J RaptorRes. 33(4):313-316 ¸ 1999 The Raptor Research Foundation, Inc. DIET OF THE MADAGASCAR HARRIER-HAWK, POLYBORO1DES RADIATUS, IN SOUTHEASTERN MADAGASCAR S• M. KARPANTY 1 Departmentof Zoology,Miami University,Oxford, OH 45056 U.S.A. STEVEN M. GOODMAN Field Museumof Natural History,Roosevelt Road at Lake ShoreDrive, Chicago,IL 60605 U.S.A. ABSTRACT.--Werecorded prey deliveries, pellets and food remains of the Madagascar Harrier-Hawk (Polyboroidesradiatus) from 10 August-5 December 1997 at Berenty and Bealoka private reservesin extreme southeasternMadagascar. The MadagascarHarrier-Hawk had a variable diet, eating at least 16 prey species,including reptiles (2.4% of biomass),birds (11.5%), mammals(85.9%) and insects(0.2%). The largest component of the diet of the harrier-hawkin terms of biomasswas Verreaux's Sifaka (Pro- pithecusverreauxi), all of which were presentedby the male to the female during courtship. KEYWO•DS: MadagascarHarrier-Hazok; Polyboroides radiatus; diet;, lemur predation. Dieta de Polyboroidesradiatus en el surestede Madagascar R•SUMEN.--Registramoslas presas,egrag6pilas y restosde comida de Polyboroidesradiatus desde Agosto 10 hastaDiciembre 5 de 1997 en las reservasprivadas de Berentyy Bealokaen el extremo surestede Madagascar.Polyboroides radiatus tuvo una diem variable consumiendopor lo menos 16 especiesdistintas de presas,las cualesincluyen reptiles (2.4% de la biomasa),aves (11.5%), mamiferos(85.9%) e insectos (0.2%). E1 mayor componentede la dieta en t6rminosde biomasafue Propithecusverreauxi, todos apor- tados por el macho a la hembra durante el cortejo. [Traducci6n de Autor] Until recently, information on the diets of Mal- Recently, there has been an increased demand agasyraptors has been limited. Langrand (1990) for information on the feeding ecologyof Malagasy summarized information based on stomachanalys- raptors. Conservationistsconcerned with protect- es of Rand (1936). More recent analysesof pellet ing what remains of Madagascar'sfragmented for- remains of several owl species (Goodman and ests need to understand the ecosystemfrom the Thorstrom 1998, Goodman et al. 1993a, 1993b, perspectiveof primary consumers.By focusing ef- Langrand and Goodman 1996) and the Madagas- forts on protecting predator populations such as car Buzzard (Buteo brachypterus)(Goodman and raptorsand the habitat containingtheir often large Langrand 1996) have expandedour knowledgeof home ranges, we can more easily preserve organ- this group. One endemic species of raptor for ismsexisting at all levelsof the food chain (Watson which little information is available is the Mada- and Lewis 1994). Primatologistshave also become gascarHarrier-Hawk (Polyboroidesradiatus). Despite interested in documenting the diet of Malagasy its abundance in a variety of habitats across the raptors as they seek to understand the historic and island ranging from montane rainforest to spiny present-day selection pressureson lemur life his- desert scrub, little is known about the ecology of tory characteristics.While predation on diurnal le- this raptor and there is no quantified information murs by the harrier-hawk has rarely been docu- on its diet. mented (Goodman et al. 1993), laboratory and field studiesof lemur vigilanceand alarm callssug- 1 Presentaddress: Department of Ecologyand Evolution- gest that it may represent an important threat to ary Biology, SUNY at Stony Brook, Stony Brook, NY free-ranging lemurs (Macedonia 1990, Sauther 11794-5245 U.S.A. 1989). Wright (1998) demonstrated in a 10-yr be- 313 314 KARPANTYAND GOODMAN VOL. 33, NO. 4 Table 1. Diet of the MadagascarHarrier-Hawk at Berenty and Bealoka Reservesin southeasternMadagascar. Data are combined from the two sites. % TOTAL % TOTAL Tm MNI a INDIVIDUALS BIOMASS Reptiliab Furdfer lateralis 1 1.6 O.1 Tracheloptychusmadagascariensis 4 6.3 0.5 Chalarodonmadagascariensis 1 1.6 0.1 Dromicodryasquadrilineatus 1 1.6 1.7 Ave sc Bubulcusibis (nestlings) 3 4.8 0.3 Numidameleagris a 1 1.6 4.3 Streptopeliapictu rata 1 1.6 O.8 Coracopsisnigra 3 4.8 2.8 Coracopsisvasa 2 3.2 2.4 A cridotheres tristis • 1 1.6 0.6 Dicrurusforficatus(nesdings) 4 6.3 0.3 Mammalia f Propithecusverreauxi verreauxi (adult) 2 3.2 29.1 Propithecusverreauxi verreauxi (young) 3 4.8 19.3 Microcebusmurinus (adult) 3 4.8 0.6 Microcebusmurinus (young) 1 1.6 0. l Tenrec ecaudatus 4 6.3 33.4 Rattus rattus• (adult) 7 11.1 3.3 Rattus rattus (subadult) 1 1.6 0.2 Insecta (Orthoptera, Coleoptera,Lepidoptera) 20 31.7 0.2 Total 63 100 100 MNI = Minimum number of individuals. Reptilebiomass estimates provided by C. Raxworthy(pers. comm.). Avian biomassfrom Goodman et al. (1997) and Telfair (1994). Probablyintroduced to Madagascar(Langrand 1990). Introduced to Madagascar. Mammal biomassestimates from Mittermeier et al. (1994) and fieldworkby S.M. Goodman (pers.comm.). havioral study that the large-bodied, diurnal rain- is characterizedby a distinct dry season(March-Novem- forest Milne-Edward's Sifaka (Propithecusdiadema ber) and highly variable amounts of annual rainfall •n the wet season (December-February). The forest frag- edwardsi)responds to aerial predators such as the ments of Berenty and Bealoka are separatedby 10 km of harrier-hawk by giving loud alarm calls, dropping sisalplantations and small fragmentsof spinyforest; thus, low in the canopy and choosing daytime resting there is no migration of primates between reserves.The siteslower than feeding sites.The purpose of this land vertebrate faunal composition of the two reserves•s essentiallyidentical. study was to describe the diet of the Madagascar One occupied Madagascar Harrier-Hawk nest was Harrier-Hawk. found in each reservein early August1997 during the courtship stageof the breeding cycle.Birds were individ- STUDY AREA AND METHODS ually identified by key features. Species-specificvocaliza- Researchwas conducted between 10 August-5 Decem- tions were used to assist in nest location. A blind was ber 1997 at the privately-ownedBerenty and BealokaRe- constructedadjacent to the nest tree within Berenty.Us- servesin southeasternMadagascar (25ø00'S; 46ø18'E, el- ing 10 X 40 binoculars and a 30X spotting telescope, evation 100 m) near Amboasary-Sud,80 km west of continuous observationsof this nest were performed at Tolagnaro (Fort Dauphin). Together, these two reserves half-dayintervals in conjunction with lemur observations composeapproximately 200 ha of gallery and spinyforest Focal hours totaled 134 at this nest. Behavioral observa- on the banks of the Mandrare River and are dominated tions were not performed on the Bealoka nest. by Tamarindusindica, a characteristictree of this semiarid Prey items brought to incubating or brooding parents thorn-scrub habitat. The climate of southern Madagascar were identified. Pellets and prey remains were also col- DECEMBER 1999 DIET OF MADAGASCAR HARRIER-HAWK 315 lOO% al) Verreaux's Sifakas making this diurnal lemur 9O% the most important prey item for the hawks in 8O% terms of biomass (48.4%). All of the sifakaswere 7O% taken during the courtshipstage of the breeding season(Fig. 1). The large lipotyphlan tenrec (Ten- w 60% rececaudatus), which weighsup to 1.95 kg, was the 5O% o secondmost important prey item at 33.4% of total m 40% biomass.A small portion of the harrier-hawk'sdiet, i• REPTILIA either by numbers or biomass, consisted of intro- 320%O% ßAVES duced species(Numida meleagris,Acridotheres tristis Ir• MAMMALIA and Rattus rattus). lO% I[] INSECTA DISCUSSION Courtship Incubatmn Nestling BREEDING STAGE Our studypresents the first detailed analysisof the diet of the MadagascarHarrier-Hawk. Good- Figure 1. Percentage biomassof reptilian, avian, mam- man and Pidgeon (1991) previouslydocumented malian and insect prey of the MadagascarHarrier-Hawk the consumptionof a flying fox (Pteropusrufus) by during courtship, incubation and nestling stagesof the harrier-hawksat Berenty and suggestedthat harri- breeding cycle based on direct observation of prey brought to the nest. er-hawksmay nest near and exploit concentrated food resources.Our study supportsthis finding. Not only did harrier-hawks eat sifakas, which are leeted from beneath the nests.Prey collectionsat Berenty two to three times their body mass,but they also were made daily and at Bealoka weekly.Remains were nestedwithin 500 m of a heronry and consistently sorted, dried and stored in plasticbags. Identification of exploitedthis resourceduring the nestlingstage of prey remains was made to genus and specieslevels at the Universit6 d'Antananarivo. Paired parts of any taxon their breeding cycle. Rand (1936) reported re- were separatedand the largest number of parts of either mains of insects,spiders, lizards, frogs and small the left or right side was considered the minimum num- mammalsin the digestivesystem of the Madagascar ber of individuals (MNI). Data from the two nests were Harrier-Hawk. Our studydemonstrates the diet of combined as data from the Bealokanest were limited by the weekly collectionschedule. The compositionof the the harrier-hawkto be extremelyvariable, contain- prey communities were identical between the two re- ing at least16 vertebrateand invertebrateprey spe- serves. cies. The continued success of this hawk in the fragmented forests of southern Madagascarwill RESULTS likely be due to its ability to exploit such a wide Reptiles accountedfor 11.1% of total individuals variety of prey. and 2.4% of total biomass, birds for 23.9% of in- One of the more strikingfindings of our study dividuals and 11.5% of biomass, mammals for was the predation of harrier-hawkson