J RaptorRes. 38(3):256-262 ¸ 2004 The Raptor ResearchFoundation, Inc.

NESTING BIOLOGY AND DIET OF THE MADAGASCAR HARRIER (CIRCUS MA CROSCELES) IN AMBOHITANTELY SPECIAL RESERVE, MADAGASCAR

LILY-ARISONRENE DE ROLAND,JEANNENEY RABEARIVONY, AND IGNACERANDRIAMANGA ThePere•ne Fund's Madagascar Project, B.P 4113, Antananarivo(101), Madagascar

RUSSELL THORSTROM 1 ThePeregrine Fund, 5668 WestFlying Hawh Lane, Boise,ID 83709 U.S.A.

ABSTP,ACT.--We studied MadagascarHarriers (Circusmacrosceles) in the central high plateau at Ambo- hitantely SpecialReserve, Madagascar during the 1997 and 1998 breedingseasons. We located11 nests and documented eight nesting attempts during the two seasons.All nestswere placed on vegetation within marshes(N = 9) and averaged43 cm above water level. Breeding commenced in late August and September,during the middle of the dry season.Egg layingoccurred from the middle of September to the end of October, peaking in late September and spanning 40 d for eight clutches.The incubation period was 32-34 d at eight nestsand nestlingsfledged at 42-45 d of age (N = 7) in Novemberand December,at the start of the rainy season.Of 23 eggslaid in eight nests(o• clutch size = 2.9), 17 of 23 (74%) hatched,and seven(41%) of thosehatchlings fledged. Overall productivity was 0.9 youngfledged per breeding attempt and nest successwas 75% (N = 8). Only one of six successfulnesting attempts fledged two young. The MadagascarHarrier diet from 272 identified prey was composedof insects (48%), snakes (21%), birds (21%), lizards (6%), rodents (3%), and domestic chickens (1%); in terms of biomassbased on prey remainsand pellets:birds (45%), reptiles(35%), and mammals(18%) made 98% of prey. This is the first breeding studyof this species,and it showsthis harrier reproducesat a relatively low rate, and has an unusual diet relative to related species. KEYWOP, DS: MadagascarHarrier, Circus macrosceles;Madagascar, nests; diet; nesting behavior.

DIETA Y BIOLOGIA DE ANIDACION DE CIRCUS MACROSCELES EN LA RESERVA ESPECIAL DE AMBOHITANTELY, MADAGASCAR RESUMEN.--Estudiamoslos aguiluchos de el Madagascar (Circusmacrosceles) en la Reserva Especial de Ambohitantely,Madagascar durante las estacionesreproductivas de 1997 y 1998. Localizamos11 nidos y documentamosocho intentosde anidaciondurante las dosestaciones. Todos los nidosfueron ubicados en vegetaci6nde pantano (N = 9) y promediaron 43 CMS sobre el nivel del agua. La reproducci6n comenz6 a finalesde Agostoy Septiembre,a mediadosde la estaci6nseca. La posturade huevosocurri6 desde mediados de Septiembre hasta finales de Octubre, con un pico hasta finales de Septiembrey expandifindosepor 40 diasy eight nidadas.E1 periodo de incubaci6nfue de 32-34 dias en ocho nidos y los pichonesemplumaro• a los 42-45 dias de edad (N = 7) en Noviembrey Diciembre, al inicio de la estaci6n11uviosa. De los 23 huevospuestos en eight nidos (tamafio de la nidada = 2.9) 17 de 23 (74%) eclosionarony seven(41%) de los pichonesemplumaron. La productividadgeneral fue de 0.9 pichonesemplumados por intento reproductivoy el fixito de anidacionfue de 75% (N = 8). En solo uno de seisintentos exitososde anidacion emplumaron dos piehones.La dieta del aguiluchode Mad- agascar(de 272 itemsde presasidentificadas) estuvo compuesta por insectos(48%), Serpientes(21%), aves(21%), lagartijas(6%), roedotes (3%) y gallinasdomesticas (1%). En tfirminosde biomasa,con baseen restosde presasy egagropilaslas aves constituyeron un 45%, losreptiles un 35% y losmamlferos un 18% para un total de un 98% de presas.Este es el primer estudiosobre la reproducci6nde csta especiey muestraque el aguiluchose reproduce a una tasalcnta y tiene una dicta inusualcomparada con las de las especiesrelacionadas. [Traducci6n de Cfisar M•trquez]

Correspondingauthor's e-mail address:[email protected]

256 SEPTEMBER 2004 NESTING BIOLOGY OF THE MADAGASCAR HARPdER 257

Our knowledge of the ecology and biology of sonal temperaturesrange from 10-25øC (Nicoll and Lan- Malagasy raptors has gradually increased based grand 1989). mostly on research in northeastern Madagascar METHODS (Thorstrom and Rene de Roland 2000); however, We searchedmarshes, valleys, and rolling hills over the the biology of the MadagascarHarrier (Circusma- whole reservedaily fkomJuly-December 1997 and Au- crosceles)is still poorly known. The cosmopolitan gust 1998-January 1999 for potential breeding pairs.We group, the harriers (Circus),is relativelywell-known watched for harriers flying near marshesto detect pair 16 speciesof medium-sizedhawks, and includes activities and courtship behavior. We followed harriers daily by sight until a nest was confirmed by a pair's be- one speciesin Madagascar.This harrier (C. macros- havior or by locating the nest. Nest observationswere celes)was recently separatedand elevated to a full made froin the ground with 10X binoculars and 20-45X speciesrank by M. Wink in Simmons(2000), based spotting scope at distances of 200 in. We observed nest on DNA evidence and morphological differences sitesfroin 0500-1830 H and rotated among nests.When accessible,we measurednest length, width, and height from its sister species,the Rfiunion Harrier (C. above water level to the nearest 1.0 cm. We considered maillardi;Bretagnolle et al. 2000). The Madagascar the incubation period to be the time (d) from the laying Harrier is a rather uncommon bird of the open of the penultimate egg (usually the second or third laid marshes and grasslands,and is observed infre- egg) to the hatching of the secondegg (Simmons 2000) We measured length and breadth of eggs to the nearest quently coursingover grassfields, fallow rice fields, 0.1 mm with vernier calipers and fresh egg massto the marshes, and wetlands (Langrand 1990, del Hoyo nearest grain with a 100 g Pesolaspring scale (Pesola, et al. 1994). Globally,it is classifiedas a vulnerable Jackson,MS U.S.A). Reproductive variables and produc- species(BirdLife International 2000). Its diet con- tivity were defined as: breeding attempt (neststhat con- sistspredominantly of birds, reptiles, mammals, tained at least one egg), laying date (when the first egg waslaid), clutch sizeper individual nest (number of eggs and insects (Rand 1936, Langrand 1990). Threats laid in nests), mean clutch size (mean number of eggs to this speciesinclude dry seasongrassland fires laid per breeding attempt), number of eggs hatched, that usuallyoccur during its nesting season,loss of youngfledged (number of youngsurviving to first flight), marsh and grasslandhabitat, and human persecu- productivity (number of young fledged per breeding at- tempt), and nest success(number of total breeding at- tion to protectpoultry and for food (Paverne1997, tempts that fledged at least one young). BirdLife International 2000). The population sta- Prey delivered by adults was identified and quantified tus and distribution is not well known. Here, we during daily nest observations.Prey remains were col- provide new information on the breeding biology lected froin nestsand were identified by S. Goodman of World Wide Fund ibr Nature (WWF) in Madagascar. of the MadagascarHarrier from the high central Madagascar Harriers were trapped with a bal-chatn plateau region of north-central Madagascar. placed near the pair's center of activityduring the breed- ing period (Thorstrom 1996). We determined body mass STUDY AREA with a 1000 g Pesolaspring scaleand measuredwing and tail length to the nearest 0.1 mm with vernier calipersor We studied the MadagascarHarrier in Ambohitantely to the nearest 1 mm with a metric tape measure. Three SpecialReserve (18ø13'S,47ø16'E) and its surrounding birds were color banded for individual identification. area during two consecutivebreeding seasons,1997-98 (Randriamanga 2000). This reserve is situated in the RESULTS high central plateau, about 130 kin northwest of Anta- nanarivo, the capital of Madagascar.In 1982, the area was Courtship activitiesbegan in August and Sep- classifiedas a SpecialReserve of 5600 ha (Nicoll and Lan- grand 1989). Of the 5600 ha, only 2800 ha are still cov- tember. Initiation of the breeding seasonbegan ered by native forests, 1960 ha by grasslandsincluding with courtshipflights and pairs defending a space marshes, and 840 ha of exotic plantations (Langrand around a site, where they intended to place a nest. 1995). The eastern-facingslopes contain the nativeforest We located seven nest sites in 1997 and four in fragments. In the higher elevations and the western sec- 1998, of which two were on the same territories as tion of the reserve,the area is coveredby grasslandscom- 1997, (total of nine sites in an area ca. 1500 ha; posed of Aristida rufescens,Loudetia sp., and Andropogon sp., and low-lyingareas of marsheswith other grassesand Fig. 1). The distance between neighboring nests reeds (e.g., Cyperaceae). On the knolls and summits in averaged1514.5 m + 674.6 m (range = 370-3720 the western section, small stands of introduced trees of m, N = 9 pairs). Pinus patula and Eucalyptussp. are common. The eleva- tion varies from 1267-1660 m. The climate is character- Courtship Behavior.Courtship behaviorsconsist- ized by two distinct seasons;a dry warin period froin ed of intraspecific vocalizations, aerial display April-October and a hot rainy seasonfroin November- flights by males,pair-formation flights, nest build- March. The mean annual rainfall is 2150 mm and sea- ing activities,courtship feedings, copulations, and 258 RENE DL ROLAND ET AL. VOL. 38, NO. 3

close to the nesting area. Also, "k•k•kf•k•k•" was the sign of alarm or emitted when a predatory an- imal was close to the nest. Three adults were trapped during this period. Measurementstaken were: wing chord = 459 mm, 461 mm, and 443 mm; and tail = 288 mm, 285 mm, and 272 mm, for two females and one male, respectively.Madagascar Harriers show moderate reversesize dimorphism. The massof two females were 850 and 910 g and the one male was 600 g. Nest Building. Nest constructionbegan 15-17 d after pair formation. The earliestnest building was observed on 26 August 1997. The male was pre- dominantly responsible for building (quantitative data not available) the nest, while the female re- mained in the nest vicinity vocalizingfor food. All nestswere built on vegetation in marshes. Herba- ceousvegetation and grassesused in nest building were collected from the ground (e.g., Eulalia vil- losa,Osmunda regalus, Scirpus spp., Kotschyaafricana, Aristida rulescerts,and Pteridiumspp.), while dry twigs were collected from Eucalyptusand Pinus trees. All nesting material was collected from 40- 200 m of the nest site. Nest building activities started early morning, Legend ? • k,• from 0530-0900 H, and continued again from 1700 H-sunset. Due to the level of water below • LargestForest • Stre• ß 1@97Nesls Block nests,only nine of 11 located nestswere accessible [----] Savannah :•Trail I 1998Nests for measurements. Nest measurements averaged 63.2 _+ 7.2 cm (range = 50.0-77.0 cm) by 44.5 _+ Figure 1. Location of study site, Ambohitantely Special 4.5 cm (range = 34.2-51.3 cm). For nine nests,the Reserve,in Madagascarand nest locationsduring 1997 mean height above water was 43.2 +- 15.5 cm and 1998 breeding seasons. (range = 24.5-81.0 cm). Pairsbuilt new nestsevery year and the mean distance between the previous nest defense.Aerial displayflights beganwith a spi- year's nest in a given territory was575 m +_205.1 raling, ascending flight by the male, then pro- m (N = 2). Nest building took 25-30 d (N = 11 gressedto a fasthard flapping flight with steepun- nests). A renesting attempt occurred at one nest dulationsand then to rapid descendingspirals with 200 m from the first nest which was destroyeddur- "koue" vocalizations.The earliest pair-formation ing the early incubation period. flight observedwas on 12 August 1997. Pair-for- Egg Laying. During the two breeding seasons, marion flights involved the male and female flying the earliest recorded laying date was 13 September slowly together in their territory and sometimes 1997 with incubation starting on 15 September stalling and grabbing at the other member of the 1997 and the latest laying date was 20 October pair. Males called "koue... koue" every 10-15 sec 1997 with incubation startingon 26 October 1997 during pair-formationflights. During the courtship (N = 8 nests). The modal clutch size was three (N period the female wasprovisioned with food by the = 5), followed by clutches of two (N = 2), and male. Gopulationsalways came after a prey delivery there wasone four-eggclutch; the mean clutch size from male to female. Gopulationsaveraged 8.9 + was 2.9 + 0.2. Mean dimensionsof 23 Madagascar 0.9 (SE) sec in duration (N = 12, range = 6-15 Harrier eggswere 48.7 _+ 1.7 mm (44.1-51.0 mm) sec) and occurred 3-4 times a day. by 37.5 _+1.1 mm (35.7-40.0 ram). Mean fresh egg The territorial defense call may be described as masswas 36.5 +_ 1.2 g (34.8-39.0 g). "ouek" given ca. 5 sec when another harrier was Incubation. Only females incubated. In 216.3 hr SEPTEMBER 2004 NESTING BIOLOGY OF THE 1VIADAGAS(LARHAP, mEt• 259

Table 1. Reproductive successof Madagascar Harriers (Circusmacrosceles) at Ambohitantely Special Reserve, Mada- gascar,during the breedingseasons of 1997 and 1998.

BREEDING MEAN No. EGGS No. YOUNG FI,EDGLINGS/ NEST ATTEMPTS NUMBER CLUTCH HATCHED FLEDGED BREEDING SUCCESS YEAR NESTS OBSERVEDOF EGGS SIZE (%) (%) A'ITEMPT (%) (N) 1997 7 4 12 3 10 (83%) 4 (40%) 1.00 (4/4) 100 (4) 1998 4 4 11 2.7 4 (37%) 3 (75%) 0.75 (3/4) 50 (2) Total 11 8 23 2.9 14 (61%) 7 (50%) 0.88 (7/8) 75 (6) of nest observations, females incubated for 196.8 nests containing 23 eggs, 14 (61%) hatched, and hr (91%) and the nest was unattended for 19.5 hr seven (50%) of those hatched fledged (Table 1). (9%). The incubation period ranged from 32-34 In total, sevenyoung fledged fkom eight breeding d (N = 8 nests). During the incubation period, the attempts, for an overall productivity of 0.9 young male's primary role was food provisioninc to the fledged. Nest successfor the 2 yr of the studywas female and nest defense. 75% (N = 8). In 1998, two nestswere destroyedby Nestling Period. At hatching, nesding massvar- a grasslandfire during incubation. Reproductive ied from 25-32 g (N = 6). Brooding and feeding losses resulted from brood reduction (59%), ad- the young was the female's responsibility.When dled eggs (26%), and Pied Crows (Corvus albus; the male arrived with prey, he circled above the 15%) among the eight nestingattempts. For each nest, and called "tou ... tou ... touff" to the fe- nest containing three young, the third individual male. Also, the female solicited food by calling was alwaysdead at lessthan 10 d of age. In nearly "kiou ... kiou ... kid .... "Prey were delivered all nesting attempts with two or more young, the directly to females mostly by an aerial transfer secondnestling hatched did not survivemore than (88.8%; N = 166). In several instances,when the 40 d, and was often dead between 14-35 d of age. female did not respond to the male and leave the Only one of eight nesting attempts was successful nest or was absent, the male delivered the prey di- in fledging two young. rectly into the nest (11.2%). The earliest the fe- Food Habits. We observed272 prey itemsbeing male was observed leaving the nest and hunting delivered to females and nestlingsduring the two for the nestlingswas when the nestlingswere 24 d study seasons.On a numerical basis, insectswere of age. Young first flew from 42-45 d of age (N = the predominant prey comprising 48.2% (N = 7). First flights of young were about 1-5 m from 131) of the diet, followed by snakes20.9% (N = the nest. 57), birds 20.6% (N = 56), chameleons(Furc•fer Post-fiedgingPeriod and Dispersal. By 48 d of spp.) 5.9% (N = 16), rodents 3.3% (N = 9), and age, young were flying 20 m from the nest. Fledg- chickens (Gallus gallus) 1.1% (N = 3). Prey re- lings were alwaysfed at the nest by the adults dur- mains (N = 12) and pellets (N = 22) left in the ing the first week. At 50 d of age, the male tried nestswere identified and composedof 24% insects, transferring food to the young in flight while the 44% birds, 14% snakes, 6% chameleons, and 12% female placed prey on tufts of grass averaging rodents and insectivores(N = 50 identified prey; 136.7 + 101.7 m (N = 6, range = 40-300 m) from Table 2). On a biomassbasis, birds (44.7%), rep- the nest. After 50 d of age, young were not ob- tiles (35.6%), and mammals (18.6%) comprised served being fed by the females and solicited food 98.9% of the estimatedbiomass from prey remains with the "kiou" call. At 55 d of age, young flew up and pellets (Table 2). to a height of 100-200 m and did not return to DISCUSSION their nests for periods of ca. 15 min. Fledglings began taking prey in flight fkom the male at 55 d In this first ecological study of the breeding and of age. The adult females had disappearedfrom diet of the MadagascarHarrier we found: nesting their nesting territories when young were ca. 65 d was limited to marshy areas of savannahhabitat in of age (N = 7 nests).Young dispersedfrom their Ambohitantely Special Reserve, breeding com- natal areasat 70 d of age alongwith the adult male. menced in the spring (August-September)similar Reproductive Success.In eight fully-monitored to other southern harriers, copulationswere always 260 RENE DE ROLAND ET AL. VOL. 38, No. 3

Table 2. List of prey speciesidentified from remains and pellets of MadagascarHarriers (Circus macrosceles)at AmbohitantelySpecial Reserve, Madagascar during 1997-98.

BIOMASS TOTAL PREYSPECIES NO. ITEMS BIOMASS(G) (%) ANDSOURCE

INSECTS 60 l.l Orthopteran ( Nomadacrisseptemfasciata) 5 Estimate a REPTILES 1850 35.6 Chameleon (Furciferlateralis') 150 Estimate a Snake ( Liopholidophislateralis') 200 Estimate a AVES 2325 44.7 MadagascarFlufftail ( Sa•vthrurainsularis) 300 Dunning 1993 Common Quail (C0turnix Coturnix) 100 Dunning 1993 MadagascarPartridge (Marga•vperdix madagascarensis) 4 220 Dunning 1993 MadagascarButton-quail (2itrnix nigricollis) 1 40 Estimate a Common Stonechat ( Saxicolatorquata) 2 15 Dunning 1993 MadagascarLark ( Miralta hova) 3 45 Estimate a MadagascarCisticola (Cistic0la cherinus) 1 10 Dunning 1993 Domestic chicken ((;allus gallus) 3 300 Estimate a Unidentified birds 5

MAMMAI.S 965 18.6 Black Rat (Rattus rattus) 4 200 Garbutt 1999 House Mouse (Mus musculus) 1 15 Garbutt 1999 Lowland-streaked Tenrec ( Hemicentetessemmispinosus) 1 150 Garbutt 1999 TOTAL 50 5200

Based on mass measurements taken in the field.

accompaniedby prey delivery by males, clutch was In R6union, a small island 700 km east of Mad- typicallysmall (2.9 eggs)for a tropical harrier, and agascar,the R•union Harrier, a sisterspecies to the substantialbrood reduction occurred. We suggest MadagascarHarrier, began the breeding seasonat that the relatively unusual diet for harriers com- the same time as the Madagascar Harrier with prised mainly of insectsmay have stimulated sib- courtship displays in August-September, nest ling aggressionamong nestlings. building in October-November and onward, egg In AmbohitantelySpecial Reserve, the topogra- layingfrom January-April, and fledglingsreported phy of the area has led to irregular formation of with adultsup to October (Bretagnolleet al. 2000). marshes and valleys, thus restricting the nesting Compared to the MadagascarHarrier, the R•union area for MadagascarHarriers. On the other hand, Harrier had a prolonged breeding season.Harriers the aggressivenessof males toward conspecifics in Madagascarat Ambohitantely Special Reserve, also seemed to result in the spacing of nesting commencedbreeding with courtshipbeginning in pairs. Grasses, herbaceous vegetation and dry August-September,and nest constructionand egg branches, from Pinus and Eucalyptusspp. trees, laying in September-October,at a time when the were utilized for nest constructionby the harriers, water level in marshes was at its minimum. Paverne sameas reported by Paverne(1997) in Madagascar. (1997) alsohad two nestsunderway in September MadagascarHarriers placed nestson grasstufts in and one in November 1996, in the same general marshes, which facilitated nest construction, simi- region asAmbohitantely. Paverne (1997) reported lar to other marsh harrier species(Simmons 2000). on two nestswith four and two eggseach, and one The Madagascar Harrier nests were about 40 cm with undetermined number of eggsor young. We above the water level. suspectthe nest located in November (Paverne SEPTEMBER 2004 NESTING BIOLOGY OF THE MADAGASCAR HARRIER 261

1997) might possiblyhave been a renesting at- man-causedthreats toward this species.First, adults tempt due to the timing of the incubation period are persecutedfor food and as a threat to domestic we recorded and the second nesting attempt we fowl. In one case in 1996, Randriamanga (2000) documented. By the time the rainy seasonbegan met a poacher who had displayedcarcasses of 13 in late November and December; and water level MadagascarHarriers he had killed. Local people in the marshes began rising, young harriers had alsoconsume the eggsand nestlingsas a sourceof fledged (first flights) or were near fledging. For protein (Randriamanga 2000). Second, during ev- the R6union Harrier, fledging appeared to occur ery dry season (April-October) the grasslaudsof from March-June, and much later than our obser- Madagascar, especially the high-plateau region vations for MadagascarHarriers. which includesthe reserve,are burned by human- For MadagascarHarriers, hatching successaver- set fires to stimulate green growth fbr cattle fodder aged 61% for both years, and was extremely low and land clearing. During this study, two harrier (37%) during the 1998 breeding season due to nestswere destroyedduring the incubation period grassland fires destroying several nests and the by uncontrolled human set fires. In 1996, in an presenceof addled eggs.The female fed and cared area near Ambohitantely,Paverne (1997) suggest- for the first-hatchednestling immediately, and we ed a wild fire destroyedone of the harrier nestshe believe this led to a decrease of incubation time of was observing.Finally, the conversionof low-lying the remaining eggs, possiblycausing the high in- marshes and wetlands to rice fields for human food cidenceof addled eggs(39%). production reducesnesting habitat for this species. Simmons (2000) showed that all harrier species Currently, the vulnerable statusof the Madagascar exhibit reverse size dimorphism, and we found Harrier is justified based on threats to its habitat MadagascarHarrier females have one of the larg- and its sparsedistribution throughout Madagascar. est body sizes,at 850-910 g, of the 16 harrier spe- This specieshas been recorded at a number of pro- cies found in the world. tected areas and national parks, but most of the Probablydue to limited food and aggressivebe- protected areas have been establishedto preserve havior from the first-hatched nestling, the second forested habitat and have limited grassland and and third hatched young died at 35 and 10 d of wetland protection. Biologists need more infor- age (N = 2 nests), respectively.The weakestyoung mation on the MadagascarHarrier population size were killed by the first-hatchednestlings, similar to and dynamics in order to provide conservation many other raptors living in food restricted envi- strategiesand protect it in the future. ronments (Meyburg 1974, Simmons1988, Gargett ACKNOWLEDGMENTS 1990). Brood reduction is well documented for other harrier speciesas well (Simmons 2000). This study was conducted by The Peregrine Fund's MadagascarProject with funding provided, in part, by Previous reports on the food habits of the Mad- John D. and Catherine T. MacArthur Foundation, the agascarHarrier came from four stomach contents Walt Disney Company Foundation, the Little Family examined by Rand (1936), who found the head Foundation, Environment Now and other important con- and feet of a MadagascarPartridge (Margaroperdix tributions. We thank Richard Ranarisoa, Director of As- sociation Nationale pout la Gestion des Aires Prot6g6es madagascarends),a rat (Ratms sp.), fur of a small at Ambohitantely Special Reserve and his staff for their mammal, two frogs, a young whistling duck (Den- support during this study, and Steve Goodman of WWF drocygnasp.), and an insectivorousmammal. for his assistancein identifying prey remains. We also In AmbohitantelySpecial Reserve the grasslands thank the Direction des Eaux et For6ts for their collab- and low-lyingmarsh valleyssupport a dense insect oration with the Peregrine Fund's Project in Madagascar. We especiallythank R. Watson and B. Burnham for their population of grasshoppers(Nomadacris septemfas- continued support of this project and L. Kiff, J. Berkel- ciata) and provide habitat for snakesand savannah man, R. Simmons, V. Bretagnolle, and J. Bednarz for birds. Grasshopperswere the predominant prey their comments and improving an earlier version of this taken by frequency,but in terms of biomassbirds, manuscript and Amy Siedenstrangfor creating the map reptiles,and mammalsmade up 90% of the harrier LITERATURE CITED diet. BraDLIFE INTERNATIONAL. 2000. Threatened birds of the The Madagascar Harrier is categorized as a vul- world. Lynx Edicions and BirdLit• International, Bar- nerable species(BirdLife International 2000). In celona and Cambridge, U.K. Ambohitantely Special Reserve, and most likely BRETAGNOLLE,V., J.M. THIOLLAY,AND C. ATTIla.2000. Sta- throughout Madagascar,there are three major hu- tus of R6union Marsh Harrier (Circus maillardi) on 262 RENE DE ROLAND ET AL. VOL. 38, NO. 3

RGunion Island. Pages 669-776 in R.D. Chancellor PAVERNE,L. 1997. Nidification et comportementde bus- and B.-U. Meyburg lEDS.], Raptors at risk. World ard de maillard (Circus maillardi) dans la rGgion Working Group on Birds of Prey and Owls. Hancock d'Ankazobe au nord-ouest d'Antananarivo. Working House, Berlin, Germany. Groupon Birds in theMadagascar Region Newsletter7:21- DUNNING,J.B., JR. 1993. CRC handbook of avian body 24. masses. CRC Press Inc., Boca Raton, FL U.S.A. RAND, A.L. 1936. The distribution and habits of Mada- GAPmUTT,N. 1999. Mammals of Madagascar.Yale Univer- gascarbirds. Bull. Am. Mus. Nat. Hist. 72:143-499. sity Press,New Haven, CT U.S.A. RANDaIAMANC,A, I. 2000. Contribution/t l'Gtude de la biol- GA•GETr,V. 1990. The Black Eagle, a study.Acorn Books ogic de la reproduction et 6cologle de busard de and RussellFriedman, Johannesburg,South Africa. Madagascar Circus macroscelesdans le tampoketsa DELHO¾O, J., A. ELLIOTT,AND J. SARGATAL(EDs.). 1994. d'Ankazobe. MGmoire de D.E.A., Universit6 Handbook of the birds of the world. Vol. 2. New world d'Antananarivo, Antananarivo, Madagascar. vultures to guineafowl. Lynx Edicions, Barcelona, SIMMONS,R.E. 1988. Offspring quality and the evolution Spain. of cainism. Ibis 130:339-357. nANGRAND,O. 1990. Guide to the birds of Madagascar. . 2000. Harriers of the world: their behaviour and Yale University Press,New Haven, CT U.S.A. ecology.Oxford Univ. Press,Oxford, U.K. . 1995. The effects of forest fragmentation on bird THORSTROM,R. 1996. Methods fbr capturing tropical ibr- speciesin Madagascar: a case study from Ambohitan- est birds of prey. Wildl. Soc.Bull. 24:516-520. tely Forest Reserveon the Central High Plateau. M.S. --AND L.A. RENE DE ROLAND. 2000. Status and con- thesis, Univ. of Natal, Durban, South Africa. servationof raptors on the MasoalaPeninsula. Pages MEYBURG,B.-U. 1974. Sibling aggressionand mortality 35-41 in R.D. Chancellor and B.-U. Meyburg [EDS.], muong nestling eagles.Ibis 116:224-228. Raptors at risk. World Working Group on Birds of NICOLL,M. AND O. LANGRAND.1989. Madagascar:revue Prey and Owls. Hancock House, Berlin, Germany. de la conservation et des aires ProtGgGes.World Wide Fund for Nature International, Gland, Switzerland. Received29July 2003; accepted16 May 2004