Competitive Interactions Within and Between Species in a Guild of Avian Scavengers

Home , Andean condor, Black vulture, King vulture, Turkey vulture

COMPETITIVE INTERACTIONS WITHIN AND BETWEEN SPECIES IN A GUILD OF AVIAN SCAVENGERS

MICHAEL P. WALLACE AND STANLEY A. TEMPLE Departmentof WildlifeEcology, University of Wisconsin,Madison, Wisconsin 53706 USA

ABSTRACT.--Weobserved Andean Condors(Vultur gryphus),King Vultures (Sarcoramphus papa),Black Vultures (Coragypsatratus), Turkey Vultures (Cathartesaura), and CrestedCara- caras(Polyborus plancus) interacting at 217 animal carcassesat two sitesin northern Peru. At 53 carcassesfor which we knew order of arrival, Turkey Vulturesusually arrived first, Black Vulturessecond, and condorsthird. On the basisof our observationsof 8,066 aggressive encountersbetween birds, we constructeddominance hierarchies by calculatingthe proportion of encounterswon by an individualof one species,sex, or age during encounterswith an individual of another species,sex, or age. Within each speciesthere was a positive relationship between a bird's dominanceand its age. In condors,males dominatedfemales of the sameage. Interspecificdominance was correlatedpositively with body mass.There are convergentsimilarities between the organizationsof guildsof Old and New World vultures. ReceivedI August1986, accepted24 November1986.

WHENlimited resourcesare found in •9idely coramphuspapa), Black Vultures (Coragypsatra- dispersed,rich patchesthat are ephemeral and tus), Turkey Vultures (Cathartes aura), and unpredictable in spatial and temporal occur- Crested Caracaras(Polyborus plancus). We have rence, competition between consumers could shown previously that during our study sev- involve either differential exploitation or in- eral of thesespecies were being limited by the terference.Differential exploitationin suchsit- availability of carrion during a prolonged pre- uationsis often basedon the differentialability E1Nifio drought (Wallace and Temple 1987). of individuals to locate the resource, and interference usually is achieved through the estab- lishment of dominancehierarchies among in- STUDY AREA AND METHODS dividuals that have found the resource (Maurer The field observationstook place between 1980 and 1984). Avian scavengersthat feed on carcasses 1984 in northern Peru in the Cerro Illescas region of large animals competefor just such a limit- andin thevicinity of •Iaupe. The Cerro Illescas isan ed, dispersed,rich, ephemeral, and unpredict- isolated mountain range of the Sechura Peninsula able food source, and differences in order of (6.0øS,81.0øW). Avian scavengersin this study area arrival at carcasses and dominance hierarchies fed almostexclusively on the carcassesof either ma- when birds are feeding together have been re- rine birds and mammals that washed ashore on the ported among members of scavengingguilds peninsula'sbeaches or fetal_ungulatesthat diedin (Kruuk 1967, Houston 1975). the surrounding desert.The Naupe area is about 150 km east of the Cerro Illescas in the western foothills The most detailed studiesof competitive in- of the Andes (5.35•S, 79.4øW).Avian scavengersin teractionsbetween avian scavengershave tak- this study area fed primarily on carcassesof domestic en place in Africa where large guilds of Old livestockthat died on the area'sdesert grasslands. World vultures (Accipitridae) feed on ungulate In both areaswe observedbirds interacting at 217 carcasses (Petrides 1959; Attwell 1963; Kruuk carcassesof large animalsat as many sites:129 burros 1967; Houston 1974, 1975). In contrast, there (Equusasinus), 57 dogs (Canisdomesticus), 12 sea lions have been few descriptions,most of them an- (Otaria byronia ), 10 goats( Capra h ir cus ), 4 horses( Equus ecdotalaccounts (e.g. Koford 1953,Stuart 1978), caballus),4 pigs (Susscrofa), and 1 green sea turtle of competitive interactions in guilds of New (Cheloniamydas). Each of these carcasseswas suffi- ciently large to be highly attractive to all members World vultures (Cathartidae). of the guild. We opened the largestcarcasses so that We observed patterns of differential exploi- they couldbe consumedby all guild members.If we tation and interference within and between 5 had not done this, smaller vultures would have been speciesthat are competingmembers of a guild forcedto wait until larger birds had first ripped the of avian scavengersin northern Peru: Andean thick-skinned carcassesopen. We watched these car- Condors (Vultur gryphus),King Vultures (Sar- cassesfor a total of 4,071 h from blinds positioned

290 The Auk 104: 290-295. April 1987 April 1987] CompetitionAmong Avian Scavengers 291

85 141 17 53 58 6 39 23 1.0 3O

2O7 131 0.5 --

132 rr 141 85 ß

-3 -2 -1 0 I 2 3 4 5 -2 -1 0 I 2 3 4 5 6 AGE OF INDIVIDUAL"A"MINUS AGE OF INDIVIDUAL'•' AGE OF INDIVIDUAL "A"MINUS AGE OF INDIVIDUAL 'B" Fig. 1. Proportionof encounterswon by a King Fig. 2. Proportionof encounterswon by a male Vulture of a certain age (individual A) when inter- AndeanCondor of a certainage (individual A) when actingwith a King Vulture of another age (individ- interactingwith a male condor of another age (in- ual B). Number above each point is samplesize. Age dividual B). Number aboveeach point is samplesize. differencesare in years. Age differencesare in years. nearby or, by using telescopes,from vantagepoints up to 0.75 km away. The number of scavengersob- During our observationswe alsohad opportunities servedat thesecarcasses ranged up to 240individuals to recordthe sequenceof arrival of Turkey Vultures, at a time. Black Vultures, and Andean Condors at 53 carcasses We identified all speciesthat visited the carcasses, we knew had not been visited previously by scav- and when possiblewe noted individual sexand age. engers.We positionedthese 53 carcassesin open areas AndeanCondors are sexuallydimorphic and undergo frequentedby all three speciesso that all birds had sequentialchanges in plumage, beak, and eye col- chancesof encounteringthem. oration until they are 6-7 yr old (K. C. Lint pers. comm., J. W. Carpenter pers. comm.). This allowed RESULTS us to distinguish7 age classesof condorsas well as their sex.King Vultures are sexuallymonomorphic, Use of carcassesby scavengers.--Weobserved but undergosequential changes in plumageand skin colorationwith age that allowed us to distinguish6 carcasseson 239 days. Turkey Vultures visited age classes(Heck 1968, C. Benevidiespers. comm.). the carcasseson 199 (83%) days, Black Vultures Turkey Vultures, Black Vultures, and Crested Cara- on 149 (62%) days, Andean Condors on 124 caras are also sexually monomorphic, but birds less (52%) days, King Vultures on 15 (16%) days, than 1 yr old were distinguishedfrom older birdson and caracarason 6 (3%) days. When Turkey the basisof plumage, skin, and beak color (Brown Vultures were present,they were accompanied and Areadon 1968). by Black Vultures on 73% of the days, by con- We recordedthe outcomesof 8,066direct aggres- dors on 58% of the days, by King Vultures on siveencounters between individuals that fed actively 8% of the days, and by caracarason 3% of the at carcasses.The winner was the bird that supplanted days.When BlackVultures were present,they the other individual, regardlessof whether the sup- planting involved physicalcontact or mere intimi- were accompaniedby Turkey Vultures on 95% dation. Becauseof the large number of Turkey and of the days, by condorson 75% of the days, by Black vultures that often interacted at the carcasses King Vultures on 10% of the days, and by cara- at the sametime, only about 50%of their interactions carason 4% of the days. Caracaraswere accom- were recorded.Condors and King Vultures, because panied by Turkey and Black vultures on 100% of their sizeand lower numbersat feedingsites, were of the days and by King Vultures and condors easier to track, and 95-100% of their interactions were on 66% of the days. On days when King Vul- recorded. From these data the frequenciesof inter- tures were present both Turkey and Black vul- actionsbetween different species,sexes, and age tures were always there; caracaraswere present classeswere calculated.Not all of the potential cat- on 47% of the days,and condorson 36% of the egoriesof interactionswere observed,but for those observedwe calculatedthe proportion of encounters days. When Andean Condors were present, won by an individual of one species,sex, or age dur- Turkey Vultures accompaniedthem on 95% of ing encounterswith an individual of a different the days, Black Vultures on 90% of the days, species,sex, or age.On the basisof theseprobabilities King Vultures on 9% of the days,and caracaras we constructed dominance hierarchies. on 6% of the days. 292 WALLACEAND TEMPLE [Auk,Vol. 104

45 40 34 16 136 64 29 11 29 18 5 1.0 92 ß ß ß ß 53 lOO

83 66

53 -5 -4 -3 -2 -1 0 I 2 3 4 5 6 MALE• AGE I,At•US FEMALE'S AGE

0 I 2 3 4 5 6 Fig. 4. Proportion of encounterswon by a male AGE OF INDIVIDUAL"A" MINUS AGE OF INDIVIDUAL'•" Andean Condor of a certain age when interacting with a female condorof anotherage. Number above Fig. 3. Proportionof encounterswon by a female eachpoint is samplesize. Age differencesare in years. Andean Condor of a certain age (individual A) when interactingwith a female condorof anotherage (in- dividualB). Numberabove each point is samplesize. spring. We never saw aggression between Age differencesare in years. adults and juveniles, even though they were often in close proximity. Order of arrivaL--Usually, Turkey Vultures King Vultures usually arrived in pairs or in arrived first at the carcasses,Black Vultures sec- what appeared to be family groups of 3, and ond, and Andean Condors last. Turkey Vul- landed in trees near the carcass.Although oc- tures arrived first at 92% of the 53 fresh car- casionallyas many as 8 King Vultures perched cassesfor which we knew the sequence of in surrounding trees, we never saw more than arrival. Black Vultures arrived second at 72% of 5 birds at a carcassat the same time during !7 the carcasses,and Andean Condors arrived third dayson which King Vultures were present.The at 70% of the carcasses. The order of arrival was proportion of encounterswon by an individual not determined solely by the relative abun- King Vulture varied with the magnitudeof the dance of the three species;Black Vultures out- differencein agesof the birds (Fig. 1). This pro- numberedTurkey Vultures by at least a 2-to-! duced an intraspecifichierarchy basedon age. ratio on the study area. The condor was the BecauseAndean Condors are sexually di- least common of the three species. morphic, we were able to record the outcome Interactionswithin each species.--Duringag- of interactions on the basis of sex as well as gressiveencounters between juvenile and adult age. The proportion of encounterswon by in- Turkey Vultures, 231 (68%) of 339 interactions dividuals of a certain age in encounters with were won by adults;the remaining encounters individuals of the samesex but a different age were won by juveniles. In contrast,juvenile increasedlinearly and leveled off when the dif- Black Vultures were much more submissive to ference between their ages was greater than 3 adult Black Vultures; adults won 528 (93%) of yr (Figs. 2 and 3). Both among males and fe- 567 interactionswith juveniles. We never saw males there was a dominance hierarchy based more than 3 caracaras feed at a carcass at one on age. time, and the threesomes behaved as if they We also observedaggressive interactions be- were mated pairs accompaniedby their off- tween condorsof different sexes(Fig. 4). Males

TABLE1. Number of interactionsobserved between speciesof avian scavengersin Peru and the proportion of interactionswon by eachspecies. Sample sizes are given in parentheses.

Percentageof interactionswon by speciesA when interactingwith: Vultur Sarcoramphus Coragyps Polyborus Cathartes SpeciesA gryphus papa atratus plancus aura Vultur gryphus -- 100%(185) 94% (818) -- (0) 100%(682) Sarcoramphuspapa 0% (185) -- 99% (2,142) -- (0) 100%(148) Coragypsatratus 6% (818) 1% (2,142) -- 17% (196) 44% (863) Polyborusplancus -- (0) -- (0) 83% (196) -- 91% (114) Cathartesaura 0% (682) 0% (148) 56% (863) 9% (114) -- April 1987] CompetitionAmong Avian Scavengers 293 generally dominated females, unless the fe- TABLE2. Ranks and body mass of members of an male was more than 1 yr older than the male. avian scavengingguild in Peru.

Interactions betweenspecies.--Juvenile and Rank adult caracaraswere generally dominant dur- in ing interactions with either Black Vultures or guild Average body Turkey Vultures (Table 1). Adult caracaraswon Speciesand classof hier- massand more encounters with Black Vultures than did individual archya range (kg)b juvenile caracaras(85% vs. 72%), but both adult Vultur gryphus and juvenile caracarashad high proportions of Adult males 1 12.5 (10.9-13.6) wins during encounterswith Turkey Vultures Older juvenile males 2 12.3 (10.9-13.6) (91% and 90%, respectively). Adult females 3 10.1 (9.6-11.4) Older juvenile females 4 10.2 (8.6-10.9) In one-on-one situations, Turkey Vultures Younger juvenile males 5 11.3 (10.9-11.6) won only slightly over half of their aggressive Younger juvenile encounterswith Black Vultures, demonstrating females 6 9.5 (9.1-10.2) very little dominance over Black Vultures (Ta- Sarcoramphuspapa ble 1). However, when Black Vultures outnum- Adults 7 3.4 (3.1-3.7) bered either Turkey Vultures or caracarasat a Older juveniles 8 3.4 (3.0-3.6) carcass,individuals of the latter two species Younger juveniles 9 3.1 (3.0-3.6) were always intimidated by the Black Vulture Polyborusplancus groups. Adults 10 1.6 (1.4-1.7) King Vultures, which weighed about twice Juveniles 11 1.6 (1.6) as much as Turkey and Black vultures (Table Cathartes aura 2), held a nearly completedominance over both Adults 12 1.4 (0.9-1.8) these vultures (Table 1). We never saw King Juveniles 13 1.3 (1.0-1.5) Vultures and caracarasinteract, although there Coragypsatratus were opportunities. At no time did we see a Adults 12 1.5 (1.2-1.8) King Vulture initiate an encounter with a con- Juveniles 13 1.3 (1.1-1.5) dor. Condors won all 185 aggressive encoun- • Basedon the win/loss data presentedin the text, Table 1, and Figs. 1-4. For groupsgiven the sameranking, there were insufficientobser- terswith King Vultures (Table 1). Condorswere vations to determine the precise relation of closely ranked birds. also dominant over the smaller vultures. bBased on measurementsof trapped individuals of each species. Dominancehierarchy within the guild.--Within each speciesthere was a positive relationship between a bird's rank and its age (Table 2). Male small- to medium-size, but not large, carcasses condorswere generally dominant over females by themselves.If they could get to smaller car- of the sameage. Between species,a bird's rank cassesbefore the other speciesarrived, they was correlated positively with its body mass could acquire a meal without interacting with (Spearman's rank correlation test, rs = 0.823; the more dominant guild members that arrive P < 0.001). later. Caracarasarrived at a carcassonly after it had been visited by other vultures. They usually DISCUSSION fed on carcassesonly when no other birds were present, often at daybreak before vultures had Organizationof the guild.--Turkey Vultures begun flying. When other scavengerswere have a well-developed olfactorysense that per- present, caracarasusually stayed one to several mits them to use odor as an aid in finding food meters away and picked up insects or small (Bang 1964, Stager 1964, Houston 1984). There scrapsof meat. Most aggressiveinteractions be- is no evidence that the other members of the tween caracaras and other species occurred scavenging guild possessthis ability. Further- when Black Vultures or Turkey Vultures at- more,Turkey Vultureshave lighter wing-load- tempted to supplant a caracarafrom a scrapof ing than the other avian scavengersin the study meat. area (Poole 1938, Brown and Areadon 1968). Although individual caracarasand, to a less- This facilitatesslow flight closeto the ground, er extent, individual Turkey Vultures were where the Turkey Vulture's olfactory abilities dominant over individual BlackVultures, they are mosteffective. Turkey Vultures can rip open were completely intimidated by Black Vulture 294 WALLACEAND TEMPLE [Auk,Vol. 104

TABLE3. Number of interactionsobserved between species of African scavengersand the proportionof interactionswon by eachspecies (based on Kruuk 1967,Anderson and Horwitz 1979).Sample sizes are given in parentheses.

Percentageof interactionswon by speciesA when interactingwith: SpeciesA Neophron Necrosyrtes Pseudogyps Gyps Trigonoceps Torgos Neophron -- 0% (4) -- (0) -- (0) 0% (I) -- (0) Necrosyrtes 100% (4) -- 0% (6) 0% (4) 0% (1) -- (0) Pseudogyps -- (0) 100%(6) -- 71% (80) 0% (1) 6% (50) Gyps -- (0) 100%(4) 29% (80) -- 0% (2) 16% (18) Trigonoceps 100%(1) 100%(1) 100%(1) 100%(2) -- 0% (15) Torgos -- (0) -- (0) 94% (50) 84% (18) 100%(15) -- groups,which often included over 50 birds. The food was consumed,and the carcasscould again largest Black Vulture group we saw included be defended by dominant individuals. Black 230-240 birds that arrived in less than 1 h. Vultures frequently fed closeto King Vultures Intact carcassesof large animals had such and condors,but were pecked constantlyand thick hides that the smaller scavengerscould displacedby the larger scavengers. gain accessonly through the anal and buccal King Vultures generally arrived in pairs or areas or places where lesions broke the skin. what appearedto be family groups.Aggressive When smaller birds congregatedat a large in- interactionswere initiated by adults when ju- tact carcass,this accessproblem permitted only veniles were present,but the young apparently 5 or 6 to feed simultaneously;subordinate in- were not their dependentoffspring. When an dividuals had to wait for dominant birds to be aggressorsupplanted its victim from a feeding sated.The waiting birds becamenoticeably ex- spot,the displacedKing Vulture usually relo- citedwhen the later-arrivingKing Vulturesand cated and resumed feeding promptly, but oc- Andean Condorsapproached the carcass.With casionally a victim counterattacked.We ob- their heavier beaks and greater strength, the servedfour suchfights between King Vultures; larger birds opened the carcassat severalpoints. each lasted only a few secondsbefore the ul- This often caused Black Vultures to swarm on timate loser was chased from the carcass area. the carcassin a "feeding frenzy." Normally Condors were usually last to arrive at a car- subordinate birds apparently associatedthe cass. Marked individuals waited hours or even presenceof condorsat a carcasswith greater daysafter they discovereda carcassbefore they access to the food and became too excited to be landed and approachedit. Their hesitancymay dominated by higher-ranking conspecifics. have been due to several factors. Condors can When a large number of subordinateindivid- easily go for several days between feedings, uals attacked the carcass at once, the resource unlike smaller vultures that feed more fre- temporarily became impossible to defend. quently (Calder 1984). Becauseit was difficult Within minutes,generally, the easilyaccessible for the larger condorsto regain flight when they were on the ground, particularly after a large meal, they possiblywere cautiousabout landing until confident they would not be dis- TABLE4. Rank and body massof speciesof an Af- turbed. Once condorsapproached the carcass, rican scavengerguild. their size and strength allowed them to easily Rank in Body displaceother membersof the guild. When guild mass condorsinteracted, males generally dominated Species hierarchya (kg)b females,probably becausethey weighed more Torgostracheliotus 1 7.5 (Fig. 4, Table 2). However, young male condors Trigonocepsoccipitalis 2 5.9 were generally subordinate to females that Gypsruppellii 3 6.4 were more than 1 yr older, even though a young Pseudogypsafricanus 4 5.7 Necrosyrtesrnonachus 5 1.9 male could weigh as much as a third more than Neophronpercnopterus 6 1.9 a female. Apparently there is a learned, social

Basedon Kruuk (1967) and Anderson and Horwitz (1979). componentto dominancein addition to the ad- From Brown (1971). vantageof greater size. April 1987] CompetitionAmong Avian Scavengers 295

Comparisonswith other scavengerguilds.- LITERATURE CITED Among Old World vultures, Kruuk (1967) and ANDERSON,D. J., & R. T. HORWITZ. 1979. Competi- Houston (1975) found that sympatric avian tive interactionsamong vultures and other avian scavengerson the African plains showeda high competitorsßIbis 121: 505-509ß degree of ecologicalsegregation. Competition ATTWELL,R. I.G. 1963. Some observations on feed- for food at carcasseswas reducedby the differ- ing habits,behaviour and inter-relationshipsof ential arrival of speciesat carcasses,the utili- northern Rhodesian vulturesß Ostrich 34: 235- zation of different parts of the carcass,and the 247. maintenance of a dominance hierarchy be- BANG,B.G. 1964. The nasalorgans of the Blackand tween specieswhen they were together at car- Turkey vulturesßJ. Morphol. 115: 153-184. casses. BROWN,L. 1971. African birds of prey. Boston, Houghton Mifflin Co. We compared the interspecific dominance •, & D. AMADON. 1968. Eagles,hawks and fal- hierarchies among Old and New World scav- cons of the worldß Feltham, England, Country engers. Combining species-interaction data Life Booksß from Kruuk (1967) and Anderson and Horwitz CALDER,W.A. 1984. Size, function, and life history. (1979), we calculatedthe proportion of encoun- Cambridge,Massachusetts, Harvard Univ. Press. ters won by one species in encounters with HECK,H. 1968. Aboutthe plumageof a captive-bred another species (Table 3). Comparison of the and raisedKing Vulture (Sarcoramphuspapa)ß Zoo African data (Table 4) with that from Peru (Ta- Gart. (NF) 35: 314. ble 2) revealed that in both guilds of avian HOUSTON,D. 1974. The role of griffon vultures Gyps africanusand Gypsruppellii as scavengersßJ. Zool. scavengersthe interspecific dominance hier- London 172: 35-46ß archy is based primarily on body size. There 1975. Ecologicalisolation of African scav- wasa strongcorrelation between rank and body enging birds. Ardea 63: 55-64ß massof African scavengers(Spearman's rank ß 1984. Does the King Vulture Sarcoramphus correlation test, rs = 0.900; P < 0.05), as there papause a senseof smell to locatefood? Ibis 126: was among Peruvian scavengers. 67-69ß There is remarkable convergencein the or- KOFORD,C. B. 1953. The California Condorß Natl. ganization of the guilds of avian scavengersin Audubon Soc.Res. Repts. No. 4. Africa (membersof the family Accipitridae) and KRUUK,H. 1967. Competition for food between vultures in East Africa. Ardea 55: 171-193. South America (mostly membersof the family MAURER,B. A. 1984. Interference and exploitation Cathartidae). Within both guilds there is dif- in bird communitiesß Wilson Bull. 96' 380-395ß ferential exploitation basedon the order of ar- PETRIDES,G.A. 1959. Competition for food between rival at carcasses and interference based on in- five speciesof EastAfrican vulturesßAuk 76: 104- tra- and interspecific dominance hierarchies 106. determined by size, sex, and age. POOLE,E.L. 1938. Weights and wing areasin North American birdsß Auk 55: 511-517.

ACKNOWLEDGMENTS STAGER,K. E. 1964. The role of olfaction in food location by the Turkey Vulture Cathartesaura. Our work in Peru was supportedby the U.S. Fish Los Angeles Co. Mus. Contrib. Sci. 81: 1-63. and Wildlife Service and the National Geographic STUART,P. 1978. Behavioral interactions and niche Society.We thank Carla Christianson,Wilman Torres separationin Blackand Turkey vulturesßLiving Arce, Leon Hecht, and Peter Schoonmakerfor help Bird 17: 79-84. in data collection and assistancethroughout the proj- WALLACE,M.P., & S. A. TEMPLE.1987. Impacts of ect. The Peruvian Ministerio de Agricultura, Depart- the 1982-83 El Nifio on populationdynamics of mento de Faunay Flora helped obtain the appropri- Andean Condors in Peru. Biotropica in pressß ate permits for our work.