P. cueruleus).He found no dominance-spe- directed toward predatorsand (2) toward higher-rank- cific differences in the partitioning of time budgets ing members of a foraging group and, thus, is influ- among feeding, scanning,and other activities, partic- enced both by the size of the foraginggroup and by an ularly immediately after exposureto a model Eurasian individual’s dominance status within the group. As Sparrowhawk (Accipiternisus). This difference in our predicted, (1) dominant Tufted Titmice spent more results may be attributable to a relatively lower per- time scanningupon arriving at the feeding tray when ceived risk of predation by the titmice in my study they were tested as isolatesthan when they were tested than by the birds in Hegner’s study. It may be then with a subordinatepresent, and (2) subordinatesspent that if the perceived predation risk is high, high- and more time scanning at the feeder when a dominant low-ranking birds scan at similar rates, whereas if the conspecificwas present than when tested alone. dangeris low, dominants may relax their scanningrate The proximate causeof the lower feeding rate (kcal when feeding in a group. hrr’) in subordinatescompared to dominants when Other studieshave demonstrated that the vigilance they were housed together remains open to question. level of foragingbirds increasedwith decreasinggroup This difference could reflect (1) a cost of higher con- size (e.g., Powell 1974, Caraco 1979, Barnard 1980, specific-directed vigilance levels in subordinates, (2) Bertram 1980, Elgar and Catterall 1981, Studd et al. restricted accessto the feeder for subordinates,or (3) 1983, Elgar et al. 1984), with increasingdistance from the lower energeticcost of social subordinance(Hog- cover (Barnard 1980, Caraco et al. 1980a), with the stad 1987). appearanceof a predator (Caraco et al. 1980b, Hegner Several studies have suggestedthat dominant indi- 1985) with a reduction in visibility (Metcalfe 1984) viduals of Pam spp. force subordinate conspecifics when foraging in microhabitats presumed risky (Jans- foragingin their companyto occupymicrohabitats with son 1982; Lendrem 1983; Ekman 1987), when for- greater predation risk (Glase 1973; Jansson 1982; Ek- aging for conspicuousrather than cryptic prey (Law- man and Askenmo 1984; Ekman 1987). Ekman and rence 1985) and when an alarm call wasgiven (Sullivan Askenmo showed that when dominant Willow Tits 1984b). These studies share the assumption that the (Pmw montanus)were removed, subordinatefirst-year detection of predators is the primary function of vig- SHORT COMMUNICATIONS 935 ilance. Indeed, Knight and Knight (1986) have pro- HEGNER,R. E. 1985. Dominance and anti-predator vided the only evidence that vigilance in birds that behaviour in Blue Tits (Purus caeruleus).Anim. feed in social groups has the additional function of Behav. 33:762-768. detecting kleptoparasitic attacks launched by conspe- HOGSTAD,0. 1987. It is expensive to be dominant. cific flock-mates.They found a positive correlation be- Auk 104:333-336. tween vigilance and group size for groups of eight to JANSSON,C. 1982. Food supply, foraging, diet and 14 Bald Eagles(Huliaeetus leucocephalus). The results winter mortality in two coniferousforest tit species. of the present study constitute the first experimental Ph.D.diss., Univ. of Gothenburg, Sweden. demonstration that vigilance levels in birds can be in- JENNINGS,T., AND S. M. EVANS. 1980. Influence of fluenced by an individual’s dominance status (but cf. position in the flock and flock size on vigilance in Waite 1986).Unfortunately, however, becausethe birds the Starling, Sturnus vulgaris.Anim. Behav. 28: in this study were not aged and the sex was only in- 634-635. ferred, I cannot eliminate the possibility that the vig- KNIGHT, S. K., AND R. L. KNIGHT. 1986. Vigilance ilance differencesobserved were partly attributable to patternsin Bald Eaglesfeeding in groups.Auk 103: differences in age and sex per se that were correlates 263-272. of dominance status. KREBS,J. R., ANDN. B. DAVIES. 1987. An introduc- tion to behavioural ecology. Sinauer Associates, I thank A. P. Marshall and R. J. Stephansfor help Sunderland,MA. trapping the birds. J. Ekman, T. C. Grubb, Jr., R. E. LAWRENCE,E. S. 1985. Vigilance during ‘easy’ and Hegner, A. S. Gaunt, and S. Lustick made valuable ‘difficult’ foraging tasks. Anim. Behav. 33: 1373- comments on an earlier version of the manuscript,and 1375. J. A. Smallwood offeredsome helpful suggestionswhile LENDREM,D. W. 1983. Predation risk and vigilance this work was in progress.The Beatty, Condron, Pride- in the Blue Tit (Purus cueruleus).Behav. Ecol. more, and Riggs families and the heirs of Sarah Sociobiol. 14:9-13. Finkbone granted permission to trap birds on their METCALFE,N. B. 1984. The effects of habitat on the properties. This study was supported by NSF grant vigilance of shorebirds: is visibility important? BSR-83 13521 to T. C. Grubb, Jr. Anim. Behav. 32:981-985. LITERATURE CITED MOORE,N. J. 1972. The ethology of the Mexican Junco (Junco phaeonotuspalliatus). Ph.D.diss., BARNARD,C. J. 1980. Flock organizationand feeding Univ. of Arizona, Tucson. budgetsin the House Sparrow (Passerdomesticus POWELL,G.V.N. 1974. Experimental analysis of the L.). Anim. Behav. 28295-309. social value of flocking in Starlings (Sturnis vul- BERTRAM,B.C.R. 1980. Vigilance and group size in saris) in relation to predation and foraging.Anim. Ostriches.Anim. Behav. 28:278-286. Behav. 22:501-505. CARACO,T. 1979. Time budgetingand group size: a PULLIAM,H. R. 1973. On the advantagesof flocking. test of theory. Ecology 60:6 18-627. J. Theor. Biol. 38:4 19-422. CARACO,T., S. MARTINDALE,AND H. R. PULLIAM. ROBINSON,J. G. 1981. Spatial structure in foraging 1980a. Avian time budgets and distance from groupsof wedge-cappedcapuchin monkeys Cebus cover. Auk 97:872-875. nigrivittatus.Anim. Behav. 29:1036-1056. CARACO,T., S. MARTINDALE,AND H. R. PULLIAM. SIEGFRIED,W. R., AND L. G. UNDERHILL. 1975. 1980b. Avian flocking in the presenceof a pred- Flocking as an antipredator strategy in doves. ator. Nature 285:400-40 1. Anim. Behav. 23:504-508. EKMAN,J. 1987. Exposure and time use in Willow SNEDECOR,G. W., ANDW. G. COCHRAN. 1967. Sta- Tit flocks:the costof subordination.Anim. Behav. tistical methods. Iowa State Univ. Press, Ames, 35:445-452. IA. EKMAN,J., ANDC.E.H. ASKENMO. 1984. Social rank STLJDD,M., R. D. MONTGOMERY,AND R. J. ROBERTSON. and habitat use in Willow Tit groups. Anim. Be- 1983. Group-size and predator surveillance in hav. 32:508-5 14. foragingHouse Sparrows(Passer domesticus). Can. EKMAN, J., G. CEDERHOLM,AND C.E.H. ASKENMO. J. Zool. 61:226-231. 1981. Spacing and survival in winter groups of SULLIVAN,K. A. 1984a. The advantages of social the Willow Tit Parus montanus and Crested Tit foraging in Downy Woodpeckers. Anim. Behav. P. cristatus-a removal study. J. Anim. Ecol. 50: 32:16-22. l-9. SULLIVAN,K. A. 1984b. Information exploitation by ELGAR,M. A., ANDC. P. CATTERALL.198 1. Flocking Downy Woodpeckersin mixed-speciesflocks. Be- and predator surveillancein House Sparrows:test haviour 9 1:294-3 11. of an hypothesis.Anim. Behav. 29:868-872. WAITE,T. A. 1986. Experimental studiesof foraging ELGAR,M. A., P. J. BURREN,AND M. POSEN. 1984. and anti-predator behavior in some captive birds Vigilance and perception of flock size in foraging of a winter bark-foraging guild of temperate de- House Sparrows (Passer domesticusL.). Behav- ciduous woodland. M.S.thesis, The Ohio State iour 90:2 15-223. Univ., Columbus. GLASE,J. 1973. Ecologyof social organization in the black-cappedChickadee. Living Bird 121235-267.