j. RaptorRes. 31 (4):308-312 ¸ 1997 The RaptorResearch Foundation, Inc.

LONG-TERM STUDY OF A GOSHAWK BREEDING POPULATION ON A MEDITERRANEAN MOUNTAIN (ABRUZZI APENNINES, CENTRAL ITALY): DENSITY, BREEDING PERFORMANCE AND DIET

VINCENZO PENTERIANI Laboratoired •Ecologie,Universitd de Bourgogne, Batiment "Mirande, "B.P. 400, 21011 Dijon Cgdex,France and StazioneRomana Osservazionee ProtezioneUccelli, (S.R.O.P.U.), % Oasi Naturale W.W.F. "Boscodi Palo," 00055 Ladispoli,Rome, Italy

ABSTRACT.-•Thebreeding density,breeding performance and breeding diet of a Mediterranean pop- ulation of Goshawks(Accipiter gentilis) were studiedin centralItaly over a 10-yrperiod. Sixteengoshawk nestingpairs were found within a 318 km2 area (5.03 pairs/100 km2).The minimumdistance between pairs averaged3.5 km. I observeda total of 51 nestingattempts at sevennest sites.Annual mean productivityof breedingpairs was 2.28 fledgings/successfulpair and 1.69 fledgings/eggqayingpair. The occurrenceof nestingfailures was correlated with rainfall during April-May (r = 0.77, P < 0.01) and brood sizeat fledgingwas related to layingdate (r -- -0.71, P < 0.05). The diet included20 species of birds (representing75% of the prey and 71% of the biomass)and six speciesof mammals(repre- senting25% of the prey and 29% of the biomass).Wood Pigeon (C0lumbapalumbus), Jay (Garrulus glandarius),Mistle (Turdusviscivorus), red squirrel (Sciurusvulgaris) and edible dormouse (Glis glis) accountedfor 67% of all prey captures.In terms of biomass,Jays (20.76%), Wood Pigeons (16.03%), red squirrels(12.64%) and edible dormice (12.29%) were the dominant prey. I found 23 pluckingsites on the ground (44.2%), 21 on rock with moss(40.4%), sevenon tree stumps(13.5%) and one on the root of an uprootedtree (1.9%). K•Y WORDS: Goshawk;Accipiter gentilis;breeding density; breeding performance, breeding diet.

Un estudio de tiempo largo de poblaci6n de cria en Accipitergentilis en un Montafia Mediterrfineo (AbruzziApennines, Centro de Italia): desnidad,exito de cria y dieta R•suMv.N.--Densidad de cria, exito de cria y dieta de cria de la poblaci6nen el Mediterr•tneode Accipiter gentilisfueron estudiadose el centro de Italia arriba de un tiempo de 10 aftos.Decisies Accipitergentilis parescon nidosfueron encontradosdentro de 318 km2 (5.03 pares/100kme). La distanciaminima entre paresfue 3.5 km. Yo observeun total de 51 intentosde poner nido en sietasitios del nido. La productividadde promedio annual de paresen cria fue 2.28 pajaritos/parescon exito y 1.69 pajaritos/ huevo-ponidodel par. La occurenciade nidos sin exito fue correlacionadocon 11uviadurante abril- mayo(r = 0.77, P < 0.01). A1tiempo de poner (r = -0.71, P < 0.05). La dietaincluyo 20 especiede pajaros(representando 75% de presay 71% de biomass)y seisespecie mamiferos (representando 25% de la presay 29% de la biomass).Columba palumbus, Garrulus glandarius, Turdus viscivo'rus, Sciurus vulgaris, y Glisglis dio cuentapor 67% de todala presa.En referenciaa biomass,Garrulus glandarius (20.76%), Columbapalumbus (12.29%), fueron la presadominante. Yo econtre23 sitiosde desplumaren el terreno (44.2%) 21 en pierdascon musgo(40.4%) sietaen un tronco de un arbol desarraigado(1.9%). [Traducci6n de Ra61 De La Garza, Jr.]

In Europe, most studies of Goshawks(Accipiter and Kostrzewa1990, Joubert 1991, Pielowski 1991, gentilis)have focusedon northern and central Eu- and Olech 1996). A few studieshave attempted to ropean populations(Thiollay 1967, Opdam 1975, estimatebreeding densitiesin Mediterranean hab- Opdam et al. 1977, Kos 1980, Marquissand New- itats but almostnothing is known concerning their ton 1982, Kalabfir 1984, Wilke et al. 1985, Gosz- food habits (Morillo and Lalanda 1972, Benussi czynski and Pilatowski 1986, B•hler and Oggier and Perco 1984, Garrigues et al. 1990, Mafiosa et 1987, Widfin 1989, Anonymous1990, Kostrzewa al. 1990, Mafiosa 1994).

308 DECEMBER 1997 GOSI-•WK BREEDING BIOLOGY IN ITALY 309

The objectivesof this study were to determine using weight data from the study area or bibliographic the breeding densityof a Goshawkpopulation on material (Geroudet 1946-57). All pelletsfrom individual visitswere pooled. I recorded the speciesthat occurred a Mediterranean mountain, to describe its breed- in samplesbut made no attempt to quantify the number ing biology in the area, to showa possiblerelation- of individuals in samples. To avoid counting the same ship betweenweather and breeding successand to prey twice in remains and pellets, prey in pellets were describe the breeding diet. computed only if they had not been found as remains•n the samevisit (Mafiosa 1994). Moreover, during each vis- STUDY AREA AND METHODS it, an effort was made to remove all prey remains. I studied use of different plucking sitesby Goshawks. This studywas conducted from 1984-93 in an area of To do so, I assigned each prey remain to one of five the Apennine mountainsin the Abruzzi region of central Italy (41ø49'N,13ø47'E) that includedAbruzzo National categories:bare rock, rock with moss, ground, root of Park and the Sirente and Majella massifs.The area uprooted tree and tree stump. A single plucking site was ranges in elevation between 800-2300 m and consists counted only once if it had been repeatedlyused by the mostly of beech (Fagussylvatica) forested slopeswith species. grazed and fallow farmland at the baseof the mountains. Above 1900 m, forestsare replacedby high-altitudepas- RESULTS tures. I found a total of 16 pairs of nesting Goshawks I located Goshawknest sites by searchingforested areas mapped on 1:10000 aerial photos and 1:25000 topo- in the 318 km2 study area for a density of 5.03 graphicmaps. Nest sites were located by walkingthrough pairs/100 km2. Minimum distancebetween pairs forestsbetween November-April, observingnuptial dis- averaged3.5 km (range 2-4.9 km, SD = 1.02). A plays and territorial flights (especiallyfrom the second G-test value of 0.99 indicated that nest sites were week of March to the first week of April) and playback of taped calls.Occasionally, I found nestsfrom pluckings, distributed regularly. droppings and moulted feathers,and adult and nestling Because I did not know of all seven breeding calls. Density of nesting pairs was estimated based on pairs at the beginning of the study, I observed a Newton et al. (1977). Regularityin nest-sitespacing was total of only 51 nestingattempts and collectedde- computedusing a G-test(Tjernberg 1985). The breeding chronology and performance of Gos- miled breeding data at only 43 nests over the 10- hawks was studied from 1984-93 at seven nest sites. From yr studyperiod. Egg laying took place betweenthe the end of March until the end of April, nestswere visited second and third week of April and fledging oc- to determine occupancy,changes of site and egg4aying curred between the first and secondweek of July. dates. Occupied nestswere observedfrom the end of May Eggswere laid during the first week of April at only until fledging to determine nesting success.Counts of one nest.Three nestsfledged one young(7%), 20 nestlingswere done from the ground (the slope was al- fledgedtwo young (46.5%) and 11 fledgedthree waysvery steep, facilitating counts) or by watchingthe young(25.6%). In 9 cases(20.9%), the pairsfailed nest with a high-powertelescope (60X). I used the definitions of Kostrzewa and Kostrzewa to fledge a single young. (1990) and laying pairs were those pairs that laid eggs, All of the nesting areas contained at least two successfulpairs were thosethat fledged at leastone chick nest structuresat the time nestingwas first initiated and unsuccessfulpairs were thosethat lost their clutches each year. Despite this, Goshawksusually used the or nestlingsat any stagebefore fledging. same nest from one year to the next. Only in seven The "Hydrographic Service" (Ministero dei Lavori Pubblici 1984-93) provided monthly valuesfor total pre- casesdid they use alternate nestsand in two cases cipitation (mm) and mean temperaturesfrom March- they actuallylaid secondclutches in the samenest. June. To investigate the relationship between spring Annual productivitywas 2.28 fledgings/success- weather and Goshawkbreeding performance, weather ful pair (SD = 0.41, range 1.75-2.75) and 1.69 data were examined for each month and for pairs of months (e.g., March-April, April-May and May-June;Kos- fledgings/egg-layingpair (SD = 0.58, range 0.6- trzewa and Kostrzewa 1990). Rainfall and temperature 2.6). In 1991, only a single pair bred successfully were used as statisticallyindependent values. fledging 3 young. From 1984-93, I collected prey remains and pellets The percentage of unsuccessfulpairs was corre- near seven nests and known plucking sites throughout the breeding seasonto determine the diet of Goshawks. lated with rainfall during April-May which coincid- Prey remainsand pelletswere identifiedby macroscopic ed with prelaying and laying, incubation and comparisonwith reference collections.For each verte- hatching periods (r = 0.77, P < 0.01). There was brate prey item identified, I attemptedto identify both no relationship between percentage of unsuccess- its speciesand whether it wasan adult or juvenile. I de- ful pairs and temperature nor wasthere a relation- termined age classbased on the size, , feather characteristicsand degree of ossificationin the prey re- ship betweenpercentage of fledged young per suc- main. The biomassof eachprey species was estimated by cessfulpair and May-Junetemperature or May-June 310 PENTERIANI VOL. 31, No. 4

Table 1. Prey items found in the diet of Goshawksin Of these, 23 were on the ground (44.2%), 21 on the AbruzziApennines, Italy. Weights of prey are given rocks covered with moss (40.4%), seven on tree •n grams. Specieswith fewer than 10 individualsin the diet are listed as other . stumps(13.5%) and one on root of an uprooted tree (1.9%). No pluckingsites were found on bare rock substrates. TOTAL WEIGHT N(%) (%) DISCUSSION Birds 582 (73.76) 103 273 (72.1) I found the breedingpopulation of Goshawksin Columbapalumbus 67 (8.49) 22 961 (16.03) the Abruzzimountains to be very regularin its dis- Strix aluco 13 (1.65) 6006 (4.19) tribution with a densitysimilar to that of other ar- Picusviridis 17 (2.15) 3378 (2.36) eas in Europe (5 pairs/100 km2 in Finland, Wik- Garrutusglandarius 194 (24.59) 29 729 (20.76) man 1977;4.5-5.5 pairs/100 km 9 in Germany,Kra- Corvusc. cornix 16 (2.03) 7760 (5.42) Turdusmerula 13 (1.65) 1066 (0.74) mer 1972,Link 1981,Dietrick and Ellenberg 1982; Turdusphilomelos 31 (3.93) 2790 (1.95) and 4.4 pairs/100 km9 in Switzerland,Biihler and Turdusviscivorus 108 (13.69) 12 028 (8.4) Oggier 1987). Unidentified thrush 23 (2.91) 1736 (1.21) Unlike otherareas of Europe(Opdam 1975, Op- Unidentified 61 (7.73) dam et al. 1977,Rust and Kechele1982, Kayser Other birdsa 39 (4.94) 15 819 (15.32) 1993) and other areas within the Mediterranean Mammals 200 (26.24) 39 956 (27.9) (Morillo and Lalanda 1972, Garrigueset al. 1990, Sciurusvulgaris 73 (9.25) 18106 (12.64) Mafiosa et al. 1990, Mafiosa 1994) where Goshawks Glisglis 88 (11.15) 17600 (12.29) have very diverse diets, Goshawksin the Abruzzi Unidentified mammals 31 (3.93) Apennineshad very low speciesdiversity in their Other mammalsb 39 (4.99) 4250 (4.11) diet. This wasprobably related to the fact that the Total 782 123160 high elevationof the studyarea limited the num- a Other birds:Buteo buteo (nestlings), Accipiter nisus, Falco tinnun- bers and typesof bird speciesavailable (20 breed- culus,Columba livia, unidentifiedpigeon, Alectoris graeca, Picoides ing specieswith a densityof 36 pairs/10 ha, Ber- major,unidentified woodpeckers,Pica pica, Saxicolarubetra, Parus noni 1995) comparedto the neighboring,lower- caeruleus,Carduelis chloris, Acanthis cannabina, Gallus domesticus. elevation piedmont (59.2 pairs/10 ha, Pandolfi bOther mammals: unidentified mouse, unidentified shrew, Lepus and Taferna 1991) and plain (158 pairs/10 ha, capensis. Bernoni et al. 1989). At lower elevations, land use wasalso more varied and habitatswere more pro- rainfall.The mean brood sizeat fledgingwas neg- ductiveresulting in a greaterdiversity of preyspe- ativelycorrelated with layingdate (r = -0.71, P < cies. Similar Goshawkdiets dominated by a few 0.05). prey specieshave been reported in coniferousfor- I identifieda totalof 782prey items representing estsin Sweden(Wid6n 1989), higher altitudesin 26 different speciesin the diet of these Goshawks Britain (Marquissand Newton 1982), mountainsof (Table 1). There were 20 speciesof birds (repre- France (Joubert 1991), and woodlands in north- senting75% of the prey and 71% of the biomass) eastern Italy (Benussiand Perco 1984). and six speciesof mammals(representing 25% of Crows (27.25%), thrushes (22.18%), medium- the preyand 29% of the biomass).Wood Pigeons sizedrodents (20.4%) and pigeons(9.63%) were ( Columbapalumbus) , Jays (Garrulus glandarius), Mis- the mostfrequently taken as prey.A large portion tie Thrush (Turdusviscivorus), red squirrels(Sciu- of the diet also consistedof nestlings,most of rus vulgaris)and edibledormice (Gtisglis) account- which were crows.Crows, thrushes, pigeons, me- ed for 67% of the identifiedprey. In termsof bio- dium-sizedrodents and nestlingshave also been mass, the Jay (20.76%), the Wood Pigeon reportedin the dietsof Goshawksin Europe (H6g- (16.03%), red squirrel (12.64%) and edible dor- lund 1964, Sulkava1964, Opdam et al. 1977, Gosz- mouse(12.29%) were the dominantprey. czynskiand Pilatowski 1986, Wid6n 1989, Mafiosa The total proportion of nestlingbirds (N = 51) et al. 1990, Kayser1993, Mafiosa1994) and in oth- that occurred in the samplewas 8.76%. Most of er regions (Reynoldsand Meslow 1984, Boal and these were nestlingsof crows (N = 28), thrushes Mannan 1994). The abundanceof forest species (N = 16) and raptors (N = 7). suchas Wood Pigeons,Jays, red squirrelsand edi- I identified52 pluckingsites used by Goshawks. ble dormice in the diet indicates that these Gos- DECEMBER 1997 GOSHAWI< BREEDING BIOLOGY IN ITALY 311 hawks hunt inside or at the edges of woodlands BENUSSI, E. AND F. PERCO. 1984. Osservazioni eco-etolo- (Opdam et al. 1977). giche sull'AstoreAccipiter g. gentilisnel carsotriestino. The productivityof Goshawksin the Abruzzi Ap- U.D.I. 9:3-25. ennines was more typical of the productivitythat BERNONI, M., L. IANNIELLO AND P. PLINI. 1989. Cens•- mento dell'avifauna nidificante in un bosco deciduo has been observed in bad years in forested areas dell'Italia centrale. Avocetta 13:25-29. in Lorraine and Limous regionsof France (Thiol- 1995. II metodo del mappaggio in una faggeta lay 1967, Nore 1977), in Finland (Cramp and Sim- del Parco Nazionale d'Abruzzo. Avocetta 19:131. mons 1980), in high elevation areas in Britain BOAL, C.W. AND R.W. MANNAN. 1994. (Marquiss and Newton 1982), in southern Bavaria diets in a ponderosa pine forest on the Kaibab Pla- (Rustand Kechele 1982), in the easternpart of the teau. Pages97-102 in W.M. Block, M.L. Morrison and SwissLowlands (B/ihler et al. 1987) and in Catal- M.H. Reiser [EDS.],The Northern Goshawk:ecology unya (Mafiosa et al. 1990). Productivity in the and management. Stud.Avian Biol. No. 16. Abruzzi Apennineswas heavily weather dependent. B(3HLER, U. AND P.A. OGGIER. 1987. Bestand und bes- During cold, wet springsbreeding startedlater and tandsentwicklungdes habichtsAccipiter gentilis in der Schweiz. Ornithol. Beob. 84:71-94. prolonged rain reduced reproductivesuccess pos- , R. KLAUS AND W. SCHLOSSER.1987. Brutbestand sibly by decreasinghunting successand food in- und jungenproduktion des habichtsAccipiter gentilis in take. Rainfall appearedto be a very significantfac- der Nordostschweiz 1979-1984. Ornithol. Beob. 84:95- tor. During the breeding seasonof 1991, nearly all 110. the nestingattempts failed owingto intenserainfall C•MP, S. AND K.E.L. SIMMONS. 1980. Handbook of the during the incubation period. The highest inci- birds of Europe, the Middle East and North Africa. dence of nest failure occurred when there was be- Vol. 2. Oxford Univ. Press, Oxford, UK. tween 60-120 mm of rain. The negativeimpact of DIETRICK,J. AND H. ELLENBERG.1982. Aspects of Gos- rainfall on Goshawk populations has also been hawkurban ecology.Pages 163-175 in R.E. Kenward stressedby Kostrzewaand Kostrzewa(1990). Based and I.M. Lindsay[EDS.], Understanding the Goshawk. I.A.F., Oxford, UK. on the correlation between weather and nest fail- GARRIGUES,R., R. MARTINEZAND J.A. MORATA. 1990. In- ures, laying date and mean brood sizeupon fledg- troducciGnal estudiode la biologia del azor (Accip,ter ing, April and May appeared to be critical in the gentilisL., 1758) en Albacete. Al-basit,Rev. Estud. Al- breeding cycle. Forest cutting during April and bacetenses27:123-162. May alsohad a marked negativeeffect on the pro- GEROUDET, P. 1946-1957. La vie des oiseaux. Vols. 1-6 ductivityof Goshawks.Forests were cut in the vicin- Collection de Poche. Les Beaut•s de la Nature. De- ity of five nesting areas during this period. At all lachaux et Niesd• S.A., Paris, France. five areas,the nesting attempt failed that year. By GOSZCZYNSKI,J. AND T. PILATOWSKI.1986. Diet of Com- contrast,forest cutting at four nestingareas during mon Buzzard (Buteobuteo L.) and Goshawks(Accip,ter the last three weeksprior to fledging had no effect gentilisL.) in the nestingperiod. Ekol.Pol. 34:655-667. HOGLUND,N.H. 1964.l•ber die Ernfhrung des Habichts on the breeding success. (AccipitergentilisLin.) in Schweden.Viltrevy2:271-328 ACKNOWLEDGMENTS JOUBERT,B. 1991. Questionssur la chronologie de la I am indebted to Fabio Liberatori for his invaluable reproduction chez l'Autour des palombes (Accip,ter cooperationin the field throughout the studyperiod and gentilis)en Haute-Loire. Nos Oiseaux41:39-54. to Gianni Marangoni for his precioushelp in determin- KALmaR,L. 1984. Note sulla biologia e 1o sviluppopos- ing the prey remains. I also thank Marina Cerasoli and tembrionale dell'Astore, Accipitergentilis, in Romania. Francesco Pinchera who contributed to the search of the Riv. Ital. Ornitol. 54:179-190. sites,as well as M. Chiavetta,Mr. Pellegrini,Ms. Pellegri- KAYSER,Y. 1993. Le r•gime alimentaire de l'Autour des ni, M. Garfagnini, M. Bernoni, the Caramanico Terme palombes,Accipiter gentilis (L.), en Alsace. Ciconia17. and Rocca di Mezzo Forest Stations. The Administration 143-166. of the Abruzzi National Park partially provided logistic Kos, R. 1980. Der habicht in der bundesrepubhk and financial supportfor this work. I thank especiallyE Tassi, the Director and C. Sulli. I thank Bruno Faivre, Deutschland.Vb•elwelt 101:161-175. Bernard Frochot, Patrick Giraudoux and Jean Secondi KOSTRZEWA,A. AND R. KOSTRZEWA. 1990. The relation- for their indepth reviewof the manuscript.Stefania Sar- ship of spring and summerweather with densityand aceni kindly helped with the Englishtranslation. breeding performance of the Buzzard Buteo buteo, GoshawkAccipiter gentilis and KestrelFalco tinnunculus. LITERATURE CITED Ibis 132:550-559. ANONYMOUS.1990. Breedingbiology of Goshawksin low- IC•MER,V. 1972. Habichtund sperber.Wittenberg-Luth- land Britain. B•it. Birds 83:527-540. erstadt, Berlin, Germany. 312 PF.NWF. m•q• VOL. 31, NO. 4

LINK, H. 1981. Zur situation des habichts im Frankischen polamento di Passeriformidi un bosco di conifere e Raum im Vergleich zu anderen mitteleuropaischen di uno di latifogliedell'Italia centrale.Pages 375-378 populationen. Okol.Vogel. 3:221-226. in M. Fasola [ED.], Atfl II Seminario Italiano Censi- MA•OSA, S. 1994. Goshawk diet in a mediterranean area menti Faunisticidei Vertebrafl. Suppl.[tic. Bid. Selvag- of northeasternSpain. J. RaptorRes. 28:84-92. gina, XVI:l-820. , J. REALAND E. SANCHEZ.1990. Comparaci6 de PIELOWSKI,Z. 1991. Liczebnoscpopulacji i efektyleg6w l'ecologiade duespoblaciones d'astor Accipitergentilis ptak6w drapieznychw krajobrazie rolniczym okolic a Catalunya:el Vall•s Moian•s i la Segarra.El Medi Czempinia (zach. Polska). Acta Ornithol.26:107-118. Natural clel Vallgs 2:204-212. REYNOLDS,R.T. ANDE.C. MESLOW.1984. Partitioningof MARQUISS,M. AND I. NEWTON. 1982. The Goshawk in food and niche characteristicsof coexistingAccipiter Britain. Br. Birds 75:243-260. during breeding. Auk 101:761-779. MINISTERO DEI LAVORI PUBBLICI. 1984-93. Annali idrol- RUST,V.R. ANDW. KECHELE.1982. Ergebnisselangj/ih- ogici. Sezione autonoma del genio Civile, Bacini con riger Bestandkontrollenbeim habicht (Accipitergentil- foce al litorale tirrenico dal salinello al fortore. Serv- is) auf zwei sfidbayerischenKontrollfl/ichen. Garmisch- izio Idrografico. er vogelkd.B• 11:10-26. MORILLO,C. ANDJ.LALANDA. 1972. Primerosdatos sobre SULKAVA,S. 1964. Zur nahrungsbiologiedes habichts, la ecologiade las Falconiformesen los montesde To- A ccipitergentilis (L.). Aqu ilo, S• Zool.3:1-103. ledo. Bol. Estac. Cent. Ecol. 2:57-70. THIOLLAY,J.M. 1967. Ecologied'une populationde ra- NEWTON I., M. MARQUISS,D.N. WEIR AND D. MOSS. 1977. paces diurnes en Lorraine. TerreVie 21:116-183. Spacingof Sparrowhawknesting territories.J. Anita. TJERNBERG,M. 1985. Spacingof Aquila Ecol. 46:425-441. chrysaetosnests in relation to nest site and food avail- NORE,T. 1977. L'Autour et l'epervier en Limousin.Or- ability.Ibis 127: 250-255. nithol. Limousin 7:25-41. WIDI•N,P. 1989. The hunting habitatsof GoshawksAccip- OLECH,B. 1996. Anthropogenicfood of the Goshawkas itergentilisin boreal forestsof central Sweden.Ibis 131: a factor influencingthe breedingphenology. Page 37 205-231. in M. Pandolfi [ED.], 2nd International Conference WIKMAN,H. 1977. Duvhokspredationpa skogsfageli sy- on Raptors, Urbino, Italy. dastraFinland 1975-1976.Foredrag fra Nordisksko- OPDAM,P. 1975. Inter- and intraspecificdifferentiation gefuglsymposium:59-72. Direktoratetfor vilt of fersk with respectto feedingecology in two sympatricspe- vannfisk,Viltrapport 5. cies of the genus Accipit• Ardea63:30-54. WILKE M., A. GOTTMANN AND H.G. SCHNEIDER. 1985. --, J. THISSEN, P. VERSCHURENAND G. Mf3SKENS. Beutetieredes habichts (Accipiter gentilis) auf drei nor- 1977. Feedingecology of a populationof Goshawk dhessischenUntersuchungsfl/ichen. Vogelkd. 11:69-77. Accipitergentills. J. Ornithol.118:35-51. PANDOLFI,M. ANDG. TAFERNA.1991. Censimentodel po- Received8 March 1997;accepted 10 August1997