(ISSN 0892-1016) Th e Journal of Raptor Research Volume 31 March 1997 Number 1 Contents Peregrine Falcon Recovery Along the West Central Coast of the Baja CALIFORNIA Peninsula, Mexico. Aradit Castellanos, Fernando Jaramillo, Federico Salinas, Alfredo Ortega-Rubio and Cerafina Arguelles 1 Winter Bald Eagle Distribution is Inversely Correlated with Human Activity Along the Colorado River, Arizona. Bryan t. Brown and Lawrence e. Stevens 7 Population Status of the Endangered Hawaiian Hawk. Linnea s. Hall, Michael L. Morrison and Peter H. Bloom 11 Nest-Site Selection by Four Sympatric Forest Raptors in Southern Norway. Vidar Selas 16 Distribution and Species Richness of a Forest Raptor Community in Relation to Urbanization. Thomas Bosakowski and Dwight g. Smith 26 Does Vegetation Structure Limit the Distribution of Northern Goshawks in the Oregon Coast Ranges? Stephen DeStefano and Jon McCioskey. ... 34 Food Habits of the Lanner Falcon (Falco biarmicus feldeggii) in Central Italy. Federico Morimando, Francesco Pezzo and Alessandro Draghi 40 Nesting Distribution and Population Status of U.S. Ospre\s 1994. Lawrence M. Houghton and Larry M. Rymon..., 44 The Delaware Bayshore of New Jersey: A Raptor Migration and Wintering Site of Hemispheric Significance, clay Sutton and Paul Keriinger 54 Food Habits of Common Barn-owls along an Elevational Gradient in Andean Argentine Patagonia. Alejandro Travaini, Jose A. Donazar, Olga Ceballos, Alejandro Rodriguez, Fernando Hiraldo and Miguel Delibes 59 Indicators of Male Quality in the Hoots of Tawny Owls (Strix aluco) . Bridget M. Appleby and Stephen M. Redpath 65 Role of Refuse as Food for Migrant, Floater and Breeding Black Kites ( MlLVUS MIGRANS ) . Guillermo Blanco 71 Short Communications Interspecific and Intraspecific Aggression Among Grifton and Cinereous Vultures at Nesting and Foraging Sites. Guillermo Blanco, Jose M. Traverse, Javier Marchamalo and Felix Martinez 77 Hunting Synchrony in White-tailed Kites. Molly F. Skonieczny and Jeffrey R. Dunk 79 Blood Parasites of Nestling Goshawks. E.P. Toyne and R.W. Ashford 81 Migration of Flocks of Honey Buzzards in Southern Italy and Malta. Nicolantonio Agostini, Daniela Logozzo and Charles Colero 84 Nesting-Tree Preference and Nesting Success of Japanese Lesser Sparrowhawks in Japan. Mutsuyuki Ueta 86 Letters 89 BOOK Review. Edited by Jeffrey S. Marks 90 Manuscript Referees 91 The Raptor Research Foundation, Inc. gratefully acknowledges a grant and logistical support provided by Boise State University to assist in the publication of the journal. Persons interested in predatory birds are invited to join The Raptor Research Foundation, Inc. Send requests for information concerning membership, subscriptions, special publications, or change of address to OSNA, P.O. Box 1897, Lawrence, KS 66044-8897, U.S.A. The Journal of Raptor Research (ISSN 0892-1016) is published quarterly and available to individuals for $24.00 per year and to libraries and institutions for $30.00 per year from The Raptor Research Foundation, Inc., 14377 117th Street South, Hastings, Minnesota 55033, U.SA (Add $3 for destinations outside of the continental United States.) Second class postage paid at Hastings, Minnesota, and additional mailing offices. POSTMASTER; Send address changes to The Journal of Raptor Research, OSNA, P.O. Box 1897, Lawrence, KS 66044-8897, U.SA Printed by Allen Press, Inc., Lawrence, Kansas, U.SA Copyright 1997 by The Raptor Research Foundation, Inc. Printed in U.SA. © This paper meets the requirements of ANSI/NISO Z39 .48-1 992 (Permanence of Paper). THE JOURNAL OF RAPTOR RESEARCH A QUARTERLY PUBLICATION OF THE RAPTOR RESEARCH FOUNDATION, INC. Vol. 31 March 1997 No. 1 J. Raptor Res. 31(l):l-6 © 1997 The Raptor Research Foundation, Inc. PEREGRINE FALCON RECOVERY ALONG THE WEST CENTRAL COAST OF THE BAJA CALIFORNIA PENINSULA, MEXICO Arad n Castellanos Centro de Investigaciones Biologicas del Noroeste, S.C., Div. De Biol. Terrestre., Apdo. Post. 128, La Paz , B.C.S., Mexico Fernando Jaramillo PG7 Consultores, S.C., Apdo. Post. 1887, Cuernavaca, Mor., Mexico Federico Salinas, Alfredo Ortega-Rubio and Cerafina Arguelles Centro de Investigaciones Biologicas del Noroeste, S.C., Div. de Biol. Terrestre., Apdo. Post. 128, La Paz, B.C.S., Mexico Abstract. — The central west coast of the Baja California peninsula was an important Peregrine Falcon (Falco peregrinus) breeding area supporting a population of about 13 breeding pairs. This population declined drastically during the 1960s and early 1970s. We conducted field surveys and compiled data on nesting Peregrine pairs from 1980-94 to address the current status of the Baja population. We found 10 pairs nesting in the area indicating the Peregrine population has recovered in the area since the late 1970s. Due to increased human activity in the area, proper management is needed to provide suitable nesting sites and to minimize human disturbances during the nesting season. Keywords: Peregrine Falcon; recovery; status ; Baja California; Falco peregrinus. Recuperation del halcon peregrino en la costa centro occidental de la penisula de Baja California, Mexico. Resumen. — La costa centro occidental de la peninsula de Baja California, ha sido un area importante de reproduction del halcon peregrino (Falco peregrinus) , manteniendo alrededor de 13 territorios de anidacion. Esta poblacion declino drasticamente entre los anos 1960s e inicios de los 1970s. Nosotros realizamos estudios en el campo y recabamos datos sobre las parejas anidantes de peregrinos entre 1980 y 1994 para determinar su situation actual. Encontramos diez parejas de peregrinos anidando en el area. Nuestros hallazgos sugieren que los peregrinos se han estado recuperando desde finales de los anos 1970s. Las actividades humanas estan creciendo en el area, lo que representa tanto amenazas como oportunidades para los peregrinos. Se requiere de un manejo apropiado que proveea de sitios adecuados de anidamiento para minimizar las perdidas reproductivas asociadas al hombre durante la estacion reproductiva. [Traduction Autores] Peregrine Falcon ( Falco peregrinus) recovery has been documented in several regions of the world (Fyfe 1988, Kiff 1988, Ratcliffe 1993, En- derson et al. 1995). Nesting territories, deserted during the decline of the species are now being reoccupied by breeding pairs. Reproduction has now returned to pre-DDT levels (Kiff 1988, New- ton 1979, Ratcliffe 1993). The ban on DDT use in many countries (Ratcliffe 1993) and the suc- cess of Peregrine recovery programs (Fyfe 1988, Ratcliffe 1993) has contributed to the Peregrine’s removal from endangered status and to a redesign 1 2 Castellanos et al. Vol. 31, No. 1 North Brccdsyipara Smgt* Prob«Wy netting Inland potential 28 * 00 ' 26 * 30 - Figure 1. Recent records of nesting Peregrine Falcons on the central west coast of Baja California, Mexico. (Shaded areas are Banks, 1969 records.) of its management strategies (Walton and Thelan- der 1988). The historical and current status of the Mexican Peregrine Falcon population is not well known (Banks 1969, Fyfe et al. 1976). Populations in the Baja California peninsula and the Gulf of Califor- nia are known to have declined between 1967-84 (Porter et al. 1988). By 1976, the Baja west coast population, once containing 38 breeding pairs, was considered on the brink of extirpation (Fyfe et al. 1976). However, recent reports of newly occupied nesting territories in this area (Daneman and Guz- man Poo 1992, Castellanos et al. 1994) suggest a recovery is in process. In this report, we provide information on current numbers, distribution, and productivity of Peregrines, and discuss their con- servation needs on the west central coast of the Baja California peninsula. Study Area and Methods The study area, a part of El Vizcaino biosphere reserve, covers about 350 km along the west coast of Baja Cali- fornia, Mexico (Fig. 1). The study area includes: desert islands, a coastal fringe of mostly sandy beaches and two big coastal lagoons with extensive wetlands (Massey and Palacios 1994). Most of the area is covered with halo- phytic vegetation, 30 cm high. Mountains are restricted to Benitos and Cedros Islands and some parts of the El Vizcaino reserve. We used three data sources on the breeding popula- tion of Peregrine Falcons in the area. Data from 1980- 85 were obtained by F. Jaramillo during wildlife invento- ries to promote the creation of the El Vizcaino Reserve and in 1994 on Cedros and Benitos Islands. Data from Ojo de Liebre and San Ignacio lagoons were obtained by March 1997 Peregrine Recovery in Baja California 3 A. Castellanos, F. Salinas, C. Arguelles and A. Ortega in 1993-94. Nesting surveys of breeding Peregrine Falcons were conducted by searching the plateaus, canyons, hills and islands on foot and inspecting all potential nesting sites using binoculars. Boats were used to explore the lagoons and to reach the islands. We also used recent records in the oological collection of the Western Foun- dation of Vertebrate Zoology (WTVZ) and personal re- ports from other individual observers. We compared our data with those of Banks (1969). Productivity of breeding pairs was estimated by checking three nests in 1993 and five nests in 1994. Results and Discussion We located at least 10 nesting pairs of Peregrine Falcons in six locations in the survey area, includ- ing small islands in the open sea, islets inside coast- al lagoons and an inland mountainous zone (Fig. 1). The maximum number of pairs seen in a single year was six pairs in 1994 when we observed pairs on San Roque islands, Laguna Ojo de Liebre (Pie- dras Island and El Chaparrito navigation channel) and San Ignacio Lagoon (Pelicano and Ballena Is- lands) (Fig. 1, Table 1). Because not all the histor- ical and potential nesting sites were surveyed, we considered this to be a minimum estimate of tire total number of breeding pairs in the area. The historical population estimate for the study area was about 13 pairs (Banks 1969). Breeding pairs were found on small cliffs, di- rectly on the ground, in Osprey nests and on arti- ficial structures such as channel markers and ship- wrecks. Ojo de Liebre (Scammon’s) lagoon nest sites were new (Castellanos et al. 1994). Lack of inland nesting records from the study area led us to believe that the cliff-nesting pairs on Pacific North Vizcaino (near Sierra San Jose de Castro) were also new to the breeding population. A nest- ing record in the WTVZ reported by B. Reitherman in 1981 documented a pair nesting at San Ignacio Lagoon for the first time. Daneman and Guzman Poo (1992) also found two pairs nesting there in 1989. In addition to the 10 nesting pairs, there were several recent single sightings and two probable nest site reports from Benitos and Cedros islands (Fig. 1, Table 1). Nearest nesting site distances var- ied between 4.4 and 60 km. Nests were 4.4 km apart in San Ignacio lagoon and 8 and 22 km apart in Ojo de Liebre lagoon (Fig. 1) We followed the reproductive success of three pairs during 1993 and five pairs during 1994, lo- cated at Ojo de Liebre and San Ignacio Lagoons. An average of 3 eggs, 1.8 nestlings, and 1.6 fledg- lings were produced per nest (Table 2). This pro- ductivity appeared to be within the range of pro- ductivity for healthy populations (Cade et al. 1988, Ratcliffe 1993). The Peregrine Falcon is defined by Mexican law as a vulnerable species (Diario Oficial 16 de Mayo de 1994). In spite of its status, lack of national re- covery programs on threatened species has failed to produce management plans in their Mexican range. The use of organochlorine pesticides in the region is not well documented but it has appar- ently been reduced since the late 1970s. Studies conducted then and in the late 1980s in Baja Cal- ifornia, the Gulf of California and in northwestern Mexican waterfowl wintering areas show pesticide levels in eggs and bird tissues are among the lowest in North America (Spitzer et al. 1977, Mora et al. 1987). This evidence leads us to believe that cur- rent organochlorine pesticide use does not repre- sent a major threat for Peregrines in the region. We did, however, identify other threats to breed- ing Peregrines in the area. Nests on the ground are accessible to dogs and cats abandoned by fish- ermen in the islands. Coyotes ( Canis latrans ) tem- porarily invade Piedras island in Ojo de Liebre La- goon destroying the nests of Peregrines, Ospreys and other birds. W 7 e found coyote tracks at one failed Peregrine nest in 1994 and presumed the coyote caused the nest failure. Avian depredation by Western Gulls ( Larus occidentalis) and Common Ravens ( Corvus corax ) is also a problem. Gulls and ravens apparently caused the disappearance of three eggs from a nest on a tower over the water in Ojo de Liebre lagoon. A third and more serious threat to nesting Per- egrines in the area is from humans when fishing, tourism and other human activities cause inciden- tal disturbances at nests (Daneman and Guzman Poo 1992). The study area is still relatively inacces- sible and partially uninhabited, thus, habitat de- struction apparently is not yet a factor of concern. The coastal lagoons, islands and wetlands within the study area support a magnificent avifauna which provides adequate numbers and variety of Peregrine prey (Massey and Palacios 1993). Nev- ertheless, the density of breeding Peregrine Fal- cons in the area seems to be rather low for such an ideal location. Most of the resident pairs re- ported are coastal or “small island Peregrines” (Ratcliffe 1993) nesting on small cliffs or even on the ground. This is apparently because of a short- age of suitable natural nesting sites on the main- 4 Castellanos et al. Vol. 31, No. 1 Table 1. Historical and recent records of nesting Peregrine Falcons in the central west coast of Baja California. Recent records are from 1980-94. Locality Sources and Date Individual/Nest Reference San Benitos Islands Walker, 1927 and 1950 4 nests Banks (1969) Tyler, 8 April 1930 Nest with 2 eggs WFVZ, this study Harrison, 1 April 1938 Nest with 2 eggs WFVZ, this study 22 March 1992 1 seen This study Cedros Island Kaeding, 1905 Common, breeding Banks (1969) Carpenter, 2 April 1932 Nest with 4 eggs WFVZ, this study 8, 31 July and 18 August 1980 1 seen This study April 1991 1 seen This study 2, 3 and 15 March 1992 1 pair, 2 sightings This study 22 March 1992 1 female, 1 male This study Natividad Island Sechrist, 12 March 1917 Nest with 4 eggs WFVZ, this study Lamb, December 1924 6 pairs resident Banks (1969) D.S.D., 6 April 1930 Nest with 3 eggs WFVZ, this study Bancroft, 3 April 1932 2 nest with eggs WFVZ, this study 31 July 1985 1 breeding pair This study Ojo de Liebre Lagoon 26 January 1982 1 seen This study June, October, November 1984 1 seen This study 29 July 1985 2 females and 2 males This study 21 September 1985 1 seen This study 1993 (Castellanos et al. 1994) 3 breeding pair This study San Roque Island Huey, 20 April 1927 Nest with 2 eggs Banks (1969) D.S.D., 5 April 1930 Nest with 3 eggs WFVZ Harrison, 6 April 1932 Nest with 4 eggs WFVZ, this study L. Flores, 1994 1 breeding pair L. Flores pers. comm. Asuncion Island Walker, 1938 1 pair Banks (1969) San Ignacio Lagoon Reitherman, 11 April 1981 Nest, 2 young, 2 eggs WFVZ, this study September 1985 1 pair, one seen This study Daneman & Guzman Poo, 1992 2 breeding pairs Daneman & Guzman Poo, 1992 March-June, 1994 2 breeding pair This study Pacific North Vizcaino 13 June, 1984 1 breeding pair This study June 1984 1 breeding pair This study 4 November 1984 2 seen This study March and September 1985 2 seen This study March 1997 Peregrine Recovery in Baja California 5 Table 2. Peregrine Falcon reproductive success during 1993-1994 on the Scammon’s and San Ignacio lagoons. 1993 1994 Scam- mon’s Lagoon Scam- mon’s Lagoon San Ignacio Lagoon Occupied nests checked 3 3 2 Productive nests 3 1 1 Total eggs laid 10 10 4 Eggs failed to hatch 1 — — Eggs disappeared — 6 a — Eggs broken — l b l b Average clutch size 3.3 3.3 2.0 Nestlings 9 3 3 Nestling disappeared 1 1 Total young fledged 8 2 3 Nesdings/ occupied nest 3 1.0 1.5 Fledglings/ occupied nest 2.6 0.6 1.5 Fledglings/productive nest 2.6 2.0 3.0 a Probably avian and coyote depredation. b Unknown causes. land coast, inland and on small islands, a factor that limits cliff-nesting raptors in other areas (Rat- cliffe 1993). Peregrines do take advantage of suit- able artificial nesting sites when they were available in the study area (Castellanos et al. 1994) Given these circumstances, we suggest the imple- mentation of a Peregrine Conservation Program in the El Vizcaino Reserve, focusing on the following: a) regular surveys of historical and new nesting ter- ritories to assess the current population status and trends; b) studies of biology and nesting ecology to enhance knowledge of this poorly known resi- dent population; c) installation of artificial nesting substrates to increase the population on the main- land coast and at coastal lagoons; and d) an edu- cation program to promote public awareness of this and other species of wild birds. Acknowledgments We thank L. Kiff for permitting us to use the Western Foundation of Vertebrate Zoology oological collection, M. Acevedo for his field assistance and L. Flores for un- published data on Peregrines. The research was support- ed by the CIBNOR, S.C., CONACYT (Project 3031N), Ex- portadora de Sal, S.A. and the SEMARNAP. Thanks to W. Wehtje, M.F. Campana for his academic support and R. Bowers and Dr. Ellis Glazier for editing of this English language paper. Literature Cited Banks, R.C. 1969. The Peregrine Falcon in Baja Califor- nia and the Gulf of California. Pages 81-91 in JJ- Hickey [Ed.], Peregrine Falcon populations. Umv. Wisconsin Press, Madison, WI U.S.A. Cade, T., J. Enderson, C. Thelander and C.M. White. 1988. The role of organochlorine pesticides in Pere- grine population changes. Pages 463-468 in T. Cade, J. Enderson, C. Thelander and C. White [Eds.], Per- egrine Falcon populations: their management and re- covery. The Peregrine Fund, Boise, ID U.S.A. Castellanos, A., F. Salinas-Zavala and A. Ortega-Ru- bio. 1994. Status and reproduction of the Peregrine Falcon at a coastal lagoon in Baja California Sur, Mex- ico. J. Raptor Res. 28:11 0-1 1 2. Daneman, G.D. and J. GuzmAn Poo. 1992. Notes on the birds of San Ignacio Lagoon, Baja California Sur, Mexico. Western Birds 23:11-19. Diario Oficial de la Federacion. 1994. Tomo CDLXXXIX, No. 10, 16 de Mayo de 1994. Secretaria de Gobernacion. Enderson, J.H., W. Heinrich, L. Kiff and C.M. White. 1995. Populations changes in North American Pere- grines. Pages 142-161 in Trans. 60th North American wildlife and natural resources conference. Wildlife Management Institute, Washington, DC U.S.A. Fyfe, R.W. 1988. The Canadian Peregrine Falcon recov- ery program 1967-1985. Pages 599-610 in T. Cade,J. Enderson, C. Thelander and C. White [Eds.], Pere- grine Falcon populations: their management and re- covery. The Peregrine Fund, Boise ID U.S.A. Fyfe, R.W., S.A. Temple and T.J. Cade. 1976. The 1975 North American Peregrine Falcon survey. Can. Field Nat. 90:228-273. Kiff, L.F. 1988. Changes in the status of the Peregrine in North America: an overview. Pages 123-139 in T Cade, J. Enderson, C. Thelander and C. White [Eds.], Peregrine Falcon populations: their management and recovery. The Peregrine Fund, Boise, ID U.S.A. Massey, B.W. and E. Palacios. 1994. Avifauna of the wet- lands of Baja California Mexico: current status. Studies in Avian Biology 15:45-57. Mora, M.A., D.W. Anderson and M.E. Mount. 1987. Seasonal variation of body condition and organochlo- rines in wild ducks from California and Mexico J. Wildl. Manage. 51 (1):132— 141. Newton, I. 1979. Population ecology of raptors. Buteo Books, Vermillion, SD U.S.A. Porter, R.D., M.A. Jenkins, M.N. Kirven, D.W. Anderson and J.O. Keith. 1988. Status and reproductive per- formance of marine Peregrines in Baja California and the Gulf of California, Mexico. Pages 105-114 in T. Cade, J. Enderson, C. Thelander and C. White [Eds.], Peregrine falcon populations: their management and recovery. The Peregrine Fund, Boise, ID U.S.A. Ratcliffe, D. 1993. The Peregrine Falcon. Academic Press, San Diego, CA U.S.A. Spitzer, P.R., R.W. Risebrough, J.W. Grier and C.R. Sin- delar,Jr. 1977. Eggshell thickness-pollutant relation- ships among North American Ospreys. Pages 13-19 6 Castellanos et al. Vol. 31, No. 1 in J.C. Ogden [Ed.] Trans. North American Osprey research conference. U.S. Nat. Park Serv. Proc. Ser. 2, Washington, DC U.S.A. Walton, B.J. and C.G. Thelander. 1988. Peregrine Fal- con management efforts in California, Oregon, Wash- ington and Nevada. Pages 587-597 in T. Cade, J. En- derson, C. Thelander and C. White [Eds.], Peregrine Falcon populations: their management and recovery. The Peregrine Fund, Boise, ID U.S. A. Received 28 February 1996; accepted 28 November 1996 J Raptor Res. 31(1):7-10 © 1997 The Raptor Research Foundation, Inc. WINTER BALD EAGLE DISTRIBUTION IS INVERSELY CORRELATED WITH HUMAN ACTIVITY ALONG THE COLORADO RIVER, ARIZONA Bryan T. Brown SWCA, Inc., Environmental Consultants, 56 West 400 South, Suite 201, Salt Lake City, UT 84101 U.S.A. Lawrence E. Stevens Applied Technology Associates, Inc., P.O. Box 22459, Flagstaff, AZ 86002 U.S.A. Abstract. — Helicopter surveys for Bald Eagles ( Haliaeetus leucocephalus) were conducted along the Col- orado River through Glen Canyon National Recreation Area and Grand Canyon National Park, Arizona, U.S.A., during winter 1990-91. Eagle abundance and distribution were examined for a possible corre- lation with human activity levels as documented in National Park Service recreational use reports. Twen- ty-two times more eagles were detected in river reaches with low human use compared to river reaches with high to moderate human use. Eagle distribution did not correspond to prey abundance, biomass patterns, or habitat conditions frequently associated with eagle foraging habitat. Moderate to high levels of human activity may have been responsible for lower eagle abundance in some reaches of the river, reinforcing the need for continued management of some areas as refugia where species sensitive to human disturbance can be protected from higher levels of human activity. Key Words: Bald Eagle, Colorado River, Arizona; Haliaeetus leucocephalus; human disturbance. La distribution invierna de el Haliaeetus leucocephalus esta correlacionada inverso con actividad humana enseguida del Rio Colorado, Arizona. Resumen. — Inspecciones de helicoptero para Haliaeetus leucocephalus fueron conducidas a lo largo del Rio Colorado por el area Nacional de Recreation del Glen Canyon y el Parque Nacional del Grand Canyon, Arizona, U.S.A., durante el invierno de 1990-91. La abundancia y distribution del H. leucoce- phalus fueron examinados por una posible correlation con niveles de actividad humana como fueron documentado en reportes de el Servicio Nacional de Parques. Vieinte-dos veces mas fueron los H. leucocephalus descubiertos en tramos del no con uso bajo de humanidad comparado con tramos del no con uso alto y moderado. La distribution de H. leucocephalus no corresponds con la presa abundante, patron de distribution de materia biologica, y condiciones de habitat frecuentemente asociada con habitat de forraje de H. leucocephalus. Niveles altos a moderado de actividad humana puede ser res- ponsable por la baja cantidad de H. leucocephalus en tramos de el rio, reforzando la necesidad para la continuation de administration de unos area de refugio donde especie delicadas a molestos humanos pueden ser protegidos de niveles alto de actividad humana. [Traduction de Raul De La Garza, Jr.] Human activities can influence Bald Eagle {Hal- iaeetus leucocephalus ) behavior and distribution (Stalmaster 1987). Numerous studies have shown that eagle distribution and foraging behavior can be detrimentally affected by unmanaged human recreational activities such as hiking (Stalmaster and Newman 1978), fishing (Knight et al. 1991, Skagen et al. 1991), and boating (Knight and Knight 1984, McGarigal et al. 1991), and by phys- ical developments within eagle habitat (Buehler et al. 1991). These aspects of eagle-human interac- tion have been addressed primarily through two types of studies: those directly quantifying eagle “flush response distance” to different categories of human activity (e.g. Stalmaster and Newman 1978, McGarigal et al. 1991), and those indirectly linking eagle distribution to secondary measures of human activity such as housing density and development set-back distance (e.g., Buehler et al. 1991). The purpose of our study was to test the null 7 8 Brown and Stevens Vol. 31, No. 1 Figure 1 . Map of the Colorado River study area in Glen Canyon National Recreation Area and Grand Canyon National Park, Arizona, showing River Reaches One through Five. hypothesis that winter distribution of Bald Eagles along the Colorado River in Arizona was not cor- related with indirect measures of human activity obtained from National Park Service (NPS) recre- ational use reports. We also considered physical variables and prey differences that may influence eagle distribution. Study Area and Methods Our study was conducted along the Colorado River through Glen Canyon National Recreation Area (GCNRA) and Grand Canyon National Park (GCNP) in northern Arizona, from 2 km downstream of Glen Can- yon Dam to the confluence of the Little Colorado River (120.2 km; Fig. 1). The completion of Glen Canyon Dam in 1963 altered the river’s physical and biological char- acteristics, and a subsequent introduction of rainbow trout ( Oncorhynchus mykiss ) benefited wintering Bald Ea- gles (Brown et al. 1989, Brown and Stevens 1992, Brown 1993). A concentration of wintering eagles began to oc- cur during the 1980s at Nankoweap Creek, 105 km down- stream from the dam, when large numbers of spawning trout were present (Brown et al. 1989). The study area was divided into five river reaches based on bedrock geology (modified after Schmidt and Graf 1990: Table 1). Reaches varied in geomorphic width, but all were in deep canyons bounded by cliffs typically >300 m high. Reach One, from Glen Canyon Dam to the Paria River, supported a productive rainbow trout fishery and was heavily used in winter by bank and boat fishermen from Lees Ferry. Boating activity upstream of Lees Ferry consisted of motorboats used by fishermen; downstream of Lees Ferry, motorized and nonmotorized rafts were on multi-day trips through GCNP. Reach Two, from the Par- ia River to Soap Creek Rapid, was moderately used by bank fishermen and hikers that accessed it from Lees Ferry or by several trails. The number of raft trips de- parting downstream to pass through GCNP was limited Table 1. Habitat features, relative human activity levels, and prey parameters potentially influencing winter Bald Eagle distribution along the Colorado River in Glen and Grand canyons, Arizona. Parameter River Reach 1 2 3 4 5 Length (km) 22.6 16.7 18.5 28.2 34.2 Surface 3 (ha/km) 12.6 9.3 5.3 5.7 9.2 % riffle /rapid 3 5.7 4.3 13.9 15.8 11.4 Boats/mo b 671 30 19 19 19 Fishermen / mo c 570 350 10 10 10 Activity level High Moderate Low Low Low Fish d (kg/ha) 190 53 69 79 82 Water fowl c 6344 157 101 35 21 a Unpubl. data, U.S. Bureau of Reclamation. b Data provided by NPS; Reach 2 estimated. c Data provided by NPS; Reaches 2-5 estimated. d Unpubl. biomass data, 1990-1994, from R.A. Valdez and R.J. Ryel, BIO/WEST, Inc., Salt Lake City, UT, U.S.A. e Number winter waterfowl detected/km 2 /hr of observation (Unpubl. data). March 1997 Eagle Distribution and Human Activity 9 Table 2. Numbers of Bald Eagles detected and expected by river reach in winter 1990-1991 along the Colorado River in Glen and Grand canyons, Arizona. Numbers/reach are presented two ways: the entire study period and that time when no spawning trout were in Nankoweap Creek in Reach Five. River Reach Parameter 1 2 3 4 5 Entire study period No. detected 3 3 36 58 43 No. expected 28.5 19.0 22.6 30.3 42.6 Detections/km 0.13 0.18 1.95 2.06 1.26 No spawning trout No. detected 3 2 20 41 31 No. expected 19.3 12.9 15.3 20.6 28.9 Detections/km 0.13 0.12 1.08 1.45 0.91 by NPS regulations. The community of Marble Canyon was immediately adjacent to the river where Navajo Bridge on U.S. Highway 89 crossed Reach Two. Reaches Three, Four, and Five extended from Soap Creek Rapid to the Little Colorado River and were remote, far from roads, and accessed only by a few foot-trails, with the number of bank fishermen and hikers limited by NPS regulations. Nankoweap Creek was located in Reach Five. Fifteen helicopter surveys of the study area were con- ducted, 1/wk from November 1990-March 1991, to de- termine eagle abundance and distribution. A special re- search permit was obtained from NPS for helicopter use because of the potential for helicopters to disturb park resources and visitors. Most surveys occurred before 1200 H. Air speed was approximately 90 km/hr at a height of 100 m above river level on a flight path directly over the river. Surveys were initiated from both ends of the study area. The pilot and front-seat observer detected eagles; a rear-seat assistant recorded observations and localities. Numbers of eagles detected were summarized by reach and compared to expected numbers of eagles/reach us- ing a Chi-square goodness-of-fit test, with statistical sig- nificance accepted at P < 0.05. Expected numbers of ea- gles/reach were derived by multiplying the total number of eagles detected by the proportion of total river-km/ reach. Winter boat launches/mo and bank fishermen/ mo for Reach One from November 1990-March 1991 were ob- tained from recreational records maintained by NPS at GCNRA; monthly totals were averaged to obtain mean use. Winter raft launches/mo for Reach Two were ob- tained from records maintained by NPS at GCNP. Reach- es Two through Five had almost identical levels of boat- ing activity since all raft trips launching at Lees Ferry must pass through them; Reach Two received slightly more overall boating use because of motorboat activity by NPS maintenance and patrol trips. Mean numbers of bank fishermen/mo for Reaches Two through Five were estimated based on interviews with NPS personnel and our personal observations. Results and Discussion Twenty-two times more eagles were detected in Reaches Three through Five than in Reaches One and Two when total numbers of eagles detected were compared (Table 2; x 2 ~ 44.0, df = 4, P < 0.001); we found no difference in the number of eagles detected among Reaches Three through Five (x 2 = 5.4, df = 2, P = 0.07). Total numbers of eagles detected/km in Reaches Three through Five were more than 10 times greater than num- bers of eagles detected in Reaches One and Two (Table 2). Eagle distribution did not correspond to esti- mates of prey distribution. Fewer eagles were ob- served in Reaches One and Two despite biomass indices that indicated fish and winter waterfowl abundance to be greater in Reaches One and Two (Table 1 ) . Reaches One and Two were wider with many pools with shallow margins all along the riv- er, They also had a smaller percentage of riffle and rapid habitats making them more suitable for ea- gles as foraging areas (Hunt et al. 1992). It appeared that negative effects due to moder- ate to high levels of human activity in these two reaches of the river may have reduced their suit- ability as eagle foraging areas or perhaps disturbed eagles from perching and roosting habitat. We did not find any eagles within 1 km of intensively used areas near Lees Ferry and Navajo Bridge at Marble Canyon. This negative correlation between human activity and Bald Eagle distribution was consistent with that reported in other studies (Buehler et al. 1991, McGarigal et al. 1991) , although reasons why eagles avoid areas of higher human activity remain 10 Brown and Stevens Vol. 31, No. 1 unknown (Buehler et al. 1991). Repeated flushing by bank fishermen, hikers, or boats could have caused wintering eagles to avoid those reaches heavily used by anglers. McGarigal et al. (1991) re- ported a similar finding on the Columbia River where eagles avoided areas with heavy boat traffic. Although apparent habituation of Bald Eagles to human activity has been reported (Knight and Knight 1984), scarcity of eagles in Reaches One and Two of the Colorado River suggested that ha- bituation did not occur in this area. An alternative hypothesis was that more eagles occurred in Reaches Three through Five because they were attracted to easily accessible spawning trout in Nankoweap Creek (Reach Five) . Abundant trout were spawning at Nankoweap from 15 Feb- ruary until the end of the study period (Brown and Stevens 1992). To test this hypothesis, we contrast- ed numbers of eagles detected/reach when spawn- ing trout were and were not present at Nankoweap. Prior to 15 February, total numbers of eagles de- tected in Reaches Three through Five were 18 times greater than in Reaches One and Two (Table 2; x 2 ~ 27.3, df = 4, P < 0.001). Therefore, we rejected the alternative hypothesis. Managed boating, hiking, and fishing activities are an established use of national parklands. How- ever, we found an inverse correlation between win- tering Bald Eagle distribution and human activity along the Colorado River within GCNRA and GCNP. Although correlation does not prove cau- sation, we suggest human activity as the most likely cause. This would represent an apparent contra- diction of NPS goals to preserve natural resources for future generations. A tradeoff of resources for recreation may be acceptable if the status quo can be maintained in Reaches Three through Five of our study area. However, the growing popularity of national parks such as GCNP increases economic and political pressures to provide more wilderness recreation opportunities (Frome 1992), and may result in long-term increases in human activity lev- els in Reaches Three through Five. Our findings reinforce the need to continue to manage some areas as refugia where species potentially sensitive to human disturbance, such as Bald Eagles, are not excluded or influenced by higher levels of human activity. Acknowledgments This study was funded by the Glen Canyon Environ- mental Studies (GCES), U.S. Bureau of Reclamation (USBR), as part of a study to determine the influence of fluctuating flows from Glen Canyon Dam on wintering Bald Eagles. We thank D. Wegner and the staff of GCES for support, and GCNRA and GCNP (particularly S. Cherry of the GCNP river office) for providing recrea- tional activity records. N. Kline, M. Sogge, T. Yates and others conducted or assisted with helicopter surveys. We thank pilots B. Barnes, B. Caldewood, S. Chubbuck and M. Santee of USBR. K. Buck and M. Yard of GCES, and E. Pemberton and T. Randle of USBR provided unpub- lished data on river surface area and percent riffle/rapid habitat. R. Valdez and R. Ryel provided unpublished data on fish biomass. Drafts of this manuscript were reviewed and improved by G. Bortolotti, R. Glinski, T. Grubb, R. McClelland and C. White. Literature Cited Brown, B.T., R. Mesta, L.E. Stevens, and J. Weisheit. 1989, Changes in winter distribution of Bald Eagles along the Colorado River in Grand Canyon, Arizona, /. Raptor Res. 23:110-113. AND L.E. Stevens. 1992. Winter abundance, age structure, and distribution of Bald Eagles along the Colorado River, Arizona. Southwest. Nat. 37:404—408. . 1993. Winter foraging ecology of Bald Eagles in Arizona. Condor 95:132-138. Buehler, D.A., TJ. Mersmann, J.D. Fraser and J.K.D. Seegar. 1991. Effects of human activity on Bald Eagle distribution on the northern Chesapeake Bay./. Wildl. Manage. 55:282-290. Frome, M. 1992. Regreening the national parks. Univ. of Ar izona Press, Tucson, AZ U.S.A. Hunt, W.G., J.M. Jenkins, R.E. Jackman, C.G. Thelander and A.T. Gerstell. 1992. Foraging ecology of Bald Eagles on a regulated river./. Raptor Res. 26:243-256. Knight, R.L. and S.K. Knight. 1984. Responses of win- tering Bald Eagles to boating activity. /. Wildl. Man- age. 48 :999— 1 004, , D.P. Anderson and N.V. Marr. 1991. Responses of an avian scavenging guild to anglers. Biol. Conserv. 56:195-205. McGarigal, KL, R.G. Anthony .and F.B. Isaacs. 1991. In- teractions of humans and Bald Eagles on the Colum- bia River estuary. Wildl. Monogr. 115:1-47. Schmidt, J.C. andJ.B. Graf. 1990. Aggradation and deg- radation of alluvial-sand deposits, 1965 to 1986, Col- orado River, Grand Canyon National Park, Arizona. U.S. Geol. Survey Prof. Pap. 1493, U.S. Govt. Print. Off., Washington, DC U.S. A. Skagen, S.K., R.L. Knight and G.H. Orians. 1991. Hu- man disturbance of an avian scavenging guild. Ecolog- ical Applications 1:215-225. Stalmaster, M.V. 1987. The Bald Eagle. Universe Press, New York, NYU.S.A. and J.R. Newman. 1978. Behavioral responses of wintering Bald Eagles to human activity. / Wildl. Man- age. 42:506-513. Received 1 April 1996; accepted 21 October 1996 J. Raptor Res. 31 (1) :1 1 — 15 © 1997 The Raptor Research Foundation, Inc. POPULATION STATUS OF THE ENDANGERED HAWAIIAN HAWK Linnea S. Hall and Michael L. Morrison Department of Biological Sciences , California State University, Sacramento, CA 95819 U.S.A. Peter H. Bloom Western Foundation of Vertebrate Zoology, 439 Calk San Pablo, Camarillo, CA 93010 U.S.A. Abstract. — We assessed the current abundance and distribution of Hawaiian Hawks (‘io; Buteo solitarius) on the island of Hawaii to determine if this federally endangered bird should be downlisted to threat- ened status. We found a density of 0.004 hawks/ha on the island. Using an estimate of 400 000 ha of suitable ‘io habitat on Hawaii, we estimated a total of 1600 hawks (1120 adults; 560 pairs) on Hawaii, Based on the wide distribution of ‘io among vegetation types on the island and litde apparent change in numbers during the past decade, we agreed with the recommendation for downlisting the hawk but suggested that researchers collect long-term demographic data to better understand the status of this species. Key Words: Buteo solitarius, Hawaiian hawk, ‘io, population status. El estado de poblacion del Buteo solitaruis en peligro de extincion. Resumen. — Nosotros fijamos la cantidad corriente y distribucion de Buteo solitarius en la isla de Hawaii para determinar si el pajaro en peligro de extincion por leyes federales debe ser reducido a estado amenazado. Nosotros los encontramos una densidad de 0.004 halcon/ha en la isla. Usando la estimacion de 400 000 ha de habitat conveniente en Hawaii, nosotros estimamos un total de 1600 halcones (1120 adultos; 560 parejas) en Hawaii. En base de la distribucion amplia de B. solitarius entre clases de vege- tacion en la isla y poco cambio aparente en la cantidad durante la decada pasada, nosotros estamos de acuerdo con la recomendacion para reducir el halcon pero sugerimos que los investigadores junten datos demograficos de larga duracion para poder entender el estado de este especie mejor. The Hawaiian Hawk ( Buteo solitarius), or ‘io, was federally listed as an Endangered Species in 1967 (37 FR 4001, 11 March 1967) based on its restrict- ed range on the island of Hawaii (hereafter Ha- waii), its low numbers at the time of listing (Berger 1981), and the perceived threats to its preferred habitat from agricultural and commercial devel- opments (U.S. Fish and Wildlife Service [USFWS] 1984) . At the time of listing, no intensive study of the ecology of the ‘io had ever been conducted, and anecdotal accounts gave differing reports on its abundance across the island (Munro 1944, USFWS 1984). Uncertainty over ‘io abundance continued through the next decade. An intensive survey ini- tiated by the USFWS in 1976 on Hawaiian forest birds was unable to estimate the ‘io population size (Scott et al. 1986). After a detailed study of ‘io breeding biology in <1% of the island’s area, Grif- [Traduccion de Raul De La Garza, Jr.] fin (1985) found that the species might be rela- tively unaffected by habitat modifications com- pared to many other native bird species after find- ing that foraging and nesting occurred in agricul- tural areas and in stands of exotic vegetation (Baskett and Griffin 1985). Griffin (1985, 1989) es- timated the population at 900 breeding pairs and a total of 2700 individuals in 1983. Because of this, the USFWS proposed downlisting the ‘io from en- dangered to threatened status (58 FR 41684, 4 Au- gust 1993). Because of questions over the validity of basing such a reclassification on 10-year-old data, the USFWS requested that an island-wide sur- vey be conducted of the ‘io population to obtain a more current estimate of the population size. Herein, we present our survey results and sam- pling design to provide a baseline for future sur- veys designed to monitor the size of the ‘io popu- lation on Hawaii. 11 12 Hall et al. Vol. 31, No. 1 Study Area and Methods The most efficient way to sample dominant vegetation types across Hawaii for the occurrence of ‘io was to con- duct unlimited distance point counts (Blondel et al. 1981) along paved and dirt roads across the island. Point counts were selected to make our methods generally comparable to those of Scott et al. (1986), who conduct- ed the most complete previous census of ‘io on Hawaii as part of the USFWS’s Hawaiian Forest Bird Survey (1976-79). We also needed a method that could sample the ‘io’s use of vegetation ranging from lowland agricul- tural areas to subalpine woodlands (Scott et al. 1986, Griffin 1989), and would be applicable to birds with home range sizes varying from 48 ha in agricultural areas to 490 ha in forests and mid-elevation pasturelands (Bas- kett and Griffin 1985). Use of roadways was the only fea- sible means of satisfying these objectives. Some studies have indicated that roadside counts can give biased esti- mates of bird densities and vegetation associations, but other studies have indicated that road counts can be use- ful and appropriate when large areas need to be sampled and monitored long term (Fuller and Mosher 1981). When possible, we used roads that crisscrossed an area to more thoroughly sample for ‘io. We conducted point counts 0.1-16 km off main roads to ensure that traffic noise did not interfere with the counts and that we more adequately sampled vegetation that could contain ‘io. Count stations were located disproportionately among vegetation types (Table 1), based on information that ‘io were unlikely (or very uncommon) in shrublands (vege- tation type 10), upper-elevation mamane-naio ( Sophora chrysophylla-Myoporum sandwicense) woodlands (vegetation type 12), and exotic pioneering lava vegetation (vegeta- tion type 5) (J. Jeffrey andj. Giffin pers. comm.). All count stations were 0.8-3. 2 km apart, and counts were conducted by 1-2 observers between 0900-1700 H. Each point count lasted for exactly 10 min, which was the same count length used by Scott et al. (1986), and in- cluded 8 min of listening and watching for hawks, plus 2 mm of playback of taped adult territorial and fledgling calls of ‘io. After the first minute of the tape elapsed, we turned it off and observed the area for any hawks for 7 min. Anytime an ‘io responded to the tape, either by calling or flying to the point, we immediately stopped the tape, but continued the count to determine if any addi- tional hawks were observed. We then played the tape again for 1 min, and watched for the last minute of the count. Although surveys have not previously used broadcast calls, Banko (1980) and Baskett and Griffin (1985) re- ported that ‘io call and defend their territories in the winter. On 12-13 December 1993, we tested if broad- casted territorial calls elicited responses from ‘io by going to areas known to have ‘io present (J. Jeffrey pers. comm.). We watched ‘io that were ^200 m away while we played the taped calls. Eighty percent of the hawks responded by taking flight, calling or coming to the tape. No counts were conducted when precipitation exceed- ed a light rain, or when wind exceeded 24 km/h. We recorded the distance from the point of initial detection of all observed hawks, the detection mode (visual, aural or both), the morph (light, dark, unknown), and the veg- etation where the bird was observed (Table 1 ) . Although Table 1. Vegetation descriptions and codes for survey transects used in analyses of Hawaiian Hawk numbers across Hawaii in December 1993. a Code Fre- quency 6 Description 1 63 Sugar cane fields with exotic and/ or native trees or shrubs at edges, as windrows. 2 41 Short or tall exotic trees with exotic shrubs, and sometimes exotic grass- es. 3 22 Macadamia nut or papaya orchard with native and/or exotic trees or shrubs at edges. 4 38 Grassland with scattered exotic and/or native trees (especially o’hia) ; scattered homes. 5 5 Pioneer exotic vegetation growing on lava. 6 68 Native trees and native shrubs occa- sionally with scattered orchard trees, or exotic understory and homes. 7 99 Mixed exotic and native trees, some- times with mixed exotic and native shrubs or grass. 8 24 Residential area with scattered exotic and native vegetation. 9 40 Native tree and mixed exotic and na- tive shrub vegetation on lava, some- times with scattered homes; a pioneer community. 10 2 Mixed exotic and native shrubs with scattered native and exotic trees. 11 16 Native trees and grassland; non-pio- neer community. 12 4 Mamane-naio vegetation, with grass and/or exotic shrub understory, or sometimes with scattered exotic trees. “Scientific names of plants listed: sugar cane ( Saccharum offici- narum), macadamia nut ( Macadamia ternifolia), papaya (Carica pa- paya), o’hia ( Metrosideros polymorpka), mamane ( Sophora chrysa- phylla), and naio (Myoporum sandwicense). b Frequency = total number of times the described vegetation was recorded along survey routes. we tried to determine age and sex of hawks in the field, it was often difficult to make a positive identification, so in our analyses we combined all sightings. We used program DISTANCE (Buckland et al. 1993, Laake et al. 1993) to estimate the densities of ‘io in the 12 major vegetation types recorded during our surveys March 1997 Status of Hawaiian Hawk 13 Table 2. Summary of ’io density estimates (all ages together) by vegetation type, calculated by program DISTANCE (Laake et al. 1993) from survey data across Hawaii, December 1993. a Vegetation CODE b To- tal Ef- fort No. Poi- nts No. To Obs- VD. Esti- ma- tor Mod- el No/ 1 Esti- mate SE Density Estimations 6 %CV 95% Cl df De- tect. Prob. Enc. Rate Density BooTSTRAP f Est. %CV Runs All veg types 399 399 98 5 0.004 0.0007 15.9 0.003-0.006 345 0.02 0.24 0.004 23.7 400 Veg 1 63 63 26 1 0.002 0.0006 29.1 0.001-0.003 86 0.08 0.41 0.002 56.2 100 Veg 2 41 41 5 1 0.0004 0.0003 77.1 0.0001-0.002 12 0.11 0.12 — * Veg 3 22 22 5 1 0.003 0.0016 58.8 0.0009-0.008 22 0.03 0.23 — Veg 4 38 38 16 1 0.004 0.0011 29.1 0.002-0.006 52 0.04 0.42 — Veg 5 5 5 0 h Veg 6 68 68 10 1 0.003 0.0014 41.3 0.002-0.007 76 0.02 0.15 — Veg 7 99 99 22 1 0.005 0.0013 26.6 0.003-0.008 118 0.02 0.22 0.005 67.6 100 Veg 8 24 24 7 1 0.009 0.0046 54.0 0.003-0.024 26 0.01 0.29 — Veg 9 40 40 1 — Veg 10 2 2 0 — Veg 11 16 16 7 1 0.005 0.0024 50.5 0.002-0.013 17 0.03 0.44 — Veg 12 4 4 0 — a For explanation of DISTANCE program estimations, see text. b For vegetation code explanation, see Table 1. c Total effort = the sum of the number of times each point and its corresponding vegetation was sampled. d Estimator Model No. = the mathematical estimator model selected by program DISTANCE to analyze the point data, where the chosen model was the one that had the smallest Akaike’s Information Criterion value (AIC). e Density estimations: Estimate = density in number of ’io/hectare; SE = standard error of the estimation; %CV = percent coefficient of variation of the estimate; 95% Cl = 95% confidence interval for the estimate; df = degrees of freedom used in the analysis; Detect Prob. = the estimate of average probability of detecting an ’io; Enc. rate = the number of animals expected to be observed per point. f Density Bootstrap values: Est. = bootstrapped density estimate; %CV = percent coefficient of variation of this estimate; Runs = total number of bootstrap runs conducted. s “ — ” = too few degrees of freedom to conduct bootstrap analyses. h “ — ” = no or too few ’io observed along this survey route, so no density analysis could be performed. (Table 1). Observations of ‘io were entered as the radial distance to the hawk from the point. We truncated the distances at 3000 m (the maximum distance at which most ‘io were observed) to allow all hawk observations to be entered into the analyses. We instructed the program to select the most appropriate density estimation model for each analysis, based on maximum likelihood ratio tests of the models vs. each other. We also instructed the program to conduct 400 bootstrap samples for the island- wide data, to obtain reliable estimates of the variances around the density estimates, and 100 bootstrap samples for each of the analyses of density by vegetation type. We estimated the current population size of the ‘io on Hawaii based on the density of hawks per vegetation type, and the estimated percent cover by each vegetation type on the island of Hawaii (Jacobi and Scott 1985, Cuddihy and Stone 1990). Results and Discussion We sampled 40 transects across Hawaii, with 399 points covering approximately 500 km of roads. Among these points, 98 different ‘io were ob- served. Thirty-three hawks were identified as adults, 7 as immatures, and 58 as unknown-aged. Forty-five hawks were light morph birds and 14 were dark morphs. Densities ranged from a low of 0 in vegetation types 5, 9, 10, and 12 to a high of 0.009 hawks/ha in vegetation type 8 (Table 2). Most densities were between 0.003 and 0.005 hawks/ha, with an overall mean of 0.004 hawks/ha. Vegetation types 5 and 9 had lava as a major ground component, and thus had poorly-developed tree cover. Type 10 vegeta- tion was dominated by shrubs, and type 12 was in mamane-naio woodland. Type 2 vegetation was typ- ified by exotic trees of various sizes and had very few hawks. Type 8 vegetation consisted of residen- tial areas with both native and exotic tree compo- nents and showed the highest hawk densities. Grasslands with scattered exotic and native trees (vegetation type 4) also had moderately-high den- 14 Hall et al. Vol. 31, No. 1 sities of birds. Bootstrapped density estimates matched the model estimates in all cases where ad- equate degrees of freedom existed. Surveys found ‘io most commonly in areas with native and/or exotic tree cover, usually with un- derstories of exotic grass, and sometimes with na- tive and/or exotic shrub understories. Although ‘io were not found frequently in small patches of mixed native and exotic forest surrounded by open fields or orchards, they were commonly observed over the open areas, or in open places, with scat- tered native and/or exotic trees. For example, in sugar cane fields with ribbons of native or exotic trees between fields, or with trees extending down from higher elevation forests; in open pasture land with scattered native trees; in orchards (especially macadamia nut) with taller native and/or exotic trees at the perimeters. This indicated that ‘io are now using areas that are not pure native forest. Based on these data and anecdotal breeding rec- ords from these more open areas, it appears that they are also able to successfully breed there (J. Jeffrey and J. Giffin unpubl. data). Griffin (1985, 1989) estimated that the popula- tion of ‘io on Hawaii was about 2700 hawks in 1983. Of this, 1800 were adults. This estimate served as the basis for the Hawaiian Hawk Recov- ery Plan developed by the USFWS (1984). It used an abundance of 2,000 hawks (the midpoint be- tween the 1500 and 2500 adult birds thought to be needed for a self-sustaining population) as the tar- get to downlist the species to threatened status. The island-wide estimate of ‘io density was based on a total forested area of 343 000 ha (J.M. Scott pers. comm., Griffin 1989). This value correspond- ed roughly to the potential ‘io habitat contained within the Hawaiian Forest Bird Survey area (Ja- cobi and Scott 1985) . Using this area, and the over- all estimate of ‘io density from our surveys (0.004 birds/ha, 95% Confidence Interval [C.I.] = 0.003- 0.006), we obtained a total density of 1372 ‘io (range = 1029—2058) on Hawaii. Much of the low- land forested areas of Hawaii, including the sugar cane, macadamia nut and other disturbed areas oc- cupied by ‘io, were excluded from Griffin’s area estimate. We therefore modified the Griffin esti- mate by adding 60 000 ha of mixed sugar cane- lowland forested area and various other minor veg- etation types (Cuddihy and Stone 1990), bringing the total potential ‘io habitat to 400 000 ha. This raised our estimate of ‘io on the island to about 1600 birds (range = 1200-2400), with 1120 adults or 560 pairs. Our estimated density of adults (1120) is about 25% below the lower end of the target range nec- essary for a stable ‘io population, according to the Recovery Plan (target = 1500-2500 adult birds). Assuming that all birds alive during our surveys survived to breed, the total number of birds we estimated (1600) is just above the lower end of the target range, but is still below the mean target val- ue of 2000. The target value of adults is not en- closed in the confidence interval around the 0.004/ha value (95% C.I. = 0.003-0.006), but is enclosed if we assume that all hawks alive breed (400 000 ha x 0.006 - 2400 hawks). We found a relatively high number of birds that were widely distributed among vegetation types on Hawaii, including heavily-disturbed areas. In addi- tion, our results were similar to those found 10 years earlier by Griffin (1985, 1989), indicating the likelihood of a relatively stable population during the past decade. Thus, we concluded that down- listing to threatened status was supported. As other biologists have suggested for the ‘io, long-term demographic studies are necessary to ac- curately assess the overall status of the population (USFWS 1984, Griffin 1989). Our fieldwork did not assess population trends, reproductive fecun- dity and success, dispersal or mortality, all of which have been shown to be problematic for other for- est birds on Hawaii (Scott et al. 1986). Thus, we think that the USFWS should initiate a long-term demographic study so future density estimations can be evaluated in light of other population data. Such a study is necessary before delisting from threatened status is considered. Acknowledgments We thank S. Johnston at the USFWS Pacific Islands of- fice in Honolulu, Hawaii; J. Jeffrey at the USFWS office in Hilo; L. Katahira at the NPS office at Hawaii Volcanoes National Park; M. Reynolds, J. Jacobi, and T. Pratt at the National Biological Service office at Hawaii Volcanoes Na- tional Park; J. Giffin at the Forestry and Wildlife office in Kamuela; and all of the associated field assistants in Hawaii for their assistance in establishing transects and conducting surveys. We also thank reviewers of earlier reports and manuscript drafts, including C.R. Griffin, D.E. Anderson, J.M. Scott, F. Duvall II, S. Fancy, and B. Harper for significantly improving our analyses and con- clusions. This research was funded by Cooperative Agree- ment No, 1448-0001-93676 between the USFW 7 S and the Western Foundation of Vertebrate Zoology. March 1997 Status of Hawaiian Hawk 15 Literature Cited Banko, W.E. 1980. Part I. Population histories-species accounts. Forest birds: Hawaiian Hawk ( l io). Natl. Park Serv. CPSU/UH Avian Rep. 6A, History of En- demic Hawaiian Birds. Hawaii Volcanoes National Park, HI U.S.A. Baskett, T.S. and C.R. Griffin. 1985. Final project re- port on the biology of the endangered Hawaiian Hawk: ecology, life history, and environmental pollu- tion problems. Missouri Coop. Wildl. Res. Unit, Univ. Missouri, Columbia, MO U.S.A. Berger, A.J. 1981. Hawaiian birdlife. Univ. Hawaii Press, Honolulu, HI U.S.A. Blondel, J., C. Ferry and B. Frochot. 1981. Point counts with unlimited distance. Pages 414-420 in C.J. Ralph and J.M. Scott [Eds.], Estimating the numbers of terrestrial birds. Stud. Avian Biol No. 6. Buckland, S.T., D.R. Anderson, K.P. Anderson, K.P. Burnham and J.L. Laake. 1993. Distance sampling: estimating abundance of biological populations. Chapman and Hall, London, UK. Cuddihy, L.W. and C.P. Stone. 1990. Alteration of native Hawaiian vegetation: effects of humans, their activities and introductions. Coop. Nad. Park Resour. Stud. Unit, Univ. Hawaii, Honolulu, HI U.S.A. Fuller, M.R. and J.A. Mosher. 1981. Methods of de- tecting and counting raptors: a review. Pages 235-246 in C.J. Ralph and J.M. Scott [Eds.], Estimating the numbers of terrestrial birds. Stud. Avian Biol No. 6. Griffin, C.R. 1985. Biology of the Hawaiian Hawk ( Buteo solitarius). Ph.D. dissertation, Univ. Missouri, Colum- bia, MO U.S.A. . 1989. Raptors in the Hawaiian islands. Pages 155-160 in K. Steenhof, M.N. Kochert and M.N. LeFranc Jr. [Eds.], Proc. Western Raptor Manage. Symp. Workshop. Natl. Wildl. Fed, Sci. Tech. Ser. No. 12, Washington, DC U.S.A. Jacobi, J.D. and J.M. Scott. 1985. An assessment of the current status of native upland habitats and associated endangered species on the island of Hawaii. Pages 3— 22 in C.P. Stone and J.M. Scott [Eds.], Hawaii’s ter- restrial ecosystems: preservation and management. Coop. Nad. Park Resour. Stud. Unit, Univ. Hawaii, Honolulu, HI U.S.A. Laake, J.L., S.T. Buckland, D.R. Anderson and KP. Burnham. 1993. Distance user’s guide, version 2.0. Colorado Coop. Fish and Wildl. Res. Unit, Colorado State Univ., Fort Collins, CO U.S.A. Munro, G.C. 1944, Birds of Hawaii. Tongg Publ. Co., Honolulu, HI U.S.A. Scott, J.M., S. Mountainspring, F.L. Ramsey and C.B. Kepler. 1986. Forest bird communities of the Ha- waiian islands: their dynamics, ecology, and conser- vation. Stud. Avian Biol. No. 9. U.S. Fish and Wildlife Service. 1984. Hawaiian Hawk recovery plan. U.S. Fish and Wildl. Serv., Pordand, OR U.S.A. Received 15 December 1995; accepted 24 October 1996. J. Raptor Res. 31 (1) :16— 25 © 1997 The Raptor Research Foundation, Inc. NEST-SITE SELECTION BY FOUR SYMPATRIC FOREST RAPTORS IN SOUTHERN NORWAY VlDAR SELAS Department of Biology and Nature Conservation, Agricultural University of Norway, P.O. Box 5014, N-1432 As, Norway Abstract. — Differences between 0.1 ha nest-site plots of Honey Buzzards ( Pernis apivorus), European Sparrowhawks ( Accipiter nisus), Northern Goshawks (A. gentilis ) and Common Buzzards ( Buteo buteo ) were compared to randomly sampled 0.1 ha control plots within a 400 km 2 area with 80% forest and <2% agricultural land in southern Norway. At Honey Buzzard nest sites, forests were more productive than in control plots and there was a higher proportion of spruce, older trees and a higher tree density at Northern Goshawk nest sites than in control plots. Nests of European Sparrowhawks were also in sites with higher tree density than expected. Common Buzzard nest sites were situated in steeper terrain than control plots and more often had a southern aspect. For sparrowhawks, nesting in forests with high tree density may be an adaptation to avoid goshawks and pine martens ( Martes martes ) which are their main nest predators. For the larger species, nest-site selection may be a response both to nest predation risk, microclimate, foraging habitat and food supply. Key WORDS: Honey Buzzard ; European Sparrowhawk; Common Buzzard ; Northern Goshawk; Accipiter nisus; Accipiter gentilis; Buteo buteo; Pernis apivorus; forest, nest-site selection; Norway. Seleccion del nido de cuatro rapaces de bosque sin que no aparean en el sur de Norway. RESUMEN. — Diferencias entre 0.1 ha parcela de sitio de nido de Pernis apivorus, Accipter nisus, A. gentilis y Buteo buteo fueron comparados con muestras alazar 0.1 ha parcelas manejadas dentro de una area de 400 km2 con 80% bosque y <2% tierra agricola en el sur de Norway. En nidos de Pernis apivorus, los bosques fueron mas productivo en las parcelas manejadas y habia una proportion alta de Picea, arboles maduros y densidad alto de arboles en nidos de A. nisus tambien estaban en sitios con densidad alta de arboles mas de lo que esperabamos. Nidos de B. buteo estaban situados mas en terreno abrupto que en parcelas manejadas y con frecuencia tenia aspecto del sur. Para A. nisus , nidos en el bosque con densidad alta de arboles puede ser un adaptation para evitar A. gentilis y Martes martes que son su principal depredador de nido. Para la especie mas grande, la seleccion del nido puede ser reaction a riesgo de depredador al nido, microclima, habitat de forraje y suministro de comida. [Traduction de Raul De La Garza, Jr.] Breeding pairs of raptors use relatively large ar- eas, and thus have a good opportunity to select nesting places that maximize the probability of suc- cessful breeding and lifetime reproduction (New- ton 1979). Interspecific differences in nest-site se- lection may be due to differences in body size and flight performance of different species, but it can also be due to interspecific differences in nest pre- dation risk, climatic conditions during breeding and feeding habits (Newton 1979, Janes 1985), or to interspecific competition for nest sites and ter- ritories (Newton 1979). For several bird species, dense foliage close to the nest both reduces the rate of detection, and impedes the ability of predators to hunt in the vi- cinity of the nest (Martin 1993). On the other hand, dense foliage may decrease the possibility for breeding birds to detect and escape from preda- tors (Gotmark et al. 1995). Thus, selection of nest site may be a trade-off between concealment and opportunities to escape or attack predators, which also depend on flight ability, body size or other characteristics of the species. Selection may also be affected by a trade-off between current and future reproduction, since short-lived species with large brood sizes have more to lose when nesting at- tempts fail than long-lived species with smaller brood sizes. Cover may also be an important factor since it can shield nests from wind or rain and limit ex- cessive nocturnal radiation loss or excessive diur- nal heat-gain from solar radiation (Walsberg 1985) . 16 March 1997 Nest-site Selection By Four Raptors 17 Protection from thermal extremes may be the most important factor in nest-site selection by medium- and large-sized raptors where nest predation is low (Newton 1979, Janes 1985). At higher latitudes, the timing of breeding in these birds should be im- portant, since early breeders are faced with more severe climatic conditions than those species which begin nesting later in spring. If prey are not evenly dispersed throughout the landscape, raptors should select nest sites closest to the best hunting areas in order to reduce time and energy connected with foraging. Thus, local varia- tion in the availability of food may influence the nest-site selection (Janes 1985), and explain inter- specific differences in nesting habitat. In Fennoscandia, four raptor species hunt and nest in forest-dominated landscapes. The Europe- an Sparrowhawk ( Accipiter nisus ; mean body mass male 150 g, female 260 g) is the main predator on small birds (Sulkava 1964a, Selas 1993) , while the Northern Goshawk (A. gentilis; mean body mass male 870 g, female 1330 g) primarily feeds upon larger bird species and mammals (Hoglund 1964, Sulkava 1964b, Widen 1987, Selas 1989). The Com- mon Buzzard ( Buteo buteo; mean body mass male 740 g, female 1100 g) is a generalist predator that responds functionally to changes in populations of its vole ( Microtus spp.) prey (Suomus 1952, Spids0 and Selas 1988), while the Honey Buzzard ( Pernis apivorus; mean body mass male 750 g, female 910 g) mainly feeds on the larvae and pupae of social hymenoptera species (Holstein 1944, Flagen and Bakke 1958, Itamies and Mikkola 1972). Several authors have described nest sites used by sparrowhawks (Tinbergen 1946, Holstein 1950, Hald-Mortensen 1974), goshawks (Holstein 1942, Dietzen 1978, Link 1986), Common Buzzards (Holstein 1956, Knuwer and Loske 1980, Solonen 1982, Jedrzejewski et al. 1988, Hubert 1993) and Honey Buzzards (Holstein 1944, Amcoff et al. 1994) in Europe. However, no one has compared nest-site selection of sympatric populations of these species in a continuous forest habitat. My aim was to study the importance of different habitat vari- ables on nest-site selection of these species by com- paring habitat variables from plots at nest sites with those from plots placed randomly in the study area. Study Area and Methods The study was conducted from 1985-93 in southern Norway (58° 43'N, 8°44'E). The study area covers about 400 km 2 and is situated 50—300 m a.s.l. and 10—30 km inland from the coast, in the boreo-nemoral zone (Abra- hamsen et al. 1977). The climate is suboceanic, and snow usually covers the ground from December-April. The study area is hilly and sharply undulating. It is dominated by forests (80%), with scattered lakes (10%), bogs (5%) and less than 2% agricultural land. Forests are character- ized by a fine-grained mosaic of young-, medium- and old-aged coniferous, mixed and deciduous stands, with Scots pine ( Pinus sylvestris), Norway spruce ( Picea abies), oak ( Quercus spp.), aspen ( Populus tremula) and birch ( Betula spp.) the dominant tree species. Forestry based on clear cutting, replanting and thin- ning of the regrowth was introduced to the area in the 1950s. At the time of my study, approximately 30% of the area had been clear-cut, with most regeneration <20 yr. The area is divided into numerous small ownerships with management of forests controlled by each of the land owners. Most of the properties are managed to provide a mosaic of forest types. Thus, there is a heterogeneous environment on a small scale, but a homogeneous, frag- mented environment on a large scale. The study area was searched for nest sites each year (cf. Forsman and Solonen 1984), and habitat variables were described at one nest site in each nesting territory located. If possible, the nest site used in 1988 was select- ed. In territories where the 1988 nest was not found, I usually described the nest site used in 1989. Alternatively, the nest site used closest in time to 1988 or 1989 was described. The breeding density of goshawks increased during the time of the study. To get a larger number of nest sites of this species, I first selected one nest site from each of the nesting territories used since 1985. Then, I selected one nest site in each of the 11 new nesting ter- ritories established during 1986-88, even though these territories substituted five of the existing ones. Since the goshawks in the five old territories had all been illegally shot by game keepers, I regarded the data to be inde- pendent. Thus, I described a total of 48 nest site plots of the sparrowhawk, 30 of the goshawk, 50 of the Common Buzzard, and 21 of the Honey Buzzard. Control plots were described during 1989. Aerial pho- tographs of the study area taken in spring 1989 (scale 1: 15000) were covered by a grid with 100 numbered inter- sections of which two were randomly selected as control plots. Out of 122 selected points, 80 (65.6%) were locat- ed in forests >20 yr old and 25 (20.5%) were in forests <20 yr old (clear-cuts and regrowth), while 9 (7.4%) were on lakes, 4 (3.3%) on bogs, and 4 (3.3%) on agri- cultural land or developed areas. Measurements were made only in control plots in habitats apparently suitable for raptors (i.e., forests >20 yr old, N = 80). Each of the nest site plots and the control plots cov- ered 0.1 ha within a circle with a radius of 17.8 m. In nest site plots, the nest was in the center of the circle. The following habitat variables were used: 1) Site type, determined from the plant community (Kielland-Lund 1981, 1994). Plots dominated by Bar- bilophozio-Pinetum or Vaccinio-Pinetum were classified as sites with poor productivity, plots dominated by Len- cobryo-Pinetum, Eu-Piceetum myrtilletosum, or Populo-Quer- cetum, were classified as sites with intermediate pro- ductivity, and plots dominated by Melico-Piceetum typi- 18 SelAs Vol. 31, No. 1 Table 1. Test results (upper, right) and P-values (lower, left) of correlation analyses of habitat variables from ran- domly-sampled control plots ( N = 80). Categorical variables were tested against each other by use of contingency table analysis (x 2 value given) and against continuous variables by use of Mann-Whitney U-test (two categories, U-value given) or Kruskal-Wallis test (more than two categories, H-value given) , while continuous variables were tested against each other by use of Spearman rank correlation (correlation coefficient given) . Site Type Forest Type Forest Age Tree Density Slope Aspect Altitude Category Site type (3 categories) 40.60 5.78 28.51 9.70 4.97 5.52 Forest type (5 categories) <0.01* 1.79 18.12 22.17 11.25 17.30 Forest age (continuous) 0.06 0.77 -0.26 0.01 770.5 2.79 Tree density (continuous) <0.01* <0.01* 0.02* 0.13 775.5 7.67 Slope (continuous) <0.01* <0.01* 0.91 0.25 778.0 4.00 Aspect (2 categories) 0.08 0.02* 0.81 0.85 0.87 1.30 Altitude category (3 categories) 0.24 0.03* 0.25 0.02* 0.14 0.52 * Statistically significant. cum, Melico-Quercetum, Alno incanae-Prunetum padi or Ulmo glabrae-Tilietum cordatae were classified as sites with the highest productivity. 2) Forest type, defined according to % pine and spruce trees with diameters >7 cm at breast height (DBH, 1.3 m above ground). Pine forest was >50% pine and spruce with pine >67%. Mixed coniferous forest was >50% pine and spruce with pine and spruce <67%. Spruce forest was >50% pine and spruce with spruce >67%. Mixed forest was 25-50% pine and spruce. De- ciduous forest was <25% pine and spruce. 3) Forest age, defined as the mean age of four trees judged to represent the age of all trees with DBH >7 cm. Ages were measured using an increment borer at breast height. 4) Number of trees, regardless of species with DBH >7 cm. Table 2. Results (P-values) from Likelihood-Ratio tests in a logistic regression model with nest-site plots and ran- domly-sampled control plots (N = 80) as responses, and all habitat variables as explanatory variables. R 2 is the pro- portion of variation that is explained by the logistic re- gression model. Sparrow- Common Honey Habitat hawk Goshawk Buzzard Buzzard Variables (N = 48) ( N = 30) (N = 50) (N = 21) Site type 0.58 0.80 0.54 0.004* Forest type 0.38 0.023* 0.23 0.35 Forest age 0.94 0.027* 0.39 0.50 Tree density <0.001* 0.013* 0.15 0.10 Slope 0.82 0.98 <0.001* 0.31 Aspect 0.92 0.06 0.037* 0.05 Altitude category 0.53 0.47 0.20 0.59 R 2 0.90 0.27 0.34 0.47 5) Slope, measured from 0-100°, 6) Aspect, defined as one of two categories: north (1- 100°, 301-400°) or south (101-300°). Nest-site plots and control plots with slopes <5° were omitted. 7) Altitude, defined as three altitude possible categories in relation to the altitude variation within a radius of 1 km from the plot. Plots were assigned to the lower altitude zone if situated in the lower third of the al- titude difference between the lowest and highest point within this area. Middle and upper altitude zones were assigned correspondingly. When considering the randomly-sampled control plots, several of the habitat variables were highly correlated (Table 1). To control for the effect of these correlations when comparing nest site plots and control plots, I used likelihood-ratio tests (SAS 1994) in a logistic regression model, with nest-site plots and control plots as responses and all habitat variables as explanatory variables (cf. Man- ly et al. 1993). Results Site Type and Forest Type. Of sparrowhawk nest- site plots, none were on sites with poor productiv- ity, 66.7% were on intermediate sites and 33.3% in the highest productivity sites. Corresponding val- ues for goshawk nest-site plots were 16.7%, 73.3% and 10.0%; for Common Buzzard 16.0%, 60.0% and 24.0%, and for Honey Buzzard 0.0%, 57.1% and 42.9% compared to 42.5%, 47.5% and 10% for control plots. When controlling for effects of cor- relations between all habitat variables, there was a significant difference between Honey Buzzard nest-site plots and control plots, while the other species did not differ from the control plots (Like- lihood-ratio tests, Table 2). The distribution of goshawk nest-site plots in dif- ferent forest types differed significantly from that * Statistically significant. March 1997 Nest-site Selection By Four Raptors 19 Figure 1. Distribution of nest-site plots of European Sparrowhawks, Northern Goshawks, Common Buzzards and Honey Buzzards, and randomly-selected control plots on different forest types. Forest types were defined according to the frequency of pine and spruce among all trees >7 cm in breast height (1.3 m above ground). The size of each plot is 0.1 ha. of control plots, with a higher proportion of nest sites in spruce forests (Table 2, Fig. 1). Sparrow- hawk, Common Buzzard and Honey Buzzard nest- site plots did not differ from control plots with re- spect to forest type when effects of correlations be- tween habitat variables were adjusted for (Likeli- hood-ratio tests, Table 2, Fig. 1). Forest Age and Tree Density. The mean forest age was 36.8 ± 18.5 (SD) yr in nest-site plots of sparrowhawks, 99.3 ± 19.1 yr in those of goshawks, 20 Selas Vol. 31, No. 1 Figure 2. Forest age and number of trees (>7 cm in breast height) in nest-site plots of European Sparrowhawks, Northern Goshawks, Common Buzzards and Honey Buzzards (solid squares), and in randomly-selected control plots (open squares). The size of each plot is 0.1 ha. 98.5 ± 20.4 yr in those of Common Buzzards and 86.7 ± 28.3 yr in those of Honey Buzzards. When using likelihood-ratio tests, only nest sites of gos- hawks differed significantly from control plots, where the mean forest age was 90.7 ± 29.0 yr (Ta- ble 2, Fig. 2) . The mean number of trees was 190.4 ± 47.7 in nest-site plots of sparrowhawks, 84.9 ± 27.7 in those of goshawks, 73.3 ± 19.9 in those of Com- mon Buzzards and 86.4 ± 33.4 in those of Honey Buzzards. The number of trees in nest-site plots of sparrowhawks and goshawks was significantly high- er than in control plots, where the mean number of trees was 62.3 ± 26.3 (Likelihood-ratio tests, Ta- ble 2, Fig. 2). Topographical Variables. The mean slope was 8.6 ± 5.6° in nest- site plots of sparrowhawks, 15.1 ± 9.1° in those of goshawks, 28.6 ± 13.3° in those of Common Buzzards, 16.3 ± 7.2° in those of Hon- ey Buzzard and 15.0 ± 10.5° in control plots. Only nest sites of Common Buzzards differed signifi- cantly from control plots (Likelihood-ratio tests, Table 2, Fig. 3). Sparrowhawk nest sites were on south-facing slopes 34.1% of the time while 63,3%, 76.0%, 33.3% and 53.8% of goshawk, Common Buzzard, Honey Buzzard, and control plots were on south- facing slopes, respectively. Only nest sites of Com- mon Buzzards were on south-facing significantly more than control plots (Likelihood-ratio tests, Ta- ble 2, Fig. 3). None of the nest sites of the four raptor species differed significantly from control plots in terms of their altitude (Likelihood-ratio tests, Table 2, Fig. 4). Discussion Site Type and Forest Type. Only the Honey Buz- zard showed a significant preference for nesting in sites with the highest productivity. This finding agreed with that of Amcoff et al. (1994). Unlike oth- er raptor species, Honey Buzzard males do not pro- vision females with food during the egg-laying and incubation period (Holstein 1944), possibly because their prey are too small to be profitably transported to the nest. Because of small size of its prey, short March 1997 Nest-site Selection By Four Raptors 21 Figure 3. Slope (0-100°) and aspect (NE, SE, SW, NW) of nest-site plots of European Sparrowhawks, Northern Goshawks, Common Buzzards and Honey Buzzards. The distance from the origin reflects the slope while the direction reflects the aspect of the plot. The size of each plot is 0.1 ha. distances between nesting and foraging areas during incubation may be especially valuable for this spe- cies. Highly productive forests may be important for Honey Buzzards because they support high densities of juvenile passerines (Tiainen 1981, Helle 1985, Stokland 1994) which appear to be important prey in early stages of the breeding season (Amcoff et al. 1994). These forests also support high biomass of invertebrates (Birkemoe 1993, Stokland 1994) on which Honey Buzzards may also rely. Preference for forest type was significant only for the goshawk, which selected spruce forest for nest- ing. This preference may be related to the larger number of important winter and spring prey spe- cies such as squirrels ( Sciurus vulgaris, Andren and Delin 1994), Hazel Grouse ( Bonasa bonasia, Swen- son and Angelstam 1993), and Capercaillie ( Tetrao urogallus, Swenson and Angelstam 1993) in spruce forest. However, preference for spruce may simply be related to the fact that it gives the best cover and thus the best protection against the main pred- ator of the goshawk, the Eagle Owl ( Bubo bubo, Ut- tendorfer 1952, Mikkola 1983). Forest Age and Tree Density. Goshawk nest sites were situated in older forests than control plots. Old forest is an important hunting habitat for the goshawk (Widen 1989) and it provides large trees for nest building (Dietzen 1978, Anonymous 1989, Siders and Kennedy 1996, Squires and Ruggiero 1996). Goshawk nests were also found in forests with a higher tree density than control plots. Gos- hawks may reduce the risk of predation by nesting in dense forests, since Eagle Owls prefer to hunt in open or semi-open landscapes (Mikkola 1983). 22 SelAs Vol. 31, No. 1 100 -t Lower zone co 0) 4 —* "(0 o c +J c o o 1— © Q_ 75 H Sparrowhawk Goshawk Common Honey n = 48 n = 30 Buzzard Buzzard n = 50 n = 21 E!!!] Middle zone 1ZI Upper zone Control Plots n = 80 Figure 4. The distribution of nest-site plots of European Sparrowhawks, Northern Goshawks, Common Buzzards and Honey Buzzards, and of randomly-selected control plots by altitude categories. A plot was assigned to the lower altitude zone if situated in the lower third of the altitude difference between the lowest and highest point within 1 km from the plot, to the middle zone if situated in the middle third of this altitude difference, and to the upper zone if situated in the higher third. For the smallest species investigated, the spar- rowhawk, the only variable that discriminated be- tween nest-site plots and control plots was tree den- sity. I obtained similar results after a thinning ex- periment, where the reuse of nest stands in thinned young forests was lower than of nest stands in young forests not thinned (Selas 1996). Place- ment of nests in dense forest could hardly be prof- itable with respect to food of sparrowhawks, be- cause the density of passerines is usually low here (Haapanen 1965, 0degaard 1982, Glowacinski and Weiner 1983, Helle 1985). Probably, predation is the most important aspect in the nest-site selection of sparrowhawks (Selas 1996), since its main pred- ators, goshawks and pine martens ( Martes martes), both prefer mature forest rather than young, dense forest when hunting (Pulliainen 1984, Widen 1989, Storch et al. 1990) . Actually, dense forest seems to be less important as nesting habitat for the spar- rowhawk when the goshawk is absent (Bomholt 1983, Newton 1986, T0mmeraas 1994). Pine mar- tens will probably find raptor nests easy, because of the smell from pellets and prey remains. Since the pine marten is also known to remember dif- ferent sites of food resources (Sonerud 1985), it is likely to be familiar with most of the old raptor nests within its home range. This may be one rea- son for why sparrowhawks rarely use nests for two successive years, unlike goshawks and Common Buzzards which are probably less vulnerable to pine marten predation due to their large size. Goshawks and pine martens are also important predators of Honey Buzzard nestlings (Kostrzewa 1991). The Honey Buzzard seems to prefer spruce, which gives best cover, as nest trees (Amcoff et al. 1994). In contrast to the other species studied, Honey Buzzards are usually silent when disturbed by humans at the nest site (Holstein 1944, Hagen 1952, Kostrzewa 1985). Rather than selecting nest- ing habitats to avoid nest predation, Honey Buz- zards appear to behave as cryptically as possible at the nest site, possibly because they are less efficient than other raptors in defending their nests against predators. In addition, low annual mortality and low clutch size of the Honey Buzzard (Holstein 1944, Kostrzewa 1985, Tjernberg and Ryttman 1994) may make nest defense less profitable than for Common Buzzards and goshawks. Topographical Variables. The only species which showed any preference for slope was the Common Buzzard, which usually nested in steep terrain. Sim- ilar results have been found for the Red-tailed Hawk ( Buteo jamaicensis, Titus and Mosher 1981, Speiser and Bosakowski 1988). Flight energetics may be more favorable on steeper slopes for larger soaring raptors like eagles and large Buteos (Speis- er and Bosakowski 1988). It may however also be important that these broad-winged species can best escape, or attack, predators in this habitat. Even though the Common Buzzard is able to rob prey from the goshawk (Fischer 1980, Jorgensen 1983), its breeding success has been found to be nega- March 1997 Nest-site Selection By Four Raptors 23 tively correlated with the distance to goshawk nests (Kostrzewa 1991). One reason for the difference in nest-site selection between the Common Buz- zard and goshawk may be that the goshawk, which is better adapted for flight and foraging in dense forest, is more dangerous to Common Buzzard in dense forest. Common Buzzards also preferred nest sites with southern aspects. There was also a tendency for a higher percentage of nest-site plots of goshawks to have southern aspects than expected, while those of Honey Buzzards tended to have northern as- pects. Common Buzzards and goshawks start their breeding nearly one month earlier than sparro- whawks and more than one month earlier than Honey Buzzards (Forsman and Solonen 1984), at a time of the year when the temperatures may still be far below freezing in southern Norway. Nests of both species w T ere most often found at sites with a southeastern aspect, which are the first heated by the morning sun when nest building occurs (Hol- stein 1942, 1956). Also in Alaska, goshawks have been found to favor southern slopes (McGowan 1975), while in more temperate areas, southern ex- posures are avoided (Dietzen 1978, Reynolds et al. 1982, Moore and Henny 1983, Link 1986, Speiser and Bosakowski 1987). A similar pattern has been observed for nest sites of Golden Eagles ( Aquila chrysaetos, Mosher and White 1976, Pfaff 1993). In Central Europe, Common Buzzards place their nest near forest edges (Knuwer and Loske 1980, Spitzer 1980, Hubert 1993), probably be- cause they hunt from perches in open areas or from forest edges (Widen 1994). In my study area, open areas were usually covered by snow when Common Buzzards arrived to breed, making the field vole ( Microtus agrestis ), which is the most im- portant prey species in this habitat (Hansson 1978, Spids0 and Selas 1988), nearly unavailable (Hans- son 1982, Sonerud 1986). Early snow-free areas available for vole hunting in the spring are found on southfacing slopes and in steep terrain, where Common Buzzard nests are usually found. Goshawks, Common Buzzards and Honey Buz- zards rarely used nest sites with southwestern as- pect, possibly because too much sun may be harm- ful to newly-hatched nesdings (c.f. Holstein 1942, Hald-Mortensen 1974, Reynolds et al. 1982, Link 1986). Unless there is good shelter, as in the dense young forests used by sparrowhawks, nest sites with a southwestern aspect are probably unprofitable re- gardless of when egg-laying begins. The observed interspecific differences in nest- site selection between the raptor species investigat- ed may be explained by interspecific differences in body size and flight performance, nest-predation risk, time of breeding and feeding habits. The risk of predation probably affects nest-site selection or breeding habits of all these species, but mostly sparrowhawks and Honey Buzzards which were most vulnerable to nest predation. Common Buz- zards, goshawks, and Honey Buzzards also showed nest-site preferences which could be explained as an adaptation to microclimate. For these three spe- cies, nest-site selection could also be connected to the availability of food in the early stage of the breeding season. These species may have a broader habitat choice and it is possible that factors other than the habitat variables I selected for study may have been of importance. This may have been es- pecially true for the goshawk, which builds larger nests than the other species and may be influenced by characters directly connected to the nest tree. Acknowledgments I wish to thank I. Selas and K.O. Selas for assistance with the field work, G.A. Sonerud for valuable advice dur- ing the analyses of the data, and T. Bosakowski, S.M. Brai- nerd, S. Dale, G.A. Sonerud and two anonymous referees for constructive comments on the manuscript. The study was supported by the Nansen Endowment. Literature Cited Abrahamsen, J., N. K. Jakobsen, E. Dahl, R. Kalliola, L. Wilborg and L. PAhlson. 1977. Naturgeografisk regioninndeling av Norden. Nordiske utredninger B 1977, 34, Stockholm, Sweden. Amcoff, M., M. Tjern- berg and A. Berg. 1994. Nest site choice of Honey Buzzard Pernis apivorus. Ornis Svecica 4:145—158. Anonymous. 1989. Goshawk breeding habitat in lowland Britain. British Birds 82:56-67. Andren, H. and A. Delin. 1994. Habitat selection in the Eurasian red squirrel, Sciurus vulgaris, in relation to forest fragmentation. Qikos 70:43—48. Birkemoe, T. 1993. Distribution of ground dwelling bee- tles along the main vegetation gradient in an old growth boreal forest. Cand. Scient. thesis, Univ. of Oslo, Norway. Bomholt, P. 1983. Population trends in Danish raptors since 1970. Proc. Third Nordic Congr. Omithol. 1981:39- 44. Dietzen, W. 1978. Der brutbiotop des habichts Accipiter gentilis in drei Gebieten Bayerns, Anz. orn. Ges. Bayern 17:141-159. Fischer, W. 1980. Die Habicht. Neue Brehm-Bucherei, Wittenberg, Lutherstadt. Forsman, D. and T. Solonen. 1984. Censusing breeding 24 Selas Vol. 31, No. 1 raptors in southern Finland: methods and results. Ann. Zool. Fenn. 21:317-320. Glowacinski, Z. and J. Weiner. 1983. Successional trends in energetics of forest bird communities. Ho- larct. Ecol. 6: 305—314. Gotmark, F., D. Blomqvist, O. C. Johansson and J. Bergkvist. 1995. Nest site selection: a trade-off be- tween concealment and view of the surroundings? /. Avian Biol. 26:305-312. Haapanen, A. 1965. Birdfauna of the Finnish forest in relation to forest succession. Ann. Zool. Fenn. 2:153— 196. Hagen, Y. 1952. Rovfuglene og viltpleien. Gyldendal, Oslo, Norway. and A. Bakke. 1958. The food of some Honey Buzzards ( Pemis apivorus (L)) in Norway. Papers of the Norwegian State Game Research, Ser. No. 2. Hald-Mortensen, P. 1974. Nest and nest site of the spar- rowhawk Accipiter nisus — and some comparisons with the Goshawk A. gentilis. Dansk Orn. Foren. Tidsskr. 68: 91-115. Hansson, L. 1978. Small mammal abundance in relation to environmental variables in three Swedish forest phases. Stud. Forest. Suec., No 147. . 1982. Daggdjurens vinterutnyttjande av kanter mellan hyggen och barrskog. Fauna och Flora 77:301- 308. Helle, P. 1985. Habitat selection of breeding birds in relation to forest succession in Northeastern Finland. Ornis Fenn. 62: 1 1 3-1 23. Hoglund, N.H. 1964. Der Habicht Acdpiter gentilis Linn in Fennoskandia. Viltrevy 2:195-270. HOLSTEIN, V. 1942. Duehogen Accipiter gentilis dubius (Sparrman). Biologiske studier over danske rovfugle I. Hirschsprungs, Copenhagen, Denmark. . 1944. Hvepsevaagen Pernis apivorus apivorus (L.) . Biologiske studier over danske rovfugle II. Hirschs- prungs, Copenhagen, Denmark. . 1950. Spurvehpgen Accipiter nisus nisus (L.). Biologiske studier over danske rovfugle III. Hirschs- prungs, Copenhagen, Denmark. . 1956. Musvaagen Buteo buteo buteo (L.). Biologis- ke studier over danske rovfugle IV. Hirschsprungs, Copenhagen, Denmark. Hubert, C. 1993. Nest-site habitat selected by Common Buzzard ( Buteo buteo) in southwestern France./. Raptor Res. 27:102-105. Itamies, J. and H. Mikkola. 1972. The diet of Honey Buzzards Pernis apivorus in Finland. Ornis Fenn. 49:7— 10 . Janes, S. W. 1985. Habitat selection in raptorial birds. Pages 159-184 in M.L. Cody [Ed.], Habitat selection in birds. Academic Press, London, UK. Jedrzejewski, W., B. Jedrzejewski and M. Keller. 1988. Nest site selection by the buzzard Buteo buteo L. in the extensive forest of eastern Poland. Biol. Cons. 43:145- 158. Jorgensen, J. 1983. Buzzard, Buteo buteo, forcing gos- hawk, Accipiter gentilis, to drop prey. Calidris 12:35. Kielland-Lund, J. 1981. Die Waldgesellschaften SO-N- orwegens. Phytocoenologia 9:53—250. . 1994. Syn taxonomy of Norwegian forest vege- tation 1993. Phytocoenologia 24:299—310. Knuwer, H. and K.H. Loske. 1980. Zur frage der habi- tat-anspriiche des Mausebussards ( Buteo buteo) bei der horstplatzwahl. Vogelwelt 101:18—30. Kostrzewa, A. 1985. Zur Biologie des wespenbussards ( Pernis apivorus) in Teilen der Niederrheinishen Bucht mit besonderen Anmerkungen zur Methodik bei Greifvogeluntersuchungen. Okol. Vogel 7:1 13-1 34. . 1991. Interspecific interference competition in three European raptor species. Ethology, Ecology & Evo- lution 3:127-143. Link, H. 1986. Untersuchungen am Habicht ( Accipiter gentilis). Deutscher Falkenorden — Schriftenreihe, Heft 2. Manly, B.F.J., L.L. McDonald and D.L. Thomas. 1993. Resource selection by animals. Statistical design and analysis for field studies. Chapman & Hall, London, UK. Martin, T.E. 1993. Nest predation and nest sites. New perspectives on old patterns. BioScience 43:523-532. McGowan, J.D. 1975. Nesting habits and reproductive success of goshawks in interior Alaska. Raptor Res. Rep. No. 3:147-152. Mikkola, H. 1983. Owls of Europe. T. & A.D. Poyser, London, UK. Moore, KR. and C.J. Henny. 1983. Nest-site character- istics of three coexisting Accipiter Hawks in northeast- ern Oregon./. Raptor Res. 17:65-76. Mosher, J.A. and C.M. White. 1976. Directional expo- sure of Golden Eagle nests. Can. FieJA-Nat. 90:356-359. Newton, I. 1979. Population Ecology of Raptors. T. & A.D. Poyser, London, UK . 1986. The sparrowhawk. T. & A.D. Poyser, Lon- don, UK 0degaard, R. 1982. Spurvefuglsamfunnet i suksesjonss- tadier blabaer-bregnegranskog (Eu-Piceetum) i Sor- Norge. Cand. Scient. thesis, Univ. of Oslo, Norway. Pfaff, A. 1993. Bestandsstprrelse, reproduksjon og naeringsvalg hos kongeprn Aquila chrysaetos i Aust-Ag- der. Cand. Scient. thesis, Univ. of Oslo, Norway. Pulliainen, E. 1984. Use of the home range by pine martens ( Martes martes). Acta Zool. Fennica 171:271- 274. Reynolds, R.T., E.C. Meslow and H.M. Wight. 1982. Nesting habitat of coexisting Accipiter in Oregon. / Wildl. Manage. 46:124-138. SAS Institute, Inc. 1994. JMP statistics and graphics guide, Version 3 of JMP. Cary, NC U.S.A. SelAs, V. 1989, Prey selection in the goshawk during the breeding season. Fauna 42:104—110. . 1993. Selection of avian prey by breeding spar- rowhawks Acdpiter nisus in southern Norway: the im- March 1997 Nest-site Selection By Four Raptors 25 portance of size and foraging behaviour of prey. Ornis Fenn. 70:144-154. . 1996. Selection and reuse of nest stands by spar- rowhawks Accipiter nisus in relation to natural and ma- nipulated variation in tree density. J. Avian Biol. 27: 56-62. Siders, M.S. and P.L. Kennedy. 1996. Forest structural characteristics of Accipiter nesting habitat: is there an allometric relationship? Condor 98:123-132. Solonen, T. 1982. Nest sites of the Common Buzzard Buteo buteo in Finland. Ornis Fenn. 59:191—192. Sonerud, G.A. 1985. Nest hole shift in Tengmalm’s Owl Aegolius funereus as defense against nest predation in- volving long-term memory in the predator. J. Anim. Ecol. 54:179-192. . 1986. Effect of snow cover on seasonal changes in diet, habitat, and regional distribution of raptors that prey on small mammals in boreal zones of Fen- noscandia. Holarct. Ecol. 9:33-47. Speiser, R. and T. Bosakowski. 1987. Nest-site selection by Northern Goshawks in northern New Jersey and southeastern New York. Condor 89:387-394. and . 1988. Nest-site preferences of Red- tailed Hawks in the highlands of southeastern New York and northern New Jersey. J. Field Ornithol. 59: 361-368. SPIDS0, T.K. and V. S et .A s. 1988. Prey selection and breeding success in the Common Buzzard Buteo buteo in relation to small rodent cycles in southern Norway. Fauna Norv. Ser. C, Cinclus 11:61-66. Spitzer, G. 1980. Zum Dispersionsmuster und dessen Bedeutung in der biologie des mausebussards ( Buteo buteo). Eine basisstudie zu einem modell seiner po- pulationsdynamik. Z. Jagdwiss. 26:11-23. Squires, J.R. and L.F. Ruggiero. 1996. Nest-site prefer- ence of northern goshawks in southcentral Wyoming. J. Wildl. Manage. 60:170-177. Stokland, J.N. 1994. Biological diversity and conserva- tion strategies in Scandinavian boreal forests. Dr. Scient. thesis, Univ. of Oslo, Norway. Storch, I., E. Lindstrom and J. Dejoung. 1990. Diet and habitat selection of the pine marten in relation to competition with the red fox. Acta Ther. 35:311-320. Sulkava, P. 1964a. On the behaviour and food habits of the sparrowhawk ( Accipiter nisus ) during the nesting season. Suomen Riista 17:93-105. . 1964b. Zur nahrungsbiologie des habichts Accip- iter gentilis. Aquilo, Ser. Zoologica 3:1-103. Suomus, H. 1952. Studies on the feeding habits of the Common Buzzard. Riistat. Julk. 8:121-127. Swenson, J.E. and P. Angelstam. 1993. Habitat separa- tion by sympatric forest grouse in Fennoscandia in relation to boreal forest succession. Can. J. Zool. 71. 1303-1310. Tiainen, J. 1981. Habitat preferences by breeding forest passerines in an area in Finland. Proc. Second Nordic Congr. Ornithol. 1979:38-42. Tinbergen, L. 1946. De sperver als roofvijand van za- ngvogels. Ardea 34:1-213. Titus, K. and J.A. Mosher. 1981. Nest-site habitat se- lected by woodland hawks in the central Appalachi- ans. Auk 98:270-281. Tjernberg, M. AND H. Ryttman. 1994. Survival and pop- ulation development of the Honey Buzzard Pernis api- vorus in Sweden. Ornis Svecica 4:133-139. T0MMERAAS, P.J. 1994. The Goshawk Accipiter gentilis m Leksvik — a sacrifice to modern forestry. Fauna 46. 180-195. Uttendorfer, O. 1952. Neue Ergebnisse fiber die Er- nahrung der Greifvogel und Eulen. Eugen Ulmer, Stuttgart, Germany. Walsberg, G.E. 1985. Physiological consequences of mi- crohabitat selection. Pages 389-413 in M.L. Cody [Ed.], Habitat selection in birds. Academic Press, London, UK Widen, P. 1987. Goshawk predation during winter, spring and summer in a boreal forest area of central Sweden. Holarct. Ecol. 10:104^109. . 1989. The hunting habitats of Goshawks Accipiter gentilis in boreal forests of central Sweden. Ibis 131: 205-213. . 1994. Habitat quality for raptors: a field exper- iment. J. Avian Biol. 25:219-223. Received 18 June 1996; accepted 1 December 1996 f. Raptor Res. 31(l):26-33 © 1997 The Raptor Research Foundation, Inc. DISTRIBUTION AND SPECIES RICHNESS OF A FOREST RAPTOR COMMUNITY IN RELATION TO URBANIZATION Thomas Bosakowski 1 Department of Biological Sciences, Rutgers University, Newark, NJ 07102 U.S.A. Dwight G. Smith Biology Department, Southern Connecticut State University, New Haven, CT 06515 U.S.A. Abstract. — We studied the species richness and distribution of the forest raptor community in a New Jersey watershed in relation to urbanization. Raptors were systematically surveyed using high volume broadcasts of conspecific and heterospecific calls during the breeding season at a total of 81 survey stations. Ten habitat variables relevant to urbanization were measured at each survey station using to- pographic maps and aerial photographs. Results showed a community composed of 10 species of breed- ing raptors. Buteo lineatus, Accipiter gentilis and Strix varia showed a significant avoidance of suburban habitat, whereas B. jamaicensis and Bubo virginianus had a greater tendency to occupy such areas. Lowland habitat was significantly selected by S. varia, B. lineatus and A. cooperii, a habitat usually most susceptible to development in the study region. Raptor species richness showed a strong positive correlation (r = 0.79, P < 0.01) with wilderness area size. No wilderness area less than 1000 ha had more than four raptor species while four to eight species were found in areas from 1000-8000 ha. Utilization of three increasing size classes of wilderness areas showed increasing trends for B. lineatus, A. gentilis and S. varia, and decreasing trends for B. jamaicensis and Bubo virginianus. Key WORDS: forest raptors ; community; urbanization; forest fragmentation; wilderness', survey. Distribucion y fertilidad especie de una comunidad bosque de rapaces en relacion a urbanizacion. Resumen. — Nosotros estudiamos la fertilidad de la especie y distribucion de una comunidad de rapaces de bosque en una linea divisoria de dos cuencas en New Jersey en relacion a urbanizacion, Los rapaces fueron inspeccionados sistematicamente usando llamadas de conespecificos y heteroespecificos trans- mitidas en alto volumen durante la temporada de cria en un total de 81 estaciones de inspeccion. Diez variables de habitat pertinente al urbanizacion fueron medidas en cada estacion de inspeccion usando mapas topograficos y fotografia aerea. Los resultados ensenaron una comunidad compuesta de 10 es- pecie de rapaces en cria. Buteo lineatus, Accipiter gentilis, y Stix varia mostraron un aversion significante al habitat suburbio, mientras B. jamaicensis y Bubo virginianus tuvieron una tendencia mayor para ocupar tales areas. Habitat de tierra baja fue sensiblemente escogido por S. varia, B. lineatus y A. cooperii el habitat por lo general mas susceptible para el desarrollo en el estudio de la region. La de la especie riquesa de rapaces mostro una correlation fuerte y positiva ( r = 0.79, p < 0.01) con la area del tamano del bosque. Ningun area de bosque menos de 1000 ha tuvo mas de cuatro especie de rapaces mientras cuatro a ocho especie fueron encontrados en areas de 1000-8000 ha. La utilization de tres clases aumentando y de diferente tamano de areas de bosque mostraron una tendencia de aumento para B. lineatus, A. gentilis y S. varia y una tendencia decreciente para B. jamaicensis y Bubo virginianus. [Traduction de Raul De La Garza, Jr.] Raptors are secondary and tertiary consumers so trophic theory suggests that they will be fewer in abundance and lesser in diversity than other breed- ing bird communities. Factors that tend to pro- mote high species diversity in raptor communities Present address: Beak Consultants, Inc., 12931 NE 126th Place, Kirkland, WA 98052. are high prey diversity and high habitat heteroge- neity (White 1974). Conversely, forest fragmenta- tion and reduced forest interior tend to reduce raptor community diversity (Thiollay and Meyburg 1988). In Maryland, Robbins et al. (1989) exam- ined the effect of forest area on breeding bird communities and found at least one raptor (the Red-shouldered Hawk, Buteo lineatus ) was impacted 26 March 1997 Forest Raptors and Urbanization 27 by habitat fragmentation and urbanization and qualified to be categorized as an “area-sensitive species.” Titus and Mosher (1981) examined nat- ural habitat variables of four sympatxic hawk spe- cies in Maryland, but urbanization and variables associated with habitat development were not in- cluded in their analysis. Nevertheless, Red-shoul- dered Hawks have been shown to avoid nesting near human dwellings (Bosakowski et al. 1992a) and Northern Goshawks ( Accipiter gentilis) nest sig- nificantly farther from paved roads and human dwellings than randomly-selected sites (Bosakowski and Speiser 1994). In this study, we examined a wider variety of ur- banization variables than previously published for forest raptors and examined their association with- in a diverse community of breeding raptors. Such data could provide further insight into the pres- ervation and management of viable forest ecosys- tems for diverse raptor communities rather than using a single species approach. Study Area The study was conducted in the Highlands Physio- graphic Region (Braun 1950) which extends southwest to northeast across the New York-New Jersey border. The study area was part of the Pequannock Watershed, which is owned and operated by the city of Newark, New Jersey. The study area includes approximately 16 100 ha and includes parts of Passaic, Morris and Sussex Counties. This hilly, mostly forested terrain is part of the Eastern Deciduous Forest Biome (Shelford 1963). Nearly all of the Highlands forests have been previ- ously cut or burned within the last 200 years (Ohmann and Buell 1968, Russell 1981), resulting in largely second growth forest dominated by oaks ( Quercus spp.) and oth- er subclimax hardwood trees (Buell et al. 1966, Russell 1981). Older forest (>100 years of age) is rare and is typically limited to small remnant stands surrounded by younger forest. The study area includes some of the last remaining wilderness in the northern half of New Jersey. Residential and commercial development is limited to about 10% of the study area. The remaining land area is composed of a mosaic of submature second growth and mature forest, typically ranging from about 40-80 years of age. About 75% of the forest is deciduous habitat, about 20% con- sists of hemlock-white pine ( Tsuga canadensis-Pinus stro- bus) stands, and 5% is mature conifer plantations. Aquat- ic habitats include five major reservoirs, several smaller impoundments, beaver ( Castor canadensis) ponds, marsh- es, shrub swamps and wooded swamps, the latter occur- ring in many areas. The study area is periodically thinned, but clear-cutting/burning is not permitted. Since access to these forests is regulated by recreational permits and motorized boats are not permitted on the reservoirs, human disturbance is greatest directly adja- cent to some suburban areas. Methods Sampling design. The spatial design of the survey sys- tematically covered the Pequannock Watershed study area with a grid pattern of 81 broadcasting stations, spaced at approximately 1.2 km intervals. The spacing of stations was not used to estimate population size, since this investigation was aimed solely at determining occu- pancy in different habitat types. The survey stations were plotted on 15 min USGS quadrangle topographic maps of the study area. Four surveys were conducted at each of the 81 broadcasting stations. Surveys were conducted on fair weather days with low wind velocity (<15 km/hr) and no precipitation. The order of sampling of calling stations was different for each of the four surveys due to weather and wind conditions, but stations in the four car- dinal quadrants of the study area were worked as equally as possible to avoid regional bias. Target species. Previous pilot studies (Benzinger et al. 1988, Bosakowski et al. 1989a) revealed 11 potentially breeding raptor species in the study area (excluding Ca- thartid vultures) . To reduce the number of surveys re- quired to survey these 11 species, vocalizations of 2-3 hawk or owl species were broadcast on each survey, re- sulting in a total of two night surveys for owls and two daytime surveys for hawks. The selection of species for each survey was based on regional nesting phenology (Bull 1964, Bosakowski et al. 1989b, Speiser and Bosa- kowski 1991, and Bosakowski 1990). Broadcasts were or- dered from the smallest to the largest raptor, to avoid potential inhibitory effects of large raptors on the re- sponse behavior of smaller species (Call 1978). About half of the raptors responded to heterospecific broad- casts, so every species was actually sampled twice during the day (or night) at all calling stations. Furthermore, some raptors were detected by visual observations or call- ing prior to broadcasts on either survey, regardless of the species that were broadcast. Broadcast vocalizations and equipment. Eastern Screech-owl ( Otus asio), Great Horned Owl ( Bubo virgi- nianus). Northern Saw-whet Owl ( Aegolius acadicus) , Barred Owl (Strix varia). Red-shouldered Hawk, Red- tailed Hawk {Buteo jamaicensis ) and Broad-winged Hawk (Buteo platypterus) vocalizations were obtained from the National Geographic Society guide to Bird Songs (Eva- tone Soundsheets, Inc., Clearwater, FL, 1983). Northern Goshawk, Sharp-shinned Hawk ( Accipiter striatus ) and Long-eared Owl ( Asio otus) vocalizations were from the Peterson Field Guide records, Cornell University, Labo- ratory of Ornithology. A Cooper’s Hawk {Accipiter coopeni) tape of the female alarm call was obtained from R.N. Rosenfield (Rosenfield et al. 1985). Tapes were broadcast from a Contec portable stereo cassette tapedeck (Model V83, Japan) rated at 10 watt output with two removable speakers. The units were pow- ered by 8 “D” cell batteries. Cassette tapes (normal bias) were broadcast at a standard volume setting, approxi- mately 95% of full capacity, but without detectable dis- tortion. The speakers were high fidelity, each containing a woofer and tweeter component for more accurate sound replication of calls. The two speakers were mount- ed back-to-back to provide bidirectional broadcasting of vocalizations at all times. Field protocol. Two surveys were conducted at night 28 Rosakowski and Smith Vol. 31, No. 1 for the five target owl species. These surveys began at least V 2 hr after sunset and were terminated by 0100 H the following morning. The first owl survey was conduct- ed from 21 March 1989-16 April 1989 for Eastern Screech-owls, Long-eared Owls, and Great Horned Owls. Species on the second owl survey, conducted from 16 April-19 May 1989, included the Northern Saw-whet Owl and Barred Owl. Hawk surveys were conducted from at least 2 hr after sunrise to no later than 2 hr before sunset. Species included on the first hawk survey conducted from 7 April-4 May 1989 included the Red-shouldered Hawk, Northern Goshawk, and Red-tailed Hawk. The sec- ond hawk survey from 18 May-20 June 1989 included the Sharp-shinned Hawk, Cooper’s Hawk, and Broad-winged Hawk. On each survey, tape-recorded vocalizations of three raptor species were played during an 18-min period. Vo- calizations of each species were recorded on cassette tape for a 3-min period followed by a 3-min period of silence. The silent period served a dual function as a listening/ watching period for vocal or visual responses of raptors and also as a refractory period prior to broadcasts of the next raptor species. The double speakers were hung on low tree branches about 1.5 m above ground during broadcasts. Macrohabitat measurements. To measure habitat at the 81 calling stations, their location was plotted on topo- graphic maps and their position noted on aerial photo- graphs (1:8000) taken during the winter of 1982. Habitat at each calling station was quantified in a 300 m radius circle centered on the calling station. This distance was chosen as representative habitat of the calling station as all raptors were detected within this distance and most were detected within 100 m. On aerial photographs, a dot-grid overlay was used to quantify the suburban area within the 300 m radius circle. To measure the length of paved roads and forest edge within the habitat circles, w T e overlayed them with a fine thread, then the thread was straightened and measured. Forest edge can be consid- ered as any abrupt change (Small and Hunter 1989) to open habitat (e.g., field, marsh, river or lake) that are easily discernable on the aerial photographs. We mea- sured distances to forest openings (>0.5 ha), paved roads, human habitation and wetlands (>0.5 ha) from aerial photographs using a metric ruler with a mm scale. Topographic maps were used to calculate slope (rise in elevation over 300 m baseline) . For the purposes of this paper, “wilderness” was defined according to a standard operational definition supplied by Webster’s Dictionary as “any uninhabited, uncultivated region.” Wilderness area was calculated as the total area of contiguous, un- cultivated, uninhabited habitats bounded by paved roads and/or housing developments. Statistical analysis. A total of 10 habitat variables rele- vant to urbanization were used to describe the 81 calling stations. Each raptor species had a unique set of stations where they were recorded and these data sets were used to calculate habitat means and standard deviations. In order to increase the sample size for Eastern Screech-owl ( N — 3) and Northern Goshawk ( N = 2), two screech- owl sightings and three goshawk nests from previous years were added to the data set. These additional raptor locations occurred in the mid- to late eighties and fell Table 1. Species richness values for breeding raptor communities studied in North America. Study Area (ha) No. OF Species New Jersey (this study) 16,100 10 Utah (Smith and Murphy 1973) Wyoming (Craighead and 20,700 11 Craighead 1956) Michigan (Craighead and 3,100 10 Craighead 1956) 9,600 7 Idaho (BLM 1979) 53,200 15 well within five of the 81 habitat circles described. No additional sites were known for the Long-eared Owl which was found at only two sites, or the Sharp-shinned Hawk which was found at one site, so they were dropped from any habitat analyses. Each species was compared to the set of stations which did not have any raptor detections during the study (un- occupied habitat, N =■ 20) to provide a measure of hab- itat selection. For statistical comparison, we used a non- parametric test (Mann-Whitney U-test) since some of the data were percentages or nonnormal in distribution (Zar 1974). Species richness was determined for all wilderness areas with four or more survey stations and a logistic re- gression curve was calculated (Excel Software, Microsoft Corp., Redmond, WA, Version 5.0) to test the strength of the relationship. To detect area relationships for in- dividual species, wilderness areas were grouped into three size class categories (0-1000, 1000-2000 and 2000- 8000 ha) for all broadcast stations. The percentage of occupied stations by a species in each size class (% usage) was subtracted from the percentage of all stations sam- pled (% availability) to determine habitat utilization (Johnson 1980) for each wilderness area size class. Usage is said to be selective if resources are used disproportion- ately to their availability (Johnson 1980). Proportions in each category were tested for increasing or decreasing trends in relation to wilderness area size using an Armi- tage (1955) proportion trend test. Results and Discussion Species Richness of the Raptor Community. We compared raptor species richness of the New Jersey raptor community with studies of other raptor communities (Table 1 ) . Of these studies, the lowest species richness was found in Michigan (Craighead and Craighead 1956) which was mostly farmland (11% wooded). New Jersey forestland and the spruce-fir-pine slopes and sagebrush benches in the Snake River of Wyoming (Craighead and Craighead 1956) had slightly higher richness but both were surpassed by Utah scrub juniper desert (Smith and Murphy 1973). Highest species rich- ness was reported for Snake River Canyon in Idaho March 1997 Forest Raptors and Urbanization 29 Table 2. Urbanization habitat variables for sites occupied by forest raptors in a northern New Jersey watershed. Top number represents the mean and bottom number represents the SD (* = P < 0.05, + = P< 0.10). Sample size for each species given in parenthesis. GHOW = Great Horned Owl, BAOW = Barred Owl, ESOW = Eastern Screech- owl, RTHA = Red-tailed Hawk, RSHA = Red-shouldered Hawk, BWHA = Broad-winged Hawk, COHA = Cooper’s Hawk, NOGO = Northern Goshawk. Variable GHOW (16) BAOW (27) ESOW (5) RTHA (22) RSHA (9) BWHA (16) COHA (10) NOGO (5) Unoccu- pied Habitat (20) Distance to Human 426.6 671.9* 778.0 477.3 888.9* 769.4 651.5 676.0+ 505.0 Habitation (m) 325.9 488.0 659.4 411.6 569.7 704.4 672.0 533.5 496.2 Number of 4.19 1.26* 0.60 3.32 0* 3.06 1.20 0* 2.35 Houses/Bldgs. (#) 7.30 3.14 0.89 5.55 0 4.35 2.25 0 4.35 Suburban Area (%) 2.7 1.0+ 0.2 2.4 0* 2.5 0.9 0* 1.5 4.1 2.5 0.3 3.7 0 3.6 1.9 0 2.4 Distance to Paved 292.8 468.0 458.0 406.6 546.1 343.8 259.0 482.0 343.8 Road (m) 258.0 465.6 393.9 345.9 549.9 293.5 141.0 243.0 294.4 Road Mileage (m) 404.7 299.4 64.0 367.5 206.7 269.4 352.5 78.0 268.4 620.2 321.4 143.1 528.7 226.3 330.5 380.5 174.4 294.1 Edge Length (m) 623.7 370.0 652.0 595.9 492.2 374.1 656.5 369.0 467.0 632.7 524.2 536.6 617.6 618.8 374.9 420.5 550.1 526.6 Distance to Wetland 563.1 197.0+ 662.0 442.0 62.2* 360.3 252.0 164.0 369.8 (m) 518.2 247.9 581.8 585.0 49.7 377.8 203.1 207.5 333.4 Distance to Forest 195.0 199.7 197.0 163.9 80.6* 141.6 112.0+ 147.0 240.2 Opening (m) 237.3 206.2 261.5 192.2 91.4 148.1 109.6 141.3 195.4 Slope (%) 6.5 6.1 10.0 6.8 5.1 + 6.9 4.4* 6.8 8.0 4.4 4.8 4.9 4.7 5.3 6.3 4.1 6.7 5.5 Wilderness Area (ha) 2291.2 3766.9 5094.0 2498.2 4388.9 3144.0 2271.6 5278.0 2729.5 2731.4 3164.4 3543.4 2971.8 3179.1 3204.8 2181.3 3320.5 2636.0 (BLM 1979) with its vertical cliffs and sagebrush desert, probably reflecting the high structural hab- itat diversity as a result of vertical and horizontal habitat partitioning. Overall, cultivated land of Michigan (circa 1950s farming techniques) ap- peared to slighdy reduce raptor diversity and abun- dance compared to contiguous forestland in New Jersey. Urbanization and Forest Fragmentation. Urban- ization and its ultimate effect in fragmenting for- ests into smaller wilderness areas is an unnatural factor which reduces raptor abundance and diver- sity. Robbins (1979) noted the disappearance of several nesting bird species including the Broad- winged Hawk after 30 yr of severe fragmentation. In our study area, the proximity of human habita- tion showed marked differences in the habitat suit- ability for several forest raptors. The Red-shoul- dered Hawk was most sensitive to human distur- bance and it occupied sites significantly further from human habitation than unoccupied habitat and it showed a complete lack of suburban habitat within the 300 m radius habitat circles examined (Table 2). The Northern Goshawk had the second largest distance to human habitation ( P < 0.10) and similarly showed a significant lack of suburban habitat within the 300 m radius habitat circles ex- amined. Additional data from 16 Northern Gos- 30 Bosakowski and Smith Vol. 31, No. 1 Figure 1. Species richness of forest raptors with respect to seven wilderness areas of varying size. Curve represents best-fit logistic regression function. hawk nest sites in the study region also showed that the species nests further from human habitation and paved roads than expected (Bosakowski and Speiser 1994). In the study region. Northern Gos- hawk nests were significantly farther from human habitation than were Cooper’s Hawk nests (Bosa- kowski et al. 1992c) or Red-tailed Hawk nests (Speiser and Bosakowski 1988). The Barred Owl was similar to the Red-shouldered Hawk and Northern Goshawk, occupying sites that were sig- nificantly farther from human habitation, with sig- nificantly fewer houses, and a tendency for less sub- urban habitat area (P < 0.10). With regard to wetlands, the Red-shouldered Hawk occupied areas significantly closer to wet- lands which resulted in a significantly lower slope percentage as well as a significantly closer distance to forest openings. The Barred Owl also had a ten- dency to be closer to wetlands (P < 0.10). The Cooper’s Hawk occupied sites with significantly lower slope percentages and was often closer to forest openings (P< 0.10), but not necessarily due to a preference for wetlands since it often used suburban forest edge as well. Overall, these three lowland species had a greater vulnerability to de- velopment pressures as valley bottoms and flat ter- rain are generally the first areas targeted for roads, houses and commercial buildings (Tiner 1985). Distance to paved roads, road mileage, edge length and wilderness area did not show any sig- nificant differences among any species compared to unoccupied habitat. Four species, Great horned Owl, Red-tailed Hawk, Eastern Screech-owl and Broad-winged Hawk did not have any variables that were significantly different from unoccupied hab- itat, suggesting that they were less sensitive to ur- banization. Forest Area Relationships. Area-sensitive species respond negatively to decreasing forest size and show predictable declines or absence as the area of the forest shrinks (Robbins 1979, Ambuel and Temple 1983, Robbins et al. 1989). Within our study area we found a strong correlation (r = 0.79, P < 0.01) for species richness of forest raptors and increasing size of wilderness areas (Fig. 1). This is likely the result of the inclusion of area-sensitive species in large wilderness areas and their exclu- sion in smaller forest fragments. No area less than 1000 ha had more than four raptor species while 4-8 species were found in areas from 1000-8000 ha in size. Thiollay and Meyburg (1988) also noted a positive correlation between the size of reserves and the abundance index of diurnal raptors (Fal- coniformes) on the Island of Java. When wilderness areas were grouped into size classes, several distinct trends emerged among the species. We calculated percent utilization of three wilderness area size classes and determined the probability of increasing or decreasing trends for raptors. Several species revealed increasing trends with increasing wilderness area size (Red-shoul- dered Hawk, P = 0.97; Northern Goshawk, P = 0,84; and Barred Owl, P = 0.82), whereas several species revealed decreasing trends (Red-tailed Hawk, P — 0.95 and Great Horned Owl, P = 0.85, Fig. 2). Craighead and Craighead (1956) noted March 1997 Forest Raptors and Urbanization 31 50 Z o § -J £ w 2 IU 0 . 40 30 20 10 -10 -20 ES0U <0.56. 0.44) sa BA0W (0.82. 0.18) ■ GHOU (0 15. 0.85) w C0HA <0.70. 0.30) m (MIA <0 33. 0 67) ■ RSHA <0.97. 0 03) G3 N0G0 <0.84. 0.16) ■ RTHA (0.05, 0.95) -30 0-1000 1-2000 2-8000 WILDERNESS AREA (ha) Figure 2. Relative utilization (use-availability) of different wilderness area size classes for eight sympatric forest raptor species (see Table 2 for species acronyms) . Numbers in parentheses represent probability of species increasing and decreasing (I, D) in relation to wilderness area size (Proportion Trend Test — Armitage 1955). that Red-tailed Hawks and Great Horned Owls were generalist predators and the results of this study indicate that they clearly benefit from forest fragmentation and urbanization (at least to the ex- tent found in the study area) . On the other hand, Red-shouldered Hawks, Northern Goshawks and Barred Owls usually prefer extensive remote areas of deep woods (Bosakowski 1989, Bosakowski et al. 1992a, Bosakowski and Speiser 1994). The northern half of our study area was virtually all wilderness and contained all 10 species. How- ever, the southern half had nearly all of the sub- urban areas and a four-lane highway, and was miss- ing three species (Northern Goshawk, Sharp- shinned Hawk and Red-shouldered Hawk) . Results of this study predict the expansion of dominant, disturbance-tolerant Great Horned Owls and Red- tailed Hawks after forest fragmentation, and a re- duction in raptor diversity. A decrease in forest area can result in decreased bird species diversity (Lovejoy et al. 1986), distur- bance to species in adjoining wetlands, reduced buffering against human disturbance and in- creased predation (Chasko and Gates 1982, Yahner 1988). It is also becoming increasingly clear that “edge effect” is beneficial to only a limited num- ber of wildlife species and, by and large, it has a strong negative impact on other members of forest communities (Robbins 1979, Wilcove et al. 1986, Yahner 1988, Robbins et al. 1989). For tropical rainforests, Thiollay (1984) found that raptors are among the first species to disappear in the process of human population growth and exploitation and are thus suitable indicators of habitat disturbance. The situation appears to be similar for temperate forests disrupted by urbanization and agriculture (Craighead and Craighead 1956). Large raptor communities are needed as popu- lation reserves to maintain genetic diversity and to provide constant recruitment to marginal habitats (White 1974, Wilcove 1987). Results of this study suggest that only large wilderness areas (2000-8000 ha) can provide the full diversity of forest raptors necessary to stock marginal habitats. Until other data on reserve size become available, the present data could have useful management implications for conservation of Northeast forest raptors, either by regulating development and recreation, or de- ciding how large an area should be set aside for future reserves. In addition, further research will be needed to determine the number and distance of other reserves (Shaffer 1985, Hunter 1990) nec- essary to support a viable population network for each species (Wilcove 1987). 32 Bosakgwski and Smith Vol. 31, No. 1 Acknowledgments This study was part of a doctoral thesis of the first au- thor that was approved by the faculty of Rutgers Univer- sity under the advisorship of Dr. D.W. Morrison. The re- search was primarily funded by a grant from the New Jersey Division of Fish, Game, and Wildlife, Nongame Wildlife Program. As such, we would like to extend our appreciation to L. Niles and M. Valent for their interest, cooperation and comments on the initial design proto- col. The New Jersey Audubon Society also provided funds for field work for which we are grateful. We also thank forester T. Koeppel for use of aerial photographs, forest- ry equipment and information on the history of Pequan- nock Watershed forestry. We also thank M.R. Fuller, G.D. Hayward, K. Titus, W.M. Block and J.A. Mosher for re- viewing various drafts of the manuscript. Literature Cited Ambuel, B. and S.A. Temple. 1983. Area-dependent changes in the bird communities and vegetation of southern Wisconsin forests. Ecology 64:1057-1068. Armitage, P. 1955. Tests for linear trends in proportions and frequencies. Biometrics I f: 375—386, Bednarz, J.C., D. Klem, Jr., L.J. Goodrich and S.E. Sen- ner. 1990. Migration counts of raptors at Hawk Mountain, Pennsylvania, as indicators of population trends, 1934-1986. Auk 107:96-109. Bent, A.C, 1937, Life histories of North American birds of prey. Part I. U.S. Natl. Mus. Bull. No. 167. Benzinger, J., P. Bacinski, D. Miranda and T. Bosakow- ski. 1988. Breeding birds of the Pequannock Water- shed. Rec. N.J. Birds 14:22-27. Bureau of Land Management. 1979. Snake River birds of prey special research report. USDI, BLM, Boise Dis- trict, Boise, ID U.S.A. Bosakowski, T. 1989. Barred Owl. Pages 114-120 in Proc. Northeast raptor manage, symp. and workshop. Natl. Wildl. Fed., Washington, DC U.S.A. . 1990. Community structure, niche overlap, and conservation ecology of temperate forest raptors dur- ing the breeding season. Ph.D. dissertation, Rutgers Univ., Newark, NJ U.S.A. and R. Speiser. 1994. Macrohabitat selection by nesting Northern Goshawks: implications for manag- ing eastern forests. Studies in Avian Biol. 16:46-49. , R. Speiser and J. Benzinger. 1987. Distribution, density, and habitat relationships of the Barred Owl in northern New Jersey. Pages 135-143 in Proc., Bi- ology and conservation of northern forest Owls, Win- nipeg, Manitoba, Canada. USDA Forest Service Gen. Tech. Rep. RM-142. , J. Benzinger and R. Speiser. 1989a. Forest owl populations of the Pequannock Watershed. Rec. N.J. Birds 15:2-8. , R. Speiser and D.G. Smith. 1989b. Nesting ecol- ogy of forest-dwelling Great Horned Owls in the East- ern Deciduous Forest Biome. Can. Field-Nat. 103:65- 69. , D.G. Smith and R. Speiser. 1992a. Status, nest- ing density, and macrohabitat selection of Red-shoul- dered Hawks in northern New Jersey. Wilson Bull. 104: 434-446. , D.G. Smith and R. Speiser. 1992b. Nest sites and habitat selected by nesting Cooper’s Hawks in north- ern New Jersey and southeastern New York. Can. Field- Nat. 106:474-479. , D.G. Smith and R. Speiser. 1992c. Niche overlap of two sympatric-nesting hawks Accipiter spp. in the New Jersey-New York Highlands. Ecography 15:358- 372. Braun, L. 1950. Deciduous forests of eastern North America. Blakiston Co., New York, NY U.S.A. Buell, M.F., A. Langform, D. Davidson and L. Ohmann, 1966. The upland forest continuum in northern New Jersey. Ecology 47:416-432. Bull, J. 1964. Birds of the New York area. Dover Publi- cations, New York, NY U.S.A. Call, M.W. 1978. Nesting habitats and surveying tech- niques for common western raptors. USDI, BLM Tech. Note 316. Chasko, G.G. and J.E. Gates. 1982. Avian habitat suit- ability along a transmission-line corridor in an oak- hickory forest region. Wildl. Mono. 82:1-41. Craighead, J-J. and F.C. Craighead, Jr. 1956. Hawks, owls and wildlife. Stackpole Publishing Co., Harris- burg, PA U.S.A. Johnson, D.H. 1980. The comparison of usage and avail- ability measurements for evaluating resource prefer- ence. Ecology 61:65-71. Lovejoy, T.E., R.O. Bierregaard, Jr. and A.B. Rylands, Jr. 1986. Edge and other effects of isolation on Am- azon forest fragments. Pages 257-285 in M.E. Soule [Ed.], Conservation biology: the science of scarcity and diversity. Sinauer Associates, Sunderland, MA U.S.A. Ohmann, L.F. and M.F. Buell. 1968, Forest vegatation of the New Jersey Highlands. Bull. Torrey Botanical Club 95:287-298. Robbins, C.S. 1979. Effects of forest fragmentation on bird populations. Pages 198-212 in Management of northcentral and northeastern forests for nongame birds. USDA For. Serv. Gen. Tech. Rep. NC-51, Min- neapolis, MN U.S.A. , D.K. Dawson and B.A. Dowell. 1989. Habitat area requirements of breeding forest birds of the Mid- dle Atlantic States. Wildl. Mono. 103:1-33. Rosenfield, R.N., J. Bielefeldt and R.K. Anderson. 1985. Taped calls as an aid in locating Cooper’s Hawk nests. Wildl. Soc. Bull. 13:62-63. Russell, E.W.B. 1981. Vegetation of northern New Jer- sey before European settlement. Am. Midi. Nat. 105' 1 - 12 . Shaffer, M.L. 1985. The metapopulation and species conservation: the special case of the Northern Spot- ted Owl. Pages 86-99 in R.J. Gutierrez and A.B. Carey March 1997 Forest Raptors and Urbanization 33 [Eds.], Ecology and management of the Spotted Owl in the Pacific Northwest. USDA Forest Service Gen. Tech. Rep. PNW-185, Pacific Northwest Research Sta- tion, Olympia, WA U.S.A. Shelford, V.E. 1963. The ecology of North America. Univ. of Illinois Press, Urbana, IL U.S.A. Small, M.F. and M.L. Hunter. 1989. Response of pas- serines to abrupt forest-river and forest-powerline edges in Maine. Wilson Bull. 101:77-83. Smith, D.G. and J.R. Murphy. 1973. Breeding ecology of raptors in the eastern Great Basin of Utah. Brig- ham Young Univ. Sci. Bull., Biol. Ser. 18:1-76. SPEISER, R. AND T. BOSAKOWSKI. 1988. Nest site prefer- ences of Red-tailed Hawks in the highlands of south- eastern New York and northern New Jersey. J. Field Ornithol. 59:361-368. and . 1991. Nesting phenology, site fidel- ity, and defense behavior of Northern Goshawks in New York and New Jersey. J. Raptor Res. 25:132-135. Thiollay, J.M. 1984. Raptor community structure of a primary rain forest in French Guiana and effect of human hunting pressure. Raptor Res. 18:117-122. AND B.U. MEYBURG. 1988. Forest fragmentation and the conservation of raptors: a survey on the is- land of Java. Biol. Cons. 44:229-250. Tiner, R.W., Jr. 1985. Wetlands of New Jersey. U.S. Fish and Wildlife Service, National Wetlands Inventory, Newton Corner, MA U.S.A. Titus, K. and J.A. Mosher. 1981. Nest-site habitat se- lected by woodland hawks in the central Appalachi- ans. Auk 98:270-281. White, C.M. 1974. Current problems and techniques in raptor management and conservation. Transac. Thir- ty-ninth North American Wildlife Conference. Wilcove, D.S. 1987. Public lands and the fate of the Spotted Owl. Amer. Birds 41:361-367. , C.P. McLellan and A.P. Dobson. 1986. Habitat fragmentation in the temperate zone. Pages 237-256 in M.E. Soule [Ed.], Conservation biology: the science of scarcity and diversity. Sinauer Associates, Sunder- land, MA U.S.A. Yahner, R.H. 1988. Changes in wildlife communities near edges. Cons. Biol. 2:333-339. Zar, J.H. 1974. Biostatistical anaylsis. Prentice-Hall, Inc., Englewood Cliffs, NJ U.S.A. Received 28 February 1996; accepted 18 November 1996 / Raptor Res. 31 (1) :34— 39 © 1997 The Raptor Research Foundation, Inc. DOES VEGETATION STRUCTURE LIMIT THE DISTRIBUTION OF NORTHERN GOSHAWKS IN THE OREGON COAST RANGES? Stephen DeStefano 1 National Biological Service, Oregon Cooperative Wildlife Research Unit, 104 Nash Hall, Oregon State University, Corvallis, OR 97331 U.S.A. Jon McCloskey 2 Bureau of Land Management, Salem District Office, 1717 Fabry Road, Salem, OR 97306 Abstract. — Northern Goshawks ( Accipiter gentilis ) breed in a variety of forested areas throughout the Pacific Northwest. Nevertheless, they were only first found breeding in the Coast Ranges of Oregon in 1995, despite apparently suitable habitat and abundant prey. We document the rarity of goshawks in the Coast Ranges by reviewing previous and current survey results for nests of goshawks and other forest birds since the 1960s, examining sightings of goshawks since 1980 and reporting on a survey we con- ducted in 1994. We suggest that nesting goshawks are rare in the Coast Ranges because of the vegetative structure of the area and its influence on prey availability. Key Words: Northern Goshawk ; Accipiter gentilis; distribution', habitat, foraging, reproduction ; Oregon. Puede ser que la estructura de vegetacion limite la distribucion de Accipiter gentilis en la sierra costa de Oregon. Resumen. — Accipiter gentilis se crian en una variedad de areas de bosque en todas partes del noroeste pacifico. Sin embargo, la primera vez que los encontraron en la sierra costa de Oregon fue en 1955, a pesar de suficiente habitat conveniente y presa abundante. Nosotros documentamos la rareza de A. gentilis en la sierra costa examinando anterior y corriente resultados de estudios para nidos de A. gentilis y otros pajaros de bosque desde los 1960s, examinando observaciones de A. gentilis desde 1980 y repor- tando un estudio que nosotros conducimos en 1994. Nosotros pensamos que los A. gentilis que hacen nido en la sierra costa es raro por la estructura de vegetacion en el area y su influencia en la disponi- bilidad de presa. [Traduccion de Raul De La Garza, Jr.] Northern Goshawks {Accipiter gentilis) are distrib- uted across northern North America and through- out forested areas of the western U.S. (Palmer 1988). They nest in a variety of forest types, includ- ing boreal, deciduous and western coniferous for- ests. In Oregon, goshawk nests are found through- out forested areas east of the Cascade Mountain Range, on east and west slopes of the Cascade Range, in the Siskiyou Mountains of southwestern Oregon and even in isolated stands of aspen ( Pop- ulus spp.) in mountain draws and valleys in the 1 Present address: USGS, Arizona Cooperative Fish and Wildlife Research Unit, 104 Biological Sciences East, Uni- versity of Arizona, Tuscon AZ 85721. 2 Present Address: Winema National Forest, 38500 High- way 97 N, Chiloquin, OR 97624. Great Basin region of southeastern Oregon (Mar- shall 1992). Goshawks were not known to breed in the Coast Ranges of western Oregon, even though their nests are found in all other areas of the state (Reynolds et al. 1982, Marshall 1992). Incidental sightings of goshawks have been reported in the Coast Ranges, but it was not until 1995 that two pairs of goshawks were found breeding there (Thrailkill and Andrews 1996). Reynolds (1975, 1978) found breeding pairs of Cooper’s Hawks {Accipiter cooperii) and Sharp- shinned Hawks (A. striatus) , but not goshawks after an extensive search of the Coast Ranges from 1968-78. He speculated that forest conditions, spe- cifically dense understories which may interfere with a goshawk’s ability to hunt, precluded gos- hawks from breeding in this area (Reynolds and 34 March 1997 Vegetation Structure and Goshawks 35 Figure 1. The Coast Ranges (shaded area) of western Oregon and areas (circles, which may represent ^ 1 sur- vey blocks) surveyed for Northern Goshawks during June-August 1994. Wight 1978, Reynolds et al. 1982) . No one has con- ducted systematic searches for goshawk nests in the Coast Ranges since, and no one has used broad- casts of goshawk vocalizations to survey for the presence of breeding pairs over large areas of the Coast Ranges (Woodbridge, USDA Forest Serv. un- publ. rep. 1990; Kennedy and Stahlecker 1993). Herein, we report results of a study conducted to document the presence of breeding Northern Gos- hawks and to assess vegetative conditions that might influence their distribution in the Oregon Coast Ranges. Study Area The Coast Ranges of western Oregon lie north of the Coquille River and west of the Willamette Valley, and are separate from the Siskiyou Mountains in the southwest- ern corner of Oregon (Fig. 1). Topography is steep and dissected by many streams. Elevations range from sea lev- el to 450-750 m at main ridge summits, with scattered peaks as high as 1250 m. Climate is characterized by mild, wet winters and relatively dry summers (Franklin and Dyrness 1973). Historically, the Coast Ranges were densely forested with sitka spruce ( Picea sitchensis), western hemlock ( Tsu- ga heterophylla) and western redcedar ( Thuja plicata) (Franklin and Dyrness 1973). Most of the mature forest (>80 yr old) has been logged or burned during the past 150 yr (Franklin and Dyrness 1973). Much of what re- mains are stands of second-growth trees with older stands occurring as islands, fragmented by clearcut logging. As a result of these disturbances, and because of tree plant- ing, Douglas-fir ( Pseudotsuga menziesii) is now the major component of the forests in this area. Western hemlock and western redcedar are common coniferous species and red alder ( Alnus rubra ) , vine maple ( Acer circinatum ) and bigleaf maple (A. macrophyllum ) are common hard- wood species (Franklin and Dyrness 1973, Forsman et al. 1996b). Methods Goshawk Surveys. We surveyed for Northern Goshawks in 24 survey blocks, totaling 3285 ha (range 60-335 ha) during June-August 1994. Surveys covered the east and west slopes of the Coast Ranges and included federally administered public lands, the MacDonald-Dunn State Forest managed by Oregon State University and areas originally searched by Reynolds (1975, 1978). To maxi- mize the potential for locating nesting goshawks, survey blocks were chosen based on our knowledge of goshawk nesting habitats (DeStefano et al. 1994), past sightings of goshawks, recommendations from agency biologists fa- miliar with local conditions and habitat and examination of aerial photographs and topographic maps to deter- mine accessibility of potential goshawk habitat. Whenever possible, older (>80 yr), larger contiguous forested blocks were surveyed. Because we were more interested in documenting the presence of breeding goshawks rath- er than calculating an unbiased estimate of nesting den- sity, we focused on areas with the greatest potential for nesting habitat, based on the published literature and our experience in Oregon. We did, however, survey a va- riety of forest types and serai stages. We used taped vocalizations of Northern Goshawks to elicit responses from adults and juveniles (Woodbridge, USDA Forest Serv., unpubl. rep. 1990; Kennedy and Stah- lecker 1993); the adult alarm call was used during the nesting period in June-July, and both adult alarm and juvenile begging calls were used during the post-fledging period in July-August. Road and foot transects were first delineated on maps and aerial photographs. Foot tran- sects were 200 m apart with broadcast stations every 300 m; stations along adjacent transects were staggered (Joy et al. 1994). Broadcast stations on roads were 250 m apart. At each station, vocalizations were broadcast in three directions (60°, 180°, and 300°) for 10 sec, with 30 sec between each call. This procedure was conducted at each station by one or two observers. Presence, location and behavior of raptors were recorded. Historical Sight in gs. To document past sightings of gos- hawks in the Coast Ranges, we searched records and data bases compiled from 1980—1995 by state and federal land management and conservation agencies and local bird- watchers. We assessed the reliability of these sightings 36 S. DeStefano and J. McCloskey Vol. 31, No. 1 and categorized them as 1 (questionable - observer had no or litde experience identifying birds, or experience could not be assessed) , 2 (reliable = experienced birder) and 3 (confirmed = experience observing raptors, pro- fessional biologist). Records were searched for date ob- served, behavior of adults, presence of immatures and other clues that might indicate reproductive activity. We also questioned biologists who have been conducting ex- tensive surveys for Northern Spotted Owls ( Strix occiden- talis) and Marbled Murrelets ( Brachyramphus mamoratus) in the Coast Ranges for the past 10-15 yr (see Nelson and Sealy 1995 and Forsman et al. 1996a for background and methods). Vegetation Surveys. We examined vegetation in the 24 survey blocks for forest stand structure and composition, understory conditions and landscape patterns. Stand can- opy structure, % canopy cover and tree species were re- corded from ground surveys and aerial photos. After con- ducting ground assessments at each survey block, we then used a Geographic Information System (GIS) to cal- culate % cover of forest type and amounts of mature (>80 yr) and second-growth forest in survey blocks. We recorded the presence and % cover of dominant under- story plant species at each survey site. Understory vege- tation was classified into six associations, according to Franklin and Dyrness (1973), and represented a gradient from dry to wet conditions. These associations were (1) ocean-spray-salal ( Holodiscus discolor-Gaultheria shallon ) as- sociation found in dry, relatively open sites, (2) Pacific rhododendron-Oregon grape ( Rhododendron macrophyl- lumrBerberis nervosa) found on dry, exposed ridge tops, (3) big huckleberry-beargrass ( Vaccinium membranaceunbXero- phyllum tenax ) on shallow, stony soils at high elevations, (4) vine maple-salal in cool, moist sites with moderately dense tree cover, (5) swordfern ( Polystichum munitum ) in moist sites associated with mature overstory conditions, and (6) swordfern-Oregon oxalis ( Oxalis oregano) along streamside slopes. Results Goshawk Surveys. We found no Northern Gos- hawks in the entire 3285 ha survey area. However, we did find two Sharp-shinned Hawks, seven Red- tailed Hawks ( Buteo jamaicensis) and two unidenti- fied raptors (too small to be goshawks). We also found two small, unoccupied accipiter nests (prob- ably Sharp-shinned Hawks built in previous years) . Vegetation Surveys. Of the 24 survey blocks cov- ered, 23 were dominated by a Douglas-fir overstory. The one remaining block was a fire-regenerated stand dominated by noble fir ( Abies procera ); this site was undisturbed by logging, except for access roads. For all 24 survey blocks combined, 63% of the area was in older (>80 yr) Douglas-fir with 85% canopy cover, 24% was conifer-hardwood mix (Douglas-fir, western hemlock, western redcedar, red alder, vine and bigleaf maple) with <65% can- opy cover, 7% was open mixed conifer (Douglas- fir, western hemlock, western red cedar) with 2^65% canopy cover, 3% was young (<80 yr) Doug- las-fir with ^85% canopy cover, 2% was older Douglas-fir with <30%) canopy cover, and 1% was clearcut areas, meadows or water. Ground cover within the survey blocks was dom- inated by understory types 4 and 5; these two types were present in 20 and 14 of the 24 stands, re- spectively. Vine maple, salal and swordfern were the most common species overall, and each survey block, with the exception of the stand of noble fir, had a dense shrub layer with 45-100% ground cov- er (* - 81%, SE = 3 %, N= 23). Twenty-three of the 24 blocks were previously disturbed by logging and fire. Historical Sightings. Records of previous sight- ings indicated that goshawks have been seen throughout the Coast Ranges every year, and dur- ing each month in any given year, for the past 15 years (Table 1), with apparent peaks in sightings in the spring and fall (Fig. 2). This would coincide with dispersing or migrating hawks. Numbers of observations were low, however, and prior to 1995, no confirmed evidence of reproductive activity had been documented. Few reports have described ag- gressive, territorial behavior or repeated sightings of adults in a particular area during the breeding season. Goshawk sightings in the Coast Ranges have been much lower than for other parts of the state (Fig. 3). Discussion Determining the “absence” or rarity of a species in a geographic region, and whether or not this is meaningful ecologically, can be a difficult task, es- pecially when the species is normally secretive and present in low densities. At least two elements are involved in making this determination. The first is that large-scale searches (both spatially and tem- porally) must be conducted because occupation of habitat by a species can vary over space and time (Morris 1987, Block and Brennan 1993, DeStefano et al. 1994, Keane and Morrison 1994). Our survey for goshawks over a single breeding season is ob- viously inadequate to address this concern, but coupled with surveys in the 1960s and 1970s, the consistent and widespread searches for Northern Spotted Owl and Marbled Murrelet nests during the 1970s-90s, and the reports of birdwatchers and agency biologists for the past 15 years, the tempo- ral and spatial scope of the search for goshawk nests in the Coast Ranges has been broad. The second element involves a determination, March 1997 Vegetation Structure and Goshawks 37 Table 1. Sightings of Northern Goshawks, grouped by year, in the Coast Ranges of western Oregon, 1980—95. Only records by experienced agency biologists or birders (levels 2 and 3; see text) were used. Unknown age or behavior indicates observer was unsure or information was not reported. Year Adult Age Imm. Unk. Observed Behaviors SOURCE a 1995 4 5 0 2 nests located on BLM lands BLM, OCWRU 1994 3 1 5 Hunting, calling, respond to owl tape, fly-by Private, OCWRU, BLM 1993 1 0 5 Unknown OB 19:57, BLM, ONHP, Private 1992 3 1 2 Fly-by, perched, soaring, unknown OCWRU, BLM, Private 1991 0 0 2 Unknown BLM, Private 1990 3 0 5 Near spotted owl nest, chasing spotted owl, unknown OB 16:94,185, BLM, Private 1989 1 0 1 Fly-by, unknown BLM, Private 1988 1 1 2 Unknown, soaring Private 1987 1 2 2 Unknown, aggressive interaction with red-tailed hawk OB 13:232,314, Private, Crannel and DeStefano 1992 1986 0 0 2 Unknown OB 12:212 1985 0 0 6 Unknown USFS, Private 1984 1 2 1 Unknown, soaring Private 1983 0 0 2 Unknown OB 10:130, Private 1982 0 0 4 Unknown Private 1981 0 0 4 Unknown Private 1980 0 0 2 Unknown Private “Source codes: BLM = Bureau of Land Management; OB = various issues (vol.:page) of the journal Oregon Birds (individual issues not listed in Lit. Cited); OCWRU = Oregon Cooperative Wildlife Research Unit; ONHP = Oregon Natural Heritage Pro- gram; Private = records of private individuals; USFS = U.S. Forest Service. probably on both objective and subjective bases, that a species is absent from a region because im- portant components of the habitat are lacking. Otherwise, the absence of a species from a region is meaningless. Our contention is that the absence, or at least rarity, of nesting goshawks in the Coast Ranges of western Oregon is an interesting phe- Figure 2. Distribution of Northern Goshawk sightings by month, consolidated for the years 1980-95, in the Coast Ranges of western Oregon. Adults and young from two nest sites located in June 1995 are not included. nomenon, and one that suggests something about vegetation structure on the distribution of this spe- cies. The Northern Goshawk has been called a forest habitat generalist (Reynolds et al. 1992). The lo- cation of the Coast Ranges of western Oregon in relation to the geographic range of the goshawk in North America, and the general forest conditions (mixture of mature and second-growth coniferous forest and openings) that exist there, appear to in- dicate that the Coast Ranges should be part of the breeding range of this species. Rarity is common along the geographic boundaries of a species’ range, yet goshawks nest in areas north, south, and east of the Coast Ranges. Why is there a gap in the breeding range of this species and what is it about the Coast Ranges of Oregon that prevents goshawks from nesting there? It could be related to climate. The Coast Ranges receive 150-300 cm of annual rainfall (Franklin and Dryness 1973), much of it falling during the spring breeding season. However, Northern Goshawks breed commonly on the 38 S. DeStefano and J. McCloskey Vol. 31, No. 1 Figure 3. Relative abundance of Northern Goshawks by county in Oregon, based on records of the Oregon Au- dubon Society and Oregon Department of Fish and Wild- life. Abundance ratings are common (black), occasional (dark gray), rare (light gray), and unknown (white). Note that coastal regions of Lane and Douglas Counties would likely be rated as having rare (light gray) sightings of goshawks, but both counties are rated as occasional (dark gray) because large portions of each reach into the Cascade Mountain foothills and range, where goshawks are seen more often. Olympic Peninsula of Washington (E. Forsman pers. comm.) and in southeastern Alaska (K. Titus pers. comm.), where annual precipitation equals or exceeds levels received in Oregon’s Coast Rang- es. A second possibility might involve predation or competition from other raptors. Great Horned Owls ( Bubo virginianus) and Red-tailed Hawks fre- quently interact with Northern Goshawks and of- ten use their nests (Moore and Henny 1983, Cran- nell and DeStefano 1992, Rohner and Doyle 1992) . It is unlikely, though, that this form of predation and/or competition is more intense in the Coast Ranges than other parts of the goshawks’ geo- graphic range. The third explanation involves vegetation struc- ture as it relates to prey availability. Goshawks tend to hunt in the ground-shrub and shrub-canopy zones of the forest (Reynolds and Meslow 1984). A dense shrub layer is characteristic of most forest areas of the Coast Ranges and disturbances such as logging and fire have decreased mature oversto- ry canopy closure, allowing more sunlight to reach the ground. These conditions, coupled with high levels of rainfall, have resulted in increased under- story stem densities and dense, lush undergrowth on many sites (Franklin and Dyrness 1973). Prey, which are varied and abundant in the Coast Rang- es, include such species as snowshoe hares ( Lepus americanus), brush rabbits ( Sylvilagus bachmani), Douglas squirrels ( Tamiasciurus douglasii ), Ruffed Grouse ( Bonasa umbellus). Mountain Quail ( Oreor - tyx pictus ) , Northern Flickers ( Colaptes auratus) and other woodpeckers, and Stellar’s and Gray Jays (Oy- anocitta stellari and Perisoreus canadensis, respective- ly) . Many of these prey species may be difficult for goshawks to capture because of the dense under- story conditions that exist throughout most of the Coast Ranges (Reynolds and Meslow 1984). In ad- dition, the larger biomass prey species (lago- morphs, grouse) may be more important to breed- ing goshawks than smaller prey (jays, woodpeck- ers) (Reynolds et al. 1992), and low availability of larger prey may depress reproductive potential (Alaska Dept. Fish and Game 1993). Dense under- story conditions would make the larger, ground- dwelling species more difficult to capture (Reyn- olds et al. 1992). Others have found the distribution of raptor for- aging to be inversely related to the density of plant cover (Southern and Lowe 1968, Wakely 1978, Ba- ker and Brooks 1981, Bechard 1982, Collopy and Bildstein 1987). Preston (1990) summarized these findings by describing raptor hunting distribution as a function of both prey abundance and avail- ability, which in turn is a function of a suite of environmental factors, including habitat character- istics (e.g., vegetation structure). In fact, suitable foraging habitat may be more important than nest- ing habitat in determining the distribution of gos- hawks in boreal forest (Widen 1989). The impor- tance of prey in the distribution and management of northern goshawks has been emphasized in the management guidelines of Reynolds et al. (1992). If a relationship between vegetation structure and availability of prey does indeed exist, then the forest conditions present in the Coast Ranges of Oregon may limit prey availability to goshawks and thus prevent or depress reproductive activity, de- spite potentially suitable nesting substrate and ad- equate populations of prey. Acknowledgments We thank W. Dean, D.S. Hopkins and J.A. Reid (Bu- reau of Land Management) , E.D. Forsman and S .J. Mad- sen (U.S. Forest Service), C. Bruce and T, O’Neil (Ore- gon Department of Fish and Wildlife) , M. Hansen, S.K Nelson, J.A. Thrailkill, and M. Wilson (Oregon Cooper- March 1997 Vegetation Structure and Goshawks 39 ative Wildlife Research Unit), and E. Schauber and S. Spean (Oregon State University) for providing field sup- port and/or information on goshawk sightings. Special thanks to R. Bayer, E. Eltzroth, D. Faxon, G. Gilson, T. Mickel and H. Nehls for sharing their private records. The manuscript benefited from reviews by R Beier, S.K. Daw, S.M. Desimone, E.D. Forsman, M.T. McGrath, M.L. Morrison and R.T. Reynolds. Additional logistical support was provided by the Oregon Cooperative Wildlife Re- search Unit and the Department of Fisheries and Wildlife at Oregon State University. Literature Cited Alaska Department of Fish and Game. 1993. A sum- mary of survey, radio-telemetry, and other results re- garding goshawk field studies in southeast Alaska. Alaska Dep. of Fish and Game, Juneau, AK U.S.A. Baker, J.A. and R.J, Brooks. 1981. Distribution patterns of raptors in relation to density of voles. Condor 83: 42-47. Bechard, M.J. 1982. Effect of vegetative cover on for- aging site selection by Swainson’s Hawks. Condor 84: 153-159. Block, W.M. and L.A. Brennan. 1993. The habitat con- cept in ornithology: theory and applications. Current Ornithol. 11:35—91. COLLOPY, M.W. AND K.L. BlLDSTEIN. 1987. Foraging be- havior of Northern Harriers wintering in southeast- ern salt and freshwater marshes. Auk 104:11-16. Crannell, D. and S. DeStefano. 1992. An aggressive interaction between a Northern Goshawk and a Red- tailed Hawk./. Raptor Res. 26:269-270. DeStefano, S., S.K Daw, S.M. Desimone and E.C. Mes- low. 1994. Density and productivity of Northern Goshawks: implications for monitoring and manage- ment. Studies in Avian Biol. 16:88—91. Forsman, E.D., S. DeStefano, M.G. Raphael and R.J. GutiErrez, [Eds.]. 1996a. Demography of the North- ern Spotted Owl. Studies in Avian Biol. 17. , RJ. Loschl, R.K Forson and D.K Barrett. 1996b. Demography of Northern Spotted Owls on the Siuslaw National Forest, Oregon. Studies in Avian Biol. 17:47-52. Franklin, J.F. and C.T. Dyrness. 1973. Natural vegeta- tion of Oregon and Washington. USD A Forest Ser- vice, Gen. Tech. Rep. PNW-8, Portland, OR U.S.A. Joy, S. M., R.T. Reynolds and D.G. Leslie. 1994. North- ern Goshawk broadcast surveys: hawk response vari- ables and survey cost. Studies in Avian Biol. 16:24—30. Keane, J-J- and M.L. Morrison. 1994. Northern Gos- hawk ecology: effects of scale and levels of biological organization. Studies in Avian Biol. 16:3—11. Kennedy, P.L. and D.W. Stahlecker. 1993. Responsive- ness of nesting Northern Goshawks to taped broad- casts of 3 conspecific calls. J. Wildl. Manage. 57:249— 257. Marshall, D.B. 1992. Status of the Northern Goshawk in Oregon and Washington. Audubon Society of Port- land, Portland, OR U.S.A. Moore, K.R. and C.J. Henny. 1983. Nest site character- istics of three coexisting accipiter hawks in northeast- ern Oregon. Raptor Res. 17:65—76. Morris, D.W. 1987. Ecological scale and habitat use. Ecology 68:362-369. Nelson, S.K and S.G. Sealy, [Eds.]. 1995. Biology of the Marbled Murrelet: inland and at sea. Northwest. Nat. 76. Palmer, R.S. 1988. Handbook of North American birds. Vol. 4. Yale Univ. Press, New Haven, CT U.S.A. Preston, C.R. 1990. Distribution of raptor foraging in relation to prey biomass and habitat structure. Condor 92:107-112. Reynolds, R.T. 1975. Distribution, density, and produc- tivity of three species of accipiter hawks in Oregon. M.S. thesis, Oregon State Univ., Corvallis, OR U.S A. . 1978. Food and habitat partitioning in two groups of coexisting Accipiter. Ph.D. dissertation, Or- egon State Univ., Corvallis, OR U.S.A. , R.T. Graham, M.H. Reiser, R.L. Bassett, P.L. Kennedy, D.A. Boyce, Jr., G. Goodwin, R. Smith and E.L, Fisher. 1992. Management recommendations for the Northern Goshawk in the southwestern Unit- ed States. USDA Forest Serv., Gen. Tech. Rep. RM-217, Fort Collins, CO U.S.A. , and E.C. Meslow. 1984. Partitioning of food and niche characteristics of coexisting A ccipiter during breeding. Auk 101:761—779. , E.C. Meslow and H.M. Wight. 1982. Nesting habitat of co-existing Accipiter in Oregon. J. Wildl. Manage. 46:124-138. , and H.M. Wight. 1978. Distribution, density, and productivity of accipiter hawks breeding in Ore- gon. Wilson Bull. 90:182-196. Rohner, C. and F.I. Doyle. 1992. Food-stressed Great Horned Owl kills adult goshawk: exceptional obser- vation or community process? / Raptor Res. 26:261- 263. Southern, H.N. and V.P.W. Lowe. 1968. The pattern of distribution of prey and predation in Tawny Owl ter- ritories./. Anim. Ecol. 37:75-97. Thrailkill, J.A. and L.S. Andrews. 1996. Presence of breeding Northern Goshawks in the Coast Ranges of Oregon. /. Raptor Res. 30:248—249. Wakely, J.S. 1978. Factors affecting the use of hunting sites by Ferruginous Hawks. Condor 80:316-326. WidLn, P. 1989. The hunting habitats of goshawks Accip- iter gentilis in boreal forests of central Sweden. Ibis 131: 205-213. Received 1 April 1996; accepted 23 November 1996 J. Raptor Res. 31(1) :40-43 © 1997 The Raptor Research Foundation, Inc. FOOD HABITS OF THE LANNER FALCON (. FALCO BIARMICUS FELDEGGI1) IN CENTRAL ITALY Federico Morimando, Francesco Pezzo Dipartimento di Biologia Evolutiva, Gruppo di Etologia e Ecologia Comportamentale, Universita di Siena, via P.A. Mattioli 4, 53100 Siena, Italy Alessandro Draghi Proeco Gestione Fauna E Ambiente Studio Associato, via Uopini 19, 53100 Siena, Italy Abstract. — The diet of the Lanner Falcon (Falco biarmicus feldeggii) was studied in an area of central Italy for one yr. Diet composition differed according to the methodology used for data collection with mammals, small passerines and insects found more frequently or exclusively in pellets. Conversely, anal- ysis based on pluckings and observations of prey taken to the nest indicated the diet consisted of only nonpasserine and large passerine birds. Overall, the birds including European Starling (Sturnus vulgaris) , Green Woodpecker ( Picus viridis) , and Jay ( Garrulus glandarius) were the most common prey taken by Lanner Falcon. Mammals taken included wood mouse (Apodemus sp.), common dormouse ( Muscardinus avellanarius ) and Savi’s pine vole ( Microtus savii ). Key WORDS: Lanner Falcon; Falco biarmicus ; diet; feeding ecology; Italy. Habitos alimenticios del Falco biarmicus feldeggii en la region central de Italia. Resumen. — La dieta del Falco biarmicus feldeggii fue estudiada durante un ano en un area de la region central de Italia. Se observo que la composicion de la dieta variaba de acuerdo con la metodologia utilzada para la recoleccion de datos, de tal manera que se encontraron pequenas aves passerinas, marmferos e insectos mas frecuentemente o exclusivamente en las egagropilas. Por el contrario, al analizar la presa en el nido y al realizar desplumamiento, se hallo que la dieta consistfa unicamente de aves no passerinas y grandes passerinas. Sobre todo aves como Sturnus vulgaris, Picus viridis, Garrulus glandarius, fueron la presa mas comun de F biarmicus feldeggii. Los marmferos incluidos en la dieta eran Apodemus sp., Muscardinus avellanarius y Microtus savii. [Traduccion de Agustina Lanusse] The European subspecies of the Lanner Falcon (Falco biarmicus feldeggii ) is limited in its distribu- tion to southern Italy and the Balkans with a pop- ulation estimated to be just a few hundred pairs (Gensb0l 1992). There are <60 pairs in the Italian peninsula but, in Sicily, the population consists of >80 pairs (Massa et al. 1991). In Tuscany, central Italy, five breeding pairs are known in a 8300 km 2 area, with a density of 1 pair/ 1660 km 2 . The biology of the Lanner Falcon in Europe is poorly studied (Cramp and Simmons 1980). Most work has been done on the African subspecies (Cade 1965, Maclean 1984, Goodman and Haynes 1989, 1992) and very few studies have been con- ducted in the Mediterranean area (Mebs 1959, Bonora and Chiavetta 1975, Massa et al. 1991). Lanner falcons nest on cliffs, making it difficult to find perches and roosts where food remains may be collected for dietary studies (Massa et al. 1991, Chiavetta 1992). Apart from studies conducted in Sicily (Mebs 1959, Massa et al. 1991), no detailed research has been done so far on the food habits of Lanner Falcons. Herein, we present results of an analysis of the diet of a pair of Lanner Falcons in central Italy based on remains collected at plucking areas and the nest site and observations of prey delivered to the nest. Study Area The study was conducted in a rural area named “Crete Senesi” in Tuscany, central Italy, at 200-300 m elevation. The climate of the area is temperate, locally subarid, with an average annual temperature of 18°C. It is very hilly, open and eroded at the bases of hills which has created an abundance of small clay/sand cliffs ranging from 10- 40 March 1997 Lanner Falcon Food Habits 41 Passeriformes □ Non Passeriformes ■ Mammalia Figure 1. Composition of the diet of Lanner Falcons during the breeding season (December-June) using three different methods: direct observations of prey taken to the nest, pellets, and plucked remains. Numbers above the bars refer to sample size for each method. 40 m in height. These small cliffs are used for nesting by Lanner Falcons (Morimando et al. 1994). The vegetation of the area is dominated by pastureland with small oak- woods ( Quercus pubescens ) interspersed with cereal crops and small patches of olive trees ( Olea europea ). Methods We observed a pair of Lanner Falcons through two re- productive seasons from spring 1994-summer 1995. We made direct observations on prey brought to the nest and we collected prey remains at the bases of nest cliffs. By watching activity patterns, we also found two main perches where the falcons plucked and ate their prey. From June 1994, we collected prey remains at these perches twice a month by searching the ground carefully under each perch. Prey were often identified in the field based on feathers and fresh remains. Remains that were not readily recognized as well as pellets were examined using a dissecting microscope and remains were identi- fied using museum specimens and identification guides (Brown et al. 1987, Svensson 1992). We also used taxo- nomic keys (Chaline et al. 1974) and a sample of speci- mens collected locally to compare bones and skeletal re- mains. We calculated the % occurrence and % biomass of each prey category in the diet based on each of the three methods. Prey biomass was determined based on the av- erage weight of prey taxa reported in the literature (Cramp 1977-94, for birds; Macdonald and Barret 1993, for mammals). We calculated the mean prey weight (MPW) using the formula: Total Biomass/Total Numbers of each prey species. Results We observed the nest 16 times for a total of 96 hr. We also collected plucked remains 26 times and collected 59 pellets under perches. Altogether, we identified 15 prey items delivered to the nest, 78 prey items from pellets, and 71 prey items from plucked remains (Table 1). The mean number of prey per collecting event was 0.94 prey for obser- vations, 1.32 for pellets and 2.73 for pluckings. Direct observations at the nest showed that pas- serine birds were delivered to the nest more fre- quently than nonpasserines (Table 1). Although most of these birds were not identified, the main prey species taken to the nest appeared to be the European Starling ( Sturnus vulgaris ) and Blackbird ( Turdus merula ) . The MPW of prey brought to the nest was 121 g (SE = 38.27). During nest obser- vations, we twice saw the male storing small prey in a cliff hole that was used as a cache. Nearly identical percentages of passerines and nonpasserines were identified from plucked re- mains (Table 1). The MPW of prey based on this type of analysis was 182.01 g (SE = 216.94), con- siderably larger than that of prey taken to the nest {t = 2.47, df = 20; P < 0.02). Based on this type of analysis, the main prey were Green Woodpecker ( Picus viridis ) , Jay ( Garrulus glandarius) and Euro- pean Starling. Most of the biomass in the Lanner diet was from nonpasserines of which the Green Woodpecker was the largest and most frequently taken species. Bone fragments in pellets were small and diffi- 42 Morimando et al. Vol. 31, No. 1 Table 1. Numbers (N), percentage of biomass (%B) and percentage of occurrence (%0) of prey items in Lanner Falcon diet in Tuscany, Italy, according to collecting methods. Pellets Pluckings Observations Prey Item N %B %o N %B %o N %B %o MAMMALS Muscardinus avellanarius 1 0.38 1.28 0 0 0 0 0 0 Apodemus spp. 3 0.96 3.85 0 0 0 0 0 0 Microtus savii 1 0.38 1.28 0 0 0 0 0 0 Rodentia ind. 4 0.9 5.13 0 0 0 0 0 0 Total mammals 9 2.62 11.54 0 0 0 0 0 0 BIRDS Falco tinnunculus 0 0 0 4 6.19 5.63 0 0 0 Phasianus colchicus 0 0 0 2 10.83 2.82 1 38.57 6.67 Columba livia 3 11.49 3.85 7 16.25 9.86 1 16.53 6.67 Columba palumbus 1 5.74 1.28 1 3.48 1.31 0 0 0 Streptopelia turtur 1 19.1 1.28 3 3.48 4.23 0 0 0 Athene noctua 0 0 0 1 1.08 1.41 0 0 0 Apus apus 0 0 0 1 0.31 1.41 0 0 0 Picus viridus 11 35.10 14.10 16 30.95 22.54 0 0 0 Total nonpasserines 16 54.24 20.51 35 72.58 49.30 2 55.10 13.33 Alauda arvensis 0 0 0 2 0.57 2.82 0 0 0 Turdus merula 6 6.13 7.69 1 0.62 1.41 3 13.22 20 Turdus philomelos 0 0 0 3 1.58 4.23 0 0 0 Garrulus glandarius 0 0 0 11 14.47 15.49 0 0 0 Pica pica 3 6.89 3.85 3 4.18 4.23 0 0 0 Corvus monedula 0 0 0 0 0 0 1 10.74 6.67 Sturnus vulgaris 37 28.33 47.44 11 5.11 15.49 5 16.53 33.33 Emberiza calandra 0 0 0 1 0.29 1.41 0 0 0 Anthsu pratensis 0 0 0 1 0.14 1.41 0 0 0 Passeriformes ind. 7 1.79 8.97 3 0.46 4.23 4 4.41 26.67 Total passerines 53 43.14 67.95 36 27.42 50.70 13 44.90 86.67 Total birds 69 97.38 88.46 71 100 100 15 100 100 cult to recognize. Nevertheless, this form of anal- ysis showed that mammals also occur in the diet of the Lanner Falcon (Table 1). Mammals identified were wood mouse ( Apodemus spp.), common dor- mouse ( Muscardinus avellanarius ) and Savi’s pine vole ( Microtus savii). Also, some pellets were made solely of insects remains, mainly Formicidae and small Coleopterans. A careful inspection of pellets showed that most insects were inside the gizzard remains of Green Woodpeckers and Starlings. Pas- serines were the most frequent prey found in pel- lets, but nonpasserine species accounted for slight- ly more biomass using this method of analysis. Main prey species represented in the pellets were the European Starling, Green Woodpecker and Blackbird. The MPW in pellets was 100.44 g (SE = 163.01) which was smaller than the MPW found in pluckings ( t — 2.80, df = 20; P < 0.01). Medium-sized passerines (x = 98.41 g) and large nonpasserines (x = 268.70 g) were the primary spe- cies found in plucked remains. In pellets, mammals (x = 22-71 g) and mainly passerines of small size (x = 63.70 g) were present. Small passerines (x = 62.69 g) and few nonpasserines of large size (x = 265.62 g) were taken to the nest. The differences among plucked remains, pellets and direct observations as to the three main categories of prey found (passer- ines, nonpasserines and mammals) , were highly sig- nificant (x 2 - 13.46, df = 4, P < 0.009). Discussion We concluded that use of a single method to determine the diet of Lanner Falcon biases the re- March 1997 Lanner Falcon Food Habits 43 suits. Data from plucking sites provided more in- formation about diet composition per searching ef- fort than other collecting methods. Pellets are also important because they provided more complete information on the composition of prey, especially that of mammals and insects. Direct observations of prey taken to the nest appear to be unnecessary since they add little to quantitative information on the Lanner Falcon’s diet. The use of plucking data alone may result in overestimates of large- and me- dium-size birds, as generally longer and heavier feathers fall directly beneath the perches, while the light feathers of small birds tend to disperse (e.g., in the wind) . Analysis of pellets alone tends to over- estimate the amount of small passerines and mam- mals. Lanner Falcons generally swallow small birds almost entirely, whereas they tear flesh and feath- ers from larger birds. This probably accounts for the high frequency of small passerine remains found in pellets. We feel that it would be best to evaluate information obtained from all the collect- ing methods to most accurately assess the diet of Lanner Falcons. We found the diet of Lanner Falcon in central Italy to be qualitatively different from that of Lan- ner Falcon in other areas. Birds comprised 88- 100% of the diet in our study area depending on the method of diet analysis used. These results were consistent with those of Cramp and Simmons (1980), except we found for the first time that Green Woodpecker is taken by Lanners. We also confirmed that they eat small mammals but that they do not appear to eat reptiles and insects as reported by Massa et al. (1991). In Sicily, birds (90.4%), reptiles (4.1%) and mammals (2.7%) are in the diet of Lanners (Mebs 1959), with Jackdaw ( Corvus monedula ) , Lesser Kestrel ( Falco naumanni) and feral Pigeon ( Columba livid) preyed upon the most. Magpie ( Pica pica) and Spanish Sparrow {Passer hispaniolensis) are also preyed upon in Sicily (Massa et al. 1991) and small mammals and rep- tiles account for only a small percentage of the diet (4% and 2.3%, respectively). Acknowledgments We are grateful to G. Fratalocchi, R. Nardi and G. Cap- pelli for their help with fieldwork. We are indebted also to G. Manganelli and N. Baccetti for their invaluable as- sistance in identifying feathers, bones and skeleton re- mains of birds. Thanks are also due to F. Cancelli for allowing us to get free access to the material of the Zoo- logical Museum of the “Accademia dei Fisiocritici” in Siena and to P. Galeotti, G. Bogliani and F. Spina for critical comments on the manuscript. S. Lovari made use- ful suggestion on the first draft of the paper and im- proved the English text. J. B. Buchanan and two anony- mous referees provided useful suggestions and improved data presentation. Literature Cited Bonora, M. and M. Chiavetta. 1975. Contribution a l’etude du Faucon lanier Falco biarmicus feldeggi en It- alie. Nos Oiseaux 33:153—168. Brown, R., J. Ferguson, M. Lawrence and D. Lees. 1987. Tracks and signs of the birds of Britain and Europe. Croom Helm, Bromley, London, UK. Cade, T.J. 1965. Relations between raptors and colum- biforms at a desert water hole. Wilson Bull. 77:340- 345. Chaline, J., II. Baudvin, D. Jammon and M.C. Saint Gi- rons. 1974. Les proies des rapaces. Doin, Paris, France. Chiavetta, M. 1992. II Lanario. Pages 674-678 in P. Bri- chetti, P. De Franceschi and N. Baccetti [Eds.], Fauna d’ltalia-AVES I. Vol. I. Edizioni Calderini, Bologna, It- aly. Cramp, S. [Ed.]. 1977—94. The birds of the Western Pa- learctic, Vol. I-IX. Oxford Univ. Press, Oxford, UK. Cramp, S. and K.E.L. Simmons [Eds.]. 1980. The birds of the Western Palearctic, Vol. II. Oxford Univ. Press, Oxford, UK. Gensbol, B. 1992. Collins guide to the birds of prey of Britain and Europe with North Africa and the Middle East. Harper Collins, London, UK Goodman, S.M. and C.V. Haynes. 1989. The distribu- tion, breeding season and food habits of the Lanner from the Eastern Sahara. Nat. Geogr. Res. 5:126-131. . 1992. The diet of the Lanner Falco biarmicus in a hyper arid region of the Eastern Sahara. J. Arid En- viron. 22:93-98. Macdonald, D.W. and P. Barret. 1993. Mammals of Britain and Europe. Harper Collins, London, UK. MacLean, G. L. 1984. Robert’s birds of Southern Africa. New Holland Publishers, London, UK Massa, B., F. Lo Valvo, M. Siracusa and A. Ciaccio. 1991. II lanario ( Falco biarmicus feldeggi Schlegel) in Italia: status, biologia e tassonomia. Naturalista siciliano 15:27-63. Mebs, T. 1959. Beitrag zur Biologie des Feldeggsfalken Falco biarmicus feldeggi.. Vogelwelt 80 : 1 42—1 49 . Morimando, F., F. Pezzo, A. Draghi and G. Frataloc- chi. 1994. Prima nidificazione di Lanario Falco biar- micus in provincia di Siena e note sulla locale distri- buzione storica. Avocetta 18:157-159. Svensson, L. 1992. Identification guide to European pas- serines. Stockholm, Sweden. Received 2 April 1996; accepted 1 December 1996 J Raptor Res, 31(1):44— 53 © 1997 The Raptor Research Foundation, Inc. NESTING DISTRIBUTION AND POPULATION STATUS OF U.S. OSPREYS 1994 Lawrence M. Houghton 1 and Larry M. Rymon Biology Department, East Stroudsburg University, East Stroudsburg, PA 18301 Abstract. — Ospreys ( Pandion haliaetus) once nested throughout most of the U.S. The decline in this population due to biocide use has been well documented, as has its recovery following the U.S. ban on DDT in 1972. A general increase in the nesting distribution and abundance of Ospreys was reported in the U.S. in 1981 but there was limited dispersal into states with low or extirpated populations. We conducted a nationwide nesting survey of nesting Ospreys in 1994, updating the 1981 data. Our data indicate a dramatic increase in the U.S. Osprey population from —8000 nesting pairs in 1981 to —14 200 in 1994. The most dramatic increases were seen in traditional nesting areas, with some new nesting in the interior U.S. Hacking projects, construction of reservoirs, nest platform management and in- creased public relations have contributed to the growth of this nesting population. Key Words: Pandion haliaetus; Ospreys ; population status ; limiting factors', dispersal ; management. Resumen. — Aguila pescadoras (Pandion haliaetus), anidaba coraunmente en la mayor parte de los EE.UU. hasta que sus poblaciones fueron afectadas severamente debido al uso de insecticidas. Sin embargo, despues de la prohibicion del uso de DDT en 1972, la especie parece estar recuperandose. En 1981, se reporto un incremento general en la distribution de sitios de nidacion y abundancia de la especie. Sin embargo, la recuperacion parece ser menor en los estados en los que las poblaciones fueron mas afectadas o extirpadas. En este estudio, reportamos el resultado de un muestreo nacional de nidos de Aguila Pescadora, que realizamos en 1994 y que reemplaza los datos de 1981. Nuestros datos indican un incremento dramatico en las poblaciones del Aquila de —8000 pares anidando en 1981 a —14 200 in 1994. Los incrementos mas dramaticos fueron observados en areas tradicionales de nidacion, y en algunos sitios neuvos en el interior de los EE.UU. Proyectos de reintroduction, construction de presas, implementation de platformas de nidacion, y un mejor entendimiento del problema por parte del publico, han sido factores importantes que han contribuido a la recuperacion de la especie. [Traduction de Jorge Vega Rivera] Ospreys generally occur along rivers, lakes, sea coast bays and estuaries, reservoirs, small streams and ponds, or any body of water where fish, their principal food, are available (Poole 1989a). Histor- ical data, though limited, indicate that Ospreys once nested in suitable habitats throughout most of the contiguous U.S., but their numbers were never equally distributed throughout the country. Records suggest that the Central States Region (Fig. 1) had the smallest population, which was nearly extirpated by the early 1900s. The entire U.S. Osprey population declined precipitously throughout the 1950s, 60s and early 70s, a result of widespread use of chlorinated hydrocarbon in- secticides and habitat destruction (Ames and Mer- 1 Present address: 134 Cheatham Hall, Virginia Polytech- nic Institute and State University, Blacksburg, VA 24061- 0321 U.S.A. sereau 1964, Ames 1966, Peterson 1969, Postupal- sky 1969, Henny and Ogden 1970). The popula- tion was “Redbook listed” under “rare and endan- gered fish and wildlife of the U.S.” and classified as “status undetermined” by the U.S. Fish and Wildlife Service in 1966 and 1968 (Henny 1977). Henny (1983) conducted a comprehensive sur- vey of the distribution and abundance of the entire U.S. Osprey population in 1981. His research in- dicated a general increase in the overall popula- tion, with limited dispersal into states with low or extirpated populations. This population enhance- ment was primarily the result of greater reproduc- tive output after the U.S. ban on DDT in 1972. The slow dispersal rate was principally a consequence of high natal site fidelity, especially for males (Spitzer et al. 1983). Such slow, limited dispersal forced hacking methods as a means of restoration in areas with low or extirpated populations. In the 44 March 1997 1994 Osprey Population 45 Wmtcni Region Central States GttffStaiatflorida MkHSoutli Atlantic Gnat Lakes Region Eastern Interior Northeast Region 1981 1476 pain 1994 2970 pain 1981 Opain 1994 2 pain 1981 1793 pain 1994 2846 pain H 1981 2263 pain 1994 4750 pain 1981 579 pain 1994 1066 pairs 1981 1291 pain 1994 2505 pain Figure 1. U.S. Osprey regions (1994). late 1970s and early 1980s, Pennsylvania, New York and Tennessee began programs to restore low or extirpated populations (Hatcher and Hammer 1983, Schaadt and Rymon 1983). Our objectives in this study were: 1 ) to survey 48 states for information concerning the general trend of U.S. Osprey nesting populations, 2) to compare the 1994 and 1981 distribution and abun- dance of U.S. nesting Ospreys, 3) to examine any changes in population numbers and range expan- sion that may have occurred since 1981, and 4) to suggest continued management options. Methods and Materials Data on the U.S. Osprey population were obtained from professional sources in each of the lower 48 states. Biologists from state and federal agencies and individual Osprey researchers were contacted by telephone or in person. A follow up questionnaire, state distribution map and completed sample questionnaire was sent to each of the contacted individuals. Survey questions addressed the distribution, abundance, historical data, nesting prefer- ences, reproductive success and hacking status of the U.S. Osprey population. State breeding bird adases and other published and unpublished sources were also re- viewed, as were state bird books, for information on the historical population and broad population and dispersal patterns and trends. To provide an estimate of the com- pleteness of nesting data, we completed a list of nesting survey methods used by each state (some states had no recent surveys) (Table 1). For purposes of evaluation, Henny (1983) divided the entire U.S. Osprey population into five regional popula- tions: Pacific Northwest, Western Interior, Great Lakes Region, Atlantic Coast, and Gulf States and Florida. An alternate approach was taken in our study. Comparable regional populations (Western Region, Great Lakes, Northeast, Mid and South Adantic Coast, and Gulf States and Florida) were established, but additional regions (Eastern Interior and Central States) were added to ex- amine dispersal patterns, recent population fluctuations and migratory movements between wintering and breed- ing areas (Fig. 1) Results and Discussion Mid and South Atlantic Coastal Region. This re- gion begins in Delaware Bay and encompasses all of Delaware, Maryland, the coastal plain of Virgin- ia, North and South Carolina and Georgia (Fig. 2). It supports one of the largest concentrations of nesting Ospreys in the world (Chesapeake Bay) (Henny 1983, Poole 1989a). It is within this major estuarine system that most Ospreys in Virginia and Maryland breed. The number of nests in the Ches- 46 Houghton and Rymon Vol. 31, No. 1 Table 1. Approximate number of Osprey nesting pairs in the U.S. (1981 vs. 1994). State Nesting pairs 1981 1994 Source Alabama 2 23 2 R. Clay, Alabama Game and Fish Arizona 4 25-35 3 G. Beatty, Bald Eagle Management Coordinator Arkansas 0 0 4 K Yaich, Arkansas California 359 (1975) 500-700 5 Ron Jurek, California Dept, of Fish and Game Colorado 9 17 2 J. Craig, Colorado Division of Wildlife Connecticut 25 95 2 J. Victoria, CT Supervisor of Wildlife Research Delaware 56 75-85 2 L. Galvin-Innvaeer, DE Div. of Fish and Wildlife Florida 1750 2500-3000 4 Mark Westall, President TIOF Georgia 95 225-275 3 J. Ozier, GA Dept, of Natural Resources Idaho 323 (1974-80) 400-425 4 W. Melquist, Idaho Fish and Game Department Illinois 0 O 4 V. Kleen, Department of Conservation Indiana 0 I s J. Castrale, Indiana Nongame Biologist Iowa 0 0 4 B. Harrisman, IA Dept, of Natural Resources Kansas 0 0 4 J. Zimmerman, Div. of Biology, KSU Kentucky 0 16 1 D. Yancy, KY Dept, of Fish and Wildlife Resource Louisiana 1 10 3 S. Shively, LA Dept, of Wildlife and Fisheries Maine 1000 1 300-1 800 5 D. Hudson, ME Fish and Game Maryland 847 (1973-75) 1 000-1 400 4 S. Cardano, MD Dept of Natural Resources Massachusetts 32 (1980) 260 2 B. Davis, MA Division of Fisheries and Wildlife Michigan 123 223 1 S. Postupalsky pers. comm. Minnesota 160 350-450 2 M. Martell, UMN Raptor Center Mississippi 40 55-65 2 M. Woodrie, MS State Ornithologist Missouri 0 0 4 W. Crawford, Raptor Res. Tyson Research Center Montana 149 500-600 4 D. Flath, Montana Fish, Wildlife and Parks, MSU Nebraska 0 O 4 J. Dinan, NE Game and Parks Commission Nevada 1 4 3 G. Herron, NV Department of Wildlife New Hampshire 8 29 2 C. Martin, Audubon Society of New Hampshire New Jersey 87 200 2 (1993) C. Clark, Endangered + Nongame Species Program New Mexico 0 2 3 S. O. Williams III, NM Dept, of Game and Fish New York 120 315 2 B. Loucks, NY Endangered Species Unit North Carolina 450 (1974) 800-1 200 4 R. Wilson, NC Wild Resource Commission North Dakota 0 0 4 C. Grondahl, ND Game and Fish Department Ohio 0 l 3 D. Case, Ohio Division of Wildlife Oklahoma 0 0 4 S. Sherrod, Dir. G.M. Sutton Avian Res. Center Oregon 308 (1976) 675-700 2 C.J. Henny, NBS Leader NW Research Station Pennsylvania 0 20 2 L.M. Rymon, Environmental Studies Dir., ESU Rhode Island 19 442 L. Suprock, Div. Fish, Wildl. and Estuarine Res. South Carolina 151 (1979) 800-1 000 4 T. Murphy, South Carolina South Dakota 0 2 3 D. Backlandi, SD Game and Fish Department Tennessee 5 66 2 B. Hatcher, TN Nongame and Endangered Species Texas 0 3 3 B. Ortego, Biologist, TX Utah 12 30 3 (1995) S. Cranney, UT Division of Wildlife Resources Vermont 0 12 2 S. Parren, VT Fish and Wildlife Virginia 722 (1973-75) 1300-1500 4 (1987) Westall 1990 Washington 229 350-400 2 K. McCallister, WA Department of Wildlife West Virginia 0 3 3 S. Butterworth, WVA Div. of Wildlife Resources Wisconsin 176 39 1 1 D. Flaspohler, WN Bureau of Endangered Res. Wyoming Estimated Total 82 (1974-81) 8000 150-200 4 12,769-15,603 B. Oakleaf, WY Game and Fish Department Completeness of data 1 Both aerial and ground surveys, this percent was established by Henny et al. (1974) for the efficiency of combined aerial and ground March 1997 1994 Osprey Population 47 Table 1 . Continued. surveys and based on visibility rates. 2 Less intense aerial, ground, or boat surveys (Osprey nests recorded during Bald Eagle aerial surveys, or incomplete aerial and ground surveys). 3 Intense local surveys, or only surveying/monitoring of suspected and/or traditional nesting areas. 4 No statewide surveys, information acquired from local biologist and/ or other individuals aware of current nesting and population trends. 5 No statewide data available, therefore, a percent increase for regional data was determined and this percent increase was applied to the most recent statewide nesting data in order to provide a current estimate. 6 Very low number of nesting pairs (<15), no statewide survey (* states with <15 nests are listed under other categories where appropriate) . apeake Bay area has increased slowly over the past 20 years. Apparendy, there has been a large in- crease in occupied nests along the Patuxent River from 22 in 1973 to 72 in 1994 (S. Cardano pers. comm.). Overall, the number of Ospreys nesdng in the Bay area appears to be leveling off and Spitz- er (1989) suggests that the population is nearing carrying capacity. Spitzer (1989) estimated the mean age at first breeding in part of the Chesa- peake Bay region to be about two years higher than that of the region between New York City and Bos- ton (5.7 vs. 3.7 yr), apparently the result of limited nest-site availability. This delay should slow popu- lation growth rate by bringing mortality into bal- ance with natality (Spitzer 1989, Poole 1989b). The number of Ospreys nesting in Virginia, the coastal Carolinas and Georgia has more than doubled since 1981. Several new interior sites have contrib- uted to this growth (T. Murphy and J. Ozier pers. comm., We stall 1990). Despite fluctuations, the overall number of nest- ing pairs in Delaware increased from 56 pairs in 1981 to 75—85 in 1994 (L. Gelvin-Innaer pers. comm.). In Delaware Bay, however, where breed- • Netting concentration! © Netting range Approiimatc taanber of netting pain In • a pacific area (eg. cxnaaty, reaervolr.etc.) Figure 2. U.S. Osprey nesting distribution and abundance (1994). 48 Houghton and Rymon Vol. 31, No. 1 ing pairs declined drastically in the 1960s and early 1970s apparently due to use of biocides, recovery has not occurred. On the Delaware side of the Bay, a region which historically supported numerous nests, only three nests were occupied in 1987 (Spitzer 1989) and approximately the same num- ber were reported in 1994 (L. Gelvin-Innaer pers. comm.). Spitzer (1989) suggests increased water turbidity may be one limiting factor. The general trend in traditional areas of these states has been rapid population growth but very limited range expansion. Northeast Region. This region has both coastal and interior populations. Maine is the only state in this region which shows a contiguous population linking the coast with the interior. New Hampshire has an interior nesting population contiguous with that of Maine. Other northeast interior popula- tions are located in northeastern Massachusetts, western Vermont, western New York and the Adi- rondack Mountain region of New York (Spitzer 1989). Expansion of the breeding range of Ospreys in this region has been slow. From 1975—87, there was only a gradual spread of breeders along the coast of Connecticut, New York (Long Island), Rhode Island and Massachusetts, and relatively few Os- preys dispersed 25 km or more into substantially different and unoccupied habitats (Poole 1989a, Spitzer 1989). Spitzer (1989) recorded eight pairs that made moderate dispersals to the interior zone and this number has increased slightly (2-5 addi- tional pairs) within the past 5 yr. Most new nests in the Northeast region are lo- cated in the vicinity of previously established Os- prey breeding habitats. In the coastal region, this trend has been due in large part to intensive man- agement, including nest-site protection and nest platform construction during the last two decades (1967-87) (Poole 1989a, Spitzer 1989). The area from Cape May, New Jersey to Cape Cod, Massachusetts was most heavily affected by bi- ocide use. The number of nesting pairs in this vi- cinity declined from over 1000 nesting pairs in the early to mid 1900s to less than 200 nesting pairs in the mid 1970s (Henny 1977, Spitzer et al. 1983). Following the 1972 ban on DDT, nesting pairs in this region have gradually increased (Poole 1989a, Spitzer 1989). In Massachusetts, nesting has increased dramat- ically as a result of nesting platform construction. Over 90% of the Ospreys nesting in Massachusetts now use such platforms (B. Davis pers. comm.). In 1994, seven out of the 12 Vermont nests were on platforms and active management should continue to play a major role in the expansion of this pop- ulation (S. Parren pers. comm.). Nest manage- ment, particularly predator guarding of natural nests, is credited for recent increases in the Osprey breeding population in New Hampshire (C. Martin pers. comm.). The number of Ospreys nesting between New York City and Boston has grown approximately 10% annually since 1975. If this trend continues, the number of Ospreys nesting there in the year 2000 should equal or exceed historical records (Spitzer 1989). New Jersey has experienced an in- crease of approximately 6% annually (Clark and Jenkins 1993). New limiting factors (loss of suitable habitat and decreased suitability of nest sites) , how- ever, may prevent the number of nesting pairs from reaching historical or pre DDT numbers (Clark and Jenkins 1993). Western Region. Statewide aerial and ground surveys have been limited in this region so avail- able data are less conclusive. Despite this, the Os- prey population is considered to be expanding and increasing in the region. In Montana, Ospreys nest primarily in western portions of the state, mostly at Flathead Lake. In 1974, approximately 23 nesting pairs were located along the Northern Valley of Flathead Lake. In 1986, there were 66 (Henny and Anthony 1989) and, in 1994, there were close to 100 nesting pairs (D. Flath pers. comm.). Overall, statewide nesting increased from 149 nesting pairs in 1981 to 500- 600 in 1994 (D. Flath pers. comm.). Wyoming Ospreys are concentrated near the Montana border, in the northwestern part of the state (Yellowstone and Grand Teton National Parks). Henny and .Anthony (1989) indicated that new nesting occurred in Johnson, Sheridan, Crook and Carbon Counties and it is estimated that the number of nesting pairs in Wyoming at least dou- bled between 1981-94 from 80 to —160 pairs (B. Oakleaf pers. comm.). In Idaho, nesting concentrations occur at Lake Coeur d’Alene, Lake Pend Oreille, Palisades Res- ervoir and Cascade Lake (W. Melquist pers. comm.). Henny and Anthony (1989) reported nesting productivity at Lake Coeur d’Alene in the mid to late 1980s to be among the highest report- ed in the literature. W. Melquist (pers. comm.) es- timated the 1994 population to be over 400 nesting March 1997 1994 Osprey Population 49 pairs, a number which greatly exceeds historical records (Larrison et al. 1967). Washington state conducted aerial and ground surveys in 1984 and 1989. Nesting data from these surveys indicated msyor nesting concentrations in both eastern and western portions of the state. These surveys showed an increasing population, with 275 nesting pairs in 1984 and 346 in 1989 (Watson and K. McCallister unpubl. data) . The av- erage annual percent increase from 1981-89 was approximately 5%. If the rate of growth remained constant between 1989-94, the number of nesting pairs in 1994 should have been approximately 450- 500. However, because of potential limiting factors such as a lack of suitable nest sites and habitat availability, a more conservative estimate for this population in 1994 is 350—400 pairs. In Oregon, the largest concentration of nesting Ospreys is located in the Central Cascade Moun- tains. The overall Osprey population in Oregon ap- pears to be expanding. Henny and Kaiser (1996) reported that the nesting population along the Willamette River (between Portland and Eugene) increased from 13 pairs in 1976 to 78 pairs in 1993. Sixty-six of the pairs were nesting on utility struc- tures in 1993, while none were nesting on them in 1976. The number of Osprey nesting within the state has increased from 308 in 1976 (Henny et al. 1978) to >700 in 1994 (C. Henny pers. comm.) California nesting populations are concentrated in northern coastal and mountain regions (P. Bloom pers. comm.). Henny and Anthony (1989) identified four major populations at Klamath-Trin- ity system, Shasta Lake, Eagle Lake and Lake Al- man, but an estimate for the state’s nesting popu- lation in 1994 was not available, R. Jurek (pers. comm.) indicated that there were 21-23 nests on Tamales Bay, 15-20 nests along the Russian River, 35 pairs along the upper Sacramento River and 52 occupied and 30 successful nests in Marin County. He also noted that the overall number of Ospreys nesting in California has risen dramatically over the past 20 yr. Numbers of nesting pairs at Eagle Lake do not appear to be increasing (Bloom pers. comm.). Henny and Anthony (1989) indicated a substantial range expansion and population in- crease on small reservoirs in extreme northeastern California (Modoc County) where the population increased from three pairs in 1980 to 10 pairs in 1987. They also noted an increase in nesting pairs at Kent Lake (Marin County) from seven pairs in 1975 to 22 in 1986, and an increase in the number of nesting pairs located within the Sierra Nevada region. Since 1975, newly reported Osprey nesting areas have included: Lake Tahoe (El Dorado County), Lake Oroville (Butte County), Basse Lake (Madera County), New Melones Reservoir (Tuol- umne County) and New Bullards Bar Reservoir (Yuba County) (Henny and Anthony 1989). The number of Ospreys nesting in Nevada re- mains low. Two of the four existing pairs are locat- ed at Lake Tahoe and the other two pairs are nest- ing along the Huntington Valley (Gary Herron pers. comm.). In Arizona, a sizable increase in nesting pairs has taken place within the past 10 yr. Most of the Os- preys nest at the White River east/west fork and the main stem of the Black River in southeastern Arizona. However, three nests are located near Flagstaff, there w r as a new breeding attempt on Lynx Lake near Prescott in 1994 and, in 1996, a pair nested for the first time in over 30 yr at the confluence of the Salt and Verde Rivers, east of Mesa (G. Beatty and R. Vahle pers. comm.). New Mexico had its only two pairs of Osprey (1994) nesting on reservoirs in the northern por- tion of the state. Both pairs began nesting in the 1990s (S. Williams pers. comm.). The Colorado Osprey population (1994) is small and concentrated in the northcentral portion of the state. Hacking has been undertaken to en- hance the already existing population and to ex- tend breeding to the front range of the Rockies. Currendy, three nesting pairs are located far from the hacking areas; two are located in La Plata County and one in Pueblo County (J. Craig and K. Luft pers. comm.). Most Ospreys in Utah nest along the Green River and Flaming Gorge Reservoir area in the northeast corner of the state (S. Cranney unpubl. data). C. Monson (unpubl. data) reports that nesting also occurs at Fish Lake (six pairs), Panguitch Lake Navajo (two pairs), and, in 1995, one pair nested on Deer Creek Reservoir and another pair nested in Highland. Construction of reservoirs appears to have increased growth of Utah’s Osprey popula- tion. Many western states continue to show an expan- sion of nesting pairs eastward, partially due to changes in inland habitat, particularly the con- struction of reservoirs (Swenson 1981, Henny 1983). Reservoirs often provide foraging advantag- es over rivers and lakes because their still, shallow, open, water areas and reduced turbidity result in 50 Houghton and Rymon Vol. 31, No. 1 increased water clarity and higher visibility of fish (Swenson 1981, Henny 1983). A comparison of foraging times and nesting den- sities between free-flowing river habitat and three river impoundments on the upper Missouri River in Montana showed food sources and Osprey nest- ing densities to be higher at impoundments (Grov- er 1984). This indicates that additional impound- ments could benefit Osprey populations and en- courage future range expansion. Florida and the Gulf Coast. Florida has the high- est number of nesting Ospreys in this region, with distinct concentrations from the St. Johns River south to Lake Okeechobee (Westall 1990). Ospreys also nest along the east and west coasts and across Florida Bay, including the Ten Thousand Islands area (southwest Florida). Ospreys nesting in pen- insular Florida south of the 29th parallel are non- migratory or resident birds and, therefore, may be subject to different biological limiting factors than Ospreys nesting further north (Poole 1989a). Food stress may be affecting the once healthy Florida Bay population (Poole 1989). Declines there have prompted Florida to designate Osprey as a species of concern in Monroe, County. Al- though Florida Ospreys are currently adapting to an exploding human population, further land de- velopment could limit food supply and nesting habitat and thus should be carefully monitored (Ogden 1978, Westall 1990). Nesting in the Gulf Coast has been extremely limited and sporadic (Lowery 1974, Imhof 1976, Henny 1983, Reinman 1984). The number of Os- preys nesting in this region has fluctuated through- out this century and only limited nesting has been documented (Henny 1983). J. and B. Jackson (un- publ. data) note that Ospreys were historically more abundant along the lower Mississippi than now. A decrease in the number of nesting pairs there has most likely resulted from human distur- bance and manipulation (change in water flow, in- dustry and pollution, and loss of nesting habitat). In 1 994, Gulf Coast Ospreys were most abundant on the gulf islands at the southern tip of Mississippi (50-55 pairs) (M. Woodrie pers. comm.). Only three nesting pairs were recorded in Texas in 1994, and though no statewide survey was conducted, it is unlikely that many more Ospreys nested there. However, many Ospreys migrate through Texas and several have been recorded during winter months (B. Ortego pers. comm.). Alabama and Louisiana have increasing populations (Table 1), but inland nests remain sparse and irregular. The low number of documented nests in these states may be partially related to a lack of survey cover- age. Why nesting remains low is unclear, but it mer- its further attention (Westall 1990). This region is an important study area for future productivity, range expansion, nesting and predation research. The Great Lakes Region. The total number of breeding pairs in the Great Lakes region has al- most doubled from an estimated 579 pairs in 1981 to approximately 1014 in 1994. Today’s distribu- tion of nests is similar to that reported for the pe- riod of 1963-71 (Postupalsky 1969, 1977); however, the aggregations are larger, additional adjunct nests exist between traditional clusters and some range expansion has occurred. Nests still remain concentrated in northcentral and northeastern Minnesota and in northern portions of Michigan and Wisconsin (M. Martell pers. comm.). The growth of the statewide population in Mich- igan observed during 1977—92 has apparently stopped and some local declines have been noted, despite high availability of nest sites. The statewide total remained near 225 pairs from 1992-94 (S. Postupalsky pers comm.). Wisconsin Osprey nest- ing data indicate that the number of nesting pairs there increased steadily during 1983-93, however, S. Postupalsky (pers. comm.) suggests that present numbers may be leveling off. Minnesota’s nesting data was inconclusive but >200 nests were unoc- cupied in 1994 (M. Martell pers. comm.). Wisconsin, Minnesota and Michigan all show signs of range expansion to the south. The expan- sion in Minnesota has been enhanced by hacking efforts initiated in Hennepin County in 1984 (M. Martell pers. comm.). Further expansion and dis- persal is expected as nesting continues and more hacked birds return. A recent surge of nesting on artificial structures (~68% of nests in Wisconsin), could affect future Osprey numbers and status in this region. This region should continue to be eval- uated for factors limiting population growth (e.g., measurements of aquatic productivity, fish popu- lation dynamics, prey accessibility to Ospreys, pred- ators and competitors, and land use) (S. Postupal- sky pers. comm.). Eastern Interior. Major nesting concentrations such as those in Florida and Chesapeake Bay may never be realized in this region, but with the ad- vent of large water-management projects such as the TAW reservoir and waterway system, and the hacking projects that have been implemented March 1997 1994 Osprey Population 51 throughout this region, a sizable increase in the number of nesting pairs and distribution may be expected (Westall 1990). Freshwater reservoirs have been most beneficial to the nesting success here, particularly in Ken- tucky, Tennessee, West Virginia and Pennsylvania where impoundments have benefited recent Os- prey reintroductions. Overall, the Osprey popula- tion in these states surged from five nesting pairs in 1981 to 105 in 1994. Many of these breeding pairs are known to be the result of several intensive hacking programs. For example, in Pennsylvania 1 7 of the 20 statewide nesting pairs include hacked birds, four of which were hacked in West Virginia (L. Rymon unpubl. data) . All three of West Virgi- nia’s pairs are the result of hacking in West Virgin- ia. Most of the Ospreys now nesting in Tennessee are either the direct result of hacking or were at- tracted by recruits from local hacking projects (B. Hatcher pers. comm.). Kentucky has had similar results but the birds in both Tennessee and Ken- tucky have not been well monitored. Central States. Both the Mississippi and Missouri Rivers seem ideal as conduits for interior nesting, yet they remain virtually unoccupied by Ospreys. Historical data suggest that Ospreys once nested in small numbers along parts of the Mississippi (J. and B. Jackson unpubl. data) . Other historical re- cords indicate that very few Ospreys nested within the Central States Region (Hicks 1935, Black 1992, Robbins and Easterla 1992). Reasons why Ospreys are not presently nesting in most of this region remain unclear but the Osprey’s slow pioneering rate may be responsible. Many of these states ap- pear to have some suitable habitat and most are adjacent to states where Ospreys currently nest. Habitat Suitability Indices could be used to deter- mine if habitat is a major limiting factor (Vana- Miller 1987). However, recent dispersals into South Dakota, New Mexico, Texas, Ohio and Indiana, all states with no known nesting pairs in 1981, indicate that it may only be a matter of time before Ospreys expand their range into this region also. If time is inhibitory, hacking may be one option for accel- erating the process (Rymon 1989a). Conclusion 1994 U.S. Osprey Distribution and Abundance. There has been a significant increase in the num- ber of nesting pairs in the U.S. from 1981-94. The overall estimate for the population in 1981 (Henny 1983) was approximately 8000 nesting pairs. Our 1994 population estimate was —14 186 ± 1417 (SD) indicating that the U.S. Osprey population has increased —75% in just over the past decade. There were many similarities in the distribution and abundance of Osprey nesting pairs between 1981 and 1994. Largest increases in numbers of nesting pairs took place in areas of traditional nest- ing: the Atlantic Coast, Pacific Northwest and the Great Lakes Region. A large increase in nesting pairs also occurred in the Eastern Interior Region where hacking has taken place. A1 though the re- cent interior population expansion can be attrib- uted to hacking efforts, hacking has only expanded the range and played a very modest role in the growth of the entire U.S. Osprey population. Overall, population growth has resulted from: 1) increased production rates following the 1972 ban on DDT, most prominent in the Northeast and Great Lakes Regions, 2) construction of numerous new impoundments, especially in the Western and Eastern Interior Regions, 3) artificial nest con- struction in nearly all regions, particularly the rel- atively recent use of utility structures and other man-made structures in the west (Henny and Kai- ser 1996), 4) hacking projects in the interior, and 5) increased public awareness and support. These dramatic changes stress the importance for region- al Osprey management while monitoring the en- tire U.S. population. Growth and Expansion of the U.S. Osprey Pop- ulation. The future growth of regional Osprey pop- ulations depends, in part, on the rate at which new breeders are recruited. Regional differences in the dispersal distances of young may be a reflection of the differences in the density and availability of nest sites (Poole 1989b). In New England, where nesting pairs were severely reduced by pesticides during the 1950s and 60s, artificial nest sites are now clustered, abundant and widely available so most new recruits find nests quickly and breed soon after arrival (Poole 1989b). In New England, dispersal distances >50 km are rare (Poole unpubl. data) . Ospreys in western and midwestern North Amer- ica seem to be dispersing much greater distances. There the breeding range may be restricted by lack of suitable nest sites and large expanses of unsuit- able habitat (Poole 1989a). Ospreys in the western U.S. have traditionally nested in trees or snags near lakes and rivers (Henny 1983); however, nesting areas are becoming saturated (C. Henny pers. comm.) For this reason, they are slowly dispersing 52 Houghton and Rymon Vol. 31, No. 1 into new breeding areas around reservoirs where breeding densities may largely exceed those along nearby free-flowing rivers (Henny 1983, Swenson 1981, Grover 1984). Lack of natural nest sites at traditional nesting areas in the western states may have caused longer dispersal and colonization at newly constructed impoundments and along utility structures (Henny and Kaiser 1996). In the Chesapeake Bay area where nest sites are saturated, Ospreys have begun to delay breeding rather than disperse to new areas (Spitzer 1989). Perhaps this is due to their high natal-site fidelity. Changes in breeding rates, proportions of non- breeders in different populations, choices of nest sites, competition for nest sites, natal dispersal dis- tances, age at first breeding and nesting dispersion should be monitored in future seasons (Spitzer 1989). An increase in the use of artificial nest structures has played an important role in the overall in- crease in the number of nesting pairs in the U.S. (Poole 1989a). Regional data on nesting structures indicate that approximately 64% of Ospreys in the U.S. nest on artificial structures, particularly artifi- cial platforms erected specifically for them (ap- proximately 50%). Excessive construction of nest- ing platforms may have drawbacks in the long run, including habituation to humans, necessary main- tenance of platforms and higher predation rates (Poole 1989a). The construction and addition of artificial nest structures on public lands has played a critical role in increasing public awareness and support for Os- prey. Several states have had volunteer Osprey nest platform projects, and some have set aside viewing areas for aesthetic and educational purposes. Nest platform management is just one example of ex- panding public awareness. Support from public utilities and media coverage appear to be enhanc- ing efforts in public relations which will continue to be important for the preservation of this species (Rymon 1989b). Acknowledgments We would like to thank all of the dedicated osprey re- searchers who contributed their statewide nesting data for the compilation of the 1994 U.S. Osprey nesting data. We are especially grateful to Charles J. Henny for his re- search efforts and his encouraging remarks toward our research. We would also like to thank A.F. Poole and S. Postupalsky for their valuable comments toward the re- vision of our original manuscript. Literature Cited Ames, P.L. 1966. DDT residues in eggs of Ospreys in the northeastern U.S, and their relation to nesting suc- cess. J. Appl. Ecol. 3:87—97. Ames, P.L. and G.S. Mersereau. 1964. Some factors in the decline of the Osprey in Connecticut. Auk 81:173- 185. Black, G. 1992. Iowa birdlife. Univ. of Iowa Press, Ames, I A U.S.A. Clark, K. and C. Jenkins. 1993. Status of Ospreys nesting in New Jersey, 1984 through 1993. Records New Jersey Birds 19:74-77. Grover, K.E. 1984. Nesting distribution and reproduc- tive status of Ospreys along the lower Missouri River. Wilson Bull . 96:496-499. Hatcher, B. and D.A. Hammer. 1983. Restoring Osprey populations by hacking preflighted young. Pages 293- 297 in D.M. Bird [Ed.], Biology and management of Bald Eagles and Ospreys. Harpell Press, Ste. Anne de Bellevue, Quebec, Canada. Henny, C.J. 1977. Research relevant to conservation, re- search, management and status of the Osprey in North America. Pages 199-222 in R.D. Chancellor [Ed.], ICBP World Birds of Prey Conference, Vienna, Austria. . 1983. Distribution and abundance of Ospreys in the United States. Pages 175-186 in D.M. Bird [Ed.], Biology and management of Bald Eagles and Ospreys. Harpell Press, Ste. Anne de Bellevue, Quebec, Cana- da. and J.C. Ogden. 1970. Estimated status of Osprey population in the United States. J. Wildl. Manage. 34: 214-217. , J.A. Collins and W.J. Deibert. 1978. Osprey dis- tribution, abundance and status in western North America II. The Oregon population. Murrelet. 59:19- 25. and R.G. Anthony. 1989. Bald Eagle and Osprey. Natl. Wildl. Fed. Sci. Tech. Ser. No. 12. and J.L. Kaiser. 1996. Osprey population in- crease along the Willamette River, Oregon, and the role of utility structures, 1976-93. Pages 97-108 in Raptors in human landscapes. Academic Press, Lon- don, UK. Hicks, L. 1935. Distribution of the breeding birds of Ohio. Univ. of Ohio, Columbus, OH U.S.A. Imhof, T.A. 1976. Alabama birds. Univ. Alabama Press. Larrison, E.J.J.L. Tucker and M.T.Jollie. 1967. Guide to Idaho birds. J. Idaho Acad. Sci. 5:93-99. Lowery, G.H., Jr. 1974. Louisiana Birds. Louisiana State Univ. Press, Baton Rouge, LA U.S.A. Ogden, J.C. 1978. Preliminary report on a study of Flor- ida Bay Ospreys. Pages 143-151 zn J.C. Ogden [Ed.], Trans. North American Osprey Research Conference. U.S. Nat. Park. Serv., Trans. Proc. Ser. No. 2., Wash- ington, DC U.S.A. March 1997 1994 Osprey Population 53 Peterson, R.T. 1969. Population trends of Ospreys in the northeastern United States. Pages 333—337 in].]. Hickey [Ed.], Peregrine Falcon populations. Univ. Wisconsin Press, Madison, WI U.S.A. Poole, A.F. 1989a. Osprey. A natural and unnatural his- tory. Cambridge Univ. Press, Cambridge, NYU.S.A. . 1989b. Regulation of Osprey Pandion haliaetus populations: the role of nest site availability. Pages in B.U. Meyburg and R.D. Chancellor [Eds.], Raptors in the Modern World. WWGBP, Berlin, London and Par- is. Postupalsky, S. 1969. The status of Osprey in Michigan in 1965. Pages 338-340 in]]. Hickey [Ed.], Peregrine Falcon populations: their biology and decline. Univ. Wisconsin Press, Madison, WI U.S.A. . 1977. Status of the Osprey in Michigan. Pages 153-165 mJ.C. Ogden [Ed.], Trans. North Amer. Os- prey Res. Conf. U.S. Nat. Park, Serv., Trans. Proc. Ser. No. 2. Reinman, J.P. 1984. The status of Osprey populations of the central and northern gulf coasts of Florida. Pages 109-117 in M.A. Westall [Ed.], Proc. Southeastern U.S. and Caribbean Osprey Symposium, Int. Osprey Found., Sanibel Island, FL U.S.A. Rymon, L.M. 1989a. The restoration of Ospreys ( Pandion haliaetus ) to breeding status in Pennsylvania by hack- ing (1980-1986). Pages 359-362 mB.U. Meyburg and R.D. Chancellor [Eds.], Raptors in the modern world. WWGBP Berlin, London, and Paris. . 1989b. Management applications of public ed- ucation. Pages 259-263 in B.G. Pendleton [Ed]), Proc. Northeast Raptor Management Symposium and Workshop. Natl. Wildl. Fed. Sci. Tech. Ser. No. 13. Washington, DC U.S.A. Robbins, M.B. and D.A. Easterla. 1992. Birds of Mis- souri, the distribution and abundance. Univ. Missouri Press, Columbia, MO U.S.A. Schaadt, C.P. and L.M. Rymon. 1983. The restoration of Osprey by hacking. Pages 299-305 in D.M. Bird [Ed.], Biology and management of Bald Eagles and Ospreys. Harpell Press, Ste. Anne de Bellevue, Que- bec, Canada. Spitzer, P.R. 1989. Osprey. Pages 22-29 in B.G. Pendle- ton [Ed.], Proc. Northeast Raptor Management Sym- posium and Workshop. Natl. Wildl. Fed. Sci. Tech. Ser. No. 13. Washington, DC U.S.A. , P.R., A.F. Poole and M. Scheibel. 1983. Initial population recovery of breeding ospreys in the region between New York City and Boston. Pages 231-241 m D.M. Bird [Ed.], Biology and management of Bald Eagles and Ospreys. Harpell Press, Ste. Anne de Belle- vue, Quebec, Canada. Swenson, J.E. 1981. Status of Osprey in southeastern Montana before and after the construction of reser- voirs. West. Birds 12:47—51. VANA-MiLLER, S.L. 1987. Habitat suitability index mod- els: Osprey. U.S. Fish and Wildl. Serv. Biol. Rep. 82(10.154). Westall, M.J. 1990. Osprey. Pages 22-28 in B.G. Pen- dleton [Ed.], Proc. Southeast Raptor Management Symposium and Workshop. Natl. Wildl. Fed. Sci. Tech. Ser. No. 14. Washington, DC U.S.A. Received 6 February 1996; accepted 31 October 1996 J. Raptor Res. 31 (1 ) : 54—58 © 1997 The Raptor Research Foundation, Inc. THE DELAWARE BAYSHORE OF NEW JERSEY: A RAPTOR MIGRATION AND WINTERING SITE OF HEMISPHERIC SIGNIFICANCE Clay Sutton 1 and Paul Kerlinger 2 Herpetological Associates, Inc., 129 Buck Ave., Cape May Court House, NJ 08210 U.S.A. Abstract. — The Delaware Bayshore of New Jersey, including the lower Maurice and Cohansey Rivers and the coastal zone of Cumberland County, is a major migration and wintering area for raptors. Over 89 days of autumn migration for three years, almost 12 000 raptors of 17 species were counted at East Point at the mouth of the Maurice River (x = 30.7 raptors/hr). In addition, the lower drainage areas of the Maurice and Cohansey Rivers and the Cumberland County coastal zone were found to support high densities of Black ( Coragyps atratus ) and Turkey Vultures ( Cathartes aura). Bald Eagles ( Haliaeetus leucocephalus) , Red-tailed Hawks (Buteo jamaicensis) , Northern Harriers ( Circus cyaneus), Cooper’s Hawks ( Accipiter cooperii), Sharp-shinned Hawks (A. striatus), Rough-legged Hawks ( Buteo lagopus) and American Kestrels ( Falco sparverius) in winter. Such high numbers and diversity of migrating and wintering raptors make the area exceptional and, perhaps, unique in eastern North America. The importance of this area for raptors emphasizes the need for its conservation, especially in light of rapid development of nearby farmlands and forests for homes and industry. KeyWords: migration ; winter population, Delaware Bayshore, conservation. El Delaware Bayshore de New Jersey: Una migracion de rapaces y invernada con importancia hemisfer- ico. Resumen. — El Delaware Bayshore de New Jersey, incluyendo los rios abajo de Maurice y Cohansey y la zona costera de el condado Cumberland, es el mayor lugar de migracion invernada para rapaces. Arriba de 89 dias de migracion en el otono por tres anos, casi 12000 rapaces de 17 especie fueron contados en el Punto Este en la boca del Rio Maurice (x = 30.7 rapaces/hr). En suma, en la boca baja del desagiie del Maurice y Cohansey Rios y la zona costera del condado Cumberland se encontro apoyo de densidad alto de Coragyps atratus y Cathartes aura, Haliaeetus leucocephalus, Buteo jamaicensis, Circus cyaneus, Accipiter cooperii, A. striatus, Buteo lagopus, y Falco sparverius en invierno. Tanta cantidad y diversidad de rapaces de migracion y sidos de invierno hace la area excepcional y, quiz.as, unico en el este de Norte America. La importancia de este area para rapaces enfasisa la necesidad para su conservacion, espe- cialmente con el rapido desarrollo de tierras agricolas y bosques para casas e industria. [Traduccion de Raul De La Garza, Jr.] The Delaware Bayshore of New Jersey, especially Cape May, has long been known as a critical area for a diversity of migrating birds (Allen and Peter- son 1936, Stone 1937, Dunne and Sutton 1986, Wiedner et al. 1992), Although much of the region has been threatened or altered by human devel- opment, areas north of Cape May along the Dela- ware Bayshore including Cumberland County, have been largely spared. Recent proposals for 1 Present Address: 129 Bucks Ave., Cape May Court House, NJ 08210 U.S.A. 2 Present Address: 31 Jane St. 14D, New York, NY 10014 U.S.A. barge port development, industrial sand mines and other projects along the Maurice and Cohansey riv- ers in Cumberland County threaten to alter aquat- ic habitats, wetlands and uplands, rendering them less suitable for wildlife. Because of these threats, we conducted an 8-yr study of raptor use during autumn migration and winter along the Delaware Bayshore of Cumber- land County, New Jersey including the Maurice and Cohansey river drainages. In this report we document a large and previously unknown aggre- gation of migrating and wintering raptors in this area that is of hemispheric significance and worthy of major conservation effort. 54 March 1997 Raptor Migration at Delaware Bay 55 Figure 1. Maps of New Jersey and Cumberland County Coastal Zone study area including the lower Maurice and Cohansey Rivers, and the Delaware Bayshore. Closed circles indicate observation sites along the rivers. East Point hawk watch is denoted by a star at the mouth of the Maurice River. Methods Migrating raptors were counted at East Point, Cum- berland County, New Jersey (Fig. 1) from 9 September- 30 November 1989-91, by a single observer during pre- determined peak flight hours (mid-morning through mid-afternoon) for an average of 4.4 hr per day (range = 1-9 hr) from the wooded edge of a salt marsh near the mouth of the Maurice River at Delaware Bay. On about one-third of all days, the observation site was moved 3 km inland from the Bay at about midday, as the flight became higher and moved farther inland. East Point is a peninsula that juts out into Delaware Bay. To assess the relative importance of this site, counts from East Point were compared with same day hawk migration counts from Cape May Point, New Jersey, 32 km to the south. Counters at both sites used standard Hawk Migra- tion Association of North America methods and record- ing forms (Bednarz and Kerlinger 1990, Kerlinger 1989). We sampled wintering raptors along tidal portions of the Maurice River during eight winters, 1987-88 through 1994—95; the Cohansey River during five winters, 1990- 91 and 1994-95; and in the Cumberland County Coastal Zone (CCCZ) in five winters 1989-94. Transport among sites was by automobile. Surveys were done between 22 November— 21 March, mostly between 0800—1600 EST. A total of 69 surveys were done on the Maurice River and 16 on the Cohansey River. On the Maurice River, a survey consisted of seven observation stations from a point 22 4 km upstream from the river mouth (Fig. 1 ) to East Point and at five stations on the Cohansey River from Bridge- ton downstream to within 2 km of the mouth of the river, a distance of 17.6 km. At each observation site, all flying and perched raptors within view were counted for 50 min. Care was taken to avoid recounting individuals by noting plumage characteristics and flight direction. If a raptor was observed flying into the adjacent sampling area it was not counted if resighted in that area. Raptors seen perched between sampling sites were included in the nearest sampling site if an individual of that species was not sighted from the adjacent stations. Ten power binoculars and a 20 power spotting scope were used. On about 96% of the surveys, two observers were present and on the remaining surveys only one observer. The same observer was present for 65 of the 69 Maurice River sur- veys and all of the others. Counts were conducted mostly in sunny weather with northwest winds. Ninety percent of the counts were done with these conditions, since they promoted soaring and hunting of raptors. Habitat along the Maurice and Co- hansey rivers and the Cumberland County coastal zone includes tidal marsh and swamp forest, as well as adjoin- ing upland forests and farms. Wetlands in these areas range from marshes typical of tidally inundated areas to tidally influenced freshwater marshes farther upstream. 56 Sutton and Kerlinger Vol. 31, No. 1 Saltmarsh cordgrass ( Spartina alterniflora) dominates the marshes at the mouth of the rivers, with saltmarsh hay (5. patens ) present. Farther upstream, although tidally in- fluenced, the marshes are comprised of freshwater and brackish wetland plants such as wild rice ( Zizania aquati- ca), common reed ( Phragtnit.es communis), pickerel weed ( Pontederia cordata ) , arrow arum ( Peltandra virginica) , spat- terdock ( Nuphar advena), waterlily ( Nyrnphaea odorata) and cattail ( Typha latifolia ) . Common reed is more prev- alent on the Cohansey River. The banks along portions of the Maurice River are quite steep and often heavily wooded, although cropland is near both rivers. Devel- oped areas are limited to single family homes and a few recreational boat yards. Both rivers average less than one- half mile wide and both are about one mile wide near the mouth. In addition to surveys on the Maurice and Cohansey rivers, one road survey per year was made in the first two weeks of January 1990-94, during which raptors were counted in the entire Cumberland County coastal zone (Fig. 1). This latter survey consisted of 12 primary obser- vation sites along a 48 km course. Standard raptor road survey methods were used between observation sites. The coastal zone count included regular counts (nearest cal- endar date) from the Maurice and Cohansey rivers, all conducted in a 3-d-period. This method was an attempt to determine roughly the total number of raptors that winter in the Cumberland coastal region. Results and Discussion Over 89 d (389.25 hr), we observed 11 944 hawks (30.7/hr) of 17 species at East Point (Table 1). Sharp-shinned Hawks ( Accipiter stnatus ) were most numerous, accounting for 39.7% of all hawks counted. This species, plus Turkey Vultures (Ca- thartes aura). Northern Harriers ( Circus cyaneus ), Red-tailed Hawks (Buteo jamaicensis) and American Kestrels ( Falco sparverius ) accounted for 84.5% of all raptors observed. In 1990, the year most obser- vations were made, a total of 9042 raptors was re- corded during 308.5 hr of observation on 60 d (9 September-7 December). This was 34.6% of the 26 164 raptors seen at Cape May Point during the same 60-d-period. Species composition differed between East Point and Cape May Point which may be attributable, in part, to differences among species in their tenden- cy to cross water. Greater numbers of vultures, Bald Eagles ( Haliaeetus leucocephalus) , Northern Harri- ers, Rough-legged Hawks ( Buteo lagopus), Red- tailed Hawks and relatively fewer Ospreys, Sharp- shinned Hawks and falcons were counted at East Point (Table 1). Kerlinger (1985) documented dif- ferences in water crossing tendency among most of the species reported here in a study at Cape May Point. Upon reaching the end of the peninsula, only some species crossed without hesitation. At Table 1. Summary of autumn migrants recorded at East Point and Cape May, New Jersey, on the same 89 days during 1989, 1990 and 1991. Total numbers of migrants counted and percentage of total hawks counted at each site are given. One Swainson’s Hawk ( Buteo swainsoni) was also seen at East Point. East Point Cape May Per- Per- Species Count CENT Count CENT Black Vulture 39 <1% 24 <1% Turkey Vulture 1219 10% 62 2% Osprey 233 2% 2012 6% Bald Eagle 97 1% 65 <1% Northern Harrier 888 7% 1110 3% Sharp-shinned Hawk 4744 40% 16 853 47% Cooper’s Hawk 754 6% 2312 6% Northern Goshawk 7 <1% 36 <1% Red-shouldered Hawk 91 1% 234 <1% Broad-winged Hawk 188 2% 1168 3% Red-tailed Hawk 1398 12% 1453 4% Rough-legged Hawk 42 <1% 4 <1% Golden Eagle 14 <1% 28 <1% American Kestrel 1854 16% 7400 21% Merlin 297 3% 1551 4% Peregrine Falcon 72 1% 855 2% Total Hawks 11 944 35 968 Total Hours 389.25 798.5 Hawks Per Hour 30.7 45.0 East Point, the vast majority of most species moved west, whereas Osprey ( Pandion haliaetus), Merlins ( Falco columbarius) and Peregrine Falcons (Falco per- egrinus ) usually flew to the southeast toward Cape May. The latter species are not reluctant to cross Delaware Bay (Kerlinger 1985, 1989). It is likely that upon seeing the end of the Cape May penin- sula, many soaring birds turn westward to follow the Delaware Bayshore, rather than proceeding to the tip of the Cape May peninsula where they would be counted. Others fly through Cape May and are counted near the Point before flying north and west along the Bayshore toward East Point. During January road surveys of the coastal zone, 14 species were observed totaling between 472 and 932 birds/survey (Table 2). This year to year vari- ation was attributable largely to variation in the number of roosting vultures and, perhaps, to weather. Species composition and numbers of most species observed along the Maurice and Cohansey rivers were similar (Table 2). The numbers of American Kestrels and Sharp-shinned Hawks along March 1997 Rartor Migration at Delaware Bay 57 Table 2. Summary of wintering diurnal raptors along Maurice and Cohansey Rivers and the Cumberland County Coastal zone area. Letters following species names indicate status of the species (T = federally threatened, t = New Jersey threatened, e = New Jersey endangered, u = unknown status in New Jersey; first of state letters indicates breeding season, second indicates nonbreeding season — migration and winter) . Totals are the mean of yearly mean totals and min-max of yearly average totals. Species Maurice River (1987-1995) Cohansey River (1990-1995) Cumberland Coastal, Zone (1989-1994) Mean Range Mean Range Mean Range Black Vulture 9.9 1-45 7.0 3-10 35.8 9-77 Turkey Vulture 73.4 49-111 54.8 41-68 272.8 165-501 Bald Eagle (T,e,e) 5.5 3-10 2.4 2-4 21.2 17-27 Northern Harrier (e,u) 20.0 15-24 19.0 16-22 122 79-171 Sharp-shinned Hawk 2.6 2-4 4.1 3-6 19.8 6-37 Cooper’s Hawk (e,e) 1.3 1-2 1.9 1-3 8.2 3-14 Northern Goshawk 0.1 0-1 0.2 0-1 1.4 0-2 Red-shouldered Hawk (e,t) 0.2 0-1 0.4 0-1 3.6 2-5 Red-tailed Hawk 37.9 33-42 37.3 34-42 134.6 86-159 Rough-legged Hawk 1.0 0-2 0.4 0-1 10.2 4-17 Golden Eagle 0.2 0-1 0.1 0-1 0.8 0-3 American Kestrel 2.2 1-3 7.2 6-11 30.6 18-43 Merlin 0.1 0-1 0.2 0-1 1.4 0-4 Peregrine Falcon (T,e,e) 0.1 0-1 0.2 0-1 1.2 0-2 Totals 154.7 117-204 134.7 115-151 663.8 72-932 Years of sureys 8 5 5 Mean number of surveys per year (range) 8.6 7-14 3.2 2-4 1 the Cohansey River were slightly greater than along the Maurice River because there is more farmland and forest edge along the Cohansey. Conversely, more Bald Eagles and Turkey Vultures were reported from the Maurice River, a function of greater waterfowl numbers, waterfowl hunting and carrion present at a livestock farm. Determin- ing absolute abundance for statistical comparison is extremely difficult in such a heterogeneous land- scape. Turkey Vultures were the most numerous of all species, followed by Red-tailed Hawks and North- ern Harriers. Bald Eagles were an important com- ponent of the winter raptor community of the Del- aware Bayshore of New Jersey; as many as 16 were seen on a survey of the Maurice River. Among the wintering raptors observed, five were on the fed- eral or New Jersey threatened or endangered lists (Table 2). Single Merlins, typically a Neotropical migrant, wintered on the study site in most years. Probably a combination of factors is attributable to the great diversity and abundance of raptors mi- grating and wintering along the Delaware Bay- shore. Delaware Bay is a topographic barrier to mi- gration. Raptors reluctant to cross water seem to either fly around the Bay or terminate their migra- tion along the Bayshore. The importance of migra- tion stopovers among songbirds, shorebirds and waterfowl has long been known, but few biologists have recognized its importance among raptors (Newton 1979, Kerlinger 1989). Many of these birds are compelled to make migratory stopovers in the area. The abundance of open space, high- quality habitat and abundant prey make the Bay- shore very attractive to both migrating and winter- ing hawks. The Bayshore has frequently been recognized as an ecosystem of hemispheric or global significance. For example, the wetlands of the Delaware Bay are a RAMSAR site, in part because they support glob- ally important populations of shorebirds and wa- terfowl. In addition, the EPA National Estuary Pro- gram has recognized the area as biotically impor- tant and has funded m*yor studies of the area (Zap- palorti et al. 1993) . Much of the Maurice River and three tributaries, the Manumuskin River, Menan- tico Creek and Muskee Creek, are now included in the U.S. National Park Service Wild and Scenic Riv- 58 Sutton and Kerlinger Vol. 31, No. 1 er program. The data presented herein support these designations. Another factor that influences the winter abun- dance of Bald Eagles, vultures and to a lesser ex- tent Red-tailed Hawks, Golden Eagles and Pere- grine Falcons are the vast concentrations of water- fowl and other birds that migrate and winter along the Delaware Bayshore. Winter waterfowl counts conducted simultaneous to these raptor counts show that more than 25 species of waterfowl in numbers totaling greater than 10 000 to 15 000 occur regularly on the Maurice River portion of the study site alone (Sutton 1988). Likewise, huge numbers of migrating songbirds, prey for some hawks, are known to frequent the Delaware Bay- shore (Wiedner et al. 1992). These birds search for stopover habitat as they move along the Bayshore. Our data indicate that the Delaware Bayshore, and Maurice and Cohansey river drainages, sup- port one of the largest and most diverse concen- trations of migrating and wintering raptors report- ed in eastern North America. We make this claim understanding that there is little comparative in- formation on abundance and winter concentration areas for raptors in eastern North America (other than for eagles and vultures). The Delaware Bay- shore is all the more important because it serves hawks during both winter and migration. We know of no other migration sites that also serve as im- portant wintering areas for raptors. It seems that many migrants end their southbound migration on the Delaware Bayshore. The conservation value of this area is obvious. Acknowledgments Financial support for this research came from Citizens United to Protect the Maurice River and its Tributaries, as well as Cohansey Area River Protection, Cumberland County Department of Planning and Economic Devel- opment and New Jersey Audubon Society’s Cape May Bird Observatory. We also thank J. Galetto, G. Ewan, D. Fauerbach, C. Zirkle, E. Zirkle, R. Zappalorti and J. Lau- bengeyer for assistance during this project. C. Shultz, V Elia and J. Dowdell assisted with the fieldwork. K. Bild- stein made helpful comments on the manuscript. Literature Cited Allen, R.P. and R.T. Peterson. 1936. The hawk migra- tions at Cape May, New Jersey. Auk 53:393—404. Bednarz, J. and P. Kerlinger. 1990. Monitoring hawk populations by counting migrants. Pages 328-342. in Proc. Northeastern Raptor Symposium, Rochester, NY. Natl. Wildl. Fed. Sci. and Tech. Ser. #13. Dunne, P-J- and C.C. Sutton. 1986. Population trends in coastal raptor migrants over ten years of Cape May Point autumn counts. Rec. N.J. Birds 12:39-43. Kerlinger, P. 1985. Water-crossing behavior of raptors during migration. Wilson Bull. 97:109-113. . 1989. Flight strategies of migrating hawks. Univ. Chicago Press, Chicago, IL U.S.A. Newton, I. 1979. Population ecology of raptors. Buteo Books, Vermillion, SD U.S.A. Stone, W. 1937. Bird studies at Old Cape May. Delaware Valley Ornithological Club, Philadelphia, PA U.S.A. SUTTON, C.C. 1988. Wintering raptors and waterfowl on the Maurice River. Rec. N.J. Birds 14:42-51. Wiedner, D.S., P. Kerlinger, D.A. Sibley, P. Holt, J. Hough and R. Crossley. 1992. Visible morning flights of Neotropical landbird migrants at Cape May, New Jersey. Auk 109:500-510. Zappalorti, R., C.C. Sutton and R. Radis. 1993. Cum- berland County Delaware estuary study; Vol. 1. Rare, threatened and endangered species study; Vol. 2: Ap- pendices and mapping. USEPA Delaware Estuary Study, Cumberland County Department of Planning and Development, Bridgeton, NJ U.S.A. Received 9 October 1995; accepted 31 October 1996 J Raptor Res. 31(1) : 59-64 © 1997 The Raptor Research Foundation, Inc. FOOD HABITS OF COMMON BARN-OWLS ALONG AN ELEVATIONAL GRADIENT IN ANDEAN ARGENTINE PATAGONIA Alejandro Travaini, Jose A. Donazar Estacion Biologica de Donana, Consejo Superior de Investigaciones Cientificas, Apartado 1056, 41080 Sevilla, Spain Olga Ceballos Grupo de Estudios Biologicos Ugarra, Carlos III 1 9, 31002 Pamplona, Spain Alejandro Rodriguez, Fernando Hiraldo, Miguel Delibes Estacion Biologica de Donana, Consejo Superior de Investigaciones Cientificas, Apartado 1056, 41080 Sevilla, Spain Abstract. — We evaluated the diet of Common Barn-owls (Tyto alba ) along an elevational gradient in Argentine Patagonia. Small mammals (mainly rodents) were the main prey accounting for 93.2% of total prey items. Consumption of rodents appeared to be dependent on their availability. Sizes of mam- malian prey were variable but most ranged from 10-100 g in body mass. We concluded that the diet of these barn owls could be used as an index of cricetid rodent populations along the gradient. Key Words: Common Barn-owl, Tyto alba; prey ; Patagonia ; selectivity, gradient. Dieta de la lechuza ( Tyto alba ) a lo largo de un gradiente altitudinal en la Patagonia Andina Argentina. Resumen. — Se estudio la dieta de la lechuza {Tyto alba ) y se la contrasto con la composicion especifica de la comunidad de micromamiferos a lo largo de un gradiente altitudinal en la Patagonia Argentina. La principal presa la constituyeron pequenos mamiferos (fundamentalmente roedores) alcanzando un 93.2% del total de presas consumidas. La masa corporal media de los mamiferos presa se concentro fuertemente en un rango comprendido entre 10 y 100 g. La dieta de la lechuza resulto ser un buen indicador de la composicion de Cricetidos a lo largo del gradiente estudiado. El consumo de cada uno de los roedores presa dependio de su disponibilidad en el terreno. [Traduccion Autores] Few of the many studies of Common Barn-owl {Tyto alba) food habits have examined dietary re- sponses to elevational distribution of prey species (Herrera 1974, Brunet-Lecomte and Delibes 1984). In northwestern Patagonia, a steep eleva- tional gradient (600-3000 m) occurs over just a few km. The abrupt change in elevation and the asso- ciated change in precipitation (300-3600 mm) causes a distinct shift in the vegetation from shrub- steppe to montaneous forest habitat within a few km. Pearson and Pearson (1982) qualitatively de- scribed the small mammal species composition along this gradient. They found six species of ro- dents {Aconnaemys fuscus, Dromiciops australis, No- tiomys macronyx, Notiomys valdivianus, Irenomys tar- salis and Akodon olivaceous) occurred in the humid forest, while another seven rodent species {Cteno- mys haigi, Akodon xanthorrinus, Reithrodon auritus, Eligmodontia typus, Phyllotis darwini, Euneomys sp, and a marsupial, Lestodelphis hally) were in the dry shrubsteppe. They also analyzed the diets of Com- mon Barn-owls at three different localities at the forest-steppe transition and found that the propor- tions of species captured by traps and by owls were different. Besides providing new information on the tro- phic niche of Common Barn-owls in Argentina, here we test the feasibility of using barn owl food habits to describe changes in composition and abundance of small mammals along an elevational- precipitation gradient. We considered the data provided by Pearson and Pearson (1982) as the ac- 59 60 Travaini et al. Vol. 31, No. 1 tual representation of the small mammal commu- nity along this gradient and the barn owl prey as its descriptor. Additionally, Pearson (1986) provid- ed data on relative abundance of cricetid species in eight different habitats ranging from steppe to forest over an 8-yr-period. We made two comparisons between estimates of rodent availability and their occurrence in the diet of common barn owls: (1) on a broad scale, along the elevational-precipitation gradient, using data from Pearson and Pearson (1982), and (2) on a fine scale, along a segment of the complete gra- dient, using abundance estimates for rodent spe- cies given by Pearson (1986) . We predicted that the barn owl diet would reflect gradient changes in ro- dent community composition (Herrera 1974, Pear- son and Pearson 1982, Taylor 1994), but it would be less accurate in reflecting availability of prey at a fine scale (Jaksic and Yanez 1979). Study Area and Methods Located in northwestern Patagonia (70°30'-71°30'W; 39 o 30-40 o 20'S), the study area constitutes a portion of the Precordillera gradient and partially overlaps with the area studied by Pearson and Pearson (1982) and Pearson (1986). The greatest distance between our site and that of Pearson was under 150 km. The climate of the area is dry and cold with frost throughout most of the year and frequent snowfall in winter. Topographically, the area consists of plains from 800-900 m above sea level that are dissected by steep rugged valleys and large rivers. In general, the study area consists of lowland which coin- cides with river valleys and highland Piedmont with an intermediate area between them. Pearson (1986, 1987, 1988) has described five different habitats in intermedi- ate and highland Piedmont: steppe or scattered bushes, usually <1 m tall, usually mixed with bunch grass, much of the ground is devoid of vegetation; bunchgrass or hab- itats with relatively pure stands of one or more species of bunchgrass; weeds or areas with dense weeds and grasses, usually growing in moist places; rocks or cliffs with tum- bled rocks large enough to provide refuge for rodents; and bare ground or large areas not vegetated with a sub- strate of fine scree or rock and very few rocky shelters. Barn owl pellets were collected from September-Feb- ruary 1991, 1992 and 1994—95. Pellets were collected at 32 isolated localities more than 2 km apart. On this basis, we assumed that barn owls at each locality were different individuals. Only fresh pellets were included in our sam- ples in order to restrict our study to the spring season and avoid seasonal variation. All pellets were dissected using standard techniques (Yalden 1977). Small mammal remains were identified using taxonomic keys (Pearson 1986), reference specimens collected in the study area and museum collections. Small mammal biomass was de- rived from Pearson (1983, 1984) and Redford and Eisen- berg (1992). Following Flerrera and Jaksic (1980), we characterized barn owl food habits by the following parameters: mean mass of all small mammals (MWSM) in the diet; H'NGG, trophic diversity in relation to the number of individuals contributed by each higher taxonomic unit (mammals, birds, amphibians, invertebrates); H'NM, trophic diver- sity in relation to the small mammal component of the diet (rodents, lagomorphs, marsupials); and H'NR, tro- phic diversity in relation to the number of individuals contributed by each rodent species. The latter three pa- rameters were computed by means of Shannon’s infor- mation function (Ludwig and Reynolds 1988). Corre- sponding values of evenness (J = H'/H'max) were also calculated. The use of the barn owl food habits to detect and de- scribe changes in the composition and abundance of small mammals along a gradient was explored using cor- respondence analysis (SAS 1987). This is a multivariate ordination method (Digby and Kempton 1987, Pielou 1984), applied on a data matrix that included frequen- cies of appearance (%) of several prey categories for each barn owl locality. This type of analysis permits one to plot points for both rows and columns (here localities and small mammal prey categories) on the same plane. Cor- respondence analysis is especially appropriate for matri- ces with numerical frequencies (Cuadras 1980) and does not normally require previous transformation of data (Digby and Kempton 1987). We used this analysis to ver- ify if the pattern in rodent species distribution derived from barn owl diets was similar to that described by Pear- son and Pearson (1982) along both the elevational and precipitation gradients. For this analysis, we restricted ourselves to those localities (N = 23) with >30 identified prey items. Selection among potential prey species was studied only in intermediate and highland Piedmont. Barn owl selectivity was evaluated by comparing the species’ rank for each cricetid rodent in the barn owl diet with the species’ abundance in the field as estimated by Pearson (1986). We used a Spearman Rank Correlation Coeffi- cient (Siegel and Castellan 1988) for this analysis. It was calculated from the sum of the ranks obtained for each species in each of the five habitats considered. Compar- ison was restricted to cricetids because these were the rodent species for which Pearson (1983) estimated rela- tive abundance. Results A total of 2447 prey items were identified from barn owl pellets. Small mammals were the main prey in northwestern Patagonia, accounting for 93.2% of the total. Birds, amphibians and inverte- brates made up the remaining 6.8% (Table 1). Cal- culated from a random sample of 71 pellets, the mean number of prey/pellet was 1.85 (SD = 1.09, range 1-6). Among mammals, rodents most frequently oc- curred in the diet, representing the 98.9% of the total (Table 1). The MWSM was 54.2 g (SD - 36.7, range 17.5 g for Calomys musculinus, 286.1 g for Microcavia australis) but most small mammals were between 10-100 g (Fig. 1). March 1997 Barn owl diet in Patagonia 61 Table 1. Composition of the Common Barn-owl in the lowland, midland, and highland piedmont, northwestern Argentine Patagonia. Prey Type Lowland Midland Piedmont (%) Piedmont (%) Highland Piedmont (%) Total N % MAMMALS Rodents Hystricognath Ctenomys haigi 13.1 3.1 3.0 134 5.5 Galea musteloides 1.7 0.3 0.0 13 0.5 Microcavia australis 0.3 0.0 0.0 2 0.1 Cricetids Akodon spp 9.6 21.4 25.0 484 19.8 Auliscomys micropus 0.2 6.3 18.7 226 9.2 Chelemys macronix 0.0 1.0 3.1 37 1.5 Eligmodontia typus 31.6 13.1 6.7 374 15.3 Euneomys sp 0.0 0.1 2.8 25 1.0 Geoxus valdivianus 0.2 0.3 1.0 13 0.5 Irenomys tarsalis 0.0 0.1 0.5 5 0.2 Oryzomys longicaudatus 6.6 8.2 6.0 171 7.0 Phyllotis darwini 4.0 12.8 5.6 198 8.1 Reithrodon auritus 23.0 16.9 24.4 522 21.3 Calomys musculinus 5.4 1.9 0.0 51 2.1 Marsupials Marmosa pusilla 1.8 1.1 0.0 22 1.0 Lagomorphs Oryctolagus cuniculus 0.0 0.2 0.1 3 0.1 Birds 1.3 0.8 0.3 19 0.8 AMPHIBIANS 0.0 0.5 0.0 5 0.2 INSECTS 0.3 11.9 2.8 143 5.8 TOTAL PREY H’NGG JNGG H’NM JNM H’NR JNR (595) (980) (872) (2447) 0.28 0.20 0.06 0.05 2.06 0.78 These barn owls had a relatively narrow diet as shown by the H'NGG value (Table 1). The low even- ness index was largely due to their concentration on mammal prey (Table 1). The diversity and even- ness of the small mammal component (H'NM) showed that the diet was based upon a small num- ber of small mammal species, most of which were rodents. Among rodents, H'NR (2.06) and evenness (0.78) reached the highest values, denoting both the high number of rodent prey species consumed (N = 14, 82% of total present in the area) and their overall even representation in the diet. Two axes generated by the correspondence anal- ysis accounted for 49.5% of the variance in the diet (Fig. 2) . Representation of localities and prey cat- egories on the plane defined by the two axes clear- ly segregated the dry, lowland Piedmont localities (which tended to the positive zone of Axis I and negative zone of Axis II) from the rest of localities (which occupied all the space defined by Axis I and the positive zone of Axis II) . Prey in the low- land Piedmont localities included Ctenomys haigi, Microcavia australis, Galea musteloides, Eligmodontia typus and Calomys musculinus. Prey in most of the other localities, included in mid- and high Pied- monts were Akodon sp, Auliscomys micropus, Chelemys 62 Travaini et al. Vol. 31, No. 1 Figure 1. Relative frequencies (%) of small mammal prey in the diet of Common Barn-owls in northwestern Ar- gentine Patagonia, ordered along a logarithmic axis of body weights. macronix and Geoxus valdivianus. In a few localities, Phyllotis darwini and Euneomys sp. also occurred. The similarity between the expected and ob- served rank order in cricetid species in the diet Figure 2. Variation in the taxonomic composition of the diet of the Common Barn-owl in northwestern Argentine Patagonia. Plot of the Correspondence Analysis, areas of each delimited by minimum poligon. See Table 2 for most species abbreviations. MAR, Marmosa pusilla ; CAL, Calomys musculinus; GAL, Galea musteloides; CTE, Ctenomys haigi. Circles correspond to highland Piedmont, quadrats to lowland Piedmont and triangles to intermediate Pied- mont. (Table 2) indicated that barn owls fed on prey ac- cording to their availability in intermediate and highland Piedmont areas ( r s = 0.91, P< 0.01). The most important difference was the greater impor- tance of Reithrodon auritus and the overall absence of Akodon sp. in the diet. Discussion Barn owls in our study preyed almost exclusively on rodents, as found in most other studies (Smith and Cole 1989, Bellocq 1990, Iriarte et al. 1990, De Santis et al. 1994, Taylor 1994). This suggests that these barns owls do not behave as opportunistic, nonselective predators as suggested by Mikkola (1983). Taking this into account, care should be taken when interpreting the unusually high predation on birds reported by Noriega et al. (1993) in the Patagonian zoogeographic domain (Ringuelet 1961). Their data may reflect individual differences of individual barn owls. Our mean body mass estimate of small mammals in the diet (56.2 g) was intermediate between that reported in Spain (21.2 g) and Chile (70.7 g) (Herrera and Jaksic 1980). Because the largest prey taken in the three areas was the European rabbit, ( Oryctolagus cuniculus), the smaller MWSM in Argentina when compared with Chile was a consequence of both the high concentration of smaller prey than in Chile (30-500), and the greater con- sumption of smaller rodents than in Chile. In Argentine Patagonia, only one of the three available rodent species weighing more than 150 g ( Ctenomys haigi ) was readily consumed by owls, while in Chile three of four such spe- cies were consumed in amounts similar to that of our C. haigi (Herrera and Jaksic 1980). H'NGG diversity and evenness indexes were similar to March 1997 Barn owl diet in Patagonia 63 Table 2. Expected composition of Common Barn-owl diet expressed as a percentage for five habitats, based on Pearson (1983, 1986). Species 1 Steppe Bunch Grass Weeds Rocks Bare Owl Diet Owl Diet Field Rank AKO 42.0 76.0 49.0 41.0 36.0 27.0 AKO AKO AUL 2.0 12.0 16.0 2.0 23.0 12.0 REI ELI CHE 0.0 0.0 1.0 0.0 10.0 1.9 ELI AUL ELI 43.0 12.0 2.0 3.0 3.0 13.0 AUL ORY EUN 0.0 0.0 0.0 1.0 15.0 2.8 PHY REI GEO 0.0 0.0 2.0 0.0 2.0 1.0 ORY PHY IRE 0.0 0.0 0.0 1.0 0.0 0.3 EUN EUN ORY 4.0 0.0 19.0 2.0 8.0 10.0 CHE CHE PHY 4.0 0.0 0.0 48.0 1.0 11.0 GEO GEO REI 5.0 0.0 11.0 2.0 2.0 21.0 IRE IRE 1 AKO, Akodon sp.; AUL, Auliscomys micropus ; CHE, Chelemys macronix; ELI, Eligmodontia typus\ EUN, Euneomys sp.; GEO, Geoxus valdi- vianus, IRE, Irenomys tarsalis; ORY, Oryzomys longicaudatus; PHY, Phyllotis darwini; REI, Reithrodon auritus. those found by Herrera and Jaksic (1980) for Chilean barn owls. The inclusion of amphibians and the absence of reptiles by the Argentine owls were the main reasons for these differences. Among mammals, the preponderance of rodents among the diets of Argentine owls made the H'NM di- versity and evenness indices very low compared to those obtained for Chilean owls (Herrera and Jaksic 1980), which were more equally represented. In Chile, Spain and Argentina, the low consumption of the European rabbit is likely related to its large body size. The correspondence analysis results concurred with re- sults obtained by Pearson and Pearson (1982) and our prediction with respect to the gradient with which cri- cetid rodents are associated. Both species and localities were segregated by an aridity gradient. Irenomys tarsalis occurred in the barn owl diet only near the Nothophagus forest. Auliscomys micropus was eaten only in forest or dense cover habitats but Oryzomys longicau- datus was eaten in all the habitats considered for this analysis. Eligmodontia typus occurred in the diet of owls only in open habitats with bare soil and scattered desert shrubs, Ctenomys haigi occurred in the diet in open areas with sandy soils, and Calomys musculinus and Reithrodon auritus occurred in the diet of birds associated with the arid portion of our gradient. Observed consumption of Reithrodon auritus was, both in percent frequency and rank order, higher than ex- pected. A similar situation was observed by Jaksic and Rau (1986) for the Great Horned Owl ( Bubo virginianus) in a Chilean Patagonian steppe with about the same mammalian composition that we report. As in Europe and North America, comparison of the abundance of small mammal species is often susceptible to problems of differential trapability. On the other hand, Akodon sp. was consumed less than expected by chance in all five habi- tats. Perhaps their more diurnal activity period and small body weight contributed to these results (Taylor 1994). Based on our knowledge, the ecology of Reithrodon auritus and Akodon sp. resembles that of voles and mice de- scribed in North America and Europe (Marti 1974, Col- vin 1984). There, barn owls frequently select voles in the presence of other prey, probably because the former are heavier and easier to catch (Taylor 1994). Three factors provide a plausible explanation. First, K auritus (80.4 g) is 3 times heavier than Akodon sp. (28.0 g), even a bigger difference to that found between voles and mice (Marti 1974, Colvin 1984). Second, Akodon sp. shows greater ac- tivity during the daytime (Rau et al. 1981). The opposite is true for R, auritus which is a typical nocturnal species (Pearson 1988), thus overlapping its activity period with that of the barn owl. Finally, R. auritus feeds in open grassy habitat where it may be more exposed to owl pre- dation, while Akodon sp. prefers dense bushes in the steppe (Pearson and Pearson 1982), perhaps gaining pro- tection against owl predation. Acknowledgments We are very grateful to Martin Funes, Obdulio Mon- salvo and Miguel Angel Pineda for help with the field work. Our thanks to Gerardo Alena, Gabriel and Miguel Anz, Raul Cordero, Victor Soleno, the administrators of the Estancias Cerro los Pinos, Serranias de Lolen, Collon- Cura and Chimehuin, who permitted us to work on their land. Logistic support was provided by the Centro de Ecologia Aplicada del Neuquen (Argentina); we thank Alejandro del Valle and Antonio Guinazu for their con- stant kind assistance. We also want to thank Marta Pian- tanida from the Museo de Ciencias Naturales Bernardino Rivadavia for providing us with a complete small mammal reference collection. D.W. Holt, C. Marti and two anon- ymous referees improved the manuscript. Financial sup- port was provided by Instituto de Cooperacion Iberoam- ericana and the Ministerio de Asuntos Exteriores (Spain) throughout the Programa de Cooperacion Cientifica con Iberoamerica. Literature, Cited Bellocq, M.S. 1990. Composicion y variacion temporal de la dieta de Tyto alba en ecosistemas agrarios Pam- peanos, Argentina. Vida Silvestre Neotropical 2:32-35. 64 Travaini et al. Vol. 31, No. 1 Brunet-Lecomte, P. and M. Delibes. 1984. Alimenta- cion de la lechuza comun Tyto alba en la cuenca del Duero, Espana. Donana, Acta Vertebrata 11:213-229. Colvin, B.A. 1984. Barn owl foraging behaviour and sec- ondary poisoning hazard from rodenticide use on farms. Ph.D. dissertation. Bowling Green State Univ., Bowling Green, OH U.S.A. De Santis, L.J.M., N.G. Basso, J.I. Noriega and M.F. Grossman. 1994. Explotacion del recurso trofico por la lechuza de los campanarios ( Tyto alba) en el oeste de Chubut, Argentina. Stu. Neotrop. Fauna. Envir. 29: 43-47. Digby, P.G.N. and R.A. Kempton. 1987. Multivariate analysis of ecological communities. Chapman and Hall, London, UK. Herrera, C.M. 1974. Trophic diversity of the barn owl Tyto alba in continental western Europe. Ornis Scand. 5:181-191. and F.M. Jaksic. 1980. Feeding ecology of the barn owl in central Chile and southern Spain: a com- parative study. Auk 97:760—767. Iriarte, J.A., W.L. Franklin and W.E. Johnson. 1990. Diets of sympatric raptors in southern Chile./. Raptor Res. 24:41-46. Jaksic, F.M. and J.L. Yanez. 1979. The diet of the barn owl and its relation to the availability of prey. Auk 96: 619-621. and J.R. Rau. 1986. Prey and trophic ecology of Great Horned Owls in western South America: an in- dication of latitudinal trends. J. Raptor Res. 20:113- 116. Ludwig, J. A. and J.F. Reynolds. 1988. Statistical ecology, a primer on methods and computing. Wiley, New York, NY U.S.A. Marti, C.D. 1974. Feeding ecology of four sympatric owls. Condor 76:45-61. Mikkola, H. 1983. Owls of Europe. T. 8c A.D. Poyser, Carlton, UK. Noriega, J.I., R.M. Aramburu, E.R. Justo and L.J.M. De Santis. 1993. Birds present in pellets of Tyto alba (Strigiformes, Tytonidae) from Casa de Piedra, Ar- gentina./. Raptor Res. 27:37-38. Pearson, O.P. and A.K. Pearson. 1982. Ecology and bio- geography of the southern rainforest of Argentina. Pages 129-142 in A.M. Mares and H.H. Genoways [Eds.], Mammalian biology of South America. Spec. Publ. Ser. 6. Pymatuning Lab. Ecol., Pittsburg, PA U.S.A. Pearson, O.P. 1983. Characteristics of mammalian fauna from forests in Patagonia, southern Argentina. /. Mamm . 64:476-492. . 1984. Taxonomy and natural history of some fossorial rodents of Patagonia, southern Argentina. /. Zool. 202:225-237. . 1986. Annotated keys identifying small mam- mals living in or near Nahuel Huapi National Park and Lanin National Park, southern Argentina. Un- publ. Manuscript. . 1987. Mice and the postglacial history of the Traful valley of Argentina. J. Mamm. 68:469-478. . 1988. Biology and feeding dynamics of a South American herbivorous rodent, Reithrodon. Stud. Neo- trop. Fauna Environ. 23:25-39. Pielou, E.C. 1984. The interpretation of ecological data. Wiley, New York, NY U.S.A. Rau,J., R. Murua and M. Rosemann. 1981. Bioenerget- ics and food preferences in sympatric southern Chil- ean rodents. Oecologia 50:205-209. Redford, K.H. and J.F. Eisenberg. 1992. Mammals of the Neotropics, the southern cone, Vol 2. Univ. of Chicago Press, Chicago, IL U.S.A. Ringuelet, R.A. 1961. Rasgos fundamentales de la zo- ogeografia de la Argentina. Physis 22:151-170. SAS. 1987. SAS/STAT Users guide version 6. SAS Inst. Inc. Cary, NC U.S.A. Siegel, S. and N.J. Castellan. 1988. Nonparametric sta- tistics for the behavioral sciences. McGraw-Hill. New York, NY U.S.A. Smith, J.D.B. and J. Cole. 1989, Diet of the barn owl, Tyto alba, in the Tanami Desert, Northern Territory. Aust. WMl. Res. 16:611-624. Taylor, I. 1994. Barn owls, predator-prey relationships and conservation. Cambridge Univ. Press, Cambridge, U.K. Yalden, D.W. 1977. The identification of remains in owl pellets. Occas. Publ. Mammal. Soc., London, UK. Received 23 April 1996; accepted 26 November 1996 J. Raptor Res. 31(1) :65-70 © 1997 The Raptor Research Foundation, Inc. INDICATORS OF MALE QUALITY IN THE HOOTS OF TAWNY OWLS (STRIX ALUCO) Bridget M. Appleby Wildlife Conservation Research Unit, Department of Zoology, University of Oxford, South Parks Road, Oxford 0X1 3PS UK Stephen M. Redpath I. T.E. Monks Wood, Abbots Ripton, Cambridgeshire PEI 7 2LS UK Abstract. — The number of songs in a male’s repertoire, and the amount of time he spends singing, have been shown to correlate with territory size and quality, reproductive success, parental care and parasite load in some passerine species. In addition, females of some species use song rate and com- plexity as a cue to mate choice and are more responsive to more frequent and complex songs. Few studies, however, have examined the influence of body size and parasitic infections on the sound fre- quency (pitch) and structure of vocalizations of birds. The Tawny Owl ( Strix aluco ) hoot is important in communication between birds at night when visual information is limited, and is simple enough to allow a quantitative analysis of its structure. Six temporal and four frequency measures of the hoots of 50 Tawny Owls were taken, and compared to body mass, wing length, breeding success and number and intensity of parasitic infections of the singers. There was a decrease in call frequency with increasing body mass and the vibrato tail of the last note was longer in larger birds, but there was no part of the call that correlated with breeding success. There was an increase in call frequency as the number of parasitic infections increased, and there was a decrease in the length of calls as the intensity of parasitic infections increased. Key WORDS: Strix aluco; hoot structure, body weight, body size, breeding success; blood parasites; female choice, assessment. Indicadores de calidad del macho en ululatos de Strix aluco. Resumen. — La cantidad de canciones en el repertorio del macho, y la cantidad de tiempo que dedica cantando, han mostrado una correlacion entre el tamano del territorio y la calidad del exito de re- production, preocupacion paterna, y carga parasitica en especies passerinus. En suma, hembras usan la velocidad y complication de la cancion como serial para escoger su pareja estan mas interesadas en canciones complicadas y de mas frecuencia. Pocos estudios, sin embargo, han examinado la influencia del tamano corporal y infeccion parasitica en la frecuencia modulada del sonido (tono) y estructura de vocalization en pajaros. El ululato de Strix aluco es importante en comunicacion entre los pajaros por la noche cuando information visual es limitada, y es lo suficientemente como simple para dejar un analisis cualitativo de su estructura. Seis temporal y cuatro frecuencias moduladas de los ululatos de 50 S. aluco fueron tornados, y comparados al tamano corporal, largo de ala, exito de cria, y numeros e intensidad de infeccion parasitica de los cantadores. Hubo reduction en frecuencia de llamadas y au- mento de el tamano corporal y la cola vibratus de la ultima nota fue mas larga en los pajaros grandes, pero no hubo ninguna parte de la llamada que hizo correlacion con el exito de cria. Hubo un aumento en la frecuencia de llamadas cuando la cantidad de infeccion parasitica aumentaron, y hubo una re- duccion en la duration de llamadas cuando la intensidad de infeccion parasitica aumento. [Traduction de Raul De La Garza, Jr.] Some bird species can recognize their offspring (Tschanz 1968, Beer 1969, Beecher et al. 1981), mates (White 1971, Brooke 1978, Jouven tin et al. 1979) and territorial neighbors (Falls 1992, Gal- eotti and Pavan 1993) through their vocalizations. Song can also give information on the position and orientation of the signaler (Richards 1981, Mcgre- gor and Falls 1984, McGregor and Krebs 1984). Song appears to also be important in sexual se- lection in many species and females have been 65 66 Appleby and Red p ath Vol, 31, No. 1 shown to be more responsive to males with higher song output (Houtman 1992, Payne and Payne 1977) whether song output was increased by lengthening songs or reducing intersong intervals (Wasserman 8c Cigliano 1991). In Willow Warbler ( Phylloscopus trochilus) males, there was a link be- tween high song output and good territories (Ra- desater et al. 1987, Radesater 8c Jakobsson 1989), and Houtman (1992) showed that Zebra Finch (Poephila guttata ) males with higher song output produced heavier offspring. Females of some spe- cies have also been shown to prefer males with more complex repertoires (Searcy 8c Marler 1984, Baker et al. 1986, Catchpole et al. 1984, Eens et al. 1991a). The complexity of song has been found to correlate with territory size, survival and reproduc- tive success (Hiebert et al. 1989, McGregor et al. 1981, Catchpole 1986). As song is used by males to attract females, it has been hypothesized that song may give information on the parasite infections of male singers. Hamil- ton 8c Zuk (1982) found bird species with more complex songs were more likely to have blood par- asite infections, but this correlation disappeared when phylogeny was taken into account (Read 8c Weary 1990). Moller (1991) studied the effect of parasites within a population of Barn Swallows (Hi- rundo rustica ) and found that males infected with mites produced less song. Despite the implications for mate choice, litde is known about how the size and weight of a bird affects its vocalizations, and whether parasite infec- tions affect the structure or frequency of a bird’s call. This is particularly relevant for nocturnal birds, which may not have detailed visual infor- mation about a potential mate or rival. In this pa- per, we present an analysis of the information con- tained in the “hoot” vocalization of the Tawny Owl (Strix aluco). Tawny Owls are nocturnal woodland birds, so visual information transferred between territorial rivals or potential mates at night is prob- ably very limited. The hoot is individually distinct and constant with time and is thought to function in communication between the sexes as well as ter- ritorial defense (Galeotti 8c Pavan 1991). Methods Recordings of Tawny Owls were made at Wytham Woods, Oxfordshire (51.46° N 1.2° W), Monks Wood, Cambridgeshire (52.24° N 0.14° W), a farmland area, (The ‘Fens’) Cambridgeshire (52.29° N 0.1° W) and Kielder Forest, Northumberland (55° 15'N 2° 35'W). Re- cordings were made using a Uher or Sony Walkman Pro- fessional tape recorder (TC-D5 PRO) with a Sennheiser MZW 816 microphone. Recordings were made of male birds on calm, dry nights from October-December 1992- 93 in Wytham Woods, and in March 1993 in Monks Wood and the Fens. Recordings of birds at Kielder Forest were made in November 1994. Birds were stimulated to hoot using playback of an unfamiliar male owl. The hoots of male owls can be distinguished from hoots of female owls by the squeaky grating quality of the female hoot. In the case of the Wytham owls, the sex of the birds was con- firmed using radiotags. Recordings were made as near to birds as possible, at distances between 5-50 m. Sonograms were produced on a Macintosh LCII com- puter. Tawny Owl hoots have a basic structure of three notes (Fig. 1) that can be clearly determined in owls from all areas. Sonograms of the whole hoot were made using Soundedit Pro software (Macromind Paracomp, Inc., 600 Townsend, San Francisco, GA 94103 U.S.A. 1992) and temporal measures were recorded. Soundedit Pro did not give accurate frequency measures, so sonograms of the first note of each hoot were made with Canary soft- ware (Cornell Laboratory of Ornithology, 159 Sapsucker Woods Road, Ithaca, NY 14850 U.S.A.), and the frequen- cy measures were recorded. Temporal measures used were similar to those de- scribed by Galeotti and Pavan (1991). Six temporal mea- sures were recorded for each call. These were: note 1 (Dl), internote interval one (14), note two (D2), inter- note interval two (15) and note 3 was split up into fre- quency modulated length (FML) and tail (Fig. 1). Fre- quency measures used were the highest and two lowest frequencies of the first note (HIGH, LOW1 and LOW2) and the middle of the highest part of the first note (MED) . All time measures were recorded in milliseconds (ms) and all frequency measures were recorded in KHz, Only clear recordings with little background noise were used for making sonograms. Any sonograms where all call parameters could not be determined were discarded. An “average” sonogram was calculated for each owl by taking the mean of a minimum of 3 calls from each owl. Breeding success of owls was measured at Wytham Woods, Monks Wood and the Fens in 1993. Breeding suc- cess was defined as the total number of owlets fledged/- breeding area and was found by monitoring nests or by locating calling chicks in June and July before mortality of fledged chicks has occurred. Fledged young can be located in June and July when they call loudly and con- tinuously for food (Muir 1954, Southern 1970). Males were caught in spring 1993 and 1994 using nest- box traps (Petty et al. 1994). Males were weighed, mea- sured for wing length and blood slides were prepared by placing a drop of blood from the branchial vein directly onto a glass slide and smearing it with a second slide to produce a blood layer one cell thick. The smear was then air dried and fixed in absolute ethanol and stained with Giemsa’s stain. Parasite species were identified by M. An- war at Oxford University and the number of each species was quantified by counting the number of parasites in 10 000 blood cells. The two measures of parasite infection used were the number of parasite species present and the total number of blood cells containing parasites (all parasite species pooled) . For some owls, the blood para- site counts and body measurements were taken at differ- March 1997 Information in Tawny Owl Hoots 67 Figure 1. Sonogram of Tawny Owl hoot showing time and frequency measures. ent times of year than recordings of their hoots. In this case, the identity of the owl was confirmed using radi- otags or by confirming that the same male was present the following breeding season. The data were examined visually using histograms and no evidence was found for a significant deviation from the normal distribution. Parametric tests were therefore used. Results The hoots of 50 male Tawny Owls were recorded at the four study sites (Table 1). Although none of the temporal measures of call varied significandy among the four sites (ANOVA, P > 0.05), there was a significant difference between the sites in all of Table 1. The number of Tawny Owls from the four study sites for which data on both call parameters and body measurements, breeding success and parasite loads were available. Fledg- # Para- Study # Body Wing ing Infec- site Site Owls Mass Length Success tions Load Wytham 20 5 5 20 4 4 Monks Wood 9 4 4 9 0 0 Fens 16 6 6 16 0 0 Kielder 5 4 0 0 4 4 the frequency measures (LOW1 P 3>46 = 1 7.03, P < 0.0001; LOW2 P 3>46 = 12.58, P< 0.0001; MED P 3f46 = 4.63, P = 0.007; HIGH P 3)46 = 6.08, P = 0.001). The Kielder owls were lighter than those of the other populations (jF 3>15 — 3.25, P — 0.051), but this was based on only 4 birds from Kielder and there was no significant difference among the sites for wing length and the breeding success of pairs (P > 0.05). Pearson Rank Correlations were not significant for body mass, wing length, breeding success and parasite number and load for any of the owls used in the analysis; however, there was a significant cor- relation between number of parasites and breed- ing success (r = 1, P < 0.001) which was based on only four Wytham owls. There was no significant correlation between the body mass of a bird and the length of any of the notes or the internote intervals of its call, but there was a negative correlation between the highest fre- quency of the song and body mass (Fig. 2, r = —0.48, P <0.05). There was no significant corre- lation between wing length and any of the fre- quency measures of the call. The only time mea- sure of the call that correlated significantly with wing length was the tail of the third note (r = 0.71, P < 0.003). Likewise, there was no significant cor- 68 Appleby and Redpath Vol. 31, No. 1 + LOW 1ST weight ▲ HIGH • LOW 2ND B MEDIUM Figure 2. Effect of body mass of Tawny Owls on the four frequency measures of their hoots. relation between average breeding success in 1993 and any of the call parameters. All four frequencies were entered into a regression model as indepen- dent variables to try to predict breeding success as the dependent variable, however the model did not significantly predict breeding success (iq^o = 0.29, P = 0.88). The number of parasite species present in an individual ranged between one and four. The par- asites found were Leucozytozoon ziemanni, Haemopro- teus syrnni, H. noctuae and Trypanosoma. All fre- quency measures increased with increasing num- X Wytham □ Kielder Figure 4. Effect of parasite load on the total length of Tawny Owl hoots. bers of parasites (Fig. 3) , but this was only signifi- cant for the lowest frequency at the start of the first note. The Kielder birds had higher numbers of parasites and higher frequency calls, but due to the small sample size it was not possible to analyze the data by site to see if the relationship held in each case. There was a significant negative correlation be- tween the length of the internote interval (r = — 0.75, P < 0.05), gaps in the song (r = —0.76, P < 0.05), and the total length of the song (Fig. 4, r = -0.87, P ^ 0.05) and the parasite load, but there was no significant correlation with any of the fre- quencies measured (P > 0.05). NUMBER OF PARASITIC INFECTIONS + Low 1st A High O Low 2nd A Medium Figure 3. Effect of parasitic infections of Tawny Owls on the four frequency measures of their hoots. March 1997 Information in Tawny Owl Hoots 69 Discussion For a bird embarking on a territorial dispute, assessing the size of the opponent is important in predicting the probable outcome of a fight. This is especially true for an aggressive bird like a Tawny Owl, where vocal disputes are often fierce and prolonged. For a female assessing a potential mate, the body mass of a male might give in- formation on resources available in his territory. The size of the syrinx is likely to affect the frequency of the call and could vary with the size of a bird. We found that the highest frequency of Tawny Owl calls to be more closely correlated with body mass than the lowest frequency. This decrease in highest note frequency with size implies that birds might be emitting their highest note to adver- tise their small size. Newton (1988) found that smaller sparrowhawk males had a higher lifetime reproductive success than large males. Although the selective pressures favoring reduced size in Tawny Owl males are probably different to those in sparrowhawks, reversed sexual di- morphism implies there might be selection for smaller males in Tawny Owls. Females might, therefore, favor small males and be using male hoots to assess this. The only call parameter that correlated significantly with wing length was the length of the vibrato tail of the third note, which increased with wing length. Wing length is an indication of body size of birds. It is possible that a larger bird might be able to sustain the last note for a longer time, perhaps due to increased lung volume. Data on more birds would be needed to ascertain wheth- er the length of the last note was a reliable signal of body size. None of the measures of the call were significantly cor- related with breeding success, so there is no detectable signal that females could reliably use to assess the breed- ing capability of a potential partner. As there was only one year’s breeding data available, and this varies with food supply and site (Petty 1992), it is possible that a correlation does exist between reproductive ability and call parameters, but it was obscured by other factors. All measures of frequency increased with increasing number of parasitic infections and this was significant for the lowest frequency at the start of the first note. Males that are able to give low frequency calls might therefore be indicating a resistance to parasites. Birds with higher parasite loads gave significantly shorter calls. Physiologi- cal reasons that might cause this are obscure, but it might imply that long calls are costly to give and are therefore an indication of good health. In frogs, heavily infested males have also been shown to have below average call durations (Read 1988), and long calls in frogs are more effective in attracting females (Rand and Ryan 1981, Wells and Schwartz 1984, Ryan 1985). Lengths of owl hoots are fairly constant with time (Hirons 1976) so it is possible that call length could be utilized to signal para- site burdens and potential breeding capabilities. Acknowledgments We are grateful to David Macdonald, Paul Johnson, Ian Newton and Nick Davies for commenting on earlier drafts of manuscripts. Parasite analysis was performed by All Anwar, and Steve Petty kindly provided information on the weights and breeding success of Kielder birds. This work was completed while B.M.A was in receipt of a NERC studentship. Literature Cited Baker, M.C., T. Bjerke, H. Lampe and Y. Espmark. 1986 Sexual response of female Great Tits to variation in size of males’ song repertoires. Am Nat. 128:491-498. Beecher, M.D., I.M. Beecher and S. Hahn. 1981. Par- ent-offspring recognition in Bank Swallows ( Riparia n- paria ): II development and acoustic basis. Anim. Behav. 29:95-101. Beer, C.G. 1969. Laughing Gull chicks: recognition of their parents’ voices. Science 166:1030-1032. Brooke, M.L. 1978. Sexual differences in the voice and individual vocal recognition in the Manx Shearwater Pnffinus puffinus. Anim. Behav. 26:622-629. Catchpole, C.K. 1986. Song repertoires and reproduc- tive success in the Great Reed Warbler Acrocephalus arundinaceus. Behav. Scot Sociobiol. 19:439—445. , J. Dittami and B. Leisler. 1984. Differential re- sponses to male song repertoires in female songbirds implanted with oestrodiol. Nature 312:563-564. Eens, M., R. Pinxten and R.F. Verheyen. 1991. Male song as a cue for mate choice in the European Star- ling. Behaviour 116: 210-238. Falls, J.B. 1992. Playback: a historical perspective. Pages 11-33 in P.K. McGregor [Ed.], Playback and studies of animal communication. Plenum Press, New York, NY U.S.A. Galeotti, P. and G. Pavan. 1991. Individual recognition of male Tawny Owls ( Strix Aluco) using spectrograms of their territorial calls. Ethology, Ecology and Evolution 3:113-126. and G. Pavan. 1993. Differential responses of male Tawny Owls Strix aluco to the hooting of neigh- bours and strangers. Ibis 135:300-304. Hamilton, W.D. and M. Zuk. 1982. Heritable true fit- ness and bright birds: a role for parasites? Science 218: 384—387. Hiebert, S.M., PJ* Stoddard and P. Arcese, 1989. Rep- ertoire size, territory acquisition and reproductive success in the song sparrow. Anim. Behav. 37:226—273. Hirons, G.J.M. 1976. A population study of the Tawny Owl, Strix aluco and its main prey species in woodland. Ph.D. dissertation, Oxford Univ., Oxford UK . 1984. Body weight, gonad development and moult in the Tawny Owl, Strix aluco. J. Zool. Lond. 202. 145-164. Houtman, A.M. 1992. Female Zebra Finches choose ex- tra pair copulations with genetically attractive males Proc. R. Soc. Lond. B. 249:3—6. Jouventin, P., M. Guillotin and A. Cornet. 1979. Le chant du Manchot Empereur et sa signification adap- tive. Behaviour 70: 231-250. McGregor, P.K. and J.B. Falls. 1984. The response of Western Meadowlarks ( Sturnella neglecta ) to the play- 70 Appleby and Redpath Vol. 31, No. 1 back of undegraded and degraded songs. Can. J. Zool. 62:2125-2128. , J.R. Krebs and C.M. Perrins. 1981. Song rep- ertoires and lifetime reproductive success in the Great Tit (Parus major). Am. Nat. 118:149—159. . 1984. Sound degradation as a distance cue in Great Tit ( Parus major) song. Behav. Ecol. Sociobiol. 1 6: 49-56. M0ller, A.P. 1991. Parasite load reduces song output in a passerine bird. Anim. Behav. 41:723-730. Muir, R.C. 1954. Calling and feeding rates of fledged Tawny Owls. Bird study 1:111—117. Newton, I. 1988. Individual performance in spar- rowhawks: the ecology of two sexes. Proc Int. Ornithol. Congress 1:125-154. Payne, R.B. and K Payne. 1977. Social organization and mating success in local song populations of Village Indigobirds Vidua chalybeata. Z. Tierpsychol. 45:133- 173. Petty, S.J. 1992. Ecology of the Tawny Owl Strix aluco in the spruce forests of Northumberland and Argyll. Ph.D. dissertation, The Open Univ., , G. Shaw and D.I.K. Anderson. 1994. The value of nest boxes for population studies and conservation of owls in coniferous forests in Britain. J. Raptor Res. 28:134-142. Radesater, T. and S. Jacobsson. 1989. Song rate cor- relation of replacement territorial Willow Warblers Phylloscopus trochilus. Ornis Scand. 20:71-73. , N. Aandbjer, A. Bylin and K. Nystrom. 1987. Song rate and pair formation in the Willow Warbler, Phylloscopus trochilus. Ornis Scand. 20:71—73. Rand, A.S. and M.J. Ryan. 1981. The adaptive signifi- cance of a complex vocal repertoire in a neotropical frog. Z. Tierpsychol 57:209-214. Read, A.F. 1988. Sexual selection and the role of para- sites. TREE 3:97-101. AND D.M. Weary. 1990. Sexual selection and the evolution of bird song: a test of the Hamilton & Zuk hypothesis. Behav. Ecol. Sociobiol. 26:47-56. Richards, D.G. 1981. Estimation of distance of singing conspecifics by the Carolina Wren. Auk 98:127—133. Ryan, M.J. 1985. The tundra frog: a study in sexual se- lection and communication. Univ. of Chicago Press, Chicago, IL U.S.A. Searcy, W.A. and P. Marler. 1984. Interspecific differ- ences in the response of female birds to song reper- toires. Z. Tierpsychol. 66:128—142. Southern, H.N. 1970. The natural control of a popu- lation of Tawny Owls (Strix aluco). J. Zool. 162:179-285. Tschanz, B. 1968. Trottellummen. Z. Tierpsychol. (Suppl.) 4. Wasserman, F.E. and J.A. Cigliano. 1991. Song output and stimulation of the female in White-throated Spar- rows. Behav. Ecol. Sociobiol. 29:55-59. Wells, K.D. and J.J. Schwartz. 1984. Vocal communi- cation in a neotropical treefrog, Hyla ebraccata : adver- tisement calls. Anim behav. 32:405-420. White, S.J. 1971. Selective responses by the Gannet Sula bassana to played-back calls. Anim. Behav. 19:125-131. Received 8 April 1996; accepted 24 November 1996 J. Raptor Res. 31(l):7l-76 © 1997 The Raptor Research Foundation, Inc. ROLE OF REFUSE AS FOOD FOR MIGRANT, FLOATER AND BREEDING BLACK KITES (. MILVUS MIGRANS) Guillermo Blanco Departamento de Biologta Animal, Universidad de Alcala de Henares, Alcala de Henares, E-28871 Madrid, Spain Abstract. — The use of refuse by breeding, floating, and migrating Black Kites ( Milvus migrans) was studied near a large garbage dump in Madrid, Spain. Refuse was an important food resource for non- breeding Black Kites, especially migrants that fed almost exclusively at the garbage dump. The dump was only of secondary and variable importance for kites during the breeding season. Pairs of breeding kites nesting in nearby wooded areas foraged mainly on a wide variety of wild prey and only sporadically ate refuse at the dump. Floaters roosted in nearby wooded areas but, unlike breeding pairs, they foraged mainly by scavenging at the dump. Exploitation of food resources other than refuse by breeding kites may be due to their need for large amounts of prey biomass for brood provisioning. Perhaps dumps augment populations of Black Kites by providing rich foraging areas where large numbers of nonbreed- ing and migratory kites can scavenge for food. Key Words: Black Kite, Milvus migrans; garbage dump ; foraging, refuse. Importancia de la basura en la dieta de Milanos Negros ( Milvus migrans) migrantes, flotantes y reprod- uctores. Resumen. — La importancia de la basura en la dieta de Milanos Negros ( Milvus migrans) reproductores, migradores e integrantes de la poblacion flotante fue estudiada en las cercanias del basurero de Madrid, Espana. La basura fue un recurso importante para los milanos no reproductores, especialmente para los migrantes, pero su papel fue secundario aunque variable para los milanos que se reprodujeron cerca del basurero. Los individuos flotantes forrajearon en el basurero pero tambien consumieron una amplia variedad de presas salvajes. La abundancia de alimento en el basurero no provoco que los milanos nidificaran en sus cercanias, ya que la parejas que asi lo hicieron consumieron basura solo esporadica- mente. Por el contrario, los individuos flotantes percnoctaron a diario en la zona arbolada mas cercana al basurero. El uso preferencial de presas salvajes frente a la basura por los milanos reproductores puede ser explicado como consecuencia de la necesidad de presas de gran biomasa para su consumo por los polios en crecimiento. Durante la estacion reproductora, el vertedero de Madrid podria tener una importancia indirecta para el mantenimiento y conservation de la poblacion reproductora a traves del reemplazo de las perdidas en los reproductores por los individuos flotantes. [Traduction del Autor] Breeding and nonbreeding segments of raptor populations often share the same foraging habitats (Newton 1979). Floaters, nonbreeding, nonterri- torial adults and subadults, have larger home rang- es and greater mobility than breeders, and usually concentrate in areas rich in food (Newton 1979, Ceballos and Donazar 1990). Interactions between floaters and breeding conspecifics has been little studied despite its potential to influence raptor population dynamics (Newton 1979). Both breed- ers and floaters of social species such as the Milvus kites coexist in breeding areas when food is abun- dant (Espina 1986, Koga et al. 1989, Heredia et al. 1991), especially where large waste accumulations occur at garbage dumps (Ceballos and Donazar 1988, Blanco 1994). Rubbish dumps and dung- heaps are frequently used by scavenging birds that normally exploit temporary and unpredictable food sources (Pomeroy 1975, Donazar 1992). The importance of refuse dumps for breeding popula- tions of scavenging birds has been repeatedly em- phasized (Pomeroy 1975, Coulson et al. 1987, Don- azar 1992) but, to date, it is not known whether these unnatural foraging places are influencing the population dynamics of scavenging species. The Black Kite ( Milvus migrans ) is a widespread 71 72 Blanco Vol. 31, No. 1 species that opportunistically exploits a wide array of food sources (Delibes 1975, Shiraishi et aL 1990). Floaters form an unknown proportion of populations, and usually share habitat with the breeding segment (Espina 1986, Koga et al. 1989, Blanco 1994). At present, little is known about the relationship between these two sectors of Black Kite populations. This study deals with the role of refuse in the diet of breeding, floating and mi- grating Black Kites living in the vicinity of a large garbage dump near Madrid, Spain. Study Area and Methods The study area was located at an elevation of 500-700 m in the flood plain of the Jarama River at its confluence with River Manzanares (40° 19'N, 3° 31'W) in southeast- ern Madrid Province, central Spain. It included a com- plex of riverine gypsum cliffs, riparian forests dominated by poplars ( Populus alba, P. nigra, P. X cultivar), willows ( Salix spp.) and elms ( Ulmus minor), but most of the area was converted to agriculture (mainly cereal crops, sun- flower, alfalfa and vegetables), cattle grazing and gravel extraction. The Madrid garbage dump was located in the northeastern portion of the study area. From 100 to >1000 Black Kites routinely gathered there to feed on the refuse (Blanco 1994). During the breeding season of 1994, the Black Kite population consisted of about 50 pairs nesting on trees, cliffs and electric pylons, and about 80-300 nonbreeding individuals forming a floating population of both adults and subadults (Blanco 1994). Each day, resident floaters roosted communally about 4 km southeast of the dump, in a portion of riverine elm forest where 6-10 pairs nest- ed Numbers of Black Kites in the area increased consid- erably after breeding and especially during migration (Blanco 1994). I determined the diet of nonbreeding kites by analyz- ing food remains collected beneath roosting and diurnal communal perching trees during the period when Black Kites are resident in the area (Blanco 1994). All prey remains (bones, fur and feathers) were collected on each visit to the roost, from the final stage of the nestling pe- riod (June) to the end of August. Prey collections were grouped into three time periods (22 July, 31 July and August) to assess possible temporal variations in the con- sumption of refuse. The first period (Period A: until 22 July) included the spring migration, breeding season, fledgling period and the start of migrant arrivals; Period B (23-31 July) coincided with an increase in the influx of migrants; and Period C (from 1 August to the end of that month) included the peak abundance of migrants and the remainder of the residency period of Black Kites in the area (Blanco 1994). Because all food remains found below the roosting trees were collected in each visit to the roost, I am confident that they were accurately grouped on a temporal basis. Because breeding Black Kites usually roosted close to their nesting sites during the breeding period, prey remains found at the com- munal roost site were mosdy from food eaten by local nonbreeders. Prey remains found at the roost after the breeding season corresponded to food of both local (breeding and floating) and migrant kites. After the breeding season, field observations suggested that breed- ing Black Kites and their fledglings joined communal roosts together with floaters and northern migrants (J. Vinuela pers. comm, and pers. obs.). The first juveniles appeared in the roost soon after the breeding season (Blanco 1994) when breeders left their territories. There- fore, some remains collected at this time might have orig- inated from prey of local breeders which represented a high proportion of the birds roosting communally from 23-31 July. Afterwards, the collecting period coincided with the major influx of migrants (>1000 kites in early August); at this time local breeders constituted a small proportion of the birds. The diet of breeding kites and their nestlings was de- termined by analysis of food remains found in and below the nests of 18 pairs. The material was collected after the breeding season to avoid disturbance at the nests. Pairs studied nested in trees and cliffs about 4-12.5 km from the dump. Food remains were identified by macroscopic compar- ison with reference collections and quantified assuming the smallest possible number of individuals (Marti 1987). Two main categories of food were considered in the anal- ysis: wild prey obtained by hunting and scavenging (e.g., wild birds, mammals, fish), and refuse (e.g., food items intentionally discarded by humans such as domestic re- fuse, offal from slaughterhouses and butcher shops, and marine fish) . To detect general differences in food habits of breeding and nonbreeding Black Kates, prey items were classified into nine major groups (Table 1). The data were likely biased in favor of the most durable prey remains and did not reflect the importance of arthro- pods and other prey (Marti 1987). However, the study was not designed to provide detailed descriptions of the Black Kite diet (Delibes 1975, Veiga and Hiraldo 1989, Shiraishi et al. 1990) but to detect broad differences in the consumption of refuse. Biomass of each prey species was estimated using mean weights of each prey taxon obtained from the literature. Weights of prey taxon identified at nests could not ex- ceed 300 g even when the mean weight of the taxon exceeded that weight because Black Kites usually do not transport heavier prey to the nest (Espina 1986, pers. obs.). In this case, I assumed that prey heavier than 300 g were not delivered to the nest whole. Weights of large prey from the roost were estimated based on the daily food requirements of adult Black Kites (100 g, Espina 1986; see also Heredia et al. 1991 for Red Kites [Milvus milvus]). Therefore, I assumed that the maximum inges- tion capacity per feeding bout to be equivalent to the daily food requirements (Blanco et al. 1990). Although Black Kites may tear off small pieces of the prey and bring them to the nest one by one (J. Vinuela pers comm.), and nonbreeding kites may eat prey over several days, I felt my criteria avoided an overestimation of the biomass of large prey up to the size of a White Stork ( Ciconia ciconia, weight = 3.5 kg). Refuse consumed both by breeding and nonbreeding kites typically consisted of small pieces of food, including parts of fish and fowl (mostly heads, wings and legs) and livestock bones, but no large fragments of food, which was estimated to yield 50 g of biomass on average. Be- March 1997 Refuse Use by Black Kites 73 Table 1 . Diets of breeding and nonbreeding Black Kites in the southeast of Madrid during and after the breeding season of 1994. Results are expressed as percentages of number of prey (NP) and biomass (B) for each prey class. Prey Items Breeding Black Kites Non-Breeding Black Kites Breeding Season Nonbreeding Season % NP % B % NP % B % NP % B Wild prey 76.4 93.1 40.6 57.4 21.4 29.4 Rabbits and Hares a 22.5 39.7 18.8 26.9 8.7 16.3 Other mammals b 6.0 4.9 3.1 4.5 1.0 1.8 Pigeons c 8.8 15.5 3.1 4.5 2.9 5.4 Other birds d 18.7 21.1 12.5 17.0 2.9 5.4 Fish e 13.2 11.0 3.1 4.5 0.0 0.0 Other prey f 7.1 0.9 0.0 0.0 5.8 0.4 Refuse 23.6 6.9 59.4 42.6 78.6 70.6 Livestock carrion 17.0 5.0 37.5 26.9 48.5 42.6 Chicken 5.5 1.6 18.8 13.5 22.3 20.8 Marine fish 1.1 0.3 3.1 2.2 7.8 7.2 Number of prey G'T 00 32 103 H' 0.87 0.73 0.66 a Oryctolagus cuniculus, Lepus granatensis. b Erinaceus europaeus, Rattus norvegicus, Apodemus sylvaticus, unidentified rodents. c Columba livia, Columba livia var. domestica, Columba palumbus, unidentified pigeons. d Ciconia ciconia, Anas platyrhynchos, unidentifeid Anatidae, Milvus migrans, Alectoris rufa, Gallinula chloropus, Larus ridibundus, Clamator glandarius, Otus scops, Athene noctua, Picus viridis, Sturnus unicolor. Pica pica, Corvus monedula, unidentified corvid, Turdus merula, Passer sp., unidentified bird. e Cyprinus carpio, Ictalurus melas, unidentified fish. 1 Lacerta lepida, unidentified Colubridae, unidentified Coleopterans, bird eggs. cause of the difficulty in accurately estimating biomass values from prey remains of scavengers and facultative predators (Marti 1987), the analysis conducted in this respect should be considered as a simple approximation of the diet. The aim of the biomass analysis was provide a comparative assessment of the energetic role of refuse with the natural prey for breeding and nonbreeding kites. Dietary diversity was calculated with the Shannon- Weaver index (H') considering the major prey classes (Table 1). Dietary overlap between breeding and non- breeding kites was calculated using Pianka’s index (Pian- ka 1973). Chi-square tests with Yates’ correction when df = 1, were used to test for differences in the number of prey items (NP) consumed by breeding and nonbreeding kites. Results General Food Habits. The diet of Black Kites in- cluded a wide range of wild prey and refuse (Table 1). Up to 300 nonbreeding kites returned to the roost every day from the garbage dump through- out the breeding season. Most of the food they ob- tained from the dump was small pieces of domestic refuse and slaughter offal ( Ovis, Sus, Bovis, Gallus) which was spread fairly evenly over large amounts of inorganic materials because of the daily treat- ment measures at the dump. Unidentified marine fishes, available only at the dump, were also con- sumed. Wild prey consumed by nonbreeders in- cluded mostly European rabbits ( Oryctolagus cuni- culus, NP = 34.3%, B = 41.4%), but hares ( Lepus granatensis) and pigeons ( Columba spp.) also were of some importance. Breeding kites ate mainly wild prey (Table 1), the main prey species being European rabbits (NP = 18.7%, B — 32.9%). Altogether, birds accounted for 27.5% by NP and 36.6% by B, and they were mosdy medium-sized corvids and pigeons. Spatial Variation in the Consumption of Refuse by Breeding Kites. The roost site used by non- breeding kites was 4 km from the dump. Pairs nest- ing near the roost site had a higher proportion of refuse in their diets than did pairs that nested 5.1- 12.5 km from the dump (x ± SE = 6.6 ±0.6 km; X 2 = 12.53, df = 1, P< 0.001 for NP; Fig. 1). How- ever, there was variation in the consumption of re- fuse among pairs that nested close to the dump (18.2-72.7% of NP, Fig. 2), with somewhat lower values for biomass (5.4-30.8%, Fig. 1). Although some kites probably obtained refuse at 74 Blanco Vol. 31, No. 1 60 50 40 % BIOMASS PREY ITEMS BREEDING FAR FROM THE DUMP BREEDING CLOSE TO THE DUMP RESIDENT NON-BREEDING Figure 1. Percentage of refuse in the diet of Black Kites. Data are presented for resident floaters and for kites nesting either far or close to the garbage dump of Madrid, respectively. sites other than the dump, the proportion of re- fuse in the diet of breeding pairs was negatively related to distance to the dump (Spearman rank correlation coefficient r s = —0.64, P = 0,0087 and r s = —0.70, P = 0.0038 for %NP and %B, respec- tively, N = 18, Fig. 2). Thus, the proportion of re- fuse in the diet of breeding Black Kites was variable for those pairs breeding in the vicinity of the dump but almost insignificant in terms of biomass for the remainder of the breeding population. Diet Differences and Overlap within the Local Population. The proportion of prey remains in- cluded in the refuse class was significandy higher for nonbreeding than for breeding kites (100 of 135 vs. 44 of 182, x 2 = 75.84, df = 1, P < 0.001). A similar conclusion was reached by comparing food remains from the breeding season only, both for all nests (x 2 = 14.58, df = 1, P< 0.001 for NP) and when nests located close to and far from the roosting area were considered separately (P < 0.05 in all cases, Fig. 1). I found a higher overlap in the diets of breeding Black Kites for prey biomass than for prey type (Table 2). Nonbreeding floaters showed a higher overlap with kites breeding close to the dump than with those breeding farther away, both in type and biomass of the prey they con- Figure 2. Proportion of prey from refuse in the diet of breeding Black Kites ( N = 18 pairs) in relation to the distance of their nest from the garbage dump of Madrid. Large dots represent two nests. March 1997 Refuse Use by Black Kites 75 Table 2. Percentage of diet overlap (Pianka’s index) be- tween breeding and nonbreeding Black Kites. % of Diet Overlap Number Comparisons of Prey Biomass Breeding close vs. far from the dump 69.8 86.0 Breeding close vs. nonbreeding 91.0 80.9 Breeding far vs. nonbreeding 55.6 70.8 sumed. Nevertheless, dietary overlap was high in all the comparisons as expected for individuals of the same species. Seasonal Variation in the Consumption of Re- fuse by Non-breeding Kites. Refuse was the main food for nonbreeding kites (Table 3) . However, the proportion of refuse increased as the season pro- gressed from the breeding season, when most kites gathering at the roost belonged to the local non- breeding population, toward the end of the col- lecting period when most kites were migrants (x 2 = 7.75, df = 2, P — 0.02 for NP; Table 3). As a result, dietary diversity declined as the season pro- gressed. Discussion In this study, refuse was a more important food resource for Black Kites, especially nonbreeding and migrant individuals. Floaters roosted in wood- ed areas near the dump and consumed refuse al- most extensively and very seldom taking wild prey. Apparently they were able to exploit this abundant and predictable food source because their move- ments were not restrained by breeding. Contrast- ingly, breeding Black Kites fed mosdy on wild prey items but there was a significant correlation be- tween the distance of nests from the dump and the consumption of refuse with pairs nearest the dump increasing their consumption of refuse. Preferen- tial use of natural food sources by breeding Black Kites is probably related to constraints placed on breeding kites by the need to provision broods of developing young with adequate prey biomass (Donazar 1988, Vinuela and Veiga 1992). In addi- tion, high percentages of rubbish/carrion in the diet could negatively affect the growth of young (Hiraldo et al. 1990, Vinuela 1991). After the breeding season, refuse is also used by local breeders and their fledglings. Inexperienced juveniles benefit greatly by using the dump and by Table 3. Seasonal variation in the consumption of re- fuse by nonbreeding Black Kites, expressed as percentage of number of prey items (NP) and biomass (B) . Period % B % NP H' n A (to July 22) 42.6 59.4 0.73 32 B (23-31 July) 64.7 69.8 0.66 43 C (August) 75.0 85.0 0.60 60 group foraging at a place where several hundreds of kites feed together. The presence of floaters may enhance the breed- ing population of a species by providing a mecha- nism for quick replacement of lost mates at nest sites. Through their support of juvenile and sub- adult, nonbreeders, garbage dumps may provide a mechanism of increasing the survival and recruit- ment of breeding individuals into the populations of scavenging birds. Indeed, garbage dumps have been highlighted for their potential importance in the conservation of Black Kite populations, es- pecially during migration (Donazar 1992, Blanco 1994). Acknowledgments I thank A. Acha, L. Blanco, M.C. Blanco, J.A. Fargallo, F. Gomez, F. Martinez and E. Soto-Largo for help with the field work. J.L. Telia discussed some aspects of the work and helped me with the analysis of data, and J. Vih- uela, C.D. Marti, J. Bustamante and an anonymous ref- eree provided many useful comments on the manuscript. Literature Cited Blanco, J.C., J.L. Gonz al ez and F. Hiraldo. 1990. Tro- phic and spatial relationships between wintering Red Kites ( Milvus milvus ) and Marsh Harriers ( Circus aeru- ginosus) in the Guadalquivir Marshes. Miscelanea Zool- ogica 14:161-166. Blanco, G. 1994. Seasonal abundance of Black Kites as- sociated with the rubbish dump of Madrid, Spain. /. Raptor Res. 28:242-245. Ceballos, O. and J.A. DonAzar. 1988. Actividad, uso del espacio y cuidado parental en una pareja de Alimoch- es ( Neophron percnopterus) durante el perfodo de de- pendence de los polios. Ecologia 2:275-291. . 1990. Roost-tree characteristics, food habits and seasonal abundance of roosting Egyptian Vultures in northern Spain./. Raptor Res. 24:19-25. Coulson, J.C., J. Butterfield, N. Duncan and C. Tho- mas. 1987. Use of refuse tips by adult British Herring gulls Larus argentatus during the week. J. Appl. Ecol. 24:789-800. Delibes, M. 1975. Alimentation del Milano Negro (Mil- 76 Blanco Vol. 31, No. 1 vus migrans) en Donana (Huelva, Espana). Ardeola 21: 183-207. DonAzar, J.A. 1988. Variaciones en la alimentacion en- tre adultos reproductores y polios en el Buho Real {Bubo bubo). Ardeola 35:278-284. . 1992. Muladares y basureros en la biologia y conservation de las aves en Espana. Ardeola 39:29-40. Espina, J. 1986. Variaciones en la alimentacion de una poblacion de Milanos Negros Milvus migrans. Tesina de Licenciatura, Universidad Complutense de Ma- drid, Madrid, Spain. Heredia, B., J.C. Alonso and F. Hiraldo. 1991. Space and habitat use by Red Kites Milvus milvus during win- ter in the Guadalquivir marshes: a comparison be- tween resident and wintering populations. Ibis 133: 374-381. Hiraldo, F.,J.P. Veiga and M. Manez. 1990. Growth of nestling Black Kites Milvus migrans : effects of hatching order, weather and time season. J. Zool. London 222: 197-214. Koga, K., S. Siraishi and T.A. Uchida. 1989. Breeding ecology of the Black-eared Kite Milvus migrans lineatus in the Nagasaki Peninsula, Kyushu. Jap. J. Ornith. 38: 57-66. Marti, C.D. 1987. Raptor food habits studies. Pages 67- 80 in B.A. Giron-Pendleton, B.A. Millsap, K.W. Cline and D.M. Bird [Eds.], Raptor management tech- niques manual. Natl. Wild. Fed., Washington, DC U.S.A. Newton, I. 1979. Population ecology of raptors. T. & A.D. Poyser, Berkhamsted, UK. Pomeroy, D.E. 1975. Birds as scavengers of refuse. Ibis 117:69-81. Shiraishi, S., K. Koga and N. Kawaji. 1990. Food habits of the Black-eared Kite, Milvus migrans lineatus, in Na- gasaki airport and its adjacent areas./. Fac. Agric. Kyu- shu Univ. 34:247-254. Veiga, J.P. and F. Hiraldo. 1989. Food habits and the survival and growth of nestlings in two sympatric kites ( Milvus milvus and Milvus migrans). Holarct. Ecol. 13: 62-71. Vinuela, J. 1991. Ecologia reproductiva del Milano Ne- gro en el Parque Nacional de Donana. Ph.D. disser- tation, Universidad Complutense de Madrid, Madrid, Spain. and J.P. Veiga. 1992. Importance of rabbits in the diet and reproductive success of Black Kites in southwestern Spain. Ornis Scand. 23:132-138. Received 10 April 1996; accepted 16 November 1996 Short Communications J. Raptor Res. 31(l):77-79 © 1997 The Raptor Research Foundation, Inc. Interspecific and Intraspecific Aggression Among Griffon and Cinereous Vultures at Nesting and Foraging Sites Guillermo Blanco Departamento de Biologia Animal, Universidad de Alcala de Henares, 28871 Alcala de Henares, Madrid, Spain Jose M. Traverso Jardines 18, 28610 Villamanta, Madrid, Spain Javier Marchamalo Tomas Borras 12, 28045 Madrid, Spain Felix Martinez Puerto Canfranc 22, 28038 Madrid, Spain Key Words: Aegypius monachus; Cinereous Vulture, Gyps fulvus; Griffon vulture, foraging, nesting aggression. Food partitioning within vulture guilds reduces the in- tensity of competition even when several species share the same carcass (Konig 1983, Hiraldo 1977, Mundy 1982). Apparently, aggressive encounters are minimized by the relative sizes and ages of individuals in these feed- ing groups (Alvarez et al. 1976, Grubh 1978, Anderson and Horwitz 1979, Mundy 1982, Blanco and Martinez 1996). Interspecific and intraspecific aggression is fre- quent when large numbers gather at large carcasses (Mundy 1982). Most interactions involve dominance dis- plays that either do not result in direct contact or, at most, result in slight disputes reduced to minor pecking and kicking. Overall, severe aggression seems to be avoid- ed. In addition to carcasses, vultures can at times also show intra- and interspecific competitive behaviors at their nest sites. Here too, competition may be passive and regulated by the abundance, spacing and timing of breeding of the different species, or it may result in active aggression and displacement among individuals of the same species (Donazar et al. 1989, Fernandez and Donazar 1991). Griffon Vultures ( Gyps fulvus) and Cinereous Vultures {Aegypius monachus ) share feeding and nesting situations across wide areas of the Iberian peninsula, but little is known of the frequency, intensity and context of agonis- tic interactions that result from competition between them (Konig 1974, Alvarez et al. 1976, Fernandez and Donazar 1991, Hiraldo et al. 1991). In this note we report two cases of intense aggression between Griffon and Ci- nereous Vultures. The first case occurred while we observed Cinereous and Griffon Vultures feeding on a cattle carcass near the Natural Park of Monfragiie, Caceres Province (western Spain) in November 1992. A maximum of 94 Griffon Vul- tures and 48 Cinereous Vultures were observed gathered to feed on the carcass. Most of them were feeding on the back of the cow, but two subadult Cinereous Vultures were feeding on the head by putting their heads inside the mouth of the cow. A first-year Griffon Vulture ap- proached and began to feed together with the two Ci- nereous Vultures. When the Griffon Vulture had its head inside the mouth of the cow, an adult Cinereous Vulture approached and attacked the Griffon Vulture by grasping its neck with its bill for >30 sec. When the Cinereous Vulture released the Griffon Vulture, it left a deep wound that bled profusely. Immediately the injured vulture abandoned the carcass, leaving a clear blood trail and disappeared into the adjacent vegetation. The wound ap- peared serious enough that it might have been lethal. The second case involved of a conflict between two adult (probably female) Griffon Vultures at a nesting site traditionally used in the gorges of the Riaza River, Sego- via Province (central Spain) in November 1995. We had observed an adult pair at this nest two months previously showing pair-bonding behavior and even copulating. The mates were sexed when copulation occurred and subse- quently were easily recognized by individual plumage fea- tures. The third bird was probably a female, based on the size and shape of the bill and head. The aggressive en- 77 78 Short Communications Vol. 31, No. 1 counter began when the two presumed female Griffon Vultures began to make rasping calls and to fight at the nest site. During the fight, one of the vultures laid on its back with its wings extended and defended itself with its legs and bill from pecking attacks by the other vulture. The two vultures continued to fight, alternating positions and pecking at each other’s head and neck for >20 min. At the end of the fight, both females remained motion- less on their backs and face to face with their wings ex- tended for about 5 min. Finally, after another 6 min of violent aggression, one of the vultures displaced the oth- er from the nest site. As a result of die fight, both vultures sustained numerous bleeding wounds on their heads and necks, especially around their eyes, mandibles, and throats. Initially, the male w r atehed the conflict from the edge of the nest site and only pecked a few times at the tips of the primaries of the fighting females. The male later flew to a nearby rock from which he observed the fight for the remainder of its duration. During many years of research, we have observed several hundred vulture gatherings at carcasses where numerous agonistic interactions have occurred. Most aggressive behaviors have been between individuals of the same species and especially between the more so- cial and abundant Griffon Vulture (Hiraldo et al. 1979). These encounters never generated wounds and usually resulted in the loss of a few neck or rarely pri- mary feathers. The more violent encounters have in- volved first-year birds assaulted by older individuals in- dicating that immature vultures may not have yet es- tablished dominance relationships or developed ap- propriate social behavior. Our first observation indicates that Cinereous Vultures are capable of oc- casionally injuring Griffon Vultures seriously enough to result in death. Undoubtedly, their larger body size and stronger bill enables them to dominate in aggres- sive encounters (Konig 1983). Nevertheless, this be- havior is exceptional and probably occurs only when large numbers of vultures, especially Cinereous Vul- tures, gather at a carcass. Competition for nest sites may be a cause of aggression between conspecifics under conditions of high popula- tion density (Martinez and Cobo 1993). The availability of nest sites for Griffon Vultures has not been docu- mented to be a factor limiting the population size in northern Spain where high densities are reached (Arroyo et al. 1990, Donazar and Fernandez 1990). Nevertheless, we observed an increasing number of aggressive inter- actions over the past 10 years (Martinez and Cobo 1993). These encounters have mainly been brief and have amounted to nothing more than attempts to steal nest material and food delivered to nestling. Our observation of actual fighting between two female Griffon Vultures was unusual but it may indicate that as die density of the nesting population of Griffon Vultures continues to in- crease in the Riaza River gorge, aggression within this species will become more violent. Resumen. — Se describe un caso de agresion violenta de un buitre negro (Aegypius monachus) adulto a un juvenil de buitre leonado (Gyps fulvus) en una carrona de vaca que congrego un numero inusual de individuos de la primera especie. El buitre leonado pudo haber muerto como consecuencia de la gravedad de la herida produ- cida en el cuello por el buitre negro, Se documenta tam- bien la observation de una pelea muy violenta entre dos hembras de buitre leonado en un lugar de nidificacion ocupado por una de ellas y su pareja, lo cual se interpreta como resultado de una fuerte competencia por el lugar de nidificacion en una colonia de muy alta densidad pob- lacional en Espana. [Traduction de los Autores] Acknowledgments The manuscript was written during a short stay by the first author in the Centro de Investigaciones Riologicas del Noroeste (Mexico) and benefited from the com- ments of P. Mundy, R. Rodriguez-Estrella, J.L. Telia and S.R. Wilbur. The stay was financed by a F.P.U grant from the Spanish Ministerio de Education y Ciencia. Literature Cited Alvarez, F., L. Arias De Reyna and F. Hiraldo. 1976. Interactions among avian scavengers in southern Spain. Ornis Scand. 7:215-226. Anderson, J. and P. Horwitz. 1979, Competitive inter- actions among vultures and their avian competitors. Ibis 121:505-509. Arroyo, B., E. Ferreiro and V. Garza. 1990. II Censo national de Buitre Leonado ( Gyps fulvus ): poblacion, distribution, demografia y conservation. ICONA, Ma- drid, Spain. Blanco, G. and F. MartInez. 1996. Sex difference in breeding age of Griffon Vultures (Gyps fuhrus). Auk 113:247-248. DonAzar, J.A., O. Ceballos and C. FernAndez. 1989. Factors influencing the distribution and abundance of seven cliff-nesting raptors: a multivariate study. Pages 545-552 in B.U. Meyburg and R.D. Chancellor [Eds.], Raptors in the modern world. WWGBP, Berlin, Germany. and C. FernAndez. 1990. Population trends of the Griffon Vulture Gyps fulvus in northern Spain be- tween 1969 and 1989 in relation to conservation mea- sures. Biol. Conserv. 53:83-91. FernAndez, C. andJ.A. DonAzar. 1991. Griffon Vultures ( Gyps fulvus) occupying eyries of other cliff-nesting raptors. Bird Study 38:42-44. Grubh, R.B. 1978. Competition and co-existence in Grif- fon Vultures: Gyps bengalensis, G. indicus and G. fulvus in Gir forest./. Bom. Nat. Hist. Soc. 75:810-814. Hiraldo, F. 1977. Relaciones entre morfologia, ecologia March 1997 Short Communications 79 y distribucion de los buitres del Viejo Mundo. Actas I Reunion Iberoamer. Zool. Vert., La Rabida: 753—757. , M. Delibes and J. Calderon. 1979. El Quebran- tahuesos Gypaetus barbatus (L). Monografias 22. ICONA, Madrid, Spain. , J.G. Blanco and J. Bustamante. 1991. Special- ized exploitation of small carcasses by birds. Bird Study 38:200-207. KOnig, C. 1974. Zum verhalten spanischer Geier an Ka- davern . J. Orn. 115:289-320. . 1983. Interspecific and intraspecific competition for food among Old World vultures. Pages 153-171 in S.R. Wilbur and J. A. Jackson [Eds.], Vulture biology and management. Univ. of California Press, Berkeley and Los Angeles, CA USA. MartInez, F. and J. Cobo. 1993. Gestion actual de AD- ENA/WWF Espana en el Refugio de rapaces de Mon- tejo de la Vega (Segovia). Alytes 6:507-521. Mundy, r-J. 1982. The comparative biology of southern African Vultures. Vulture Study Group, Johannesburg, South Africa. Received 1 April 1996; accepted 14 November 1996 J. Raptor Res. 31(1):79-81 © 1997 The Raptor Research Foundation, Inc. Hunting Synchrony in White-Tailed Kites Molly F. Skonieczny Department of Biology, Humboldt State University, Areata, CA 95521 U.S.A. Jeffrey R. Dunk Department of Wildlife, Humboldt State University, Areata, CA 95521 U.S.A. Key Words: White-tailed Kite, hunting, Elanus leucurus; California. Hunting behavior of White-tailed Kites {Elanus leucu- rus) has been extensively studied (Bammann 1975, War- ner and Rudd 1975, Mendelsohn andjaksic 1989). Kites primarily use open to semi-open habitats for hunting (Waian 1973, Bammann 1975, Dunk and Cooper 1994). In California, kites almost exclusively hover while they hunt (Mendelsohn andjaksic 1989, Dunk 1995) and prey on small mammals, primarily voles ( Microtus spp.) (Haw- becker 1940, 1942, Stendell 1972). In previous studies of kites, we observed that groups of kites (2-20 individuals) appeared to hunt relatively synchronously and the probability of an individual kite hunting appeared to be related to whether other kites were hunting. Hunting synchrony could result from kites advertising their presence to conspecific territory holders to potentially decrease subsequent interactions, or to more easily patrol and defend a territory, or from kites responding to variability in prey availability as a function of prey activity rhythms. Shields (1976, cited in Madison 1985) found California voles (M. californicus) exhibited ultradian rhythms in activity varying from 2-6 hr. Daan et al. (1982) reported positive correlations between vole activity and timing of hunting by raptors in Europe. We examined whether kites hunted independent of other in- dividuals hunting. Study Area and Methods This study was conducted at the Mad River Slough Wildlife Area, Areata, California. The area consists of ap- proximately 185 ha of ungrazed grassland with little to- pographic relief. It contains very few trees and shrubs, and fence posts and T-bars (ca. 3-m tall) provide most of the perches for raptors. The climate is maritime with mild winters and cool summers. The study took place from 20 November 1994-30 Jan- uary 1995. We made observations during seven 1-2 hr periods. Random points within the study area were estab- lished from which observations were made. Scan sam- pling (Altmann 1974) was used to record number of kites hunting. Using a landmark on the horizon as a starting point, we slowly scanned (x time per scan = 90 sec, range = 45-135 sec) 360° and recorded the number of kites hunting (hovering). Scans were made using binoculars. All kites were within approximately 400 m of the observ- er. We waited 5 min between scans based on Bammann’s (1975) findings that mean hunting time for kites was 5.04 min (N = 674 hunts) in this area. To determine whether kites hunted independent of other hunting kites, we compared observed numbers of kites hunting during each scan to expected numbers un- der the Poisson distribution using Chi-square analysis. We used this analysis because comparison of observed events 80 Short Communications Vol. 31, No. 1 Table 1. Number of White-tailed Kites observed vs. ex- pected (using the Poisson distribution) to be hunting si- multaneously at Areata, GA 1994-95, Number Hunting per Scan Number of Times Observed Number of Times Expected 0 43 29 1 31 41 2 21 29 3 13 14 4+ 11 5 to the Poisson distribution allows a test of the indepen- dence of the individual events (Zar 1984). Groups of 4, 5, 6 and 7 kites hunting simultaneously had expected frequencies less than 5, so they were combined into one group (>4 individuals). Results We observed 168 kites hunting during 119 scans. Mean number of kites hunting per scan was 1.41 (range = 0- 7) Individual kites hunted in a nonrandom fashion with respect to the presence of other hunting kites (x 2 = 19.36, df = 3, P = 0.0002). Few kites (1 or 2) hunted less often than expected whereas many kites (>4) hunted simultaneously more often than expected (Table 1). Discussion We found that kites hunted in a nonrandom way with respect to the presence of other hunting kites. Our statistical test was probably conservative because data were analyzed as if there were always ^4 kites on the study area that could have hunted simultaneously. On some occasions there were only a couple of kites in the sampling area. Thus, our anal- ysis was probably biased against observing ^4 kites hunting simultaneously more than expected. It appears that kites responded to either other individuals hunting or, very rap- idly, to some other factor. It was beyond the scope of this study to examine whether kites were ultimately responding to vole activity rhythms, to other individual kites for terri- torial reasons, or both. Proximately, however, kites appeared to respond to other hunting kites. Species such as Common Ravens ( Corvus corax), Bald Eagles ( Haliaeetus leucocephalus) , and many vultures are known to forage together and “obtain information” about food resources from one another (Knight and Knight 1983, Heinrich 1988), often at communal roosts, but also through local enhancement (Knight and Knight 1983) where searchers are attracted to actively foraging individ- uals, or by actually recruiting other individuals (Heinrich 1988). Although kites are also communal roosters during the nonbreeding season (Bolander and Arnold 1965, Waian 1973, Clark and Wheeler 1989), their diurnal be- havior during the nonbreeding season contrasts with other species in that kites hold hunting territories (Dunk and Cooper 1994). Moreover, amongst Common Ravens, Bald Eagles, vultures and kites, kite prey items are relatively more predictable spatially because they are not nearly as patchily distributed and are of much smaller size. We pro- pose that the information that nonhunting kites obtain from hunting individuals is not where to forage, as has been found for other species, but when to forage. If kites were ultimately responding to variations in vole activity, nonhunting kites probably benefit by observing whether a hunting kite makes an attempt at prey (wheth- er successful or not because any attempt is presumably in response to a prey detection) . If no prey attempts are made, then nonhunting individuals should remain perched and thus conserve energy. If attempts at prey are made, then nonhunting kites should use this infor- mation to decide whether to hunt. If this “vole rhythm” hypothesis is correct, then voles should exhibit synchro- nous activity rhythms over relatively large areas, at least the size of our study area. Subsequent studies should ex- amine the relationship between vole activity and the number of kites (and other raptors) hunting. Resumen. — Nosotros examinamos si Elanus leucurus en el norte de California cazaba en sincronizacion mas que es- peraban bajo el pronostico Poisson. Nosotros encontra- mos que pocos E. leucurus cazaban con menos frecuencia de lo que esperaban y muchos (>4) E. leucurus cazaban con mas frecuencia de lo que esperaban segun parecen responder con la presencia de otros E, leucurus cazando. Finalmente, E. leucurus podia estar respondiendo a activ- idad ritmica de presa, a otros E. leucurus por razones ter- ritorial, o los dos. [Traduccion de Raul De La Garza, Jr.] Acknowledgments We thank P.H. Bloom and J.B. Buchanan for helpful comments on an earlier draft of the manuscript and the Department of Wildlife, Humboldt State University, for logistical support during manuscript preparation. Literature Cited Altmann, J. 1974. Observational study of behavior: sam- pling methods. Behaviour. 49:227-267. Bammann, A.R. 1975. Ecology of predation and social interactions of wintering White- tailed Kites. M.S. the- sis, Humboldt State Univ., Areata, CA U.S.A. Bolander, G.L. and J.R. Arnold. 1965. An abundance of Wdiite-tailed Kites in Sonoma County, California. Condor 67:446. Clark, W.S. and B.K. Wheeler. 1989. Unusual roost site selection and staging behavior of Black-shouldered Kites./. Raptor. Res. 23:116—117. Daan, S.W. Altenburg, G. Boedeltje, C. Dijkstra, M. Gerkema, F. de Hass, R. Hohe, P. Koene, H. van der Leest, D. Masman, M. Schoenmakers, H. Waterbolk, P. Wildschut and M. Zijlstra. 1982. Timing of vole hunting in aerial predators. Adam. Review 12:169-181. Dunk, J.R. 1995. WTiite-tailed Kite ( Elanus leucurus). In A. Poole and F. Gill [Eds.], The Birds of North Amer- ica, No. 178. The Academy of Natural Sciences, Phil- adelphia and The American Ornithologists’ Union, Washington, DC U.S.A. March 1997 Short Communications 81 AND R.J. Cooper, 1994. Territory size regulation in Black-shouldered Kites. Auk 111:588-595. Hawbecker, A.C. 1940. The nesting of the White-tailed Kite in Southern Santa Cruz county, California. Con- dor 42:106-111. . 1942. A life history study of the White-tailed Kite. Condor 44:267-276. Heinrich, B. 1988. Winter foraging at carcasses by three sympatric corvids, with emphasis on recruitment by the Raven, Corvus corax. Behav. Ecol. Sociobiol. 23:141—156. Knight, S.K. and R.L. Knight. 1983. Aspects of food finding by wintering Bald Eagles. Auk 100:477-484. Madison, D.M. 1985. Activity rhythms and spacing. Pages 373-419 in R.H. Tamarin [Ed.], Biology of New World Microtus. Spec. Publ. No. 8, Am. Soc. Mammal. Mendelsohn, J.M. and F.M Jaksic. 1989. Hunting be- haviour of Black-shouldered Kites in the Americas, Europe and Australia. Ostrich 60:1-12. Shields, L.J. 1976. Telemetric determination of the ac- tivity of free-ranging rodents; the fine structure of Mi- crotus calif ornicus activity patterns. Ph.D. dissertation, Univ, California, Los Angeles, CA U.S.A. Stendell, R.C. 1972. The occurrence, food habits, and nesting strategy of White-tailed Kites in relation to a fluctuating vole population. Ph.D. dissertation, Univ. California, Berkeley, CA U.S.A. Waian, L.B. 1973. The behavioral ecology of the North American White-tailed Kite ( Elanus leucurus majusculus ) of the Santa Barbara coastal plain. Ph.D. dissertation, University of California, Santa Barbara, CA U.S.A. Warner, J.S. and R.L. Rudd. 1975. Hunting by the White-tailed Kite. Condor 77:226-230. Zar, J.H. 1984. Biostatistical analysis. Prentice Hall, En- glewood Cliffs, NJ U.S.A. Received 23 April 1996; accepted 29 November 1996 J. Raptor Res. 31(1) :81 — 83 © 1997 The Raptor Research Foundation, Inc. Blood Parasites of Nestling Goshawks E.P. Toyne 1 Department of Biology, Imperial College, London SW7 2BB UK R.W. Ashford Department of Parasitology, School of Tropical Medicine, Pembroke Place, Liverpool L3 5QA UK Key Words: Northern Goshawk ; Accipiter gentilis; nestling survival ; blood parasites. There are a few studies which have investigated patho- genic effects of haematozoa on wild raptors. Ashford et al. (1990, 1991) were unable to demonstrate any effect of Leu- cocytozoon toddi on the mortality of nestling or adult Euro- pean Sparrowhawks ( Accipiter nisus) in England and Korpi- maki et al. (1993) showed no effect of L. ziemanni on body mass or molting progress in almost 200 Tengmalm’s Owls (, Aegolius Junereus) in Finland. They did, however, find that four of six females which laid unusually small clutches had relatively heavy infections. In a second large study, Korpi- maki et al. (1995) found that mates of male European Kes- trels (Falco tinnunculus) infected with Haemoproteus produced smaller clutches earlier than mates of uninfected males. It 1 Present address: WWF-UK, Panda House, Weyside Park, Catteshall lane, Godaiming, Surrey GU7 1XR, UK. is unfortunate that the precise ages of these birds were un- known as this was likely to be the confounding variable. The occurrence of blood parasites in nestling North- ern Goshawks (Accipiter gentilis) is unknown and, if they do occur, their role as a factor of regulation of goshawk reproduction is unclear. The aim of this preliminary study was to investigate occurrences of blood parasites and to assess whether they are a significant mortality fac- tor in nestling goshawks. Methods Nests of Northern Goshawks were studied from March through late July 1994. Clutch sizes were determined by viewing nest contents in late April and early May 1994 with a mirror attached to a telescopic pole. We climbed into nests between 11-14 June to sex, band, weigh and measure wing lengths (standard B.T.O. maximum chord: Spencer 1984) of nesdings. Body mass was adjusted for crop contents by subtracting 60 g if nesdings had full crops and 15 g if they had crops that were half-full or less. Wing-length measure- ments were used to age nesdings (±4 d) from growth equa- tions of Swedish goshawks (Kenward et al. 1993). The first 82 Short Communications Vol. 31, No. 1 Table 1. Distribution of infections of Leucocytozoon toddi in nestling Northern Goshawks by brood size and sex. Brood Size 1 2 3 Total Nests # Infected broods 1 5 2 8 # Uninfected broods 4 6 5 15 # Infected males 1 3 0 4 # Infected females 0 2 3 5 egg-laying date, if not known from direct observations, was calculated by backdating from the age of the oldest nestling in a brood. The incubation period used was 38 d (36 d + 2d), as incubation does not start until at least the second egg is laid, two days after the first egg is laid (Cramp & Simmons 1980). We are not aware of any instances of gos- hawks removing unhatched eggs from their nests, so any eggs that were unaccounted for, after thorough searches of nest material, were assumed to have hatched and the re- sulting young to have died. In late July, nesting territories were revisited to check for occupancy and fledgling mortal- ities. Juveniles were classified as having dispersed when they were >400 m from the nesting territory and its immediate vicinity. Blood samples were taken at the same time as nestlings were banded. This was done by clipping the tip of the talon on the inner toe. All nestlings were handled in the same manner by the same observers. Blood smears were prepared in the field, immediately air-dried, fixed with ab- solute methanol and later stained with Giemsa stain in a laboratory. Slides were examined under a microscope and the parasite load was estimated on a logarithmic scale of 0 to 4, where 0 = no parasites seen in the entire film examined; 1 = fewer than 1 parasite per 100 high-power fields (X400); 2 = 1-10 parasites per 100 high-power fields; 3 = 11-100 parasites per 100 high-powered fields; 4 = more than 100 parasites per 100 high-power fields. Results and Discussion We examined the blood of 48 nestlings from 23 nests. Five female (26% of females) and four male (14% of males) nestlings from eight nests were infected with Leu- cocytozoon toddi (Table 1). One male was infected with a trypanosome, presumably Trypanosoma avium, but not with L. toddi. The single trypanosome infection suggested that trypanosomes had a negligible impact on nestling mortality. In Britain, avian trypanosomes are also found in blackflies (Simulium sp.) and European Sparrowhawks (Pierce and Marquiss 1983, Dirie et al. 1990). The median date for first egg laying of infected broods (median =13 April, range = 7-9 April, N = 8) was sim- ilar to that of uninfected broods (median = 9 April, range = 31 March-18 April, N = 14; Mann-Whitney U-test, U 8 14 = 35.0, P > 0.05) suggesting parasite infec- tion and date of egg laying were not related. L. toddi in- fections also did not appear to be associated with body mass or sex of nestlings, as male and female nestlings of differing body masses and ages were infected and, in gen- Figure 1. Effect of Leucocytozoon toddi infection on growth of male and female nestling Northern Goshawks in Wales in 1994. Males: Y= —805.9 + 1010.01og(x), fe- males: Y= -1434.5 + 1632.81og(x). eral, the mass of infected individuals of both sexes were within ranges of uninfected nestlings (Fig. 1 ) . There was no apparent association between infections and brood size (Likelihood ratio test, x 2 2 = 1-182, P > 0.05, Table 1), nor was there any apparent clustering of infections by sex with brood size (x 2 2 = 6.352, P> 0.05). Four male nestlings had infection loads of 1 and 2 and none had loads of 3. Conversely, only two female nest- lings had infection loads of 1 and 2 and three had loads of 3. Although higher infection loads were higher in fe- males this was not significant (x 2 2 = 5.094, P> 0.05). The mean clutch size, brood size at banding and fledg- lings at dispersal in infected and uninfected broods were similar (Table 2). The mortality of young goshawks up to time of banding (>13 d) in infected broods (5 from 19 broods) was not significantly higher than in uninfected broods (4 from 28 broods; Fisher’s exact 1-tail test, P = 0.255, Table 2). Parasites take around 14 d to appear in the blood (Pierce and Marquiss 1983), but infected nestlings Table 2. Productivity and survival of Northern Goshawk eggs and nestlings in nests infected (N = 6) and unin- fected (N= 10) with the blood parasite Leucocytozoon toddi in Wales in 1994. Uninfected Nests (Number) Infected Nests (Number) Eggs laid 33 22 Unhatched eggs 5 3 Nestlings hatching 28 19 Nestlings dying 4 5 Nestlings fledging 24 14 Fledglings dying 1 0 Fledglings dispersing 23 14 March 1997 Short Communications 83 may be ill and possibly die during this prepatent period. Further studies involving a larger sample size are needed to examine if there is a higher mortality of young goshawks in infected broods. If so, this might explain why only light in- fections were found since heavy infections, if they occurred, would have resulted in death before 13 d of age. There was no association between infection and mor- tality in a brood between banding (13-39 d) and fledging as nestlings from both infected and uninfected broods died (x 2 i = 0.166, P > 0.05). This suggested that infec- tions did not reduce the survival of nestlings past the young nestling stage (>14 d). The geographic distribution and physical characteristics of nesting territories of infected goshawks were compared to those of uninfected goshawks (Mann-Whitney U-test). There were no statistically significant differences (P > 0.05) between distances to nest trees in other nesting territories and running water, or between nesting territory elevations of infected and uninfected goshawks. The distribution of infected nestlings within the study area did not suggest any clustering. Physical characteristics of nesting territories of infected and uninfected territories were also similar. This was not surprising because most nests were built in larch (Larix sp.) trees >25 yr old and larch was the most common tree providing a suitable nesting substrate for goshawks. Our results suggest that parasitic infections are likely to cause no short-term mortalities in goshawks in the post-dis- persal period. Infection loads were light compared with those of sparrowhawks (Ashford et al. 1990, 1991). Resumen. — Nosotros encontramos nueve pajaritos de Ac- cipiter gentilis (cinco hembras y cuatro machos) en ocho nidos de 48 pajaritos en 23 nidos en Wales que estaban infectados con Leucocytozoon toddi y un macho que estaba mfectado con Trypanosoma. La mortalidad de los pajaritos hasta el tiempo de ser marcados (>13 d) en crias infec- tados no fue considerablemente mas alto que en crias que no estaban infectados, ni hubo ninguna asociacion entre infecciones y mortalidad en una cria entre los mar- cados (13-39 d) y los pajaros jovenes porque los pajaritos de los infectados y sin infeccion murieron. [Traduccion de Raul De La Garza, Jr.] Acknowledgments We wish to thank the Countryside Council for Wales for the provision of the necessary licenses (SB:6:94) and Herman Ostroznik, Herman Ostroznik, Jr., Steve Binney and staff of a Forest Enterprise district for help with field- work and logistics. We are also grateful to the British Eco- logical Society (SEPG No. 1119) and the Liverpool School of Tropical Medicine for funding the project. Literature Cited Ashford, R.W., I. Wyllie and I. Newton. 1990. Leuco- cytozoon toddi in British Sparrowhawks Accipiter nisus. observations on the dynamics of infections. J. Nat. Hist. 24:1091-1100. , E.E. Green, P.R. Holmes and AJ. Lucas. 1991. Leu- cocytozoon toddi in British Sparrowhawks Accipiter nisus. pat- terns of infection in nestlings. J. Nat. Hist. 25:269-277. Bennet, G.F., M.A. Pierce and R.W. Ashford. 1993. Avi- an haematozoa: mortality and pathogenicity. J. Nat. Hist. 27: 993-1001. Dirie, M.F., R.W. Ashford, L.M. Mungomba, D.H. Mo- lyneux and E.E. Green. 1990. Avian trypanosomes in Simulium and Sparrowhawks ( Accipiter nisus) . Para- sitology 101:243-247. Kenward, R.E., M. Marquiss, and I. Newton. 1981. What happens to goshawks trained for falconry. J. Wildl. Manage. 45:802-806. , V. Marcstrom and M. Karlbom. 1993. Post-nest- ling behaviour in Goshawks ( Accipiter gentilis) I. The cause of dispersal. Anim. Behav. 46:365—370, Korpimaki, E., H. Hakkarainen and G.F. Bennet. 1993. Blood parasites and reproductive success of Teng- malm’s Owls: detrimental effects on females but not on males. Fund. Ecol. 7:420-426. , P. Tolonen and G.F. Bennet. 1995. Blood par- asites, sexual selection and reproductive success of European Kestrels. Ecosdence 2:335—343. Marquiss, M. and I. Newton. 1982. The goshawk in Britain. Brit. Birds 75:243—260. Pierce, M.A. and M. Marquiss. 1983. Haematozoa of Brit- ish birds. VII. Haematozoa of raptors in Scotland with a description of Haemoproteus nissi sp. nov from the Sparrowhawk (Acdpiter nisus) . J. Nat. Hist. 17:813-821. Spencer, R. 1984. The Ringer’s manual, 3rd ed. B.T.O., Norfolk, UK Received 16 March 1996; accepted 30 November 1996 84 Short Communications Vol. 31 , No. 1 J Raptor Res. 31(1):84— 86 © 1997 The Raptor Research Foundation, Inc. Migration of Flocks of Honey Buzzards in Southern Italy and Malta Nicolantonio Agostini Via Carlo Alberto n.4, 89046 Marina di Gioiosa Jonica ( RC '), Italy Daniela Logozzo Via A. Oramsci n.26, 89046 Marina di Gioiosa Jonica (RC), Italy Charles Colero Birdlife Malta (MOS), P.0. Box 498 Valletta, CMR 01 Malta Key WORDS: Honey Buzzards; Pernis apivorus; migration; flocking behavior, orientation; navigation. Honey Buzzards ( Pernis apivorus) frequently migrate in flocks that become concentrated in narrow coastal areas (Cramp and Simmons 1980, Kerlinger 1989). During mi- gration over the central Mediterranean, they use mainly two routes. One route is from Sicily to Tunisia and the second route goes across a larger stretch of sea via the islands of Malta (Moreau 1953, Beaman and Galea 1974, Brown et al. 1982, Agostini et al. 1994b). In autumn, a major concentration of buzzards occurs in the Calabrian Apennines (southern Italy), where the distance between the Tyrrhenian and Jonian coasts is narrowest (Agostini and Logozzo 1995a, 1995b). As at the Strait of Gibraltar and the Bosphorus Porter and Willis 1968, Bernis 1973), peak numbers of buzzards are observed between the end of August and the beginning of September. Adult Honey Buzzards migrate using this route and probably use the same route in spring, crossing the cen- tral Mediterranean between Sicily and Tunisia (Agostini et al. 1994a, Agostini and Logozzo 1995b). This assump- tion is supported by observations on the island of Mar- ettimo (western Sicily), where large flocks of buzzards are seen between the end of August and the beginning of September (Zangirolami pers. comm.). Unlike adults, young buzzards appear to concentrate on the island of Malta (Agostini and Logozzo 1995b) after the second week of September (Beaman and Galea 1974) . This study was undertaken to determine if flocks of immature buz- zards on Malta result because young birds stopover on the island during periods of bad weather or if they con- Figure 1. Seasonal occurrence of migrating Honey Buzzards in the Calabrian Apennines in summer and autumn of 1995. March 1997 Short Communications 85 September 1993 □ M. Covello ■ Malta Figure 2. Variations in migration of Honey Buzzards from 14-30 September 1993 on Mount Covello and Malta. gregate until adult buzzards can show them the shortest route across the Mediterranean. Study Area and Methods Migrating Honey Buzzards were counted in the Cala- brian Apennines and from a post on the slopes of Mount Covello (elevation 700 m) from 24 August-5 October 1993 and 1995. Buzzards were also counted on the Island of Malta. There, the observation post was situated on the highest point on the island (approximately 250 m eleva- tion) from 6 September-5 October 1993 and 1995. A total of 334 hr of observations were tallied at Mount Cov- ello using 10 X 50 binoculars and a total of 188 and 177 hr of observation were tallied using 10 X 40 binoculars on the Island of Malta in 1993 and 1995, respectively. Results A total of 1095 Honey Buzzards were counted in the Calabrian Apennines and more than 90% were observed between 24 August-12 September. The number of mi- grating buzzards showed a bimodal distribution with peaks occurring in late August and after the first week of September (Fig. 1). In both years, there was overlap in the migration periods of adult and juvenile buzzards but over twice as many adults were observed in both years (adults = 346, juveniles = 136). In 1995, the number of adults observed was also significantly greater than in 1993 (X 2 = 6.38, P< 0.05). A total of 483 Honey Buzzards were counted on the Island of Malta in 1993 but only 88 were counted in 1995. More than 97% of them were observed in September and this corresponded with counts in the Calabrian Apen- nines (Fig. 2). Discussion Some authors have shown that migrating birds orient better when they fly in groups (Keeton 1970, Rabol and Noer 1973, Wallraff 1978, Von Helbig and Laske 1986), especially when groups contain adults that can show to correct migration route (Kerlinger 1989). Our observa- tions suggest that juvenile Honey Buzzards learn the shortest route to cross the central Mediterranean by mi- grating in flocks of adults. This would explain why the Honey Buzzard is commonly seen in Malta, where the African coasts are more than 400 km away. Flocking behavior might have another function. Stud- ies carried out on the Cap Bon promontory (Tunisia) during the spring migration of Honey Buzzards and Black Kites ( Milvus migrans) have shown that these rap- tors cross the Channel of Sicily more frequently when migrating in large flocks (Agostini and Duchi 1994, Agos- tini et al. 1994a). This behavior suggests that flocking is important for water crossing because an increase in flock size increases the probability the journey over water will be successfully completed (Agostini and Duchi 1994, Agostini et al. 1994b). Resumen. — Observaciones de Pernis apivorus emigrando a traves de el centro Mediterranean fueron hechos en los Calabrian Apennines (el sur de Italy) del 24 de agosto- 5 de octubre 1995, y en Malta del 6 de septiembre-5 de octubre 1993 y 1995. En los Calabrian Apennines, 1095 P. apivorus fueron observados, con un maximo de 237 pajaros emigrando e el 7 de septiembre. Casi todos los P. apivorus fueron observados entre el 24 de agosto-12 de septiembre. En Malta, 483 y 83 P. apivorus fueron con- tados en 1993 y 1995, respectivamente, con casi todos contados despues del 15 de septiembre. Estos resultados proponen que P. apivorus jovenes aprenden la ruta mas corta para cruzar el centro Mediterranian durante el primer ano de migracion cuando van en bandadas con adultos que han hecho la migracion antes. [Traduccion de Raul De La Garza, Jr.] Acknowledgments We are greatly indebted to Paul Portelli, Francesco Cecere and Alberto Zangirolami for their valuable collab- oration. Our particular gratitude goes to Allen Fish, Keith Bildstein and Javier Bustamante for their useful comments on the manuscript. 86 Short Communications Vol. 31, No. 1 Literature cited Agostini, N. and A. Duchi. 1994. Water-crossing behav- ior of Black Kites ( Milvus migrans) during migration. Bird Behav. 10:45-48. Agostini, N. and D. Logozzo. 1995a. Observations on the autumn migration of raptors over the Calabrian Apennines. Riv. Ital. Om. 64:117-120. and . 1995b. Autumn migration of Honey Buzzards in southern Italy. J. Raptor Res. 29:275-277. Agostini, N., G. Malara, F. Neri, D. Mollicone and S. Melotto. 1994a. La migrazione primaverile del Fal- co pecchiaiolo ( Pernis apivorus) a Cap Bon (Tunisia) e sullo Stretto di Messina. Atti VI Convegno Ital. di Ornit.:451-452. Agostini, N., G. Malara, F. Neri, D. Mollicone and S. Melotto. 1994b. Flight strategies of Honey Buzzards during spring migration across the central Mediter- ranean. Avocetta 18:73-76. Beaman, M. and C. G al ea. 1974. The visible migration of raptors over the Maltese Islands. Ibis 116:419-431. Bernis, F. 1973. Migracion de Falconiformes y Ciconia spp. por Gibraltar, verano otono 1972-1973: Primera parte. Ardeola 19:152-224. Brown, L. H., E.K. Urban and K. Newman. 1982. The birds of Africa, Vol. I. Academic Press, London, UK. Cramp, S. and K.E.L. Simmons. 1980. The birds of the western Palearctic, Vol. II. Oxford Univ. Press, Oxford, UK. Keeton, W.J. 1970. Comparative orientational and hom- ing performances of single pigeons and small flocks. Auk 87:797-799. Kerlinger, P. 1989. Flight strategies of migrating hawks. Univ. Chicago Press, Chicago, IL U.S.A. Moreau, R.E. 1953. Migration in the Mediterranean area. Ibis 95:329-364. Porter, R.F. and I. Willis. 1968. The autumn migration of soaring birds at the Bosphorous. Ibis 110:520—536. Rabol, J. and H. Noer. 1973. Spring migration in the Skylark ( Alauda arvensis ) in Denmark: influence of environmental factors on the flocksize and the cor- relation between flocksize and migratory direction. Vogelwarte 27:50-65. Von Helbig, A. and V. Laske. 1986. Zeitlicher Verlaf und Zugrichtungen beim Wegzug des Stars ( Sturnus vul- garis) in nordwestdeutschen Binnenland. Vogelwarte 33:169-191. W at .i .r aff, H.G. 1978. Social interrelations involved in migration orientation of birds: possible contributions of field studies. Oikos 30:401-404. Received 18 June 1996; 30 November 1996 J. Raptor Res. 31(l):86-88 © 1997 The Raptor Research Foundation, Inc. Nesting-Tree Preference and Nesting Success of Japanese Lesser Sparrowhawks in Japan Mutsuyuki Ueta Research Center, Wild Bird Society of Japan, 2-35-2 Minamidaira, Hino, Tokyo 1 91, Japan Key Words: Accipiter gularis; Pinus densiflora; Japanese Lesser Sparrowhawk; Japanese red pine, nesting success', nest tree preference. Japanese Lesser Sparrowhawks ( Accipiter gularis) breed throughout northeastern Asia (Brown and Amadon 1968). It has been shown that tree plantings in lowland areas are important breeding sites for this hawk in Japan (Endo et al. 1991) and these hawks mainly nest in Japa- nese red pines ( Pinus densiflora) (Endo and Hirano 1990, Hirano and Kimizima 1992). Because prey abundance does not seem to be limiting (Hirano and Kimizima 1992, Ueta 1992), the availability of nest sites may be an im- portant factor limiting the population of Japanese Lesser Sparrowhawks (Ueta 1996). Because nest sites of Japa- nese Lesser Sparrowhawks are important in predicting future populations of this hawk in Japan, I examined the nest-tree preferences of Japanese Lesser Sparrowhawks and determined whether such preferences influence nesting success. Methods The study was conducted from 1987-94 at 16 groves of trees in suburban areas of Tokyo. The groves were isolated and ranged from 1-4 ha in area. They were mainly coppices composed primarily of Japanese chest- nut oak ( Quercus acutissima), Storax ( Sty rax japonica) and Sawara cypress ( Chamaecyparis pisifera) . To determine nest-site preference, use by sparrow- hawks and the availability of different tree species were compared. I excluded nests in which the hawks did not March 1997 Short Communications 87 CO 1000 ha). The stand was approximately 800 m from a reservoir and 5 km from the nearest town. In 1989, the heronry was used by approximately 25 pairs of breeding herons. The herons began to use the heronry again in April 1990 but, shortly after their eggs were laid, strong winds blew them from the nests and heron eggs were found littering the ground. After the herons had left the heronry, a goshawk was found nesting in one of the abandoned heron nests. Over the next seven breeding seasons, goshawks bred in a total of six abandoned heron nests laying an average of 3.3 eggs (range = 2-4) and fledging an average of 2.5 young (range = 1-4). All nest trees were Sitka Spruce and all heron nests were made of larch ( Larix sp .) branches. Goshawks refurbished these nests with larch branches and lined the nest cups with foliage from larch, spruce and Douglas-fir ( Pseudotsuga menziesii). Nest measurements were similar to other goshawk nests built on whorls of larch and conifer trees (length: 0.94—1.20 m, breadth: 0.83-1.00 m, depth: 0.25-0.50 m, N = 3). Five of the six nests that were used were well concealed in trees with dense canopies. The other nest was in an open, thin-crowned tree. It was unclear why the herons moved from the heronry and where they moved. Herons often relay after egg loss (C. Voisin 1991, The herons of Europe, T. & A.D. Poyser, London, UK) so it was unlikely they deserted their colony after losing their clutches during the wind storm. Although there is no record of goshawks in Britain killing herons (M. Marquiss and I. Newton 1982, British Birds 75:243-260; E.P. Toyne 1994, Ph.D. dissertation, Imperial College, London, UK), it is plausible that the pair of goshawks disturbed the herons enough to make them move elsewhere. At another site within the study area, herons moved after goshawks first nested within 500 m of a heronry indicating that the presence of goshawks may cause herons to desert their nesting areas. Goshawks are known to use existing nests, including artificial ones (P. Saurola 1978, Pages 72-80 in T.A. Geer [Ed.] , Bird of prey management techniques, British Falconer’s Club, Oxford, UK; S .J. Petty 1989, Forestry Commission Bulletin, Her Majesty’s Stationery Office, London, UK). In our study, larch appeared to be the preferred nest tree of goshawks with 61.2% of 116 nesting attempts in larch as opposed to 15.5% in Sitka spruce (E.P. Toyne 1994, Ph.D. dissertation, Imperial College, London, UK) . While larch trees were nearby the heronry, the presence of heron nests appears to have attracted goshawks to the spruce site. Data for this paper were collected during a larger study of goshawk ecology supported by the Science and Engi- neering Research Council, Imperial College, Forest Enterprise, The British Ecological Society and The Hawk and Owl Trust. We are grateful to the Countryside Council for Wales for providing the necessary licences. We thank Steve Petty, Thomas Bosakowski and an anonymous referee for useful comments to an earlier draft. We thank Herman Ostroznik Jr., Mike Coleman, Steve Binney, Steve Galleozzie and staff of the Forest Enterprise for aid with this fieldwork. — E.P. Toyne, 1 Department of Biology, Imperial College, London SW7 2BB, UK and H. Ostroznik, % The Hawk and Owl Trust, Zoological Society of London, Regent’s Park, London NW1 4RY, UK 1 Present address: WWF-UK, Panda House, Weyside Park, Catteshall Lane, Godaiming, Surrey GU7 1XR, UK. 89 BOOK REVIEW Edited by Jeffrey S. Marks J. Raptor Res. 31(1):90 © 1997 The Raptor Research Foundation, Inc. Handbook of the Birds of the World, Volume 2. New World Vultures to Guineafowl. Edited by Jo- sep del Hoyo, Andrew Elliott and Jordi Sargatal. 1994. Lynx Edicions, Barcelona. 638 pp., 60 color plates, 302 color photographs, 590 distribution maps. ISBN 84-87334-15-6. Cloth, $175.00.— The volumes in this series are magnificent in both ap- pearance and content. This volume covers the Fal- coniformes and Galliformes, and volume 3 (Hoat- zin to alcids) will appear before this review is pub- lished. Twelve volumes are planned in all. The price tag might seem excessive until one puts things in proper perspective. Volume 2 weighs in at 8 pounds, for a cost of about $1.37 per ounce. By comparison, Cody’s Habitat Selection in Birds (1985) now costs $136.00 in cloth, or about $4.00 per ounce (and even in 1985, it cost more than $2.00 per ounce). Given the shear quantity of in- formation and quality of presentation, I believe that The Handbook is an exceptional deal in today’s market of inflated book prices. The book is organized by family, each of which contains a general review of the topics “Systemat- ics,” “Morphological Aspects,’’ “Habitat,” “Gen- eral Habits,” “Voice,” “Food and Feeding,” “Breeding,” “Movements,” “Relationships with Man” and “Status and Conservation.” Following the family introductions are the species accounts, typically two to four per page. Virtually every ex- tant species is covered, with each account contain- ing notes on taxonomy, subspecies and distribution (including a range map), habitat, food, breeding biology, movements and conservation status. Each account concludes with a list of recent references. In addition, each species is depicted in a color plate (averaging more than 20 individuals per plate), often with multiple paintings to show dif- ferences in sex, color morph and subspecies. The color plates appear to be excellent, although I am not qualified to evaluate all of them. The color photographs scattered throughout the text are ab- solutely outstanding. A tremendous range of spe- cies is presented, and each photograph is sharply focused and pleasingly composed. Moreover, many of the photographs depict individuals that are ac- tually doing something besides posing for portraits. For example, a Verreaux’s Eagle ( Aquila verreauxii) is seen capturing a hyrax (p. 78), an African Har- rier-Hawk ( Polyboroides typus ) hanging from a weav- er nest (with a nestling in its bill; p. 81), and a female Montagu’s Harrier ( Circus pygargus ) deliv- ering a prey item to its nestlings (p. 101). I should mention that the family overviews and species accounts were written by well-qualified rap- tor biologists, including Richard Bierregaard, Wil- liam Clark, David Houston, Alan Kemp, Lloyd Kiff, Bernd-Ulrich Meyburg, Penny Olsen, Alan Poole, Jean Marc Thiollay and Clayton White. The depth of treatment in the species accounts does not sur- pass Brown and Amadon’s Eagles, Hawks and Fal- cons of the World (1968), but the updated informa- tion and excellent plates and photographs provid- ed in The Handbook make it an excellent compan- ion piece to Brown and Amadon. As a sound, general reference, The Handbook will be indispen- sable. I cannot imagine that anyone with a keen interest in raptors could be disappointed with this book. As a bonus, the material on the Galliformes is just as impressive as that on the raptors. I urge you to obtain your own copy, or at the very least, to prod your library into acquiring the entire se- ries. — Jeff Marks, Montana Cooperative Wildlife Research Unit, University of Montana, Missoula, MT 59812 U.SA. 90 J. Raptor Res. 31(1):91 © 1997 The Raptor Research Foundation, Inc. Manuscript Referees The following people reviewed manuscripts for the Journal of raptor Research in 1996, Peer review plays a vital role in the publishing process and in improving the quality of the Journal. The names of individuals who reviewed two or more manuscripts are followed with an asterisk. Dean Amadon, David J. Anderson*, James C. Bednarz, Maria Isabel Bellocq*, James R. Belthoff*, Keith L. Bild- stein*, David M. Bird*, William Block, Peter H. Bloom, Clint B. Boal, Gary R. Bortolotti*, Thomas Bosakowski, Buchanon, Javier Bustamente*, Tom Cade*, Richard J. Cannings, Carss, Richard J. Clark, Stephen DeStefano, Gary E. Duke, David H. Ellis, Susan Ellis, James H, Enderson*, Sidney England, Allen Fish*, Eric Foresman*, Frederick Gehlbach*, James A. Gessaman, Teryll Grubb*, Ralph J. Gutierrez, Patricia A. Hall, Alan H. Harmata, Elizabeth Haug, Gregory D. Hayward, Julie Heath, Charles J. Henny*, Denver Holt*, David Houston*, C. Stuart Houston, C. Hubert, Fabian M. Jaksic*, Paul C. James, Alan Kemp, Brian Kimsey, Jon S. Kirkley, Michael N. Kochert*, Robert Lehman, Carl Marti*, Riley McClelland, Michael McGrady*, Kenneth D. Meyer*, Heimo Mikkola*, Brian A. Milsap, Peter Mundy*, Robert Nero, Ian Newton*, Torgeir Nygard, Susan M. Patla, Steve Petty, Alan F. Poole, Sergej Postupalsky, Richard Reynolds, Gary Ritchison*, Robert J. Ritchie, Robert N. Rosenfield*, Josef K Schmutz*, Dwight Smith*, Karen Steenhof, Ted Swem, Jean Marc Thiollay, Sergio Tiranti, Kimberly Titus*, Mutsuyuki Ueta, Helen Ulmschnei- der, Daniel Varland*, Ian G. Warkentin*, David Whitacre*, Clayton M. White, Sanford Wilbur*, Christian Wiklund, Joseph W. Witt, Petra Bohall Wood. 91 THE RAPTOR RESEARCH FOUNDATION, INC. (Founded 1966) OFFICERS PRESIDENT: David M. Bird SECRETARY: Betsy Hancock VICE-PRESIDENT: Michael N. Kochert TREASURER: Jim Fitzpatrick BOARD OF DIRECTORS EASTERN DIRECTOR: Brian A. Millsap CENTRAL DIRECTOR: Robert N. Rosenfield MOUNTAIN & PACIFIC DIRECTOR: Karen Steenhof CANADIAN DIRECTOR: Gordon S. Court INTERNATIONAL DIRECTOR #1: Jemima ParryJones INTERNATIONAL DIRECTOR #2: Michael McGrady DIRECTOR AT LARGE #1: Patricia L. Kennedy DIRECTOR AT LARGE #2: John A. Smallwood DIRECTOR AT LARGE #5: Keith L. Bildstein DIRECTOR AT LARGE #4: CEsar MArquez DIRECTOR AT LARGE #5: Petra Bohall Wood DIRECTOR AT LARGE #6: Katherine McKeever jjc EDITORIAL STAFF EDITOR: MarcJ. Bechard, Department of Biology, Boise State University, Boise, ID 83725 U.S.A. ASSOCIATE EDITORS Keith L. Bildstein Fabian Jaksic Gary R. Bortolotti Daniel E. Varland Charles J. Henny Javier Bustamente BOOK REVIEW EDITOR: Jeffrey S. Marks, Montana Cooperative Research Unit, University of Montana, Missoula, MT 59812 U.S.A. The Journal of Raptor Research is distributed quarterly to all current members. Original manuscripts dealing with the biology and conservation of diurnal and nocturnal birds of prey are welcomed from throughout the world, but must be written in English. Submissions can be in the form of research articles, letters to the editor, thesis abstracts and book reviews. Contributors should submit a typewritten original and three copies to the Editor. All submissions must be typewritten and double-spaced on one side of 216 X 278 mm (8 Vi X 11 in.) or standard international, white, bond paper, with 25 mm (1 in.) margins. The cover page should contain a tide, the author’s full name(s) and address(es). Name and address should be centered on the cover page. If the current address is different, indicate this via a footnote. A short version of the tide, not exceeding 35 characters, should be provided for a running head. An abstract of about 250 words should accompany all research articles on a separate page. Tables, one to a page, should be double-spaced throughout and be assigned consecutive Arabic numer- als. Collect all figure legends on a separate page. Each illustration should be centered on a single page and be no smaller than final size and no larger than twice final size. The name of the author(s) and figure number, assigned consecutively using Arabic numerals, should be pencilled on the back of each figure. Names for birds should follow the A.O.U. Checklist of North American Birds (6th ed., 1983) or another authoritative source for other regions. Subspecific identification should be cited only when pertinent to the material presented. Metric units should be used for all measurements. Use the 24-hour clock (e.g., 0830 H and 2030 H) and “continental” dating (e.g., 1 January 1990). Refer to a recent issue of the journal for details in format. Explicit instructions and publication policy are outlined in “Information for contributors,”/. Raptor Res., Vol. 27(4), and are available from the editor. 1997 ANNUAL MEETING The Raptor Research Foundation, Inc. 1997 annual meeting will be hosted by Georgia Southern University and will be held October BO through November 2 at the Marriott Riverfront in Savan- nah, Georgia. Details about the meeting and a call for papers will be mailed to Foundation members in the spring of 1997. For more information, contact Michelle Pittman (912/681-5555, e-mail: meeden@gsvms2.cc.gasou.edu) or Steve Hein (912/681-0831) at Georgia Southern Uni- versity. Raptor Research Foundation, Inc., Awards Recognition for Significant Contributions 1 The Dean Amadon Award recognizes an individual who has made significant contributions in the field of systematics or distribution of raptors. Contact: Dr. Clayton White, 161 WIDE, Department of Zoology, Brigham Young University, Provo, UT 84602 U.SA. Deadline August 15. The Tom Cade Award recognizes an individual who has made significant advances in the area of captive propagation and reintroduction of raptors. Contact: Dr. Brian Walton, Predatory Bird Research Group, Lower Quarry, University of California, Santa Cruz, CA 95064 U.SA. Deadline: August 15. The Fran and Frederick Hamerstrom Award recognizes an individual who has contributed significantly to the understanding of raptor ecology and natural history. Contact: Dr. David E. Andersen, Department of Fisheries and Wildlife, 200 Hodson Hall, 1980 Folwell Avenue, University of Minnesota, St. Paul, MN 55108 U.SA. Deadline: August 15. Recognition and Travel Assistance The James R. Koplin Travel Award is given to a student who is the senior author of the paper to be presented at the meeting for which travel funds are requested. Contact: Dr. Petra Wood, West Virginia Cooperative Fish and Wildlife Research Unit, P.O. Box 6125, Percival Hall, Room 333, Morgantown, WV 26506-6125 U.SA. Deadline: established for conference paper abstracts. The William C. Andersen Memorial Award is given to the student who presents the best paper at the annual Raptor Research Foundation Meeting. Contact: Ms. Laurie Goodrich, Hawk Mountain Sanctuary, Rural Route 2, Box 191, Kempton, PA 19529-9449 U.SA. Deadline: Deadline established for meeting paper abstracts. Grants 2 The Stephen R. Tully Memorial Grant for $500 is given to support research, management and conservation of raptors, especially to students and amateurs with limited access to alternative funding. Contact: Dr. Kimberly Titus, Alaska Division of Wildlife Conservation, P.O. Box 20, Douglas, AK 99824 U.SA. Dead- line: September 10. The Leslie Brown Memorial Grant for $500-$l,000 is given to support research and/or the dissemination of information on raptors, especially to individuals carrying out work in Africa. Contact: Dr. Jeffrey L. Lincer, Sweetwater Environmental Biologists, Inc., 3838 Camino del Rio North, Suite 270, San Diego, CA 92108 U.SA. Deadline: September 15. 1 Nominations should include: (1) the name, title and address of both nominee and nominator, (2) the names of three persons qualified to evaluate the nominee’s scientific contribution, (3) a brief (one page) summary of the scientific contribution of the nominee. 2 Send 5 copies of a proposal (^5 pages) describing the applicant’s background, study goals and methods, anticipated budget, and other funding.