ORNITOLOGIA NEOTROPICAL 13: 365–379, 2002© The Neotropical Ornithological Society

FORAGING ECOLOGY OF REINTRODUCED CAPTIVE-BRED SUBADULT HARPY EAGLES (HARPIA HARPYJA) ON BARRO

COLORADO ISLAND, PANAMA

Janeene M. Touchton, Yu-Cheng Hsu, & Alberto Palleroni

The Peregrine Fund, World Center for Birds of Prey, 566 West Flying Hawk Lane, Boise, Idaho 83709. E-mail: touchtonj@eudoramail.com

Resumen. – Dieta de Águilas Arpía (Arpia harpyja) reintroducidas en la isla de Barro Colorado,Panamá. – Un macho y una hembra de Águila Arpía (Harpia Harpyja), ambos sub-adultos criados encautividad y reintroducidos, fueron observados en la isla de Barro Colorado, Panamá, durante 89 y205 días, respectivamente, desde Junio de 1999 a Agosto del 2000. El macho capturó 25 presas denueve especies de mamíferos, y la hembra capturo 46 presas de 8 especies de mamíferos y una especiede reptil. El 52% de las capturas del macho fueron perezosos de dos dedos (Choloepuus hoffmanni) yde tres dedos (Bradypus variegatus). El 54% de las presas de la hembra fueron perezosos. Por términomedio, el macho capturó una presa cada 3,6 días, con un consumo medio diario de 888 g. Lahembra capturó, por término medio, una presa cada 4,4 días, con un consumo diario de 812 g. Laságuilas atacaron especies arbóreas solitarias a una distancia de 1–50 m, algunas veces tras variosintentos precalculados a unos 5 m. Especies arbóreas sociales fueron atacadas la mayoría de las vecespor sorpresa y a menos de 30 m, mientras que especies terrestres fueron atacadas por sorpresa y amenos de 10 m. El águila hembra fue observada capturando más especies arbóreas solitarias, demanera significante, durante tiempo soleado y durante la temporada seca. Del mismo modo, fue observadacapturando más especies arbóreas sociales durante tiempo nubloso y en la temporada de lluvias.Los promedios de capturas con éxito para el águila macho y el águila hembra fueron de 38% y 49%respectivamente.

Abstract. – A male and female Harpy Eagle (Harpia harpyja), both reintroduced captive-bred subadults,were observed on Barro Colorado Island (BCI), Panama for 89 and 205 days, respectively, betweenJune 1999 and August 2000. The male captured 25 individuals from nine different mammalianspecies and the female captured 46 individuals from 8 different mammalian and one reptilianspecies during the period of observation. Fifty-two percent of captures by the male were oftwo-toed (Choloepus hoffmanni) and three-toed (Bradypus variegatus) sloths. Fifty-four percent ofcaptures by the female were of sloths. On average, the male made a capture every 3.6 days witha daily average consumption of 888 g. The female made a capture on average every 4.4 days witha daily average consumption of 812 g. The eagles attacked solitary arboreal prey species from 1–50 mdistance, sometimes with several calculated attempts from within 5 m. Social arboreal prey specieswere most often attacked by surprise from less than 30 m, and terrestrial prey species were attackedby surprise from less than 10 m. The female eagle was observed to capture solitary arborealprey significantly more during sunny weather and the dry season. She was also observed to capturesocial arboreal prey significantly more during cloudy weather and the wet season. Capture successrates of observed predations for the male and female eagle were 38% and 49%, respectively. Accepted 14March 2002.

Key words: Harpy Eagle, Harpia harpyja, Barro Colorado Island, foraging ecology, predation.

365

TOUCHTON ET AL.

INTRODUCTION

The Harpy Eagle (Harpia harpyja) is a largepredator, member of a guild foraging prima-rily on mid-sized mammals in the forest can-opy, and considered the world's mostpowerful raptor (Brown & Amadon 1968).Harpy Eagles formerly inhabited tropical low-land rain forests from southern Mexico tonorthern Argentina (Wetmore 1965). Due tothe destruction and fragmentation of rainfor-ests and heavy hunting pressure by humans,the Harpy Eagle is presently rare or extinct inmuch of Central America north of Panama.The Harpy Eagle is currently considerednear-threatened throughout its current range(Collar et al. 1994).

Raptors have some of the lowest densitiesof all rainforest birds (Thiollay 1986). Pres-ence of Harpy Eagles has been said to indi-cate an intact ecosystem (Albuquerque 1995)as top predators are often among the firstspecies to disappear when pristine habitatundergoes human alteration or fragmentation(Leck 1979, Willis 1974, 1979a; Thiollay1985b, 1985c; Noss & Cooperrider 1994, Ter-borgh et al. 1997). More recent findings showhowever, that Harpy Eagles sometimes occurin forest near recently disturbed areas (Alva-rez-Cordero 1996). The generation of data onNeotropical raptor ecology and populationdensities is crucial to evaluate the impact offragmentation on raptors in tropical commu-nities (Thiollay 1980, 1985a, 1985b, 1985c;Whitacre & Thorstrom 1992). Terborgh(1988, 1990, 1991, 1992, 1997) has arguedthat top predators such as jaguars, pumas, andHarpy Eagles are keystone predators and thattheir removal from an ecosystem can causeprofound changes in Neotropical forest com-munities (See also Wright et al. 1994). It is notyet clear how important raptors are as regula-tors of herbivore populations.

All previous field observations of HarpyEagles have focused on nest behavior or brief

observations of a single behavioral incident.Few natural history descriptions (mostly anec-dotal or fragmentary) exist for the HarpyEagle (Bond 1927, Friedmann 1950, Goch-feld 1978) and sightings are scarce (Chebez etal. 1990, Albuquerque 1995, Galetti et al.1997). Predation records of the Harpy Eaglehave included two observed attempts (Eason1989, Peres 1990) and one post-predationencounter (Sherman 1991). Fowler & Cope(1964) initiated more detailed studies of theHarpy Eagle at two sites in British Guianathat described nests and provisioning behav-ior. Rettig (1977, 1978) continued studies atone of these nest sites and was able to providea more detailed account of breeding behavior.More recently, Alvarez-Cordero (1996) con-ducted a study that documented initial investi-gations of the current status of Harpy Eaglesand nesting home ranges for the Harpy Eaglein Venezuela and Panama. Despite these stud-ies, Harpy Eagle foraging behavior had notbeen observed. Due to the potentially impor-tant role of this predator in tropical commu-nity dynamics, studies of the Harpy Eagleforaging behavior merit further attention.

The release of five radio-tagged captive-bred juvenile Harpy Eagles into SoberaniaNational Park, Panama, commenced in Janu-ary 1998 with a male and female. The thirdrelease was a male in May 1998, the fourth afemale in August 1998, and the fifth a male inSeptember 1998. The third and forth of thesereleased eagles were transported to BarroColorado Island in June 1999 and October1999, and became the focus of this study. As aresult of the ease and finesse with which theseeagles captured their first prey in SoberaniaNational Park without any human or parentalguidance, we have made the assumption thesesubadult Harpy Eagles would exhibit thesame behavioral ontogeny as wild HarpyEagles.

With the reintroduction of these twoeagles onto Barro Colorado Island we had the

366

SUBADULT HARPY EAGLE FORAGING ECOLOGY

rare opportunity to collect extensive first-hand observations of foraging ecology andbehavior in a natural setting where the lastHarpy Eagle sighting was in 1950 (Willis &Eisenmann 1979b). Objectives of this studywere to collect numbers and species of preytaken by these subadult eagles, hunting meth-odology used, prey response, and how thesefindings could impact the prey communitydynamics on Barro Colorado Island.

STUDY AREA AND METHODS

This study was conducted on Barro ColoradoIsland (BCI), Panama (9º9'N, 79º51'W), aprotected 1500-ha island which rises 137 mabove Lake Gatun in the Panama Canal. BCIis covered by primary and secondary lowlandtropical moist forest (Croat 1978, Windsor1990). Annual rainfall averages 2638 mm andtemperature averages 26°C. BCI experiencesa pronounced wet season which begins inApril or May and usually ends in Decemberfollowed by a dry season ending in April. Thedry season averages only 293 mm of rainfallper year (Paton per. com.). Fruitfall peaksbetween March and June, and between Sep-tember and October (Foster 1982a), therebyproducing seasonal changes in food availabil-ity for frugivorous mammal populations.Mammal descriptions for BCI may be foundin Leigh et al. (1982). Extensive accounts ofthe ecology and history of BCI were pro-vided by Croat (1978), Leigh et al. (1982),Gentry (1990), and Leigh (1999).

The subjects of this study were two cap-tive-bred Harpy Eagles (a two-year old maleand female) hatched and reared at the Zoo-logical Society of San Diego, California andtransferred to the World Center for Birds ofPrey, Boise, Idaho at 118 and 87 days of age,respectively. At the World Center for Birds ofPrey they were puppet-provisioned andplaced in an enclosure with an adult femaleHarpy Eagle. Near fledging age (165 and 161

days), the eagles were transferred separatelyto an enclosure in Soberania National Park,Panama in 1998 where they were habituatedto the area during four and five weeks priorto being released. Eagles were supplied withdead native mammalian prey until they ceasedto visit the provisioning site (11 months). Themale eagle did not stray far from the releasesite. Due to security concerns, he was shortlytaken back into captivity until his release ontoBCI on 16 June 1999 at 20 months of age.The female took long dispersals in intervalsaway from the release site starting at 13months, traveling along a drainage system.She was captured 5.5 km from the release sitewhen 20 months old and re-located to BCIwith the male on 10 October 1999. Eagleswere equipped with both Biotrack backpackharness mount and Merlin tail feather mountVHF radio transmitters operating in the 216MHz range. Transmitter mass was 90 g.

The male was observed for 623 h on BCIduring 89 days from 16 June to 11 October1999. The female was observed for 1204 h onBCI during 205 days between 01 January and24 August, 2000. Eagles were located on adaily basis by telemetry using a Telonics TR-2or Wildlife Materials TRX-1000 receiver and3-element hand-held yagi antennae. Onceeagles were located, observations on eaglebehavior, and the behavior of other animalspecies within sight, earshot, or vision wasrecorded. Additional data were collected onmap position and weather. Tracking andobservations made on the eagles occurredfrom dawn to dusk. Observations were takenwhen activity changed, or every 5 min. Dur-ing predation events, data were also recordedon estimated weight and sex of prey species,and the method of attack the eagles used tocapture their prey. Observations were madeeither unassisted or with 10x25 or 8x32 bin-oculars. Observers took care to avoid disturb-ing the eagles and remained as distant aspossible consistent with accurate observa-

367

368

TOU

CHTO

N E

T AL.

TABLE 1. Number of captures of each prey species, the percentage of total captures consumed of each prey group, and percent of total mass consumed foreach species during the observation period of 89 days for the male eagle and 205 days for the female eagle.

Female predations

ass N %N % total massMar

CXen

B H

PrimM

WIgua

GCarn

W

KArti

R

C

RodMC

TOT

1

5191

2113

2

1

1

46

2.2

10.941.32.2

4.423.96.5

4.4

2.2

2.2

1.1

340.82.6

2.436.14.5

5.4

1.8

2.4

1J =

Male predations

Age1 Ave. weight (kg) N %N % total msupialaommon opossum (Didelphis marsupialis)arthrarown throated three-toed sloth (Bradypus variegates)

offman's two-toed sloth (Choloepus hoffmanni)atesantled howler monkey (Alouatta palliate)

hite-faced capuchin (Cebus capucinus)nidaereen iguana (Iguana iguana)Ìvorahite nosed coati mundi (Nasua narica)

inkajou (Potos flavus)odactylaed-brocket deer (Mazama americana)

ollared peccary (Tayassu tajacu)

entiaexican porcupine (Coendou mexicanus)entral American agouti (Dasyprocta punctata)

AL 12

A

JAA

JAA

A

JA

JAJA

AA

1.8

0.3-2.03.2-4.0

4.3

25.52.5

4.5

133

428

1.5-4.012.0-20.0

2.0-4.02.0-4.0

1210

3

3

1

1

2

11

25

4840

12

12

4

4

8

44

1.39.556.4

7.9

3.9

3.9

5.3

3.9

3.93.9

juvenile; A = adult.

SUBADULT HARPY EAGLE FORAGING ECOLOGY

tion. The male was observed throughout hisentire foraging cycle (from prey capturethrough to next prey capture), includingwhen he was eating. For the first threemonths of this study the female was observedin the same manner to document feedingbehavior. Afterwards, she was left alone to eatwhen she had captured prey and was only

checked on momentarily until she finishedher carcass.

RESULTS

Prey species captured and foraging rates. During89 days of observation the male captured25 prey individuals, 52% of which were

FIG. 1. Prey composition comparison for the eagles over 89 days observation time in the wet season.

369

370

TOU

CHTO

N E

T AL.

TABLE 2. Prey preference determined by calculating expected prey numbers form prey abundance estimates (Giacalone & Willis pers. com). Actual preydata from 89 days of observation of the male eagle and 205 days of observation for the female eagle.

Female

Expected (N) c21 P

ComBrow

HofMan

WhiGreWhi

KinRed

Coll

MexCen

3.13.313.34.40.55.31.313.3

1.3

0.0

1.10.36.32.10.92.02.38.3

0.4

1.0

0.300.560.010.140.340.160.130.004

0.84

0.32

1J =

Age1 Male

Actual (N) Expected (N) c21 P Actual (N)

mon opossum n throated three-toed sloth

fman's two-toed slothtled howler monkey

te-faced capuchinen iguanate nosed coati mundi

kajou-brocket deer

ared peccary

ican porcupinetral american agouti

AJAAJAAAJAAJAJAAA

12103

3

11

2

11

2.610.43.50.4

0.1

0.90.0

0.1

1.75.2

0.75.73.12.0

2.7

0.01.0

1.7

0.22.8

0.400.020.080.16

0.10

0.940.32

0.19

0.670.09

1519121132

1

1

juvenile; A = adult.

SUBADULT HARPY EAGLE FORAGING ECOLOGY

two-toed (Choloepus hoffmanni) or three-toed(Bradypus variegatus) sloths (Table 1). During205 days of observation the female captured46 prey individuals, 54% of which were two-toed or three-toed sloths (Table 1). Actualisland abundance values of prey species (Gia-calone & Willis pers. com.) were used todetermine prey preference. Both eaglesshowed a strong preference for three-toedsloths. The male eagle also showed a prefer-ence towards two-toed sloths (Table 2).

An 89-day comparison between the maleand female during the rainy season showedthat the male preyed upon nine species, whilethe female preyed on four (c2

1 = 1.92, P >

0.05) (Fig. 1). Forty percent of the individualsthe male captured and 17% of the individualsthe female captured during this period werejuveniles. This difference, however, is not sig-nificant (c2

1 = 2.57, P > 0.05). During this

time, the male captured significantly moreterrestrial prey than the female (c2

1 = 4.5, P <0.05).

The male captured one prey item per 3.7days (SD = 1.9, n = 24). The two juvenilepeccaries captured by the male were consid-ered one item for this calculation as they werecaptured at the same time, one in each foot,after their mother ran off. The female cap-tured one prey item per 4.4 days (SD = 2.2, n= 46). These foraging rate differences are notsignificant (U = 463, df = 1, P > 0.05). Themaximum number of days between success-ful attacks was 8 for the male, and 9 for thefemale. A significant positive correlation wasfound for the female eagle between captureday intervals and the weight of the animalcaptured (r2 = 0.14, P = 0.01). The male didnot show the same positive correlation (r2 =0.07, P = 0.26).

Based on entire carcass mass estimates,the male consumed a daily average of 888 g,which amounted to roughly his entire bodymass (6.0 kg) per week. The female con-sumed a daily average of 812 g, eating roughly

70% of her body mass (8.0 kg) per week.These consumption rates are significantly dif-ferent (c2

1 = 8.67, P < 0.003).According to recent population estimates

from island census work of Harpy Eagle preyspecies (Giacalone & Willis, pers. com.), com-bined data of the male and female suggestthat a pair of Harpy Eagles would annuallyconsume 1.4% of BCI’s standing crop ofavailable Harpy Eagle prey. Extrapolatingfrom these data two-toed sloths lose a largerproportion of their population than three-toed sloths, due to the male Harpy Eagle’sdepredations. Adult primate prey wouldexperience a projected annual loss less than2% (Table 3).

Foraging classes and success rates. We separatedHarpy Eagle prey species into three generalclasses: social arboreal, solitary arboreal, andterrestrial prey. Species within each class dis-play similar patterns in daily activity, move-ment, and location within the forest. Thefollowing dry season data pertain only toobservations of the female eagle as the malewas only observed during the wet season.

Social arboreal prey included howlermonkeys, white-faced capuchins (Cebus capuci-nus), and tamarins (Saguinus geoffroyi). Out of30 observed predation attempts on socialarboreal prey by the eagles (seven of whichwere successful), all but three were aban-doned after the first attack if unsuccessful.The longest sequence of attacks on socialarboreal prey was by the male eagle on ahowler monkey, totaling four attempts lastingfrom 14:41 to 16:11 and all unsuccessful.Prior to this sequence of attempts, only onehowler monkey had successfully been cap-tured on BCI. Predation attacks by the eagleswere made within 30 m, horizontally or at adownward angle, attempting to surprise theprey within the canopy. Successful captureswere made by the eagle grabbing the preyindividual out of the tree and either flying to

371

TOUCHTON ET AL.

the ground with it, or dropping it. In all cases,the eagles would pierce and squeeze their preyuntil movement ceased. None of the attemptson tamarins, all made by the male eagle, weresuccessful.

Solitary arboreal prey species consistedprimarily of three-toed and two-toed sloths. Atotal of 34 attacks were observed, 19 of whichwere successful. Ten out of 11 two-toed slothpredations were by the male eagle. Attackswere made from one to 50 m away. An eaglecould attack the same sloth several times overa period of many hours. The longest observedsequence of attacks on solitary arboreal preyspecies was by the male eagle: it lasted nearlynine h (08:39 to 17:25). To attack a three-toedsloth, the eagle would fly in or walk in on abranch and grab the sloth on top of theirbranch, often in a tree crown. They would

then squeeze the sloth in the same positioncaptured until movement of the sloth ceased,or they would move to a nearby branch withthe sloth if unstable. In the case of two-toedsloths, who generally hung underneath abranch, the eagles would fly in underneath thesloth, invert in the air, and attempt to pull thesloth off the branch. If successful, the eaglewould fly down to the ground while still hold-ing onto the sloth, or drop the sloth to theground.

Five observations were made of attackson terrestrial prey species, four of which weresuccessful. All terrestrial prey were either cap-tured successfully or abandoned after oneattempt by the eagles aside from an attemptby the male on a Mexican porcupine (Coendoumexicanus). Four attempts were made from12:54 to 14:51 before the porcupine was suc-

TABLE 3. Annual estimates for captures made by both eagles and estimated percentage of BCI prey spe-cies standing crops, based on BCI census data in 2000 (Giacalone & Willis pers. com.).

Age1 BCI population

estimate2 (N)

Male (N)

Female (N)

Male +

Female

Male + Female % of BCI population

Common opossomBrown throated three-toed sloth

Hoffman's two-toed slothMantled howler monkey

White-faced capuchinGreen iguanaWhite nosed coati mundi

KinkajouRed-brocket deer

Collared peccary

Mexican porcupineCentral American agouti

TOTAL 12

AJAAJAAAJAAJAJAAA

7007503000100012012003003000303002501075201205001500

12875

4.18.240.912.3

12.3

4.14.1

8.2

4.14.1

102.3

1.88.933.71.83.619.55.33.6

1.8

1.8

81.7

1.813.041.942.715.819.55.33.612.31.84.15.9

8.2

4.14.1

183.9

0.31.71.44.313.21.61.80.140.90.61.658.7

40.9

0.80.3

1.43

372

SUBADULT HARPY EAGLE FORAGING ECOLOGY

cessfully captured.Seasonal and weather patterns affected

foraging success of the female eagle differ-ently for each prey class. Foraging success isdefined as the number of prey successfullycaptured. The female eagle was observedattempting predations for all three classesduring both the wet and dry seasons. How-ever, successful captures of social arborealprey occurred significantly more often in thewet season and solitary arboreal prey signifi-cantly more in the dry season (Table 4). Dur-ing observation weather was categorized assunny, cloudy (including times of light rain),or heavily raining. No eagle activity wasobserved during heavy rain. Social arborealprey were captured more during times ofcloudy weather, whereas solitary arborealprey were captured far more during times ofsunny weather (Table 5).

Observed successful versus unsuccessfulcapture attempts made on social arborealprey occurred more during the afternoonhours of the day for both eagles while capture

attempts on solitary arboreal prey occurredmore during the noon hours for the female(Table 6). For all observed capture attemptsmade by both eagles, the female exhibited ahigher success rate than the male (Table 7)yet not significantly higher (c2

1 = 0.43, P

< 0.51). No significant change in successrates were observed for either eagle fromthe first third to the last third of study.Success rate is defined as the ratio ofobserved successful captures versus observedtotal number of attempts.

DISCUSSION

Prey captured and foraging rates. Differences inprey choice observed between the male andfemale Harpy Eagle could result from indi-vidual differences, different levels of develop-ment, or gender-specific differences relatedto morphology. Prey diversity collected for amale Harpy Eagle provisioning a nest (Rettig1978) are similar to this study’s observationson the subadult male except for the Guyanamale’s captures of Cebus sp. which couldreflect an adult’s greater hunting experience.Birds of prey often display gender specificforaging behavior and niche partitioning(Snyder & Wiley 1976). Male raptors, typicallysmaller than females, are thought to be of asize that allows them to catch the most avail-able sizes and types of prey (Cade 1982).Male Harpy Eagles may capture quicker spe-cies like Cebus sp., and animals in less accessi-ble places such as sloths hidden in lianas.

TABLE 4. Successful observed captures of each prey class by the female eagle in 65 days in the wet and dryseasons (n = 34).

N - wet season1 N - dry season1 c21 P

Social arborealSolitary arborealTerrestrial

1151

1142

8.34.30.3

0.0040.040.56

TABLE 5. Successful observed captures of eachprey class by the female eagle in different types ofweather (n = 17).

Cloudy1 Sunny c21 P

Social arborealSolitary arborealTerrestrial

41

0110

4.08.31.0

< 0.05< 0.004< 0.32

1Over 65 days observation.

1Includes light rain.

373

TOUCHTON ET AL.

Female Harpy Eagles may capture bigger ani-mals such as male howler monkeys. Thesegender differences indirectly may reduce pres-sure on preferred prey within a Harpy Eaglehome range and provision a chick with agreater diversity of food items (White & Cade1971). Moreover, having a broader range ofprey species reduces vulnerability from popu-lation fluctuations in any one prey species(Foster 1982b, Glanz 1982, Milton 1982, Rus-sell 1982, Giacalone pers. com.).

Both eagles displayed great variability inpredation rates that was consistent with ratesof adult breeding birds in Guyana (Rettigpers. com.), suggesting that success rates maybe little affected by age and experience andthat hunting behavior is innate. Alternately,this study may have not been long enough toobserve changes in predation success rates ofthe eagles as they grew older. At this point wemay not ascertain that any other factors aside

from prey size affect predation rates.

Effects on prey populations. Yearly consumptionestimates for the male and female eagle of thisstudy are 364 kg-1 year and 296 kg-1 year,respectively. This is roughly consistent withCrowned Hawk Eagles (Stephanoaetus corona-tus), primate hunters of the Kibale nationalPark, Uganda, that reside in a home rangesimilar to some Harpy Eagles (3.8 km2)(Mitani et al. 2001). Nutrition requirementsestimate that Crowned Hawk Eagles wouldconsume approximately 430 kg-1 year (Brownet. al. 1982). An adult Verreaux’s Eagle (Aguilaverreauxi), another large eagle of Africa knownto take prey between 1.8 and 5.5 kg (Kingdon1997) is estimated to consume approximately120 kg-1 year (Gargett 1993). Emmons (1987)estimated food consumption for a wild ocelot(Felis pardalis), (approximately the same weightas a female Harpy Eagle), to be 175–263 kg-1

TABLE 6. Observed capture attempts of each prey class made by both eagles during different times of theday.

06:00–10:00 10:00–14:00 14:00–18:00 c22 P

Male eagle (n = 34) Social arboreal Solitary arboreal Terrestrial

Female eagle (n = 35) Social arboreal Solitary arboreal Terrestrial

111

020

651

5150

1162

751

8.33.50.5

6.523.42.0

0.020.170.77

0.040.0010.37

TABLE 7. Success rates of both eagles for observed capture attempts made on each prey class.

Social arboreal Solitary arboreal Terrestrial TotalMale eagle

Total individuals attacked % Successful captures

Female eagle Total individuals attacked % Successful captures

1816.7

1233.3

1258.3

2254.5

475.0

1100.0

3438.2

3548.6

374

SUBADULT HARPY EAGLE FORAGING ECOLOGY

year. Latest observations have characterizedup to 30 potential resident ocelots on BCI(Giacalone & Willis pers. com.). Despite theslightly lower daily consumption rate of oce-lots, given these estimations, they must have afar greater impact on overlapping prey popu-lations than Harpy Eagles given their homerange.

Because Harpy Eagles normally reside inpairs within home ranges comparable to theentire size of BCI (Rettig per. com., Alvarez-Cordero 1996), we believe it is reasonable toconsider the male and female eagle of thisstudy equivalent to a territorial pair whenconsidering their impact on prey populations.Combining data collected on the male andfemale eagle from this study provides roughestimates of prey compositions. We do notknow whether male eagles have seasonalshifts in hunting preferences, as our data forthat individual only encompassed wet seasoncaptures. Therefore, possible errors areinflated in the numbers of Table 3. Althoughthe prey population numbers given for BCIare based on latest census data, populationsof several species such as red brocket deer,peccary (Tayassu tajacu), coati (Nasua narica),and agouti fluctuate greatly from year-to-year,and others, such as sloths, have never beenadequately surveyed. Despite the room forerror, we feel that this exercise is empiricallyuseful for gauging the impact of a pair ofHarpy Eagles on prey populations.

The estimated annual predation for two-toed sloths is the only adult population with agreater loss than 2% of their numbers. Ifmale Harpy Eagles consume fewer two-toedsloths per month in the dry season than inthe rainy season, this estimated loss would belower.

The highest estimated losses were juve-nile red brocket deer, juvenile coati, and juve-nile collared peccary. Juveniles oftenexperience high mortality rates and observefluctuating population numbers. For example,

the number of juvenile coatis born each year,and the proportion surviving varies accordingto the amount of fruit available the precedingOctober (Russell 1982). Juvenile howler mon-keys experience a minimum of 39% annualmortality rate from starvation, falls, or para-sites (Milton 1982). It is therefore difficult todetermine to what degree the predation onjuveniles made by Harpy Eagles may affectoverall population dynamics.

The impact on the collared peccary, how-ever, is likely grossly overestimated as thispredation was due to observer influence. Weincluded these data to demonstrate that theeagles may take advantage of a situation suchas a mother peccary being scared away fromher young due to a large cat, for example.Normally, observation influence was minimalas a majority of the mammals inhabiting BCIare habituated to human presence.

Factors affecting foraging methods and trends. Bothchanges in weather and seasonal patternsaffect location and anti-predator behavior ofpotential prey. The location and anti-predatorbehavior of potential prey influence aneagles’ ability to capture them, as was seenwith the female eagle. We assume that thecombination of these factors, in addition toprey abundance, influence overall foragingsuccess and therefore foraging patterns ofHarpy Eagles.

The location of prey animals (influencedby their behavior) greatly affects the ability ofthe eagles to find them. Social arboreal ani-mals are located visually, acoustically, or bychance encounter. Eagles may see howlermonkeys moving through the canopy insearch of fruiting trees, or sleeping in a groupin the sunshine. They may also be heardwhen chorusing at dusk and dawn. Eaglesattacked monkeys resting, grooming, or for-aging. Under attack the howler monkeysalarm called and ran toward the trunks oftrees while large males ran towards the eagles

375

TOUCHTON ET AL.

swinging their forelimbs. Capuchins droppedout of the tree and fled the area on theground. Tamarins would run to the ground,wait, and then return.

Because most solitary arboreal animalsmove and vocalize little, the eagles sight themfrom above the upper canopy. If a sloth isobscured by dense foliage and lianas theeagles may have a more difficult time findingthem. Both species of sloth exhibited formi-dable anti-predator strategies. Aside from hid-ing in inaccessible areas, they growled andswung fore-claws at the eagles. No evidencewas ever observed that the eagles experiencedany harm by prey individuals.

Physiological condition of prey plays arole in their behavior and therefore vulnera-bility to predation by Harpy Eagles. Three-toed sloths drop their body temperaturenightly to ambient temperature and raise itcloser to normal mammalian temperature bybasking in sunlight during the day (Montgom-ery & Sunquist 1978, McNab 1978). Becausethey usually thermoregulate by basking in theupper canopy, during times of bright sunshinethree-toed sloths are more easily spotted byHarpy Eagles. Two-toed sloths are nocturnaland maintain a more constant body tempera-ture in lowland forests than three-toed sloths(Montgomery & Sunquist 1978). Therefore,they are often found sleeping in masses oflianas or hanging from lower branches cov-ered by vegetation during the day, makingthem harder to see during times of bright sun-shine than three-toed sloths. In this study,more attempts were made by the female eagleat capturing sloths during the noon hours,when they would be more likely to be baskingin the sunlight of the warmest part of the day.

Weather may further influence preybehavior and affect the likelihood of HarpyEagle predation. In this study, the femaleeagle captured more howler monkeys duringcloudy weather. Howler monkeys not onlyvocalize at dusk and dawn, but also when rain

begins and sometimes after, increasing theopportunity for eagles to locate howler troopsby sound.

Although certain weather patterns aremore prominent in the wet or dry season, for-aging differences observed for the femaleeagle in each season may be a result of anoverall seasonal effect. The female eagle'sgreater success in capturing primates in thewet season might be attributed to a weakenedstate of the howler monkeys during this time.Mid-late rainy season (Aug.–Nov.) is the timewhen higher quality food for howler monkeyson BCI is most scarce (Milton 1982). Duringthis time howler monkeys often experiencedietary stress and may therefore be more vul-nerable to Harpy Eagle predation. Similarly,puma (Felis concolor) were found to have ahigher frequency and biomass of primates intheir diet [howler monkey, spider monkey(Ateles geoffroyi), and white-faced capuchin]during the wet season in Corcovado NationalPark, Costa Rica (Chinchilla 1997).

The ease of encountering prey by theseHarpy Eagles was influenced by weather andseasonal patterns that affected prey speciesbehavior due to physiology. The change inprey behavior during different weather classesand seasons not only affected the eagles' abil-ity to encounter prey, but also the success ofthe eagles’ predation efforts. Prey encounterrates by the eagles also varied according toprey species abundance. Foraging patterns bythese eagles were therefore indirectly ordirectly affected by weather and seasonal pat-terns, and prey species abundance.

Conclusions. Our observations of Harpy Eaglescapturing prey have provided several newinsights to the foraging ecology of this toppredator. Gender-specific differences in preyselection were observed as the male took amore diverse set of prey species, includingmore juvenile and terrestrial prey animals,than the female. We were able to measure for-

376

SUBADULT HARPY EAGLE FORAGING ECOLOGY

aging rates and consumption rates to estimatethe potential impact Harpy Eagles may haveon prey populations in comparison to otherpredators in the same area, such as the ocelot.

Seasonal and weather patterns that mayinfluence Harpy Eagle foraging behaviorwere also discovered. The female eagle wasable to capture solitary arboreal prey signifi-cantly more successfully during sunnyweather and the dry season, and social arbo-real prey significantly more during cloudyweather and the wet season.

Future studies of Harpy Eagles shouldinclude the observation of an increased sam-ple size of both sexes at the same time of yearto eliminate individual differences whenexamining gender differences, weather influ-ences, and seasonal differences. Additionally,the developmental foraging changes over alonger period of time into adulthood shouldbe studied. This study, in addition to theseproposed studies, will provide the knowledgenecessary to determine how great a role theHarpy Eagle plays in tropical forest dynamics,as well as contribute to the conservation ofthis rare species.

ACKNOWLEDGMENTS

This study was supported by The PeregrineFund. Permission was granted by the Smith-sonian Tropical Research Institute for thereintroduction of Harpy Eagles onto BCI.Field and technical support were provided byAngel Muela, Francisco Barrios, Kathia Her-rera, and Eduardo Santamaria. Jackie Giaca-lone and Greg Willis kindly provided recentmammal census data. The Smithsonian Trop-ical Research Institute’s Terrestrial-Environ-mental Sciences Program (T-ESP) providedweather data. The manuscript was improvedby comments from Lloyd Kiff, Egbert LeighJr., Jean Marc Thiollay, Richard Watson, andan anonymous reviewer. We extend our grati-tude to all of these individuals and institu-

tions for their assistance.

REFERENCES

Albuquerque, J. L. B. 1995. Observations of rareraptors in southern Atlantic rainforest of Bra-zil. J. Field Ornithol. 66: 363–369.

Alvarez-Cordero, E. 1996. Biology and conserva-tion of the Harpy Eagle in Venezuela and Pan-ama. Ph.D. diss., Univ. of Florida, Gainesville,Florida.

Bond, J. 1927. Nesting of the Harpy Eagle (Thrasa-etus harpyia). Auk 44: 562–563.

Brown, C. B. 1876. Canoe and camp life in BritishGuiana. Volume 1. E. Stanford, London, UK.

Brown, L., & D. Amadon. 1968. Eagle, hawks andfalcons of the world. McGraw-Hill, New York,New York.

Brown, L. B., E. K. Urban, & K. Newman. 1982.The birds of Africa. Volume 1. Academic PressInc., New York, New York.

Cade, T. J. 1982. The falcons of the world. CornellUniv. Press, Ithaca, New York.

Chebez, J. C., M. C. Silva, A. Serret, & A. Taborda.1990. La nidificación de la harpia (Harpia har-pyja) en Argentina. Hornero 13: 155–158.

Chinchilla, F. A. 1997. La dieta del aguar (Pantheraonca), el puma (Felis concolor) y el manigordo(Felis pardalis) (Carnivora: Felidae) en el ParqueNacional Corcovado, Costa Rica. Rev. Biol.Trop. 45: 1223–1229.

Collar, N. J., M. J. Crosby, & A. J. Stattersfield.1994. Birds to watch 2: The world list ofthreatened birds. BirdLife International,Smithsonian Institution Press, Washington,DC.

Croat, T. B. 1978. Flora of Barro Colorado Island.Stanford Univ. Press, Stanford, California.

Eason, P. 1989. Harpy eagle attempts predation onadult howler monkey. Condor 91: 469–470.

Emmons, L. H. 1987. Comparative feeding ecol-ogy of felids in a Neotropical rainforest. Behav.Ecol. Sociobiol. 20: 271–283.

Foster, R. B. 1982a. Seasonal rhythm of fruitfall onBarro Colorado Island. Pp. 151–172 in LeighJr., E. G., A. S. Rand, & D. M. Windsor (eds.).Ecology of a tropical forest. Smithsonian Insti-tution Press, Washington, DC.

377

TOUCHTON ET AL.

Foster, R. B. 1982b. Famine on Barro ColoradoIsland. Pp. 201–211 in Leigh Jr., E. G., A. S.Rand, & D. M. Windsor (eds.). Ecology of atropical forest. Smithsonian Institution Press,Washington, DC.

Fowler, J. M., & J. B. Cope. 1964. Notes on theHarpy eagle in British Guiana. Auk 81: 257–273.

Friedmann, H. 1950. The birds of North and Mid-dle America. U.S. Natl. Mus. Bull. 50, pt. 11,Washington, DC.

Galetti, M., P. Martuscelli, M. A. Pizo, & I. Simão.1997. Records of Harpy and Crested Eagles inthe Brazilian Atlantic forest. Bull. Br. Ornithol.Club 117: 27–31.

Gargett, V. 1993. The Black Eagle: Verreaux’sEagle in South Africa. Academic Press, SanDiego, California.

Gentry, A. H. 1990. Four Neotropical rainforests.Yale Univ. Press, New Haven, Connecticut.

Glanz, W. E. 1982. The terrestrial mammal faunaof Barro Colorado Island: censuses and long-term changes. Pp. 455–468 in Leigh Jr., E. G.,A. S. Rand, & D. M. Windsor (eds.). Ecology ofa tropical forest. Smithsonian Institution Press,Washington, DC.

Gochfeld, M., & M. Kleinbaum. 1978. Observa-tions on the behavior and vocalizations of apair of wild Harpy Eagles. Auk 95: 192–194.

Leck, C. F. 1979. Avian extinction in an isolatedtropical wet forest preserve, Ecuador. Auk 96:343–352.

Leigh, E. G., Jr., A. S. Rand, & D. M. Windsor.1982. The ecology of a tropical forest. Smith-sonian Institution Press, Washington, DC.

Leigh, E. G., Jr. 1999. Tropical forest ecology: Aview from Barro Colorado Island. OxfordUniv. Press, New York, New York.

Kingdon, J. 1997. The Kingdon field guide to Afri-can mammals. Academic Press, London, UK.

McNab, B. K. 1978. Energetics of arboreal foli-vores: physiological problems and ecologicalconsequences of feeding on ubiquitous foodsupply. Pp. 153–162 in Montgomery, G. G.(ed.). The ecology of arboreal folivores. Smith-sonian Institution Press, Washington, DC.

Milton, K. 1982. Dietary quality and demographicregulation in a howler monkey population. Pp.273–289 in Leigh Jr., E. G., A. S. Rand, & D. M.

Windsor (eds.). Ecology of a tropical forest.Smithsonian Institution Press, Washington,DC.

Mitani, J. C., S. J. Sanders, J. S. Lwanga & T. L.Windfelder. 2001. Predatory behavior ofCrowned Hawk-eagles (Stephanoactus coronatus)in Kibale National Park, Uganda. Behav. Ecol.Sociobiol. 49: 187–195.

Montgomery, G. G., & M. E. Sunquist. 1978. Hab-itat selection and use by two-toed and three-toed sloths. Pp. 329–359 in Montgomery, G. G.(ed.). The ecology of arboreal folivores. Smith-sonian Institution Press, Washington, DC.

Noss, R. F., & A. Y. Cooperrider. 1994. Savingnature’s legacy: protecting and restoring biodi-versity. Island Press, Washington, DC.

Peres, C. A. 1990. A Harpy Eagle successfully cap-tures and adult male red howler monkey. Wil-son Bull. 102: 560–561.

Rettig, N. 1977. In quest of the snatcher. Audubon79(6): 26–49.

Rettig, N.L. 1978. Breeding behavior of the HarpyEagle (Harpyia harpyia). Auk 95: 629–643.

Russell, J. K. 1982. Timing of reproduction by coa-tis (Nasua narica) in relation to fluctuations infood resources. Pp. 413–431 in Leigh Jr., E. G.,A. S. Rand, & D. M. Windsor (eds.). Ecology ofa tropical forest. Smithsonian Institution Press,Washington, DC.

Sherman, P. T. 1991. Harpy eagle predation on ared howler monkey. Folia Primatol. 56: 53–56.

Snyder, N. F. R., & J. W. Wiley. 1976. Sexual sizedimorphism in hawks and owls of NorthAmerica. Ornithol. Monogr. 20: 1–96.

Terborgh, J. 1988. The big things that run theworld – A sequel to E. O. Wilson. Conserv.Biol. 2: 402–403.

Terborgh, J. 1990. The role of felid predators inNeotropical forests. Vida Silvestre Neotrop. 2:3–5.

Terborgh, J. 1991. An overview of research atCocha Cashu Biological Station. Pp. 48–59 inGentry, A. H. (ed.). Four Neotropical rainfor-ests. Yale Univ. Press, New Haven, Connecti-cut.

Terborgh, J. 1992. Maintenance of diversity in trop-ical forests. Biotropica 24: 283–292.

Terborgh, J., L. Lopez, J. Tello, D. Yu, & A. R.Bruni. 1997. Transitory states in relaxing eco-

378

SUBADULT HARPY EAGLE FORAGING ECOLOGY

systems of land bridge islands. Pp. 256–274 inLaurance, W. F. & R. O. Bierregaard (eds.).Tropical forest remnants. Univ. of ChicagoPress, Chicago, Illinois.

Thiollay, J. M. 1980. Stratégies d’exploitation parles rapaces d’un écosystème herbacé Néotropi-cal. Alauda 48: 221–253.

Thiollay, J. M. 1985a. Falconiforms of tropicalrainforests: a review. Pp. 155–165 in Newton,I., & R. D. Chancellor (eds.). Conservationstudies on raptors. ICBP Technical PublicationNo. 5, International Council for Bird Preserva-tion, Cambridge, England.

Thiollay, J. M. 1985b. Species diversity and com-parative ecology of rainforest falconiforms onthree continents. Pp. 167–174 in Newton, I., &R. D. Chancellor (eds.). Conservation studieson raptors. ICBP Technical Publication No. 5,International Council for Bird Preservation,Cambridge, England.

Thiollay, J. M. 1985c. Composition of falconiformcommunities along successional gradientsfrom primary to secondary habitats. Pp. 181–190 in Newton, I., & R. D. Chancellor (eds.).Conservation studies on raptors. ICBP Techni-cal Publication No. 5, International Council forBird Preservation, Cambridge, England.

Thiollay, J. M. 1986. Structure comparée du peu-plement avien dans trois sites de forêt primaireen Guyane. Rev. Ecol. Terre Vie 41: 59–105.

Wetmore, A. 1965. Birds of the Republic of Pan-ama. Part 1. Smithsonian Institution Press,

Washington, DC.Whitacre, D. F., & R. K. Thorstrom. 1992. Maya

Project: use of raptors and other fauna as envi-ronmental indicators for design, management,and monitoring of protected areas and forbuilding local capacity for conservation inLatin America. Progress Report V, The Pere-grine Fund, Boise, Idaho.

White, C. M., & T. J. Cade. 1971. Cliff-nesting rap-tors and ravens along the Colville River in Arc-tic Alaska. Living Bird 10: 107–150.

Willis, E. O. 1974. Populations and local extinc-tions of birds on Barro Colorado Island, Pan-ama. Ecol. Monogr. 44: 153–169.

Willis, E. O. 1979a. The composition of aviancommunities in reminiscent woodlots in south-ern Brasil. Pap. Avulsos Dep. Zool. (Sao Paulo)33: 1–25.

Willis, E. O., & E. E. Eisenmann. 1979b. A revisedlist of birds of Barro Colorado Island, Panama.Smithsonian Contributions to Zoology 291,Smithsonian Institution Press, WashingtonDC.

Windsor, D. W. 1990. Climate and moisture vari-ability in a tropical forest: long-term recordsfrom Barro Colorado Island, Panama. Smith-son. Contrib. Earth Sci. 29: 1–146.

Wright, J. S., M. E. Gompper, & B. DeLeon. 1994.Are large predators keystone species in Neo-tropical forest? The evidence from Barro Col-orado Island. Oikos 71: 279–294.

379